Vous êtes sur la page 1sur 531

Competing Equalities 

LAW AND THE BACKWARD 
CLASSES IN INDIA 
Marc Galanter 
University of California Press 
Berkeley Los Angeles London 
­iii­ 
University of California Press 
Berkeley and Los Angeles, California 
University of California Press, Ltd. London, England 
© 1984 by The Regents of the University of California 
Printed in the United States of America 
1 2 3 4 5 6 7 8 9 
Library of Congress Cataloging in Publication Data 
Galanter, Marc, 1931­ 
Competing equalities. 
Bibliography: p. 
Includes index. 
1. Minorities­­Legal status, laws, etc.­India. 2. Untouchables­­
Legal status, laws, etc.­India. 3. India­­Scheduled tribes­­Legal 
status, laws, etc. 4. Civil rights­­India. I. Title. 
LAW 342. 54'0873 82­2017 
ISBN 0­520­04289­1 345. 402873 AACR2 
­iv­ 
TO MY FATHER JACOB GALANTER 1896­1972 
Justice, justice you shall pursue ( Deut. 16:20) 
­v­ 

Contents 
List of Tables 

Foreword by M. 
Hidayatullah 
xiii 

Preface 
xvii 

Acknowldegements 
xxi 

Note on Citations and 
References 
xxiii 

Note on State Names 
xxiv 

List of Abbreviations 
xxv 

Introduction 

PART ONE: 
COMPENSATORY 
DISCRIMINATION: 
THE SETTING AND 
THE POLICIES 

Chapter 1. The 
Compartmental 
Society 

Chapter 2. Reform, 
Mobility, and Politics 
under British Rule 
18 

Chapter 3. 
Compensatory 
Discrimination 
Programs and Their 
Implementation 
41 
A. Legislative 
reservations 
44 

B. Education and 
other schemes 
55 

C. Administration 
64 

D. Critics, costs, and 
benefits 
72 

Chapter 4. 
Reservations in 
Government 
Employment: A Closer 
Look 
at the Paradigm 
Program 
84 
A. The scope of 
reservations 
86 

B. The effectiveness of 
reservations 
93 

C. Assessing the costs 
and benefits 
105 

PART TWO: 
IDENTIFYING THE 
BENEFICIARIES 
119 
Chapter 5. The 
Designation of the 
Scheduled Castes and 
Scheduled Tribes 
121 
A. The invention of the 
Scheduled Castes 
122 

B. Scheduled Castes 
under the Constitution 
131 

C. The Scheduled 
Tribes 
147 

­vii­ 

Chapter 6. The Other 
Backward Classes: The 
Process of Designation 
154 
A. The pre­history of 
the backward classes 
category 
154 

B. The constitutional 
design 
159 

C. The failure to 
centralize policy 
167 

D. The varied state 
patterns 
179 

Chapter 7. The 
Backward Classes and 
the Judiciary: The 
Selection 
of Beneficiary Classes 
188 
A. Castes as classes 
188 

B. The permissibility of 
non­communal classes 
204 

C. The utility of 
communal units 
209 

D. Forbidden 
classifications and the 
burden of proof 
215 

Chapter 8. The 
Backward Classes and 
the Judiciary: The 
Measure 
of Backwardness: 
222 
A. Educational 
backwardness 
223 

B. Social backwardness 
229 

C. Cumulative 
backwardness 
240 

D. Under­representation 
243 

E. Absolute vs. relative 
backwardness 
246 

F. The sufficiency of 
data 
249 

G. Economic tests 
261 

H. Geographical 
criteria 
277 

I. Non­inclusion claims 
279 

Chapter 9. Membership 
in Groups Entitled to 
Preferential Treatment 
282 
A. Reading the 
schedules 
283 

B. Caste and sect 
290 

C. Tribe 
293 

D. Tribe and caste 
295 

E. Religion 
305 

F. Group integrity: 
admission, expulsion, 
re­admittance 
326 

G. Non­communal 
criteria: income and 
occupation tests 
341 

H. Conclusion: 
pragmatic and formal 
solutions 
348 

Chapter 10. Social 
Identity and Judicial 
Action 
352 

PART THREE: 
COMPENSATORY 
DISCRIMINATION 
AND THE JUDICIAL 
PROCESS 
361 
Chapter 11. The 
Constitutional Scope of 
the Compensatory 
Discrimination Policy 
363 
A. Preferential 
treatment in the context 
of 
Fundamental Rights 
364 

­viii­ 

B. The scope of the 
constitutional 
provisions 
for special treatment 
368 

C. The exceptional 
character of 
compensatory 
discrimination 
377 

D. A new 
constitutional vista? 
Thomas and after 
382 
Chapter 12. The 
Scope and Extent of 
Preferences: 
Constitutional 
Limits and Judicial 
Control 
396 
A. How much 
reservation? The 
minimum 
396 

B. How much 
reservation? The 
maximum 
399 

C. The intensity of 
preference 
420 

D. The wider use of 
the reservation 
device 
425 

E. The number of 
beneficiaries 
437 

F. Benefits other 
than reservations 
441 

Chapter 13. The 
Design and 
Operation of 
Compensatory 
Discrimination 
Programs: 
Constitutional 
Limits and Judicial 
Control 
444 
A. Rulefulness and 
discretion 
444 

B. Reservations as 
guaranteed 
minimums 
455 

C. Compartments 
and layers 
463 

D. Relations between 
multiple reservations 
472 

Chapter 14. The 
Setting and 
Incidence of Judicial 
Intervention 
477 
A. Courts and judges 
477 

B. Authority and 
precedent: formal 
structure 
and operating norms 
490 

C. The pattern of 
litigation 
498 

D. Lawyers 
513 

Chapter 15. The 
Uses of Judicial 
Action 
522 
A. The impact on 
policy and 
administration 
522 

B. Judicial review: 
the doctrinal 
accomplishment 
533 

C. The limits of 
judicial action 
541 

CONCLUSION 
Chapter 16. "The 
Little Done, the Vast 
Undone" 
547 
A. A Costly Success 
547 

B. Fairness and 
History 
552 

C. Secularism and 
Continuity 
559 

D. Ex Orient Lux? 
562 

Appendix: Relevant 
Sections of the 
Indian Constitution 
569 

Bibliography 
575 

General Index 
613 
Index of Cases 
621 

­ix­ 

Tables 
1. Annual Per 
Capita 
Expenditure on 
Scheduled Castes 
and Scheduled 
Tribes for the five 
Five­Year Plans 
57 

2. School 
Enrolment as a 
Percentage of 
Corresponding 
Age Group, 
by Communities, 
1973­1974 
60 

3. Views of 
Reservation in 
Government 
Employment for 
Scheduled 
Castes: Three 
North Indian 
Localities in 1968 
70 

4. Alleged 
Benefits and 
Costs of Policy of 
Compensatory 
Discrimination 
81 

5. Scheduled 
Caste and 
Scheduled Tribe 
Employment in 
Central 
Government 
Service, 1953 and 
1975 
89 

6. Scheduled 
Castes, Scheduled 
Tribes, and Other 
Backward 
Classes in 
Government 
Service: Posts 
Reserved and 
Posts 
Held, about 1976 
90 

7. Approximate 
Effectiveness of 
Reservations for 
Scheduled 
Castes and 
Scheduled Tribes 
in Central 
Government 
Services over a 
21½­Year Period 
(1953­1975) 
96 

8. Comparison of 
Scheduled Caste 
and Scheduled 
Tribe 
Employment in 
Government 
Service and 
Public Sector 
Undertakings, 
Mid­1960s 
102 

9. Comparison of 
Scheduled Caste 
and Scheduled 
Tribe 
Employment in 
Government 
Service and 
Public Sector 
Undertakings, 
1975 
103 

10. Alleged 
Benefits and 
Costs of 
Preferences in 
Government 
Employment 
106 

11. Prior 
Educational 
Performance of 
Successful 
Candidates for 
All­India Higher 
Non­technical 
Services, 
1960­1965 
112 

12. Percentage of 
Untouchables in 
British India 
124 

13. Estimates of 
Untouchable 
Population in 
British India by 
Various 
Official Bodies, 
1917­1941 
125 

14. Population of 
Scheduled Castes 
in 1941 
131 

15. Scheduled 
Castes Population 
(in millions) and 
Percentages, 
1950, 1951, 1956, 
1961, 1971 
133 

16. Scheduled 
Tribes Population 
150 

17. Scheduled 
Areas Population 
151 

18. Concessions 
for Other 
Backward 
Classes, 
1951­1952 
163 

19. Estimates of 
Population of 
Other Backward 
Classes, Early 
1930s 
169 

­x­ 

20. A Profile of 
Other Backward 
Classes in the Late 
1970s 
216 

21. Relative 
Success of Selected 
Castes on Merit 
and on Selection 
by Income­cum­
Occupation Test in 
Mysore 
Engineering 
Colleges, 1963 
271 
22. Contrasting 
Results of Different 
Tests of 
Membership 
299 

23. Possible 
Methods of 
Declaring Winners 
in a Double­
Member 
Constituency 
458 

24. A Comparison 
of Upper and 
Lower Strata of 
Backward 
Classes in Tamil 
Nadu 
464 

25. Subject Matter 
of Reported 
Litigation on 
Compensatory 
Discrimination, 
1950­1977 
501 

26. Initiating and 
Successful Parties 
in Compensatory 
Discrimination 
Litigation, 
1950­1977 
502 

27. Regional 
Distribution of 
Litigation about 
(Other) 
Backward Classes 
506 

28. Subject Matter 
of Supreme Court 
Litigation about 
Compensatory 
Discrimination 
506 

­xi­ 
Foreword 

PROFESSOR MARC GALANTER has written an informative and educative treatise on the Indian 
experience with what he terms "compensatory discrimination," a subject which deals with the 
Indian Constitutional policies to safeguard the interests of the historically disadvantaged sections of 
the population. At the present moment there is considerable debate in India whether this 
compensatory discrimination is not an unfair measure, as it keeps out deserving persons and 
promotes others not so qualified. In fact, there is growing evidence of a confrontation between 
persons who have received the benefit of reservation and those who have not. The issues resemble 
recent cases in which the United States Supreme Court has attempted to balance preferred treatment 
for blacks against the claims of whites, complaining of a sort of discrimination in reverse. 
This book is divided into three parts. The first part describes these policies and their background. 
The second part deals with the problem of identifying the recipients of this discriminatory 
advantage. The various aspects of these problems are treated very fully, and the concluding chapter 
summarises "the implications of the governmental superimposition of official categories of 
beneficiaries in the complex process of group and personal identity for the role of the courts in 
implementing the compensatory discrimination policy." The third part deals with the Constitutional 
provisions which embody this policy and with the laws which have implemented it and the judicial 
decisions which apply these provisions. Finally, the Conclusion tries to assess the achievements of 
these programmes in India in the light of the problems that are inherent in such a policy. 
The problem of disadvantaged sections of the population has existed in India for centuries, but it 
acquired special importance after the First World War when a system of protections began to be 
evolved. Later that system came to be supported by the concept of human rights and fundamental 
freedoms in the United Nations Charter and the Universal Declaration of Human Rights. 
In India the special position of the minorities, for a long time, has 
­xiii­ 
been an ineluctable fact of political life. From the time of the Indian National Congress session held 
in Karachi in 1931 and the Nehru and Sapru Reports, it was consistently felt that in order to create a 
homogenous egalitarian society and to achieve national unity, balance between the various 
communities in India was essential. Jawaharlal Nehru posed the problem thus: "How shall we 
promote the unity of India and yet preserve the rich diversity of our inheritance?" The Preamble of 
our Constitution resolves to secure "JUSTICE, social, economic and political," to all its citizens. It 
further proclaims as one of its aims the promotion among them of "FRATERNITY assuring the 
dignity of the individual and the unity of the nation." 
Minorities in India enjoy the guarantees of equality before the law (Article 14), prohibition of 
discrimination on grounds of religion, race, caste, sex or place of birth (Article 15); equality of 
opportunity in matters of public employment (Article 16), the well­known rights to freedom (Article 
19), the right to freedom of conscience and free profession, practice, and freedom in religion or 
sections thereof (Article 26). The Constitution also confers certain rights specifically on minority 
groups: the right to conserve a minority group's distinct language, script, or culture, and the right to 
establish and administer educational institutions of the group's choice (Articles 29 and 30). In 
addition to religious minorities and linguistic minorities the Constitution specifically recognises 
Scheduled Castes (untouchables), Scheduled Tribes, and socially and educationally backward 
classes and guarantees to them protection. 
Article 17 decrees the abolition of untouchability. Article 46 requires the State to promote with 
special care the educational and economic interests of the backward citizens and in particular of the 
Scheduled Castes and Scheduled Tribes and imposes a duty on the State to protect them from social 
injustice and all forms of exploitation. The Governors have been given special responsibility in this 
connection. The Constitution also provides that the Commissioner for Scheduled Castes and 
Scheduled Tribes and the Commissioner for Minorities have to submit periodic reports, which are 
placed before Parliament and which are open to discussion. Reports have been submitted by many 
commissions, making suggestions for protecting the interests of the minorites. 
The rights of minorities have been implemented by statutory enactments. These rights have been 
supported and their scope and ambit clarified by decisions of the Supreme Court and the various 
High Courts from time to time. In spite of the consistent commitment to minority rights, there is 
insufficient awareness of the laws on the subject, not only among the people directly affected, but 
also among some of those who are concerned with the administration of these laws. The 
­xiv­ 
absence of discrimination and the protection of minority groups are the tests of the political and 
moral achievement of the society. The creation of a just social order as envisaged in our Constitution 
(Article 38) requires a social consciousness and active cooperation of all the sections of the society. 
Professor Marc Galanter has written exhaustively of the history of the disadvantages and the revolt 
against these disadvantages. In his account, he has mentioned all the cases decided in the different 
High Courts and the Supreme Court which provide valuable help in understanding the objectives 
embodied in our Constitution in the light of our commitments to democracy and socialism. 
This book will be of immense use not only to the members of the legal profession but also to those 
who have at heart the achievement of national integration. It will contribute to the shaping of public 
opinion as well as to the enhancement of discussion of these matters by professionals and policy­
makers. It comes none too soon, as a great debate is going on, on the various aspects of the 
reservation policy. 
Vice President's Residence M. Hidayatullah 
New Delhi B.A. (Nag.), M.A. (Cantab.), LL.D. 
16 March 1981 Bencher, Lincoln's Inn 
­xv­ 
Preface 

IN 1957, NOT LONG after graduating from law school, I went forth on a Fulbright Scholarship to 
study India's efforts to abolish untouchability. Ready to see these efforts through my American 
"civil rights" lenses, I was struck by the fact that measures to enforce civil rights of a sort familiar to 
me were overshadowed by programs, of preferential treatment. The principle of systematic 
preference in favor of historically deprived groups was incorporated into a constitutional order with 
a strong emphasis on evenhanded application of law. 
The notion that commitment to equality might entail policies of preferential treatment was at that 
time hardly conceivable in the American setting. A few years later, it was with a sense of familiarity, 
that I first encountered advocacy of compensatory preference to enable American Blacks (then 
Negroes) to overcome their historic handicaps. Today such special treatment is a familiar feature of 
American life­although the legitimacy and efficacy of such programs is still highly controverted. 
India's long and elaborate experience with kindred policies may contain lessons for American 
policy. 
It was not the hope of extracting lessons for America that drew me to this topic, but curiosity about 
the course of India's pursuit of equality by group preference. Surely this concern reflected events in 
America, but I experienced it first as part of a challenge to understand India, then (as my 
perspective shifted) as an intriguing instance of using law to reshape recalcitrant social patterns. 
Many interruptions and digressions later, I find that this topic has over the years absorbed far more 
of my time and energy than I ever planned. This book has been long in the writing, and the 
discerning reader may find outcroppings of the understandings and perspectives that informed 
various stages of the work. I console myself that such disparities are the price of deepening 
awareness of the perplexities of the subject. 
The shape the book has taken over these years is not what I would project if I were to start now with 
my present views and interests. Were I starting now, I would undoubtedly pay more attention to 
politics and administration and somewhat less to the courts. As a lawyer, courts 
­xvii­ 
were my entrèe into Indian society. The use of courts as an observation point introduces a special 
bias, but the cases give a texture and immediacy to the account of policy. I have tried to deal with 
legal materials in sufficient depth to preserve the detail and complexity that make them meaningful 
to the immediate participants (Indians and mostly, lawyers, judges, and government officials) while 
making them accessible to those at a distance. Technical legal questions, I submit, should not be 
dismissed as "merely" technical. It is in a host of detailed questions that the tensions among 
conflicting commitments must be resolved­­and this includes all those commitments which are 
crystallized into the styles of the participants. At the same time that we see large principles 
illuminated by their concrete applications, we can observe the interplay of institutional routines and 
inventiveness that gives the law its kaleidoscopic sense of simultaneous rigidity and fluidity. 
The courts proved not only a convenient vantage, but interesting to observe in themselves, giving 
this book what became its second subject. Focussing on the intersection between courts and the 
compensatory discrimination policy, I use the courts as a way of learning about the policy, and the 
intersection as one way of learning about the function of courts in Indian society. The courts turned 
out to be more limited as a vantage point than I naively assumed at the outset. They act as a balance 
wheel channelling the compensatory policies and accommodating them to other commitments; but 
it is the political process that shapes the larger contours of these policies and gives them their 
motive force. Official doctrine­­judicial pronouncement or administrative regulation­­proved an 
insufficient guide to the shape of the policies in action and the results they produced. For a time I 
entertained the hope of including a systematic Portrayal and appraisal of the working of each of the 
various programs, but it belatedly dawned on me that such an enormous undertaking would 
postpone­­perhaps permanently­­the completion of this study. So I have contented myself with 
sketching the lines along which an appraisal loyal to the complexities of these Multigoaled policies 
might proceed. I have throughout tried to identify the points at which legal arguments depend on 
empirical assertions, to tease out what evidence I could find relevant to such assertions, and to 
indicate the yawning gaps that would have to be filled to arrive at a definitive answer to many of 
these questions. 
Compensatory discrimination policies are not only an instrumental mechanism but a medium 
through which the Indians who devise, apply, evade, and attack them express their understanding of 
their own society. This book, then, may be read as a portrayal of some contemporary Indian 
discourse about the tensions of hierarchy and equality, group and individual, policy and social 
reality. 
­xviii­ 
To this discourse I, as an outsider, have a curious and ambivalent relation. I have been intermittently 
immersed in these questions, my sympathies engaged and my intellect provoked. By writing about 
these policies from time to time, I have become willy­nilly a participant in this discourse­­a queer 
sort of participant who shares neither an insider's understanding of Indian life nor the necessity of 
living with the consequences of my observations. Nevertheless, Indian colleagues have been 
generously supportive and receptive over the years, inducing the conceit that my relative detachment 
(and my capacity to stumble over what is obvious to the insider) may produce a contribution that 
usefully complements those of insiders. I do not pretend to a total and aloof neutrality. I do not 
refrain here from drawing conclusions and proposing interpretations of the Indian Constitution. 
These are advanced in the spirit of friendly argument. 
There are no conventional categories for organizing the discussion of this sector of Indian law. Of 
the alternatives that suggested themselves, I preferred to organize the chapters around logical 
clusters of related problems, rather than around separate constitutional provisions, specific programs 
of preferential treatment, or specific problems of administration. Although many sections are 
organized around a chronological movement, I have made no attempt to adhere to chronological 
sequence overall. Instead, there are successive cuts into the same deposits from different points on 
the surface. Since the various topics are closely interrelated, I was faced with a choice between 
repetition or copious cross­reference. I have preferred the latter except where the repetition was 
small or the matter merited emphasis. 
This is a dynamic field which became a focus of political controversy in the late 1970s. There are 
bound to be new policy developments and judicial responses before this is printed; some of the 
things reported here are bound to be dated or proved wrong in a short time. I hope, though, that the 
delineation of major problems and the building of conceptual tools for approaching them will 
outlast some of the more detailed conclusions. 
In the main I have attempted to trace legal developments through materials available to me through 
1978, supplemented by more selective resort to new materials through mid­1980. Although the 
coverage of the last few years is less comprehensive than for earlier years, I believe the book 
pictures the major contours of policy as they stood at the replacement of the Janata Government by 
a resurgent Mrs. Gandhi in early 1980. 
Between the completion of the manuscript and the checking of the proofs in mid­1983, controversy 
about compensatory discrimination took on a new intensity. The violent and protracted protests 
against 
­xix­ 
reservations in medical colleges that racked Gujarat in early 1981 gave sanguinary expression to the 
widespread resentment of these policies and their beneficiaries. As conflict has mounted, the 
fundamental­­if not wholly consistent­­lines of policy have remained unchanged. The courts have 
elaborated some lines of doctrine in the earlier caselaw. The second Backward Classes Commission 
(Mandal Commission) reported ( Backward Classes Commission, 1981), proposing a major 
extension of preferential treatment to a vastly enlarged group of beneficiaries and reviving the varna 
classifications dismissed by its predecessor a quarter century earlier. The report was tabled in 
Parliament in April, 1982, and subsequently received parliamentary endorsement, but has not yet 
been implemented. The scope and intensity of the compensatory policy is very much a live and 
open question. The Mandal Report revives for replay on the larger national stage all of the questions 
and perplexities traced in this book. That stage, too, has undergone change. There has been a 
significant enlargement of the judicial repertoire by the addition of a more proactive and innovative 
style associated with the rubric of "public interest litigation." There have emerged new advocacy 
groups with aspirations and capacities to mobilize this kind of judicial response on behalf of 
constituencies of the dispossessed ( Galanter, 1984). Until now, none of this "public interest law" 
has been directly concerned with compensatory discrimination. The coming major round of 
adjudication may involve courts and litigants equipped with enhanced capabilities to address the 
fundamental questions of policy implicit in the commitment to the compensatory principle. 
­xx­ 
Acknowledgments 

THIS BOOK HAS gone through several cycles of growth. In each it has benefitted from many 
stimuli and encouragements. The United States Educational Foundation in India administered the 
Fulbright Scholarship which gave me my first taste of Indian law and life. A year working at 
Stanford Law School provided the occasion to immerse myself in Indian constitutional law. A 
decade of participation in the vigorous South Asianist community at the University of Chicago 
enlarged my appreciation for the complexities of India. A Senior Fellowship ( 1965­66) and a travel 
grant ( 1978) from the American Institute of Indian Studies enabled me to explore preferential 
policies in the field. 
The Indian Law Institute provided a congenial base during my 1965­66 fieldwork. From time to 
time, opportunities for helpful exchange were afforded by a number of institutions, including 
notably the Faculty of Law and the Department of Sociology of the University of Delhi, and the 
Indian Council of Social Science Research. My studies of this cluster of problems received valuable 
encouragement from the Joint Committee on South Asia (of the American Council of Learned 
Societies and the Social Science Research Council) and from the Institute for Research on Poverty 
of the University of Wisconsin. 
Between the inception of this project and its completion there emerged a vigorous scholarly 
community devoted to the study of law in its social context. I have enjoyed the privilege of working 
at two foci of this community­­at Buffalo and Madison. My involvement in this "law and society" 
community both delayed and enriched this book. 
At various times the International Legal Studies program of Stanford University, the Committee on 
Southern Asian Studies, the College, and the Law School of the' University of Chicago, the 
Research Committee of the Graduate School of the University of Wisconsin and the Ford 
Foundation have provided support enabling this work to reach fruition. 
Over the years a number of colleagues were kind enough to read various parts of the manuscript. 
Their comments have enriched the book and have helped eliminate some of its flaws (not without 
making 
­xxi­ 
me aware of others not so readily eradicable). For this I am grateful to Phyllis Rolnick Arora, 
Upendra Baxi, Boris I. Bittker, Rajeev Dhavan, Clarence Dias, George H. Gadbois, Jr., John 
Mansfield, S. P. Sathe, R. A. Schermerhorn, and especially to M. G. Narasimha Swamy. 
I have been the beneficiary of many forms of kindness and fellowship from colleagues who helped 
me to locate materials, shared their work in progress, and gave generously of their time and 
intelligence to educate me. Such a list is necessarily incomplete, but I should like to take this 
occasion to express my thanks to Partap Aggarwal, Vimal Chandra, Suma Chitnis, P. A. Choudhary, 
Bhagwan Das, J. D. M. Derrett, V. V. Deshpande, Lelah Dushkin, Joseph Elder, Marcus Franda, 
Henry C. Hart, L. G. Havanur, B. R. L. Iyengar, S.K. Kaul, P. R. Mehendirotta, N. R. Madhava 
Menon, N. G. Nag, N. P. Rege, Vimal Shah, G. S. Sharma, Parmanand Singh, M. N. Srinivas, 
Gregory Stanton, Richard Taub, and Eleanor Zelliott. The late Professor Max Rheinstein provided 
guidance and encouragement as well as his towering example in the comparative study of law. 
I was particularly fortunate in having access to the South Asian collections of the University of 
Chicago and the University of Wisconsin, and the assistance and support of Maureen L. P. Patterson 
and Jack Wells, who preside over those collections. 
At various times in the life of this project invaluable help in preparing the manuscript was provided 
by Betti Arnett Johnson, Doreen Blank, Cynthia Burns, Judith Frank, Eve Galanter, and Mary L. 
Galanter. Marge Eiseman did the final round with great skill and verve. Seth Galanter 
conscientiously numbered the manuscript sheets. Janet Guthrie, Catherine S. Meschievitz, and 
Hugh A. Ross ably provided crucial assistance. Robert Hayden was a source of encouragement and 
helpful criticism as he worked with me in reviving and revising the manuscript. In the arduous and 
seemingly endless task of reading the proofs and preparing the Index, I was fortunate to have the 
capable help of Susan Bissegger, Michael Macy, and Sylvan Sobel. 
The title of the concluding chapter is borrowed from Upendra Baxi's 1967 article. 
I want to thank Jean Lettofsky for the Hebrew calligraphy. 
My wife, Eve, and my children, Seth, Rachel and Sarah, graciously abided the protracted intrusion 
of this incubus. 
The views set forth here­­even when they contradict one another­are my own and should not be 
attributed to my benefactors. 
­xxii­ 
Note on Citations and References 

LAW REPORT CITATIONS of the judgments of courts are given in standard Indian legal form. 
Thus citations of official reports give the names of the first party on each side, the notation I.L.R. 
(indicating Indian Law Reports), the volume number, the title of the series (indicating the court), the 
page on which the case begins, and, finally the date of publication; for example Gopal v. Hanmant, 
I.L.R. 3 Bom. 273 (1879). Citations to the All­India Reporter, the most popular series of unofficial 
reports, follow a slightly different pattern; for example, Sunder Devi v. Jheboo Lal, A.I.R. 1957 All. 
215 (names, A.I.R. series, year of publication, name of court, page on which case begins). There is a 
separate A.I.R. volume published each year for each court. The coverage of A.I.R. and I.L.R. is 
overlapping but not identical. Occasionally, cases not found in either of these series are reported in a 
local or specialized series. The citation of these follows a similar pattern. 
Articles, books, monographs, etc., cited by author and year in the notes, are listed in the 
bibliography. A similar style is used for government documents (except Census of India 
publications, which are cited in standard style in the footnotes). 
Pre­Independence government documents are cited by the name of the committee, commission, or 
other government body from which they issued (identified in some cases by the name of the 
chairman). PostIndependence government documents, reports of commissions, study teams, 
committees, and other non­serial documents are cited by the name of the body (identified in some 
cases by the name of the chairman) and the year of publication. In some cases where such a body 
operated under the auspices of a Ministry, the name of the body has been added to the author listing. 
Other government documents are cited by the Ministry or other government body under whose 
auspices they are issued and the year of publication. Where a series is listed in the references under 
a specific agency, a citation of that body and a specific year means the report for that year. 
Government publications with named authors are cited by author. Some reports and series that are 
cited frequently are indicated by abbreviations found in the list below. 
­xxiii­ 
Note on State Names 

IN ADDITION TO the three presidencies ( Bengal, Bombay, and Madras) and the other provinces 
of' British India, there were hundreds of princely or native states which recognized the paramountcy 
of the British Crown. Most were small, but a few were the magnitude of provinces (e.g., Mysore, 
Hyderabad). When, at Independence, British India was partitioned into the new nations of India and 
Pakistan, various princely states joined the former provinces as units of the Indian federation, the 
smaller states being combined for this purpose. The resulting array of states included units quite 
disparate in scale. 
A surge of linguistic sentiment led to the separation of Andhra from Madras in 1953. In 1956 there 
was a general reorganization of states; boundaries were redrawn largely on linguistic lines, and the 
vestiges of the princely states were largely effaced. For example, old ("Princely") Mysore was 
combined with the Kannada­speaking parts of Madras, Bombay, and Hyderabad to produce a new 
and large Mysore; the composite (former Princely) State of Travencore­Cochin was combined with 
the Malabar District of Madras to form the single Malayalamspeaking State of Kerala. The 1956 
changes produced the basic map of India as it is today. Subsequently, two remaining bilingual states 
bifurcated: Bombay divided into Maharashtra and Gujarat in 1960; Punjab divided into Punjab and 
Haryana in 1966. A series of small tribal states have since been established in the northeast. 
Two name changes should be kept in mind: Madras changed its name to Tamil Nadu in 1969; 
Mysore became Karnataka in 1973. 
­xxiv­ 
List of Abbreviations 
A.I.R. 
All­India Reporter 
BC 
Backward Classes. 
BCC 
Report of the Backward Classes Commission ( 1955). 
CAD 
Constituent Assembly Debates ( 1947­50). 
CWSCST 
Reports of the Parliamentary Committee on the Welfare of the
Scheduled Castes and Scheduled Tribes. The Reports are de­
signated by the number of the Lok Sabha, the number of the
report (page no. if any after colon), date and the year issued. 
E.g.,
CWSCST (4th) 18: 27 ( 1970). 
G.O. 
Government Order 
LSD 
Lok Sabha Debates. Materials from these reports is cited by
series, volume, number, column, and date. E.g., LSD (4th 
Series)
Vol. 86, No. 3, Cols. 142­43 (12 June 1963). 
[NIC] 
Cases so noted are not included in the computations in chapter 
14
which form the basis of aggregate statements throughout the 
text. 
OBC 
Other Backward Classes. 
RCSCST 
Reports of the Commissioner for Scheduled Castes and
Scheduled Tribes. These reports are designated by the period
covered by the report, rather than by the report number or by
date of publication. E.g., RCSCST 1957­58: I, 35. 
RSD 
Rajya Sabha Debates. Material from these reports is cited vol­
ume, number, column, and date. E.g., RSD Vol. 102, No. 6, 
Cols.
118­19 (14 Aug. 1970). 
SASTC 
Reports of the Scheduled Areas and Scheduled Tribes 
Commis­
sion ( 1961). 
SC 
Scheduled Caste(s). 
ST 
Scheduled Tribe(s). 
STTDP 
Report of the Study Team on Tribal Development Programs
( 1970). 
Lakh 
A unit of one hundred thousand (100,000). 
Crore 
A unit of ten million (10,000,000). 

­xxv­ 
Introduction 

INDIA'S SYSTEM OF preferential treatment for historically disadvantaged sections of the 
population is unprecedented in scope and extent. India embraced equality as a cardinal value against 
a background of elaborate, valued, and clearly perceived inequalities. Her constitutional policies to 
offset these proceeded from an awareness of the entrenched and cumulative nature of group 
inequalities. The result has been an array of programs that I call, collectively, a policy of 
compensatory discrimination. If one reflects on the propensity of nations to neglect the claims of 
those at the bottom, I think it is fair to say that this policy of compensatory discrimination has been 
pursued with remarkable persistence and generosity (if not always with vigor and effectiveness) for 
the past thirty years. 
These compensatory discrimination policies entail systematic departures from norms of equality 
(such as merit, evenhandedness, and indifference to ascriptive characteristics). These departures are 
justified in several ways: First, preferential treatment may be viewed as needed assurance of 
personal fairness, a guarantee against the persistence of discrimination in subtle and indirect forms. 
Second, such policies are justified in terms of beneficial results that they will presumably promote: 
integration, use of neglected talent, more equitable distribution, etc. With these two ­­ the anti­
discrimination theme and the general welfare theme ­­ is entwined a notion of historical restitution 
or reparation to offset the systematic and cumulative deprivations suffered by lower castes in the 
past. These multiple justifications point to the complexities of pursuing such a policy and of 
assessing its performance. 
Independent India firmly and explicitly embraced the compensatory discrimination principle. This 
book is about the way in which this commitment has been incorporated into a regime of 
constitutionally guraranteed rights, including notably rights to formal equality, protected by courts 
equipped with broad powers of judicial review. It is an account of both institutional and intellectual 
accommodation, of government practice, and of legal doctrine. Among the newer (and almost 
­1­ 
all older) nations, India is distinctive in the major policy role played by an independent judiciary. 
Indian courts have been called upon both to resolve disputes among contending claimants and to 
provide a way of harmonizing these competing principles. The focus of this study is on the response 
of the courts to these unprecedented demands. 
This study is divided into three parts. Part One sketches the social setting and historical background 
of the compensatory discrimination policy, describes the major programs and their administration, 
and opens the question of assessing their performance. Part Two takes up questions of who: who are 
those people whose entitlement to equality requires departures from formal equality? Identifying the 
beneficiaries turns out to be a complicated and difficult process involving many intellectual 
perplexities. Once there is a commitment to these groups, we come to the cluster of questions 
addressed by Part Three: what is the scope of the commitment? what means are allowable? how 
much is too much? and what institutional arrangements shall be established to strike the balance? 
A Note on Terminology: Why "Compensatory Discrimination"? 
It is not obvious what these policies should be called, for names are not neutral. (Consider, for 
example, the contrast between "affirmative action" and "reverse discrimination" in the American 
setting.) Indian discussion usually refers to "special treatment," "concessions," "privileges," "special 
provision," or "preferential treatment" for named groups. Scholars have inclined to use "protective 
discrimination" (which I first encountered in Alexandrowicz' 1957 book.     1. More recently, scholarly 

writing has adopted the term "compensatory discrimination."     2. 

Common usages like "special treatment" are accurate but omit any reference to the principles that 
animate the policies and distinguish them from other distributive schemes. "Protective 
discrimination" has a paternalistic and static quality; it suggests a one­way flow of beneficence and 
lacks any implication that it is a policy of change. Thus, of the many unsatisfactory candidates, I 
prefer "compensatory discrimination" for the following reasons. By using "discrimination" it does 
not blink at the fact that some are left out, that we are dealing with something more than a benign 
process of inclusion. At least where 
____________________ 
1. Alexandrowicz 1957. Among candidates that have failed to gain currency are "progressive 
discrimination" ( Rudolph and Rudolph 1967: 54n. ­­ on analogy to the progressive income tax) 
and "preferential opportunity" (M. D. Nandajundaswamy in Government of Karnataka 1975, Vol. 
1, Pt. II, p. 5) 
2. Parmanand Singh 1976. 
­2­ 
scarce resources are distributed, it employs a principle of selection that is akin to the old 
discrimination. But the purpose is different: it is not exclusion and relegation but inclusion and 
recompense both for historic deprivations and to offset present handicaps. And it carries an 
implication that discrimination will cease when compensatory treatment has remedied these 
conditions. 
The drawback of the term ­­ apart from its being ten syllables ­­ is that it draws attention away from 
the non­discrimination and general welfare themes in these policies. If this name would be 
inappropriate in a setting where the element of historical compensation is less prominent,    3. it does 
seem to fit the Indian situation better than the other candidates. To minimize repetition, I have freely 
resorted to a variety of synonyms and abbreviated forms, such as "preference policy" or simply 
"preference." 
Discussions of preferences often founder on the problem of the somewhat confusing names of the 
groups entitled to receive them. In common Indian usage, the generic term for the groups entitled to 
receive preferences is the "backward classes." ("Weaker sections" is a rough equivalent for this 
usage.) This group in turn may be broken down into the Scheduled Castes ("untouchables"), the 
Scheduled Tribes, and a diverse and less well defined residuum known as the "Other Backward 
Classes" or sometimes the "backward classes." To minimize confusion I have attempted to indicate 
which usage of "backward classes" is meant by omitting capitalization when using the term in its 
broad and inclusive sense and capitalizing it when it is used in the narrower sense. Thus "backward 
classes" = Scheduled Castes + Scheduled Tribes + Other Backward Classes; "Backward Classes" = 
Other Backward Classes. 
____________________ 
3. E.g., to describe "preferential" programs for "Oriental Jews" in Israel, where the programs are 
justified in terms of integrating groups with little previous contact; the more advanced are not 
deemed responsible for the deprivations of the less fortunate groups. 
­3­ 
Part One COMPENSATORY DISCRIMINATION: THE SETTING AND THE POLICIES 

This is a book about a set of policies and about the role of the Indian judiciary in developing, 
elaborating, criticizing, effectuating, and legitimizing them. This Part begins with a sketch of the 
social setting that frames the policy of compensatory discrimination and is the ultimate object of it. 
Chapter 2 traces the political developments that link this setting to India's modern constitutional 
order. These broad introductory chapters are intended to make the subject accessible to those who 
are not specialists on modern Indian society and politics, as well as to introduce themes that recur in 
later detailed analysis of these policies. Chapter 3 provides a brief overview of the compensatory 
programs and their administration. Chapter 4 examines in more depth the most prominent of these 
programs (reservation of government jobs). Chapters 3 and 4 provide a sense of the scope, design, 
and implementation of these programs and introduce the problems of assessing their performance 
and of estimating their wider effects. 
­5­ 
1 The Compartmental Society 

INDIAN SOCIETY has been described as a "compartmental" society; within it a vast number of 
groups maintain distinct and diverse styles of life. The system by which these groups are related and 
mutually accommodated is so complex as to defy general description. However, to place in their 
setting the legal developments we shall discuss, it is necessary to sketch some of the principal 
features of India's social order and its recent history. 
1. Caste in 
One of the distinctive and pervasive features of Indian society is the division into caste.    
the narrower sense applies primarily to the Hindus, who make up 85% of India's population. Many 
features of caste are also found among non­Hindu groups, and the term "caste" as used here applies 
to these groups as well as to Hindus. There are exceptions to almost any statement that can be made 
about caste, but only the most prominent exceptions will be noted here. The term "caste" itself has a 
variety of meanings. It takes on different shadings in the context of village, locality, region, and 
nation.  2.   It is in effect a set of 
____________________ 
1. The word "caste" is not an indigenous Indian term but a graft via English (from the Portuguese 
casta). It is used to correspond to several Indian terms, but has no exact equivalent in Indian 
languages. Serious students of Indian society regard it as suspect on grounds of both its 
foreignness ( Marriott and Inden 1974:983) and its ambiguity. Thus Mandelbaum ( 1970: 29) 
says: "It [the term 'caste'] has been pinned to too many social entities ­­ endogamous group, a 
category of such groups, a system of social organisation. It seems best to use the term, apart from 
direct quotation, as an adjective that refers to the highly stratified social system." But in the 
inquiry that we are embarking on there is no way to avoid it, since caste is a key element of 
contemporary public discourse about the past, present, and future shape of Indian society. 
Awareness of the ambiguities and multiple referents of the caste notion will illuminate that 
discourse. Its theoretical insufficiencies do not allow us to dispense with it if we are to know how 
Indians (especially those educated in English) see their own society ­­ any more than we could 
eliminate equality from a discussion of contemporary American society on grounds of multiple 
and contradictory meanings attached to it by the natives. 
2. And, of course, it has special meanings in the setting of social scientific discourse where it is 
generally used to describe a system of social organization in which society comprises ranked 
hereditary kinship groups associated with a division of labor and organized into a unified and 
integrated whole. The literature is vast. Whether the caste notion is usefully employed outside the 
South Asian setting is a matter of controversy. 
­7­ 
3. To start with, a caste may be taken to mean 
shifting idioms for talking about affiliation and status.    
a jati  4.   ­­ an endogamous group bearing a common name and claiming a common origin, 
membership in which is hereditary, linked to one or more traditional occupations, imposing on its 
members certain obligations and restrictions in matters of social intercourse, and having a more or 
5. It has been estimated that there are 2,000 
less determinate position in a hierarchical scale of ranks.    
or 3,000 such castes (or subcastes, as they are sometimes called) in present­day India.     6. The 
population of these endogamous caste groups may range from hundreds to millions. An estimate 
places the median size of such groups at between 5,000 and 15,000.     7. 

Membership in the caste group is ordinarily conferred by birth and is unalterable, although it may 
be lost by expulsion from the group. Members normally marry within the group; most social 
intimacy occurs within it. Ordinarily, one does not dine with those who are outside the caste. 
Members of the caste follow a common round of life. They have a characteristic diet, domestic 
routine, style of dress, and round of religious observance. The caste may have distinct forms of 
religious expression: a preference for certain gods, shrines, and modes of worship, its own fasts and 
feasts. Some cultural characteristics may be shared by all or most castes over a wide geographical 
area, but others may be distinctive to one or a few castes. 
Although most castes have a traditional occupation, in only a few ­especially the skilled artisans ­­ is 
it followed by most members today. Agriculture is in effect open to all groups and is engaged in 
widely by persons belonging to all groups. In many cases, caste tradition precludes entry into 
certain occupations, while the network of caste ties provides access to others. 
The caste is held together by ties of kinship, by a cycle of group observances, by bonds of mutual 
assistance and support, and finally by the power of the group to exact obedience to its rules. Most 
castes formerly had (and many today have) councils, or panchayats, formal or informal, which 
adjudicate disputes among caste member's and enforce 
____________________ 
3. The multiplicity of usages not only is apparent to the observer but is internalized in the users 
who can shift categories (e.g., from gross to fine or local to regional) as the occasion requires. 
Cf. Mandelbaum 1970: 22. 
4. Jati is a term of wide application meaning kind or genus. Jatis are seen as natural units of 
society, each manifesting a specific kind of human possibility. 
5. The notion that hierarchic ordering is the central organizing principle of Indian civilization is 
expressed in the title of Dumont Homo Hierarchicus ( 1970). Kindred notions are shared by 
scholars who hold otherwise "disparate views of Indian society­­e.g., Mandelbaum 1970 and 
Marriott 1955. 
6. Ghurye 1969: 27 (2,000); Hutton 1961: 2 (3,000). The latter figure is frequently cited. 
7. Marriott and Inden 1974: 983. 
­8­ 
the received etiquette of relations with other castes. (For example, they take action against those 
who offend against taboos on eating, drinking, smoking, or sexual intercouse with members of other 
castes.) Their decisions, backed by the ultimate sanction of excommunication, may involve penalties 
ranging from fines and expiations to one or another degree of boycott, temporary or permanent. 8. 
While formal caste sanctions have declined in recent years, especially among the educated, caste 
opinion and caste loyalty remain powerful cohesive and regulatory forces. 
Our compartmental image requires correction, for castes are not self­sufficient but interdependent; 
caste unites while it divides. Although there is a greater admixture of conflict and competition than 
the harmonious articulation of functions depicted in idealizations of caste, their occupational and 
ritual specialization does make caste groups interdependent in practice. Caste groups live not in 
isolation but in the midst of a set of other groups with which they have fixed and customary 
relationships ­­ economic, political, social, and religious. In much of India the village was organized 
into a network of hereditary patronclient relations known as the jajmani system.    9. The various 
nonlandholding castes (priests, washermen, carpenters, field laborers) provided customary services 
to landowners in return for fixed portions of agricultural produce. The relationship was not a mere 
economic exchange, but implied general duties of support and prerogatives of control. The 
relationships between castes are hierarchic; transactions among groups express principles of 
precedence, obedience, and subservience. Each group has its duties and disabilities; each, even the 
lowest, has its special prerogatives and privileges. 
The number of caste groups in a given locality may vary from a few to more than thirty ­­ from five 
to twenty would be typical.  10.   There are no all­India castes and there is no nationwide hierarchy of 
castes. In 
____________________ 
10. Mandelbaum 1970: 15. 
8. On the mechanism of control within castes, see Mandelbaum 1970, chaps. 16, 17. Although 
castes have characteristically enjoyed a wide autonomy in making and applying rules for conduct 
of their own social life, the ruler in classical Hindu theory (and intermittently in practice until 
recent times) exercised a supervisory control over changes in custom and over matters of 
expulsion from and re­admittance to caste. On medieval practice, see Derrett 1968: 171­224. The 
requirement of royal approval of re­admittance, retained by the Moghul government and in many 
princely states until this century, was relinquished by the British in the late 18th century 
( O'Malley 1932: 57­61; Hutton 1961 93­97). On British control over the internal affairs of castes, 
see Galanter 1968 and 1972. 
9. On the Jajmani system, see Mandelbaum 1970, chap. 9; Wiser 1958; Beidelman 1959. The term 
jajmani was introduced into the scholarly literature by Wiser. Many other terms are current to 
describe regional varieties of the system ­­ e.g., the baluta, or shares from village lands given 
hereditary village servants in Maharashtra. 
­9­ 
any local area, the relative standing of the various castes is more or less fixed, though subject to 
slow change, particularly in the middle ranges of the scale. Castes commonly lay claim to higher 
rank. Thus Marriott and Inden refer to the "mild but ubiquitous inclination of castes to improve their 
ranks [that] can be estimated from the finding that the average villager believes the correct rank of 
his caste to be about half a rank higher than what non­members of this caste are willing to concede." 
 11.   
There is local variation both in the number and kind of castes and in the rigidity with which the 
distinctions between them are maintained. Closely related castes may vary in standing from one 
region to another, or even within the same area. The top and bottom of the scale tend to be more 
easily identifiable and generally, if not invariably, accepted. Brahmins are at the top, and the 
"untouchables" are at the bottom, although there may be disagreement about the relative rank of 
various Brahmin and untouchable castes. 
Highness and lowness in the local hierarchy correspond roughly to the ritual purity of the several 
castes. However, a caste's standing also expresses its political and economic power. To the extent 
that these fail to correspond, there will be ambiguity in the ranking of castes, and changing 
economic and political power may be followed by an improvement in ritual standing. Although there 
are local differences, highness and lowness are expressed in a variety of interactions. Standing is 
commonly manifested in the willingness with which food and water are accepted by other castes, by 
reluctance to accept food and water from other castes, in the giving and receiving of ceremonial 
gifts, and in reciprocal patterns of precedence and deference. 
Relative positions in the local caste hierarchy are often conceptualized in terms of the theoretical 
division of Hindu society into four classes, or varnas. According to this theory, society is divided 
into Brahmins, the priestly and scholarly class; Kshatriyas, rulers and soldiers; Vaishyas, merchants 
and agriculturists; and Sudras, the menial and servant class.  12.   The power of the varna theory is 
shown by the common tendency to describe the caste system as made up of four "castes" which the 
endogamous caste groups are merely subdivisions.  13.   (Hence the usage "sub­caste".) Groups 
seeking to better their position in the local hierarchy often phrase their claims in terms of a more 
prestigious varna identification by adopting names and practices associated with one of the three 
higher varnas.  14.   Sudras are clearly 
____________________ 
11. Marriott and Inden 1974: 985. 
12. There has not always been agreement about the relative standing of the three higher varnas. 
13. The classical theory attributes multiplicity to miscegenous mixing of castes. 
14. Except for Brahmins, the historical association of caste groups with particular varnas 
­10­ 
accounted as inferior to the three "twice­born" varnas, but it is misleading to think of the latter as 
ranked on a single ascending scale of precedence. Instead, they represent a set of divergent and 
incompatible social ideals around which claims for standing may be organized.  15.   A caste group 
may cultivate its identification with one or the other, but does not progress from one to the other. 
Nor are all groups belonging to the same varna rated as equals. 
Between the local caste (i.e., jati) hierarchies and the pan­India varna ranking lie a bewildering and 
irregular set of regional and subregional groupings of castes and conceptualizations of the caste 
hierarchy.  16.   There are "caste clusters," "composed of several endogamous groups," spread over a 
region or sub­region, sharing "a common name and functions . . . often [with] stylized interaction in 
commensality, hypergamy and differential status between their constituent jati groups."  17.   There are 
what Fox calls "sub­regional varna categories"­­i.e., categories embracing several otherwise 
divergent jatis, or clusters, who share a claim to a common status. And there are regional versions of 
comprehensive varna schemes ­­ e.g., the "tripartite division into Brahmin, non­Brahmin and 
untouchables" in much of South India.  18.   
Local, regional, and pan­Indian schemes all draw on a common ideological fund. The notion of 
graded inequality is explained and justified by the traditional Hindu notions of dharma and karma. 
Each caste group has its own dharma ­­ the path which each of its members should follow in 
accordance with his nature and his station in life. The inequality of stations and the more onerous 
duties incumbent on some is explained and justified by the theory of karma and rebirth ­­ that is, 
every human action has a positive or negative worth, and the moral balance of an individual's 
actions in previous lives is manifested in the station into which he is reborn. Inequality is thus 
explained and justified as the reflection of relative spiritual merit accumulated in past existences. 
Hope for the future is provided, for one may progress through successive rebirths to higher positions 
and to the ultimate equality of absorption or reunion in Brahman, the unindividuated cosmic reality. 
Such spiritual betterment is the result of doing properly the things traditionally done by persons of 
one's station ­­ by fulfilling the dharma determined by one's caste, stage of life, and family position. 
Of course, this 
____________________ 
is often doubtful. In many places there are few representatives of the Kshatriya and Vaishya 
varnas, while in others Kshatriyahood has tended to be conferred on those who hold temporal 
power. Since varnas are taxonomic categories rather than functional groups, attempts to 
enumerate their members have been unavailing. The 1931 Census did account as Brahmins 4.5% 
of the total (and 6.4% of the Hindu) population ( Schwartzberg 1978: 106) 
15. Marriott 1968: 110. 
16. Based on Fox 1969 and sources cited there. 
17. Fox 1969: 32. 
18. Id., at 33. 
­11­ 
ideology is not fully accepted by all.  19.   Nor does acceptance of these general notions imply 
agreement that one's caste has been accorded its rightful station.  20.   So the ideology of hierarchy 
justifies claims of superiority and rationalizes resignation to inferior status; but it also provides 
leverage for (and incitement to) claims to higher status. 
In spite of the idealization of stability and fixity of social position, the compartmental society has 
not been static: caste groups do change their relative positions. Caste as a principle of social order 
has persisted over millennia, if much more disorderly and asymmetrical in practice than classical 
Hindu socio­legal theory depicted it.  21.   Something very like the compartmental society of the 
recent past existed a thousand years ago.  22.   Political, religious, and economic changes have all 
affected the caste order. The principle of a compartmental society endured while the identity of the 
compartments underwent many changes. Conquest, migration, emulation, isolation and segregation, 
occupational specialization, conversion and sectarianism, the incorporation of tribal groups ­­ all led 
to the additon, fission, and fusion of castes and to changes in their relative standing. Although it 
stressed ideals of stability and fixity, the caste society did have mechanisms for incorporating new 
groups, accommodating changing practices, legitimating changes in group standing to accord with 
changing political economic and socioritual attainments. The compartmental society permitted 
some individual mobility within these communities as well as the slower movement of whole 
communities.  23.   But what endured was a pattern of 
____________________ 
19. Berreman 1963: 253­54. 
20. Berreman 1963: 223­24. Berreman observes the dilemma of the low caste person whose religion 
enjoins acquiescence low place but whose impulse is to improve that position: "In Sirkanda and, 
I believe, throughout India, this dilemma is often neatly resolved by exactly the mechanisms 
described above: accepting caste dharma but denying that one's apparent dharma is his real 
dharma. That is, the individual takes the position that a person should behave according to his 
caste status but that he or his kin group or jati is not really of the caste or status to which others 
ascribe him (or them). He is of a higher caste. He aspires to live as a member of his 'true' caste in 
order to fulfill his true Dharma. Therefore his mobility aspirations are legitimized and even made 
imperative. He rationalizes his own mobility aspirations without challenging the system." 
21. In classical theory the function of political authority was to hold the balance between the groups 
in the compartmental society. It was the duty of the ruler to uphold the caste order. The king was 
enjoined to make the four varnas fulfill their particular duties and to preserve society from the 
evil of "confusion of castes." The confirmation (and conferring?) of varna status and control over 
the order of precedence among castes by Indian rulers continued until the recent past ( Hutton 
1961: 49; Srinivas 1966: 41). The British eschewed such prescriptive assignments; but attempts to 
obtain governmental confirmation of status were channelled into the nominally descriptive 
arenas of the courts and the census ( Hutton 1961: 97; Srinivas 1966: 16; O'Malley 1932: 63). 
22. Karve 1961, chap. 2. Nothing like jatis are described in the most ancient classical literature. 
23. Silverberg 1967. 
­12­ 
graded inequality of corporate groups in which differential access to life chances corresponded in 
large measure to membership in different communities. 
At the bottom of the caste hierarchies are found the "untouchables" ­­ castes with the lowest ritual 
standing and usually the lowest economic position and traditionally subject to onerous social and 
civic disabilities. Discussion of untouchability propels us into problems of nomenclature which 
condense intense controversies over the identity of these groups and over the characterization of 
their relation to the bulk of the Hindu population.  24.   We shall trace the recent history of their 
identification in chapter 2 and take up some current controversies in chapter 6. For the moment it 
suffices to note that there is no other general term for the practice of setting apart a segment at a 
lower end of the caste order and that "untouchability" has been adopted as a salient category for 
public policy: the Constitution declares its abolition, and successive governments have promised to 
end it summarily. 
The origins of untouchability as practiced in recent times are unknown. Something like it has 
existed for at least a millennium,  25.   and many of its features were present more than a millennium 
earlier than that.  26.   Untouchability as such is not dealt with in classical Hindu learning. However, 
the low state of untouchables is reflected in and supported by the classical view of society as 
consisting of four varnas, or classes, with different natures, different functions, and correspondingly 
different stations in life. Rights and duties, procedures and penalties, were graded accordingly. The 
three higher varnas were "twice­born" by virtue of investiture with the sacred thread, while the 
Sudras were subject to numerous disabilities. They were restricted to certain occupations, forbidden 
to learn the veda, and enjoined to be deferential and subservient to their betters. 
In recent times, "untouchables" have been identified as "unclean Sudras" in some parts of India; 
elsewhere they are regarded as outside 
____________________ 
24. Current terminology includes the official term "Scheduled Castes"; also used officially is the 
Gandhian euphemism "Harijans" (lit., "people of god"), favored by politicians and resented as 
condescending by many so designated. "Dalits" (i.e., oppressed people) is favored by militants 
and eschewed by politicians and officials. "Backward classes" has a wider connotation (see chap. 
6 below), as does "weaker sections." "Depressed classes," no longer in vogue, was the 
predecessor official term, replaced by Scheduled Castes in the mid­1930s. "Untouchables" is 
regarded as callous or insulting. Isaacs' ( 1965) usage, "ex­untouchables," has not achieved 
currency. On origins of some of these terms, see notes 21 and 46 in chap. 2 below. 
25. Alberuni (c. A.D. 1020) noted certain endogamous occupational groups who lived outside 
villages and had limited intercourse with others ( Ghurye 1969: 316). He mentioned Dom and 
Chandala as "two groups not reckoned among any caste" ( Ghurye 1969: 313). A 12th­century 
grammarian reports that Chandalas warn people to avoid their touch (id.). 
26. R. S. Sharma 1958: 125. 
­13­ 
(beneath) the four varnas.  27.   Hence untouchables are often referred to as "outcastes," and there is a 
persistent tendency for foreign observers to describe them as people who are not members of any 
caste.  28.   But they are members of castes­­those which have the lowest ritual standing and often the 
most depressed economic condition. Generally, they are also lowest in income, health, education, 
and cultural resources. In the 1971 census those officially classed as untouchables (Scheduled 
Castes) by the government constituted nearly 80 million, or about 14.6% of India's total population 
of 548 million (or over 17% of India's 453 million Hindus). Untouchables are often associated with 
the most unclean and degrading occupations­­sweepers and scavengers, attendants at cremation 
grounds, hide and leather workers­­but include other artisan groups such as weavers and toddy­
tappers. In most regions a single large caste of field servants (like the Mahars in Maharashtra, the 
Paraiyans in Tamil Nadu, and the Chamars in much of North India) makes up a large portion of the 
total untouchable population, with the remainder divided among various untouchable artisan castes. 
 29.   The groups considered untouchables differ from locality to locality. A caste such as washermen 
or weavers may be considered untouchable in one part of India but not in another. Actually, only a 
small minority of today's untouchables practice these stigmatized occupations; most are landless 
agricultural laborers, often ensnared in debt. They are usually, but not invariably, associated with 
certain habits whch are unclean and loathsome to higher caste Hindus­­such as eating beef and other 
forbidden food and in some cases eating carrion. 
The relations between castes throughout the hierarchy are expressed in terms of relative social 
distance and exclusiveness corresponding to the necessity of avoiding pollution (i.e., ritual 
uncleanliness). Pollution might derive from association with death, menstruation, childbirth, impure 
food, etc. But while everyone is subject to and causes temporary and expiable pollution, 
"untouchables" are permanently and inexorably polluted and polluting. Pollution is transmitted by 
contact with their persons, by food, and in some places in the south, by proximity (and historically 
in few cases, sight) as well. Different untouchable castes might be polluting to different degrees and 
subject to disabilities to varying extents. They are not a unified group, but recognize such 
distinctions among themselves and sometimes impose similar disabilities on each other. 
____________________ 
27. R. S. Sharma ( 1958: 208, 261, 279) detects comparable ambivalence in the relation of 
untouchables to the Sudra varna two thousand years ago. 
28. "Outcaste" also means expelled from one's caste. Its use to describe untouchables flows from the 
identification of castes and varnas; i.e., not being included in a varna is equated with not having 
a caste. 
29. Patankar and Omvedt 1979: 410. 
­14­ 
Untouchable groups are set off more or less distinctly below all of the other caste groups.  30.   While 
each caste enjoys certain precedence and is subject to certain disabilities, the "untouchable" castes 
are subject to more extensive and severe disabilities and enjoy few privileges. 
Disabilities vary from place to place and in their incidence on specific castes. The imposition of 
disabilities is not unique to untouchables. It exists to some extent­­particularly in regard to religious 
facilities, commensality, and intermarriage­­throughout the system of castes. But it is found in a 
more exaggerated form for untouchables and to a lesser degree for other groups; some disabilities 
are peculiar to untouchables or more intense in regard to them. The following list of disabilities 
applied to untouchables is by no means a description of the practices in any single locality, nor is it 
an exhaustive listing of practices throughout India. It is instead a catalog of the disabilities which 
have in one place or another been typically associated with untouchability: 
Denial or restriction of access to public facilities, such as wells, schools, roads, post 
offices, and courts. 
Denial or restriction of access to temples where their presence might pollute the deity as 
well as the higher­caste worshippers, and from resthouses, tanks, and shrines connected 
to temples. Untouchables and Sudras were ineligible to become sanyasis (holy men) and 
forbidden to learn the Vedas (the earliest and most sacred books of orthodox Hinduism). 
Exclusion from any honorable, and most profitable, employment and relegation to dirty 
or menial occupations. 
Residential segregation, typically in a more extreme form than the segregation of other 
groups, by requiring them to remain outside the village. 
Denial of access to services such as those provided by barbers, laundrymen, restaurants, 
shops, and theaters or requiring the use of separate utensils and facilities within such 
places. 
Restrictions on style of life, especially in the use of goods indicating comfort or luxury. 
Riding on horseback, use of bicycles, umbrellas, footwear, the wearing of gold and 
silver ornaments, the use of palanquins to carry bridegrooms­­all of these were 
forbidden in many areas. 
Requirements of deference in forms of address, language, sitting and standing in 
presence of higher castes. 
Restrictions on movement. Untouchables might not be allowed on roads and streets 
within prescribed distance of the houses or persons of higher castes. 
Liability to unremunerated labor for the higher castes and to the performance of menial 
services for them. 
____________________ 
30. Although untouchables might readily be identified in a given locality, considerable difficulty was 
experienced in devising pan­Indian criteria to distinguish them. See chap. 5 below. 
­15­ 
Restrictions were severe and ubiquitous, but custom did not always prescribe complete exclusion. 
Instead, there was a complex pattern of exclusion, separation, and restricted use. For example, wells 
might be entirely closed or might be used with separate buckets or only through a higher caste 
intermediary. Schools might be entirely closed, or untouchable children might be required to sit 
either outside the classroom or inside but away from other children. Untouchables might be 
excluded from temples, or they might be permitted to enter but required to remain a specified 
distance from the idols. 
The opportunities for group mobility within the caste system were more or less closed for them. 31. 
Systematically degraded, subjected to paternalistic control where not harshly victimized, the 
untouchables endured a kind of perpetual slavery with little hope of betterment in this world. 
Caste is not the only basis of affiliation in this compartmental society. Sometimes enveloping and 
sometimes cutting across caste are many other sorts of groupings. Indian society was and is divided 
into a multitude of religions, each containing its own orders, denominations, sects, and movements. 
The great family or cluster of religious assemblages known as Hinduism (containing some 85% of 
the present population of India) includes numerous "unorthodox" and "reform" sects as well as 
various "orthodox" Vedic and Brahmanic communions, embracing an amazing diversity of beliefs 
and practices. The contours of Hinduism are blurred in many areas. The line between Hinduism and 
tribal religions is indistinct. Major groups like Lingayats, Jains, Sikhs, and Buddhists are accounted 
Hindus for some purposes, but are in other respects regarded as non­Hindus. The adherents of the 
religions that originated outside India ­­ Muslims, Christians, Parsis, and Jews ­­ are seen as 
distinctly outside the Hindu fold, but here too there are instances of syncretism and overlap. There 
are caste groupings that cut across religious lines to include adherents of several religions­­Jains and 
Hindus, Sikhs and Hindus, Christians and Hindus. The movement into non­Hindu religions of 
members of low­ranked castes, a recurrent recourse of the oppressed, has been accompanied by the 
development of communal distinctions in the new fold. Similarly, sects within Hinduism opposed to 
caste distinctions have been regularly absorbed into the compartmental order. Even where they are 
distinct from Hinduism, all of these religious groups form "communities"  32.   ­­ 
____________________ 
31. Srinivas 1966. But even this is not without exception. The Nadars of southern Tamil Nadu and 
the Ezhuvas of Kerala provide two examples of groups that were regarded as untouchables early 
in this century and managed by strenuous efforts to raise their social standing as well as their 
political and economic position ( Hardgrave 1969, chap. 5; Aiyappan 1965: 133­46.) 
32. "Community" in recent Indian usage is used to refer to a section of the population 
­16­ 
components of the compartmental society. They either constitute or contain endogamous castes or 
caste­like groupings. 
Outside of and overlapping with both caste and religious groupings are the tribal communities, 
concentrated in isolated hilly regions. Some follow tribal religions, most are accounted as Hindus, 
and a significant minority are Christians. Thus, although the endogamous Hindu caste (jati) is the 
paradigmatic unit of the compartmental mosaic, in practice it includes communities which are 
neither Hindu nor castes. 
The compartmental structure of communities is crosscut by other bases of cleavage and affinity ­­ 
village and regional loyalties, occupational and class ties, factional and political alliances. In recent 
generations all of these have been overlaid by differences in exposure to and acceptance of 
"modern" or "Western" cultural influences. Our picture of the compartmental society is static and 
understates the impact of recent changes. Important qualifications must be made for settings where 
communal identity has declined in salience ­­ as among educated urban elites and where class 
formations may have eclipsed community as a basis of group mobilization. The compartmental 
society's notions of prestige and models of affiliation have been both enlarged and effaced by 
"Western" models.  33.   These and other elaborations will be introduced as we proceed. For the 
moment, a sense of a society comprising ranked communal compartments provides a useful base 
line to understand the policies whose development is described in this book. 
____________________ 
differentiated by religion or caste or both (e.g., the Muslim community, the Brahmin community, 
the Agarwal Jain community). The meaning shifts according to the context (e.g., Hindu 
community comprises both Brahmin and non­Brahmin communities and these in turn may be 
divided into communities). "Communal" does not mean things held in common, but refers to any 
arrangement or view which emphasizes religious or caste groupings as social or political units. 
"Communalist" and "communalism" refer to persons or ideologies that support such 
arrangements, and they are ordinarily pejorative. 
33. Thus in a seminal series of works, M. N. Srinivas contrasted "westernization" and as methods for 
achieving higher standing. See Srinivas 1966, chaps. 1 and 2. 
­17­ 
2 Reform, Mobility, and Politics under British Rule 

EARLIER INVADERS and conquerors had accommodated and eventually been absorbed into the 
compartmental social order. The British came to India as traders and conquerors, not as social 
reformers. By the early 19th century, when they had fastened their rule over most of India, British 
policy was to recognize and accommodate the Indian social order ­­ a policy they termed one of 
non­interference. But the impact of British rule was to alter the social order profoundly, if 
inadvertently. Conditions of peace, new communications, new economic activities, new kinds of 
employment, a new legal system, a new system of property relations, and new ideas brought in their 
train new opportunities and new modes of mobility. Old powers and prerogatives were abolished; 
occupations and learning were rendered obsolete or marginal; new opportunities for gain and 
advancement were introduced; power and access to it were redistributed. Groups rose by association 
with the British Government, by trade with the British, by utilizing the new educational 
opportunities that the British introduced, by manipulation of the new legal facilities. Advantages 
gained in these ways tended to be cumulative, for utilization of these new opportunities was biased 
by the existing distribution of resources and by the network of kinship and community which 
1. 
offered access to them.    
____________________ 
1. In addition to works specifically cited, the account in this chapter relies generally on the 
following sources: on social reforms, Heimsath 1964 and S. Natarajan 1959; on changes in Indian 
society, Ghurye 1957, Srinivas 1966, and Rudolph and Rudolph 1967 ; on the untouchables and 
their movements, Zelliot 1969 and Dushkin 1957; on the non­Brahmin movements, Irschick 1969 
( South India), Omvedt 1976 ( Bombay), and Dushkin 1974 ( Mysore); on Gandhi's views and 
activites, Gandhi 1954, Dalton 1967, and Pyarelal (Nair) 1932; on Dr. Ambedkar, Keer 1962, 
Zelliot 1969, and Ambedkar 1945, 1946, and 1968; on legal developments regarding caste 
generally, Galanter 1968, and specifically on legal enforcement of and abolition of social 
disabilities, Galanter 1972a; on the drafting of the Constitution, Austin 1966 and the Constituent 
Assembly Debates. 
­18­ 
Non­interference proved elusive, for whatever the British did or refrained from doing was bound to 
have an impact on the social order. The paradoxical aspect of the policy of non­interference is 
plainly revealed if we consider the disposition of the legal system to the compartmental social order. 
Non­interference implied doing what rulers in India had always doze ­­ actively upholding and 
2. But it also implied an aloof neutrality. The outcome was a changing 
supporting the caste order.    
patchwork of support and aloofness, threaded by impulses of reform given intermittent and reluctant 
expression, all bent and reshaped in unplanned ways by forces released by the British presence. 
The establishment of a nationwide legal system brought a general movement of disputes from caste 
and village tribunals, responsive to the locally powerful, into the government's courts and spread a 
consciousness of rights which might be vindicated independently of local opinion. The 
government's courts espoused a norm of equality before the law. With few exceptions, the same 
rules applied to all. Outside the "personal law" fields, sastric and customary law was supplanted by 
a general law applicable to all. The British undertook to apply separate bodies of law in matters of 
family, property, and religion according to the religious affiliation of the parties. Application of 
these personal laws in the government's courts, with common­law procedure and commonlaw 
trained judges, introduced inadvertent changes. In the formal setting of the British­style court, the 
classical textual element of the law was elevated to a supremacy over the customary element greater 
3. (In Hindu law, for example, the salience of the varna categories was 
than it had enjoyed earlier.    
greater in the late 19th century than earlier.) 
The general features of the legal system were not articulated to a system of graded inequality. Legal 
enforcement of slaver was withdrawn.     4. The use of caste in the general civil, criminal, and 
commercial law was sharply restricted and soon abandoned. The British law did not recognize or try 
to maintain the caste order as such. The law recognized the autonomy of castes as groups. But, after 
an initial flirtation, the British were reluctant to provide affirmative support for their internal powers 
of governance. Affirmative support was confined to a narrower ambit than in the past, when Hindu 
rulers gave redress for grievances which the Anglo­Indian courts excluded as involving only social 
pri­ 
____________________ 
2. Derrett 1968, chap. 7; Jackson 1907 ; O'Malley 1932. 
3. Derrett 1968, chaps. 8, 9; McCormack 1966 ; Rudolph and Rudolph 1967, pt. 3; Galanter 1968. 
4. Legal enforcement of slavery was outlawed by the Indian Slavery Act (Act V of 1843); possession 
of slaves was made a criminal offense by the Indian Penal Code (Act XLV of 1860) § 370. On the 
way that agrestic slavery in southern India compounded the diabilities of the lowest castes, see 
Adam 1840: 171 ff.; Banaji ca. 1933:85; Kumar 1965. 
­19­ 
vileges. Yet the British courts were more ubiquitous and accessible, more efficient and relentless in 
5. 
operation; even their attenuated and indirect support had great impact.    
The British were less willing than earlier rulers to lend the civil power to uphold prerogatives 
claimed in inter­group relations. The assertion of precedence and the imposition of disabilities 
6. Caste groups enjoyed the 
received limited and for the most part indirect support from the law.    
active support of the courts in upholding their claims for precedence and exclusiveness respecting 
the use of religious premises. Exclusionary practices did not enjoy the same active judicial support­­
at least not from the higher courts ­­ in regard to "secular" public facilities such as schools, wells, 
and roads. 
However, where members of higher castes themselves undertook to "enforce" their prerogatives 
against lower castes, the courts were reluctant to provide remedies against such "self­help"­­
including boycott and secondary boycott. In their respect for autonomy of castes, the courts were 
reluctant to interfere with the disciplinary powers of castes against violators of received etiquette of 
inter­caste relations. While at the upper levels of the legal system, support for high caste dominance 
was mainly indirect and passive, there is some indication that at the lower levels of the legal system 
there was often active governmental support for the claims of the higher castes. 
Even where lower castes enjoyed rights that were formally enforceable, they could be vindicated 
only by engaging in the expensive and uncertain process of litigation, a process only partly insulated 
from other sources of advantage. Surveying a range of studies, Cohn concludes: 
The lower castes have generally been unsuccessful when, through the use of police or of 
the urban courts, they have sought to redress what they believe to be the corporate 
wrongs done to them by upper castes. The upper castes maintain their economic 
position, their knowledge of the courts and the intricacies of the law and better access to 
officials have thwarted attempts to change the position of the lower castes in the village 
7. 
society and economy.    
If in practice legal institutions often adapted themselves to prevailing patterns of disability, the 
lowest castes did enjoy a new equality in the 
____________________ 
5. Galanter 1968. 
6. Anandan 1911: 529 observes that the famous proclamation of 1858 in which Queen Victoria 
assured Indians that none would be favored, molested, etc., by reason of their religion or 
observances was "tantamount to saying that the state will not interfere in the most provoking 
inter­caste observances and the consequent ill­treatment of the depressed classes. In other words, 
the state cares to respect only the feelings of the high caste people. . . ." 
7. Cohn 1965: 108. 
­20­ 
eyes of the law and had access to it, at least formally. The law opened some possibilities for 
advancement and change to the lower castes, as it did to others. But it did not provide any special 
leverage for the lowly to use these opportunities, so use of them tended to correspond to the existing 
distribution of resources. Just as, for example, new opportunities for advancement through education 
and government service were grasped largely by Brahmins and other high castes with a literate 
tradition, so those who were already supplied with other advantages tended to make use of the new 
opportunities afforded by the legal system. The law did not provide the higher castes with an 
instrument for aggressively suppressing their inferiors (as, e.g., Jim Crow laws did in the post­
Reconstruction southern United States), but the law provided another resource which higher castes 
could use to protect their claims to precedence (and in some instances perhaps even tighten their 
8. 
hold on valued resources to the exclusion of lower castes).    
Most of the new opportunities in education and government services were grasped by those already 
9. but some members of the lower castes did find new opportunities as 
supplied with advantages,    
menial government servants, in army service, and in trade and contracting (sometimes protected by 
their monopoly in a degraded occupation like leatherwork or sweeping). British courts declined to 
enforce customary service relationships. But while this meant that weaker parties could not invoke 
the assistance of the courts when traditional expectations were disappointed, such assistance was 
hardly required by dominant groups. Many artisans and village servants were displaced from their 
traditional callings and prerogatives by new technologies and new structures of government.  10.   
The problems faced by the lowest castes in utilizing these new opportunities are exemplified in their 
struggle to obtain schooling. In 1856 a Mahar boy, refused admission to a government school, 
appealed to the Bombay Education Department, but was rebuffed on the ground that 
it would not be right for the sake of a single individual, the only Mahar who had ever yet 
come forward to beg for admission into a school attended only by pupils of caste, to 
force him into association with them, at the probable risk of making the institution 
practically useless to the great mass of natives.  11.   
____________________ 
10. Zelliott 1969: 53; Galanter 1972a: 233 and n. 41. 
11. Quoted at Zeillott 1969: 47. 
8. Chidambaram Pillai ( 1933) argues that British law recognized and promoted more restrictive 
claims of exclusiveness regarding temples in South India than had prevailed earlier. 
9. Ghurye ( 1957: 190) notes that government service was staffed by higher castes and that their 
traditional attitude to lower castes carried over into government work. 
­21­ 
In 1858 the Bombay Government announced that it reserved the "full right of refusing . . . support . . 
. to any . . . school in which the benefits of education are withheld from any class of persons on 
account of caste or race," and further resolved "that all schools maintained at the sole cost of 
Government shall be open to all classes of its subjects without distinction."  12.   Nevertheless, 
reported the Department of Public Instruction, "the opposition of the higher castes to the admission 
of [Untouchable] boys to a public school was often so strong that, even with the best will in the 
world, the Department could do very little in the matter."  13.   
A few special schools were opened by reformers, by missionaries, and by the government. In regular 
school, the few "untouchable" children were subject to invidious segregation­­being forced to sit 
away from others on the veranda was a typical experience. Some concessions in primary school fees 
and special scholarships in secondary schools and colleges were made available in the late 19th 
century. But not until the 1920s was there more than a trickle of untouchable students. 
Thus, non­interference combined with rising aspirations and expanding opportunities to produce a 
complex and uneven movement; it both liberated the lower castes from official imposition of 
disabilities and enabled a fortunate fraction to raise themselves, while it added richly to the array of 
advantages of the higher groups. The law, education, government service, and the new economy 
carved out and defined a sphere of public life in which caste distinctions were of little import; but 
the thrust of British policy was to give the higher castes more resources and opportunities to enter 
and succeed in that sphere. 
The growth of a class of Indians educated on Western lines, the spread of British political and social 
ideas, and the insistent criticism of missionaries all contributed to an intellectual ferment which 
questioned many aspects of Indian life. Practices which had appeared natural and inevitable now 
seemed problematic. Religious reformers responded to the humanist and rationalist critique of 
Indian institutions with programs to reconsturct and purify Hindu society. Groups of educated 
Hindus devoted themselves to refashioning social institutions to accord with new notions of 
humaneness, justice, and equality. 
The great reform issues throughout most of the 19th century were the abolition of sati, remarriage 
of widows, education for women, and opposition to child marriage­­issues which reflected higher 
caste practices and options. Caste restrictions (like ostracism after sea voyages) were an issue, but 
not the system of caste itself. Reformers were not indifferent to the strictures of caste, of which they 
were painfully 
____________________ 
12. Ghurye 1957: 189. 
13. Quoted at Zelliot 1969: 48. 
­22­ 
aware, since the strictures often bore with fearsome force on those who outraged conservative 
opinion by putting into practice such reforms as widow remarriage. 
The currents of reform stirred others than emancipated intellectuals. The changes under British 
power gave new scope to the impulse to self­improvement and advancement in many groups. In the 
late 19th century, many castes organized, held conferences, published newspapers, and undertook 
internal reform efforts. At first they tended to seek advancement in traditional terms of ritual status, 
vying for prestigious entries on the census, putting forth claims to exalted varna position, and 
upgrading their practices in terms of received notions of respectability. (This sometimes involved 
changes like adopting child marriages and banning widow remarriage­­the very practices opposed 
by reformers in the highest castes!) The interest of caste organizations turned increasingly to secular 
betterment: they founded hospitals, hostels, and loan funds; they sought representation, government 
jobs, and scholarships. 
Caste organization brought with it two important and related changes in the nature of castes. The 
salient groups grew in size from the endogamous jatis into region­wide alliances. Concomitantly, 
the traditional patterns of organization and leadership in the village setting were displaced by 
voluntary associations with officials whose delimited authority derived from elections. 
New forms of organization were taken up by some castes at the lower end of the spectrum.  14.   Some 
embraced the opportunities afforded by missionaries; others were influenced by Hindu reform 
movements; still others developed their own organizations. The first militant action by low castes 
was the campaign of the Shanars (later Nadars) to enter temples in the Tinnevelly District in 
Madras, a demand connected to their claim to be Kshatriyas. The Ezhuvas (Iravas) of Kerala took a 
different tack: by the early years of the 20th century, they had developed their own system of 
temples and schools. 
The first efforts to assist the low castes had come at mid­century under the auspices of a Poona 
reformer, Jyotirao Phule, who united a program of education and self­help with a thoroughgoing 
attack on caste,  15.   from missionaries who proffered conversion as a solution,  16.   and from Hindu 
religious reformers like Dayananda Saraswati, founder of the Arya Samaj, who preached a purified 
varna and attacked caste exclusiveness (in its existing degenerate form).  17.   
Although anti­caste associations had earlier voiced opposition to 
____________________ 
14. Lynch 1969: 67­69; Hardgrave 1969: 130­32. 
15. Keer 1964; Ghurye 1957: 200­1; Omvedt 1976: 106­23. 
16. Heimsath 1964; 53­54. 
17. Heimsath 1964: 321, 330; S. Natarajan 1959: 70; Ghurye 1957: 258 
­23­ 
caste distinctions, it was not until the end of the century that mainstream reformers saw caste 
hierarchy and inequality as problems in their own right. Now reform of the caste system, rather than 
of the practices of individual castes, became a major plank of the reformers. (The issue derived 
dramatic impact from its juxtaposition with the discriminations suffered by Indians in South Africa, 
which had become a public issue in India in the 1880s.) In 1895, Justice M.G.Ranade of the 
Bombay High Court circularized social reformers regarding the claims of aboriginals and 
untouchables on Hindu society. In the next dozen years, missions, schools, and scholarships for 
Depressed Classes were founded. Not only did amelioration of the lowest castes become one of the 
most prominent items of reform, but it was accompanied by a shift in attitude toward the caste 
system. 
Earlier reformers sought improvement from within. At first they had concerned themselves with 
sastras, trying to provide textual authority for such reforms as widow remarriage and abolition of 
child marriage. Gradually, they had come to realize that sastric learning provided little leverage to 
change current custom. It was the adherence of the caste group that upheld custom, not sastric 
learning. And, improvement from within by caste associations engaging in self­reform, it was 
discovered, cut both ways, for a caste might embrace orthodox usage as well as depart from it. This 
kind of piecemeal approach came to seem less inevitable as the image of Indian society changed. 
The notion of reforms proceeding from and directed to a single national community transcending 
divisions of religion, caste, and sect was not self­evident in the late 19th century. Reform of social 
practice was considered appropriate for each constituent group within the society. Thus, in his 
presidential address to the second meeting of the Indian National Congress in 1886, Dadabhai 
Naoroji explained his objections to making social reform part of the Congress program: 
How can this gathering of all classes discuss the social reforms needed in each 
individual class? What do any of us know of the internal home life, of the customs, 
traditions, feelings, prejudices of any class but our own? How could a . . . cosmopolitan 
gathering like this, discuss to any purpose the reform needed in any one class? Only the 
members of that class can effectively deal with the reforms therein needed. A National 
Congress must confine itself to questions in which the entire nation has a direct 
participation, and it must leave the adjustment of social reforms and other class 
questions to class Congresses.  18.   
But as nationalism grew, this division into "classes" seemed less natural and commendable. The 
caste institution itself fell into bad 
____________________ 
18. Naoroji n.d., p. 8 
­24­ 
repute, blamed for disunity, weakness, and lack of civic spirit. Increasingly, reformers saw the caste 
system itself as the enemy. In the early years of the century, pleading for the relaxation of 
restrictions among "sub­castes" was replaced by opposition to caste itself. Among the most radical, 
the aim was not merely to raise the conditions of the lowest castes, but to abolish the system of caste 
altogether.  19.   
The ferment within Hindu society has to be seen in the context of changing Hindu­Muslim relations. 
British rule shattered the earlier patterns of dominance and accommodation, permitting, if not 
fostering, open rivalry that found outlets in new political arenas.  20.   The concern of Muslims over 
the loss of their former preeminence was compounded by the feeling that they were falling behind 
the Hindus in modem callings and opportunities and were being reduced to a subordinate position. 
As the movement toward greater Indian participation in government gathered momentum in the late 
19th century, it threatened to accentuate this imbalance, compounding Muslim anxiety about being 
dominated. In 1906 a deputation of Muslim notables petitioned the Viceroy to safeguard Muslim 
interests in the forthcoming electoral regime, reserving seats to be chosen by separate Muslim 
electorates. The British were responsive to Muslim apprehensions. In 1909 when the Minto­Morley 
reforms were promulgated, separate electorates were provided for Muslims. 
At this point, the condition of the untouchables­­as they were coming to be called  21.   ­­which had 
become a major item of reform interest, was propelled from the realm of philanthropy and social 
uplift into the political arena. Rising concern about the untouchables was augmented by the first 
appreciation of their political significance. Eleanor Zelliot observes: 
____________________ 
19. K. Natarajan, a prominent Bombay reformer, first formulated the notion that untouchability was 
an inevitable concomitant of the caste system; to remove it, caste must itself be removed ( S. 
Natarajan 1959: 120). 
20. Barrier 1968 ; Jones 1968. 
21. In 1909 the problem of the lowest castes was for first time conceptualized under the rubric of 
"untouchability"­­a general term which opened the possibility of visualizing the problem not as 
that of a congeries of depressed groups, but as a stratum of all­India dimension with shared 
characteristics. The vicissitudes of the search for that stratum are detailed in chap. 5, § A, below. 
Use of the English term "untouchable" in print can be traced with fair precision to the year 1909. 
The Maharaja of Baroda in his remarks to the Depressed Classes Mission of Bombay on October 
18, 1909, uses the term and provides an explanation to his audience ( Sayaji Rao 1928: 244­45). 
The abstract form "untouchability" is used by Sridhar V. Ketkar ( 1909: 86) in a footnote to his 
study of the caste system; the preface of the volume is dated September 1909. Internal evidence 
suggests that the author did not have this term at his disposal when he wrote the text (cf. pp. 99 
and 121­22). The Maharaja of Baroda was the patron of Mr. Ketkar and supported him during the 
years at Cornell University when he wrote this book. (See Ketkar 1911: xxvi). Weitbrecht­Stanton 
( 1920: 173) attributes the term's prominence to the Gaikwad [Maharaja]. 
­25­ 
The granting of an electorate for Muslims in which they alone would vote brought the 
idea of communal electorates for minorities to the forefront in the minds of all 
communities which feared for their submersion in a government run by the dominant 
caste Hindu community. The granting of special electorates to the Muslim community 
also made numbers important. Whether the­vast numbers of Untouchables were truly 
Hindu and to be counted as such, or not, became an important question for the first time. 
 22.   
Concern about Hindu numbers was accentuated when the Census Commissioner (for 1911) bruited 
the suggestion that the untouchables (then known as Depressed Classes) should be enumerated as a 
group separate from the Hindus, a suggestion promptly endorsed by the Muslim League. 23. 
Suddenly there was quickened interest in the Depressed Classes from previously indifferent Hindu 
leaders who were now insistent that Depressed Classes were within the Hindu fold. 24. 
Apprehensions about a continued Hindu majority must have been seen in the light of then current 
estimates that the Depressed Classes numbered 50 to 60 million  25.   ­­from a sixth to a fifth of the 
total population and roughly the magnitude of the plurality of Hindus in the population.  26.   
"Untouchability" was thus propelled into a prominent (but not central) place on the Indian political 
stage, which it was to occupy until after Independence. This new position was soon reflected by the 
reversal of the Indian National Congress's long­standing policy of excluding "social reform" from its 
program. At its 1917 annual meeting, the Congress for the first time passed a mild and hesitant anti­
disabilities resolution.  27.   As Congress had avoided social reform because of its divisive potential, it 
now embraced it (in part) because of its unifying 
____________________ 
22. Zelliot 1969: 141. 
23. "Since the publication of the recent Reform Measures [the Minto­Morley reforms of 1909] the 
Mahomedans have been loud in declaring that, properly speaking, the outcastes are beyond the 
pale of Hinduism, and therefore their strength should not go to swell the numerical force of the 
Hindus. . . ." (Saint Nihal Singh [in Indian Review, September 1910, reprinted] in Anonymous ca. 
1913: 89.) 
24. A contemporary observer reports that "a question that is agitating Hinduism at the present 
moment is as to whether these [untouchable] classes should be counted as Hindus or not. Ten 
years ago the answer would have been emphatically in the negative. Even now the conservative 
feeling of the country is for their exclusion. But the conscience of the more advanced section of 
the educated Hindus is a little more sensitive on this point" ( Holderness 1911:101­2). Cf. Tinker 
1962:131, who reports that "as late as 1910, upper caste politicians were demanding that 
[untouchables] should be excluded from the Hindu fold in the Census." 
25. On these estimates, see chap. 5, § A, below. 
26. The Hindu percentage of the population had been falling slowly but steadily for decades. See 
n.45 below. 
27. "The Congress urges upon the people of India the necessity, justice and righteousness of 
removing all disabilities imposed by custom upon the Depressed Classes, the 
­26­ 
potential. From this point on, the social question was indissolubly linked to the political. 
At the time of Congress's first gesture of interest in the untouchables, there was only a faint scatter 
of political activity among untouchables themselves. There was, however, more noticeable stirring 
among the castes above them but below the social peaks ­­ i.e., in the middle ranges of Hindu 
society. Movements of the middle castes against the cultural preeminence of Brahmins had emerged 
in the 19th century; now, focussing on Brahmin predominance in education and government service, 
they took on political form. The Justice Party ( South Indian Liberal Federation), the organ of the 
Non­Brahmin movement in Madras, was founded in 1916; the organization of the Non­Brahmin 
party in Bombay got under way in 1917. The Princely State of Mysore instituted a system in which 
all communities other than Brahmins were denominated "Backward Classes;" from 1918, places 
were reserved for them in colleges and state services and thus was inaugurated the first modern 
regime of communal quotas. Non­Brahmin demands for a greater share of political representation 
were quickly recognized by the British. The promised post­war reforms (the Government of India 
Act, 1919) accorded them special representation. During the 1920s, the nonBrahmin parties 
flourished in the Madras and Bombay presidencies, erecting a system of preferences in educational 
admissions and recruitment to government posts. At first, these non­Brahmin, or Backward Classes, 
movements undertook to speak for the lowest strata as well, but as they prospered, their voices 
increasingly reflected their higher echelon constituents.  28.   
The 1919 reforms provided only a few nominated seats for the Depressed Classes,  29.   but for the first 
time some untouchable spokesmen were heard in political assemblies. As untouchables increased 
their political activity and as concern for their welfare became a plank of Congress policy, 
legislatures for the first time took an interest in their problems. There was a scatter of undertakings 
by government to provide land, housing, schooling, and government posts to the Depressed Classes. 
The illegality of denying access to schools, wells, and roads was declared in legislative resolutions 
and administrative orders ­­ which 
____________________ 
disabilities being of a most vexatious and oppressive character, subjecting those classes to 
considerable hardship and inconvenience" (quoted in S. Natarajan 1969: 144). The Congress 
resolution omitted a crucial passage of the original, as submitted by the Depressed Classes 
Mission Society, which specified "disabilities imposed by religion and custom" (italics supplied). 
28. On the isolation of untouchables from non­Brahmin movements, see Irschick 1969:188ff; Zelliot 
1969:143 ff; Dushkin 1974:252 ff; Ramaswamy 1978:296. 
29. Depressed Classes representation was increased (and extended to include a nominated member in 
the Central Legislative Council) following the findings of the Muddiman Committee in 1925 
( Zelliot 1969: 158). 
­27­ 
were honored largely in the breach.  30.   Special government officers were appointed to look after 
their welfare. There were fee concessions and scholarships and some pressure on schools to be 
accessible. School enrolments increased, and a few untouchables entered public services.  31.   By the 
close of the 1920s, as Eleanor Zelliot observed, "the principle of special attention was firmly 
established."  32.   
Untouchability (and the excesses of caste hierarchy) presented a problem whose solution was 
required to unlock India's national destiny. The question remained how this was to be done. The 
various answers might roughly be classified by their resemblance to two polar types that we may 
call the "evangelical" and the "secular." The "evangelical" approach stresses the uplift of 
untouchables to higher Hindu standards and the penance of caste Hindus for the injustices of 
untouchability, which is seen not as an integral part of Hinduism but as some external impurity. 
Uplifted untouchables and repentent Hindus will join together in a purified and redeemed 
Hinduism. At the other extreme, the secular approach stresses the denial of civil rights and 
economic opportunities and seeks to combat them by vigorous government intervention, and 
political action to obtain it. Untouchability is seen as an integral part of Hinduism. The secular 
approach does not seek to restore Hinduism to some original purity, but seeks instead to liberate all 
by dissolving the graded inequalities of caste into the common bond and common rights of 
citizenship in a secular state. In the "evangelical" approach, the emphasis is on moral regeneration 
and philanthropic uplift; in the secular approach, it is on civic and economic improvement under 
government auspices. 
These divergent approaches were present from the first stirrings of concern about the untouchables. 
In the 19th century, the evangelical 
____________________ 
30. E.g., in 1923 the Bombay Legislative Council resolved that untouchables be allowed to use all 
public watering places, wells, schools, dispensaries, etc. The Provincial Government did not, 
however, undertake direct responsibility for enforcement as to local facilities, but requested 
"collectors . . . to advise the local bodies in their jurisdiction to consider the desirability of 
accepting the recommendations made in the Resolution so far as it relates to them" ( Bombay 
1930 [ Starte Committee]: 52). The flavor of enforcement activity is conveyed by the action of a 
District Board which, seven years later, resolved to post notices that facilities were open to 
untouchables "at only those villages in the district where the public opinion is favorable for such 
action" (quoted at Sanjana 1946: 237). At that time, the Starte Committee concluded that the 
policy was a "complete failure" when it could not "find a single instance where Depressed 
Classes are continuously using the same public well as the higher classes. There may be such 
wells, but if so, they must form an infinitesimal proportion of the hundreds of thousands of 
public wells in the Presidency" ( Bombay 1930 [ Starte Committee]: 52). 
31. Ghurye 1957: 260; Hutton 1961:208. The first untouchable gazetted officer was appointed in 
1927 ( Zelliot 1969: 162). 
32. Zelliot 1969: 166. 
­28­ 
approach was exemplified in the activities of the Arya Samaj; the secular, by Jyotirao Phule. During 
the 1920s, each acquired a champion of immense stature. Their individual struggles illuminate the 
problems inherent in these approaches; the wrestling between them cast the mold of policy about 
untouchability for independent India. 
In 1916 Mohandas K. Gandhi returned to India from South Africa and assumed leadership of the 
Indian National Congress. In his autobiography, Gandhi tells us that from the age of twelve he was 
convinced that untouchability could not be sanctioned by Hinduism. From his first public statements 
on the subject, he describes untouchability as a sinful excrescence upon Hinduism. He calls upon 
caste Hindus to do penance, to purify themselves and Hinduism, while urging the untouchables to 
observe cleaner habits. The problem is one for voluntary private action, not for legal coercion. 
While rejecting caste in its modern manifestations, Gandhi accepted varnashrama dharma as a 
healthy division of labor by birth. Untouchability should be removed by abolishing the unwarranted 
fifth division and restoring the untouchables to their rightful estate as Sudras in a purified varna 
expressing not inequality but a horizontal and non­invidious division based on hereditary qualities. 
The notion that some are by birth polluting must be abolished, and those self­imposed restrictions 
on intercourse that are conducive to spiritual development should be retained. 
If Gandhi personified the evangelical approach, the secular approach was personified by the most 
gifted and wide­ranging spokesman to appear among the untouchables, Dr. Bhimrao Ramji 
Ambedkar.  33.   Dr. Ambedkar (born 1890) was younger and less experienced than Gandhi when he 
stepped into the political stage in the early 1920s.  34.   While Gandhi's appeals are directed principally 
to the 
____________________ 
33. These giants throw long shadows. Using them to illuminate the conflict over untouchability 
should not obscure the fact that each had rivals and critics within his own camp. The 
untouchables were (and are) a vast and heterogeneous collection. Dr. Ambedkar, throughout his 
career, received the most intense and unswerving support from his own Mahar community. 
Notwithstanding his subsequent veneration, his command of the allegiance of other groups of 
untouchables varied considerably, and there were intermittent attacks on his capacity to represent 
them. On other untouchable leaders and movements, see Zelliot 1969: 183­85, 193­94, 206­7; 
Patankar and Omvedt 1979:415. 
34. These two men shared a number of qualities but displayed a number of striking contrasts. Both 
hailed from the hinterlands of Bombay: Gandhi from Gujarat, Ambedkar from Maharashtra. 
Gandhi was a bania (merchant) by caste, while Ambedkar was a Mahar. The Mahars are a 
populous caste of field laborers and village servants in western India, which had a military 
tradition, having formed many units in the East India Company's Bombay army. Both came from 
families with a tradition of government service. Gandhi's elders had traditionally served as 
Diwans of princely states in Kathiawad, Ambedkar's as non­commissioned officers in the Indian 
army. Both grew 
­29­ 
caste Hindus, Ambedkar addresses the untouchables and directs their attention to the potentialities 
of action to be taken by government. His programs express conflicting impulses between 
assimilation into Hindu society and separation from it, a pattern that recurs throughout his career. 
These conflicting impulses are displayed in two successive conferences over which he presided in 
1929. At the first, the conferees were invested with the sacred thread by a Brahmin priest to the 
chanting of vedic hymns, in a classical instance of elevation in "sanskritic" terms. At the second, the 
conferees resolved to embrace "any other religion" if the Hindus continued obdurate in their 
mistreatment.  35.   
The next major round of bargaining over the future of India began in 1928. For the first time the 
untouchables were full­fledged participants. The negotiations began with the visit of the Simon 
Commission, which was empowered to make recommendations for a new round of constitutional 
reforms. Dr. Ambedkar, by this time recognized as an important spokesman, appeared before the 
Commission to demand reserved seats for the untouchables in legislative bodies, special educational 
concessions, and recruitment to government posts, recommendations substantially accepted in the 
Commission's Report.  36.   
When this Report proved unacceptable to the major contenders, the British Government convened a 
Round Table Conference in London to 
____________________ 
up in an atmosphere of devotional religion: Vaishnavism in Gandhi's case, while Ambedkar's 
family were adherents of the 15th­century poet and religious reformer Kabir who had Muslim as 
well as Hindu followers. Both were married young to younger and less educated wives, and both 
were sent overseas to complete their education. Gandhi went to London, though the sea journey 
meant ostracism from his caste, to join the Bar. Ambedkar, under the patronage of the Gaikwad 
of Baroda, was sent to America ­­ the first prominent Indian leader to be educated there. Both 
returned to Bombay and joined the Bar. When Ambedkar returned to India, he joined the Baroda 
civil service in a responsible job under his patron, but lack of acceptance of the untouchable by 
his fellow officers, clerks, and landlords drove him back to Bombay. Under the patronage of 
another Hindu prince, the Maharajah of Kolhapur, he entered politics. 

Ambedkar was a scholar, who managed to earn several doctorates, and a bibliophile. Unlike 
Gandhi, he was an aloof, solitary, and domineering man who found sustained organizational 
efforts trying. His impatience with organizational detail and his unwillingness to delegate 
authority made each of the many organizations he founded a personal organ, run along 
paternalistic lines and lacking in solid leadership in the middle ranks. Unlike Gandhi, whose 
devotions never ended, Ambedkar was for the most of his life a mild theist or indifferent to 
religion. Unlike the mild Gandhi, he was given to outspoken and caustic attacks on his 
opponents. Gandhi's Western heroes were Ruskin, Tolstoi, and Thoreau; Ambedkar's favorite 
historical figure, according to his biographer ( Keer 1954: 214), was Napoleon. 

35. Keer 1954: 128­30. 
36. Reported in Zelliot 1969: 170­72. 
­30­ 
which it invited delegates from all parties and prominent interest groups. The Congress, claiming 
that it alone represented Indian opinion, refused to participate. In the course of the Conference, Dr. 
Ambedkar abandoned his earlier opposition to separate electorate.  37.   His demands for safeguards in 
a future constitution now included separate electorates as well as recruitment to government 
services, laws against discrimination, and a special department to look after the Depressed Classes. 
The Civil Disobedience Campaign, which had kept Congress aloof from the first Round Table, 
ended with an arrangement for a Second Round Table Conference. Gandhi, representing Congress, 
attacked all the other delegates as unrepresentative. In particular, he attacked Ambedkar's claim to 
represent the untouchables. "I claim myself in my own person to represent the vast mass of the 
untouchables." Opposed to separate electorates for any group, he grudgingly conceded them to 
Muslims, Christians, Sikhs, and Anglo­Indians, but the demand of the untouchables for separate 
electorates he termed "the unkindest cut of all" and vowed to resist it with his life.  38.   
When the Conference's Franchise Committee was unable to reach agreement on representation for 
minorities, the delegates authorized Prime Minister Ramsay MacDonald to make an award which 
would be binding on all parties. The Communal Award, announced on August 17, 1932, gave the 
untouchables regular votes in the general electorate and granted their demand for separate 
electorates in areas where they were concentrated in addition to the regular votes they would cast as 
members of the general electorate.  39.   
Gandhi, who was imprisoned as a result of his civil disobedience campaign, now undertook to carry 
out his vow by fasting until death unless the separate electorates for untouchables were revoked. His 
objection was that separate electorates would signify a perpetual split in Hinduism, perpetuating the 
stigma of untouchability and making impossible the assimilation of the untouchables into the Hindu 
community. It is evident that Congress opposition, if not Gandhi's personally, was inspired by fear 
that the great Hindu base of Congress support would be weakened and that candidates returned 
from separate electorates would be advocates for group interests with little concern for the Congress 
program. Finally, many Congressmen, including Gandhi, 
____________________ 
37. Dr. Ambedkar's oppsition to separate communal electorates was shaken by the stubborn 
attachment of other communites to this device, which was also favored by most untouchable 
activists. His sense of obligation to represent his constituency compounded his fears that the 
claims of the Depressed Classes would be submerged without their own representatives ( Zelliot 
1969: 176­77; Patankar and Omvedt 1979: 419). 
38. Zelliot 1969: 181. 
39. Zelliot 1969: 185. 
­31­ 
did not have a very high regard for the political capacities of the untouchables.  40.   
The British Government declined to alter the Award without the consent of the affected groups ­­ 
i.e., without the consent of Ambedkar as the representative of the Depressed Classes. During the 
twenty­one days of Gandhi's fast, Ambedkar was subjected to immense pressures, not least the sense 
of the cataclysm that might engulf the untouchables if Gandhi were to die. Wresting as much as he 
could from the Congress leaders who negotiated with him, Ambedkar relinquished the separate 
electorates in return for a system of reserved seats to be elected from "joint" or general electorates. 
The final agreement reached on September 28, 1932, known as the Poona (or Yeravda) Pact, set the 
subsequent pattern of representation for untouchables.  41.   The Poona Pact provided 148 seats for 
untouchables, instead of the 78 separately elected members given by the Communal Award. It also 
provided a system of primary elections for those reserved seats; a panel of four candidates was to be 
chosen by electors from the Depressed Classes. 
In the light of the agreement, it is clearer just what was really at stake. Gandhi argued for joint 
electorates so that caste Hindus would be forced to seek votes from untouchables. But, of course, 
this was the case under the Communal Award, since untouchables also had votes in the general 
constituencies. Gandhi's principal objection had been to having separate rolls of untouchable voters, 
but this was preserved by the Poona Pact's arrangements for nomination of the candidates. 42. 
Whatever the motives of the fast, the political effect was to dramatize the issue of untouchability 
while preventing arrangements that would foster an independent political movement sufficiently 
strong to belie the Con­ 
____________________ 
40. After meeting Ambedkar for the first time in 1931, Gandhi told his secretary, "I did not know that 
he was a Harijan. I thought he was some Brahman who took deep interest in Harijans and 
therefore talked intemperately" (quoted by Zelliot 1969: 178). Since Ambedkar had been a 
prominent spokesman for several years, this suggests inattention to untouchable movements as 
well as an assumption that untouchables were incapable of providing their own leadership. Cf. 
Patankar and Omvedt 1979: 419. 
41. For Ambedkar's account, see Ambedkar 1946: 322­28; Zelliot 1969: 182­88. For a Gandhian 
account, see Pyarelal (Nair) 1932: 165­69, Gandhi 1954 b: f: 177, 65­67; also Dushkin 1957, chap. 
4. 
42. The Poona Pact provided a system of primary elections in which eligible voters of the Depressed 
Classes would "form an electoral college" to elect "a panel of four candidates belonging to the 
Depressed Classes" to stand for each reserved seat. This system was to end after ten years unless 
terminated sooner by mutual agreement (Poona Pact, reprinted at Ambedkar 1946: 88­89). 
Subsequent disagreement over this clause was resolved when the Hammond Committee (on 
construction of the constituencies) adopted Ambedkar's interpretation that "four" meant "no 
more than four" and that voting in multi­member constituencies was to be cumulative rather than 
distributive ( Ambedkar 1946: 91­94). In the 1937 elections, primaries were held in only 43 of 
the 151 reserved constituencies. Ambedkar complained that the primaries were little used 
­32­ 
gress claim to represent all of India.  43.   Ironically, this fast is often referred to as Gandhi's fast 
against untouchability. 
In a setting of Hindu­Muslim divergence, this sequence of events ­the Communal Award, the fast, 
and the Poona Pact ­­ raised concern with untouchability to new heights.  44.   It was recognized as a 
problem that demanded a solution as part of the question of Indian nationhood. The new salience of 
untouchability was not unconnected with concern about political alignments.  45.   As one Congress 
sympathizer put it: 
____________________ 
because the double election was prohibitively expensive for Scheduled Caste (as they had come 
to be called) candidates. And, analyzing the elections, he argued that when primaries were held, 
they did not insure that those successful in the general election were genuine representatives of 
the Scheduled Castes ( Ambedkar 1947: 44). This displeasure was shared by other untouchable 
leaders. Ambedkar had earlier proposed modifying the scheme to require that candidates could 
stand in the general election only if they had received 25% of the Scheduled Caste vote in the 
primary, a suggestion rejected by Gandhi ( Zelliot 1969: 188). A proposal to continue the primary 
system with a 20% requirement may be found in a report to the 1945 Sapru Committee ( Sapru 
1945 : lxviii), but after partition and the adoption of Article 325 barring electoral rolls on the 
basis of caste, the possibilities for such primaries faded from sight as well as from memory. 
43. Lelah Dushkin ( 1957: 74) observes that "the basic solution (except for the panel system) was 
what Ambedkar had advocated in 1929, what Sapru had advised at the second R[ound] T[able] 
C[onference], and what Rajah and Moonje had proposed in their pact. The whole affair would 
not have occurred if Gandhi had not treated the D[epressed] C[lasses] delegates with such scant 
courtesy at the second R.T.C." 
44. A conference of Hindu leaders, convened in Bombay on September 25, 1932, to ratify the Poona 
Pact, unanimously adopted the following resolution: "This Conference resolves henceforth, 
amongst Hindus no one shall be regarded as an untouchable by reason of his birth, and that those 
who have been so regarded hitherto will have the same right as other Hindus in regard to the use 
of public wells, public schools, public roads and all other public institutions. This right shall have 
statutory recognition at the first opportunity and shall be one of the earliest Acts of the Swaraj 
Parliament, if it has not received such recognition before that time. It is further agreed that it 
shall be the duty of all Hindu leaders to secure, by every legitimate and peaceful means, an early 
removal of all social disabilities now imposed by custom upon the so­called untouchable classes, 
including the bar in respect of admission to temples" (quoted in Rajagopalachari 1933: 1). 
45. Anxiety about the departure of the untouchables should be seen against a background of the 
much­discussed, long­term decline in the Hindu proportion of the total population of India, 
roughly summarized in the following table: 
Date 

Percentage 

Source 

1881 
74.3 

1891 
72.3 

1901 
70.3 

1911 
69.3 

1921 
68.4 

1931 
68.2 

1941 
65.9 

Sources: a: Census of India, 1911; b: Census of India, 1941: Vol. I, Part I, 102 ff. 
­33­ 
With the Harijans within the Hindu political group,  46.   the future of the Hindus is safe. 
Without the Harijans, the destiny of the Hindus will remain uncertain. . . . We do want 
to have the Harijans on our side in the political struggle, and even if it were for this 
purpose alone, the Caste­Hindus must raise this curse of untouchability and bridge the 
gulf that divides them from their co­religionists. There is a real crisis in the political life 
of the Hindu community. . . .  47.   
Efforts to remove untouchability and help its victims proliferated in the wake of the Poona Pact. 48. 
Congress had earlier committed itself to a program of fundamental rights in the future republican 
India in which there would be no recognition of civil and social disabilities.  49.   But only after 
Gandhi's fast were Congress leaders willing to countenance the affirmative use of the law to abolish 
disabilities ­­ and in particular to obtain by law the admission of untouchables to Hindu temples, 
thereby symbolizing their inclusion within the Hindu community. While temple entry absorbed 
advanced Hindu opinion, Dr. Ambedkar turned away from it to emphasize economic and 
educational advancement. Temple entry, he asserted, was up to the Hindus; it was through political 
power that the untouchables were to find their salvation, not through acceptance by the Hindus. 
Both Gandhi and Ambedkar, each in his own way, attacked the Hinduism of the day. In a 
charismatic transvaluation of Hindu 
____________________ 
46. One enduring product of these events was a new terminology. "Untouchable" was in general use 
by 1920, but "Depressed Classes" remained the official term. In 1931, Dr. Ambedkar and R. 
Srinivasan, (the Depressed Classes' representatives at the Round Table Conference) observed that 
the term "Depressed Classes" is "degrading and contemptuous" and proposed as alternatives 
"Non­caste Hindus," "Protestant Hindus," or "Non­conformist Hindus" (Indian Round Table 
Conference 1932: 76). But events took a different course. Shortly afterwards, Gandhi began to 
popularize "Harijan" ( Gandhi 1954: 16­17), and then the Government of India Act, 1935, denoted 
those who would enjoy the special electoral arrangements for untouchables as "Scheduled 
Castes," a term destined to remain the official designation for untouchables down to the present. 
Just as "Harijan" fossilizes the evangelical fervor of Gandhi's anti­untouchability campaigns, so 
"Scheduled Castes" summons up the bureaucratic rigidification and torpor that overtook 
Ambedkar's measures for secular improvement. 
47. Agarwal 1934: 22. Mr. Agarwal was a Cambridge­educated lawyer and professor in Poona. 
48. These included the formation of the Harijan Sevak Sangh, an organization devoted to the uplift 
of the Harijans. Its board was composed entirely of non­Harijans, giving rise to reproaches from 
Dr. Ambedkar and others. 
49. In 1931, the annual meeting of the Indian National Congress at Karachi propounded a program of 
fundamental rights for future republican India which included: "(vi) no disability to attach to any 
citizen by reason of his or her . . .caste . . .in regard to public empolyment, office of power or 
honors, and in the exercise of any trade or calling. (vii)equal rights of all citizens in regard to 
public roads, wells, schools and other places of public resort" ( All­India Congress Committee, n. 
d., p. 66). 
­34­ 
practice, Gandhi dramatized his identification with Harijans, and particularly with the bhangis 
(sweepers), the lowest and most polluting of castes in traditional terms.  50.   During his tours, he 
stayed in the sweepers' quarters. Calling himself a Harijan and a Bhangi, he expressed the wish to be 
reborn as an untouchable. Gandhi advocated a purified varnashrama dharma in which untouchables 
would be restored to their rightful place as Sudras.  51.   But to Ambedkar, the four­varna system is the 
very root of untouchability. For Hinduism to accommodate the untouchables, it must be purged of 
the varna system. Where Gandhi proposed to rid Hinduism of its imperfections and restore it to its 
ideals, Ambedkar was skeptical of these ideals themselves. To him, those who observed 
untouchability were not bad Hindus, but merely observant ones. The only way to change their 
conduct was to destroy the authority of the sastras. Thus, reform of Hinduism meant extirpating the 
offending doctrines and abolishing the hereditary priesthood; in their place a new standard book of 
doctrine should be promulgated and a licensed, certified priesthood established.  52.   
Dr. Ambedkar increasingly despaired of the possibilities of reform within Hinduism. He alternated 
proposals for radical reform of Hinduism with recommendations that untouchables convert to a 
religion that would guarantee them equality. In 1936 he conducted an extended flirtation with 
Sikhism.  53.   Proposals for conversion aroused a storm of opposition. Gandhi charged that the 
religious feelings of the untouchables were being exploited in a political game. In a characteristic 
combination of genuine solicitude and uncomprehending condescension, he expressed doubt 
whether the untouchables could 
____________________ 
50. "A Bhangi does for society what a mother does for her baby. A mother washes her baby of the 
dirt and insures his health. Even so the Bhangi protects and safeguards the health of the entire 
community by maintaining sanitation for it. The Brahman's duty is to look after the sanitation of 
the soul, the Bhangi's that of the body of society" ( Gandhi 1954b: 215­18 [from Harijan of 
November 28, 1936]). 
51. Gandhi's program was by no means moderate in its day. C. B. Agarwal is probably closer to 
popular feeling among the educated when he complains that Gandhi is moving too fast. While 
favoring civil rights and educational help, Agarwal rejects interdining and inter­marriage and 
would restrict temple entry to those untouchabales who give up meat­eating. 
52. This was a program so calculated to outrage respectable Hindu sentiment that the organization 
cancelled their invitation when they saw advance copies of the speech (see Ambedkar 1945). But 
it was not entirely without precedent; e.g., in 1928, a rural Depressed Classes Conference had 
called for proscription of the Manu Smrti, the pre­eminent Hindu law text ( Zelliot 1969: 170). 
53. One interesting sidelight of these negotiations is that Ambedkar called in the Hindu Mahasabha, 
the foremost of the Hindu communalist parties, to seek their approval. Apparently the aim of this 
was to remain Hindu for purposes of the reserved seats awarded by the Poona Pact. The Hindu 
Mahasabha agreed on the condition that conversion was to Sikhism rather than to Christianity or 
Islam. Ambedkar himself 
­35­ 
distinguish the relative merits of the contending religions any more than a cow could.  54.   
The new popular governments which took office in 1937 avowed their opposition to untouchability. 
Madras passed the first legislation imposing criminal penalties for denying to untouchables the use 
of public facilities. Temple entry had been proclaimed earlier in a few progressive princely states. In 
1938, for the first time in British India, governments intervened to secure the opening of temples 
when Bombay and Madras passed temple­entry acts. Encouragement of education for untouchables 
increased, and there were measures to secure them some posts in government service. 
The non­Brahmin movements that had emerged so dramatically a generation earlier now bifurcated, 
as lower castes insisted on provisions separate from those for the upper non­Brahmin groups. 55. 
(The former emerge as the "other backward classes," whose story we take up in chapter 6.) To a 
large extent, these movements were absorbed into the Congress. Congress governments continued 
and extended the regime of communal quotas in public services and college admissions that these 
movements had produced, as did non­Congress governments elsewhere.  56.   In the central services, 
the Government of India reserved quotas for Muslims and other religious minorities. 
During the Second World War, when the Congress opposed cooperation with the British war effort, 
Dr. Ambedkar served as a member of the Viceroy's executive council, the first untouchable to attain 
such a lofty station. He secured for the Scheduled Castes (as they were now called) reservations of 
posts in central government services and various other concessions, including recruitment into the 
army and the police.  57.   
Gandhi, who was in prison along with the Congress leadership for 
____________________ 
agreed that it would be preferable to stay within Hindu culture, thereby avoiding Muslim 
domination. As it happened, after the conversion of a few of his followers, bad feeling developed 
between Ambedkar and the Sikhs, and the matter dropped out of sight ( Zelliot 1969: 224­26; 
Keer 1962: 288). But the issue of the untouchables enjoying special treatment after abandoning 
Hinduism is a persistent one that resurfaced dramatically in the 1950s and has remained 
unresolved. See chap. 9, §§E and F, below. 
54. Gandhi's response to charges that this statement was demeaning was that neither could his wife 
( Gandhi 1954b: 154 [from Harijan of March 13, 1937]). 
55. The Madras Provincial Backward Classes League, consisting of the less forward non­Brahmin 
communities, was founded in 1934 for the purpose of securing separate treatment from "the 
forward non­Brahmin communities." See chap. 6 below. On the tension between the upper and 
lower strata within the non­Brahmin movement, see Saraswathi 1974, chap. 4; Ramaswamy 1978: 
297 (Andhra districts). 
56. On communal quotas in public services and educational institutions under the Unionist 
government in the Punjab, see Tinker 1962: 29. 
57. On wartime developments, see Zelliot 1969: 262­65. 
­36­ 
obstructing the war effort, had undergone a radical change in his thinking about untouchability. He 
was now willing to countenance legal compulsion to permit untouchables to assert their rights, and 
envisioned a future constitution in which all forms of untouchability would be abolished by law. He 
no longer espoused personal exclusiveness, but opposed all caste distinctions. He no longer 
defended varna while denouncing caste. Untouchability could be eradicated only if Hindu society 
became casteless. To this end he now encouraged interdining and intermarriage, which he hoped 
would lead to a situation in which "finally there will be only one caste, known by the beautiful name 
of Bhangi, that is to say, the reformer or remover of all dirt."  58.   
When Independence came in 1947, caste was in bad odor. It was widely viewed as an impediment to 
individualism and to broad national loyalties and thus inimical to progress and democracy. The 
hardships inflicted on the lowest castes inspired humanitarian revulsion. It was widely accepted that 
caste would have no place in independent India and that efforts to ameliorate the effects of past 
inequalities were in order. As power passed into Indian hands, the exclusion of untouchables from 
public facilities and from Hindu temples were made statutory offenses throughout most of India. 
Reservations for untouchables were established in the central services, and a program of educational 
assistance was begun. 
Ambedkar was one of the principal architects of the Constitution that emerged from almost three 
years of deliberations.  59.   It is a federal constitution: authority is explicitly apportioned between the 
center and constituent territorial units. There is a greater subordination of the units to the center 
than in the United States, but the states have considerable autonomy and are more than mere 
administrative divisions. It is a written constitution ­­ with the distinction of being the, longest in the 
world. It had to be written in de to be federal ­­ in order to arrange the division of powers between 
the center and the states. But also ­­ and this accounts for its length ­­ because there were a number 
of other matters which the constitution­makers wished to put beyond the reach of temporary or 
narrow majorities. These include safeguards for various minorities and a large part of the 
fundamental 
____________________ 
58. On development of Gandhi's thinking on caste, see Gandhi 1954b: 72­85; Dalton 1967: 171­74. 
59. Ambedkar's hopes of leading an independent political party in which untouchables played a 
central role foundered in the 1945 provincial elections as they had in the 1937 elections. Through 
the good offices of the Muslim League, he became a member of the Constituent Assembly, 
which became the governing body of India when the country was divided and given 
independence on August 15, 1947. In a gesture of reconciliation, Prime Minister Nehru appointed 
him Law Minister, and he was chosen as chairman of the Drafting Committee of the Constituent 
Assembly. 
­37­ 
structure of the government apparatus. Thus, detailed provisions regarding the judiciary, the civil 
service, and the proceedings of legislatures, adopted without fundamental change from the British 
administration, are all entrenched in the Constitution. There is a written and enforceable set of 
Fundamental Rights ­­ modelled in large measure after the United States Bill of Rights. There is a 
set of Directive Principles of State Policy, which specify the general goals of governmental policy ­­ 
the right to work, higher living standards, free and compulsory education, etc. ­­ and anticipate a 
welfare state which was described by one skeptical observer as "Fabian socialism without the 
socialism."  60.   Finally, judicial review is explicitly provided: the Constitution is to be interpreted and 
its provisions enforced by a unitary national judiciary. 
The chapter on Fundamental Rights contains a number of familiar features: prohibitions of 
governmental restriction of the freedoms of religion, speech, press, association, and movement; 
minimal standards of criminal justice; protections for property ­­ all analogous to their American 
counterparts. The draftsmen, cognizant of the difficulties in interpretation and in application 
encountered in America and Australia, specified exceptions and limitations to each right. The listing 
of exceptions and the deliberate omission of a "due process" clause reflect an attempt to reduce the 
scope for judicial discretion in determining the extent of these rights. In general, the draftsmen 
wished to confine judicial review to the application of express principles of law and tried to avoid 
giving the courts a role of continuous constitution­making. 
The Fundamental Rights also contain bans on discrimination on grounds of religion, caste, sex, 
language, etc. The attempt to secure equality goes beyond guarantees of equal treatment by 
government: crucial provisions regulate not only "state action" but also the behavior of private 
persons and groups. Caste discrimination by government or by private individuals is banned. But it 
was recognized that evenhanded treatment would not suffice to overcome the cumulative 
disadvantages of those at the bottom of the social hierarchy. To promote the advancement of the 
untouchables, tribals, and other "socially and educationally backward classes," the Constitution 
allows protective or compensatory discrimination in their behalf. The notion of communal quotas 
and safeguards which had become an accepted and prominent principle of administration during the 
last forty years of British rule was decisively rejected as a general principle of governance. There 
are no safeguards for religious minorities apart from the guaranteed fundamental rights. Only for the 
Scheduled Castes and Scheduled Tribes (and to a lesser extent, for other backward classes) are these 
devices 
____________________ 
60. Jennings 1953: 31. 
­38­ 
countenanced. There is a deliberate departure from formal norms of equality in order to offset the 
historic inequalities of these groups. 
Ambedkar's role in the Constituent Assembly underlines the decisive rejection of Gandhian notions 
of village self­rule. While Gandhi idealized village India and its handicraft technology, Ambedkar 
despised Indian villages as "dens of ignorance, narrow­mindedness and communalism"  61.   and 
ardently supported machine technology which would provide leisure, cultural advancement, and 
finally equality. To the same end, he rejected Gandhi's ideal of trusteeship by the rich in favor of a 
kind of state socialism which would promote rapid industrialization. And in spite of his suspicion 
that India might require benevolent autocracy, he was a supporter of centralized parliamentary 
government rather than of village autonomy. In all of this Ambedkar stood closer to the left wing of 
the Congress than either stood to Gandhi. 
The Constitution incorporated much of the secular program he had championed, but Dr. Ambedkar 
was dissatisfied with the performance of the Congress government. After a bitter political 
disappointment in the 1952 election, he disowned his role as designer of the Constitution. 62. 
Throughout his career, Dr. Ambedkar had written on the historical aspects of the caste system and 
displayed a strong interest in Buddhism, which he considered a religion of equality and social 
reform. This was reinforced by his belief that the untouchables were former Buddhists who had been 
degraded by the Hindus for refusing to give up Buddhism and for continuing to eat beef. 63. 
Ambedkar's interest in Buddhism intensified after Independence, and he dedicated himself to the 
revival of Buddhism in India. After his political disillusionment was complete, he returned to the 
notion that the problems of untouchability could be solved only by the abandonment of Hinduism. 
In 1955, a year before his death, he launched a mass movement for conversion to Buddhism that 
inspired millions of untouchables to abandon Hinduism 
____________________ 
61. VI CAD 39. 
62. Dr. Ambedkar's bitter disappointment is understandable, but in retrospect and from afar, his 
sense that the constitutional protections which he designed or at least concurred in were a failure 
seems less than fully warranted. The untouchables, unlike the Muslims, were a scattered minority 
with no territorial base, no unity among themselves, no rallying point in religion or historical 
glories, and with little capacity and fewer resources for organization. If ultimately they had little 
choice but to accept what they were given by the majority, their leadership did well by them in 
obtaining safeguards and preferences out of the independence struggle. Although their 
constitutional position might have been even stronger, the major shortfall has been not in 
constitutional protection but in organization and imagination to utilize effectively the protections 
and opportunities that are available. 
63. Ambedkar 1948: 76, 80­83. The idea that Untouchables or low castes were former Buddhists had 
gained popularity among a scattering of low caste groups ( Zelliot 1969: 231 and n. 83). It had 
been expounded earlier by Colonel Olcott of the Theosophical Society and others. 
­39­ 
and embrace Buddhism.  64.   A generation later, the Buddhists were a significant element of both 
political militancy and intellectual leadership; and Dr. Ambedkar was the venerated "patron saint" 
of the untouchables.  65.   
The wrestling of Gandhi and Ambedkar displays an ironic reversal. Gandhi, for decades the 
embodiment of the evangelical approach of repentance and personal growth, acknowledged the need 
for external controls and eventually adopted all the planks of the secular reform platform 
championed by Ambedkar. Dr. Ambedkar, on the other hand, lost faith in the efficacy of secular 
political reform without a cultural transformation with its roots in new (or restored) religious 
identity. 
What emerged was a national commitment to reduce invidious distinctions among groups, coupled 
with a conviction that it was permissible and perhaps necessary to employ these discredited caste 
notions to effectuate the equalizing policies that we survey in the following chapters. We shall then 
return to explore the evolution of these categories and to trace their career as instruments of the 
equalizing policies that we call compensatory discrimination. 
____________________ 
64. The number of Buddhists reported by the census increased from 180, 823 in 1951 (of whom only 
2,500 were in Maharashtra) to 3,250,327 in 1961. Of these, 2,789,501 were in Maharashtra 
( Census of India 1961; Paper No. 1 of 1963). Zelliot ( 1969: 239) estimates that 80% of the 
Mahar caste reported themselves as Buddhists in 1961. The 1971 census numbered 3,874,942 
Buddhists. These figures represent a considerable understatement: many who regard themselves 
as Buddhists would not publicly identify themselves as such for fear of losing educational and 
other benefits conferred on Scheduled Castes. See chap. 9, §E, below. 
65. Although respected as a "founding father," Ambedkar is the prime symbol of self­assertion by 
low castes. The renaming of Marathwada University in his honor in 1978 touched off widespread 
rioting throughout that region of Maharashtra. Resentment of reservations and concessions for 
untouchables was given murderous expression; Mahars were especial targets of attack ( 
Morkhandikar 1978 ; Abraham 1978 ). 
­40­ 
3 Compensatory Discrimination Programs and Their Implementation 

INDIA'S POLICY OF compensatory discrimination comprises a wide array of preferential schemes. 
In this chapter we will survey some of the most prominent programs, consider some general features 
of their administration, and explore the question of how they should be assessed. 
These programs are authorized by constitutional provisions that permit departure from formal 
equality for the purpose of favoring specified groups. We shall later examine closely the meaning of 
these provisions and the tensions between them and other constitutional commitments. For the 
present it may be helpful to identify them as the constitutional anchorage of the various policies 
described here. 
Article 46, a "Directive Principle of State Policy," declares: 
The State shall promote with special care the educational and economic interests of the 
weaker sections of the people, and, in particular, or the Scheduled Castes and the 
Scheduled Tribes, and shall protect them from social injustice and all forms of 
1. 
exploitation.    
The Constitution's ban on discrimination in government employment is qualified by Article 16(4), 
which permits the State to make 
any provision for the reservation of appointments or posts in favour of any backward 
class of citizens which, in the opinion of the State, is not adequately represented in the 
2. 
services under the State.    
____________________ 
1. The Directive Principles of State Policy, contained in Part IV of the Constitution, are not 
themselves justiciable, but the Constitution prescribes it as "the duty of the State to apply these 
principles in making laws" (Art. 37.) The jurisprudential status of the Directive Principles is 
discussed below in chap. 11. 
2. In addition, Article 335 provides: "The claims of the members of the Scheduled Castes and the 
Scheduled Tribes shall be taken into consideration, consistently with the maintenance of 
efficiency of administration, in the making of appointments to services and posts in connection 
with the affairs of the Union or of a State." The relation of these provisions is discussed in chap. 
11, §B, below. 
­41­ 
The general provisions banning discrimination by government (Art. 15) and banning discrimination 
in government­aided educational institutions (Art. 29 [2] ) are similarly qualified by Article 15(4), 
which provides: 
Nothing in Article 15 or Article 29(2) . . .shall prevent the State, from making any 
special provision for the advancement of any socially and educationally backward 
classes of citizens or for the Scheduled Castes and the Scheduled Tribes. 
There are specific provisions that we shall encounter later. But these, along with specific provisions 
for reservation of seats in legislative bodies (which will be discussed shortly), are the core of the 
constitutional commitment to compensatory discrimination. 
The benefits of "compensatory discrimination" are extended to a wide array of groups. There are 
3. First, there are those castes designated as Scheduled Castes (SC) on the basis 
three major classes.    
of their "untouchability." They numbered nearly 80 million (14.6% of the population) according to 
4. Second, there are the Scheduled Tribes (ST), who are distinguished by their 
the 1971 census.    
tribal culture and physical isolation and many of whom are residents of specially protected 
Scheduled Areas. They numbered more than 38 million (6.9% of the population) in 1971.     5. Third, 

there are the "Backward Classes" (or, as they are sometimes called, "Other Backward Classes,")  6.   a 
heterogeneous category, varying greatly from state to state, composed for the most part of castes 
(and some non­Hindu communities) low in the traditional social hierarchy, but not as low as the SC. 
Also included among the Other Backward Classes (OBC) are a few tribal and noma­ 
____________________ 
3. Temporary preferences for yet another group, the Anglo­Indians, were not based on their 
backwardness but were designed to cushion the loss of the privileged position they had formerly 
enjoyed. Reserved posts in certain services and special grants to Anglo­Indian educational 
institutions were guaranteed by Arts. 336(1) and 337. These preferences were on a diminishing 
scale and ceased to be operative on January 26, 1960. Reservation for Anglo­Indians in the 
services has ceased. The states have, with some reductions, continued to make grants to Anglo­
Indian schools. Since the constitutional grants were not merely authorized but guaranteed, there 
was no occasion for this group to be treated under the general provisions the preferences, Arts. 
15(4) and 16(4). On the limits of state power to place conditions on the recipients of these grants, 
see Bombay Education Society v. State of Bombay, A.I.R. 1954 S.C. 561; In re Kerala Education  
Bill, 1957, A.I.R. 1958 S.C. 956. The Constitution not only provides a formula for determining 
the amount of grants and reservations, but also defines membership in the Anglo­Indian 
community (Art. 366[2]). On the status of Anglo­Indians and the background and effects of the 
Bombay Education Society case, see Grimshaw 1959. 
4. On the selection of these groups, see chap. 5, §A, below. 
5. On these groups, see chap. 5 §C, below. 
6. On the selection of these groups, see chaps. 6, 7, 8, below. 
­42­ 
dic groups, as well as converts to non­Hindu religions from the SC, and in some areas the 
7. It has been estimated that there were approximately 60 million persons under 
Denotified Tribes.    
the Other Backward Classes heading in 1961     8. ­­ roughly the magnitude of the SC population at 
that time (64 millions). (Today the portion of the population designated under this heading is 
probably larger.) 
For the most part, preferences have been extended on a communal basis. Members of specified 
communities are the beneficiaries of a given scheme, and all members of the community, however 
prosperous, are entitled to the benefits. However, some schemes use a means test to supplement the 
communal on ­­ only members of the listed communities with incomes below the specified ceiling 
are eligible. In a few instances, the communal test has been replaced by an economic one ­income or 
occupation or a combination of the two ­­ and a few schemes use tests neither communal nor 
economic.  9.   
Preferences are of three basic types: First, there are reservations, which allot or facilitate access to 
valued positions or resources. The most important instances of this type are reserved seats in 
legislatures, reservation of posts in government service, and reservation of places in academic 
institutions (especially the coveted higher technical and professional colleges). To a lesser extent, the 
reservation device is also used in the distribution of land allotments, housing, and other scarce 
resources. Second, there are programs involving expenditure or provision of services ­­ e.g., 
scholarships, grants, loans, land allotments, health care, legal aid ­­ to a beneficiary group beyond 
comparable expenditures for others. Third, there are special protections. These distributive schemes 
are accompanied by efforts to protect the backward classes from being exploited and victimized. 
Forced labor is prohibited by the Constitution,  10.   and in recent years there have been strenuous 
efforts to release the victims of debt bondage, who are mostly Scheduled Castes and Tribes. 
Legislation regulating money lending, providing debt re­ 
____________________ 
10. Art. 23(1). 
7. "The Denotified Trbes, or Vimukta Jatis, are the former Criminal Tribes. They became 'ex­
criminal' when the Criminal Tribes Act of 1924 was repealed in 1952. Their population is 
estimated at upwards of 40 lakhs [4 million], of whom more than half live in [Uttar Pradesh.] 
Although U.P. lists several of them as Scheduled Castes, most States count them all as Other 
Backward Classes. They pose special psychological, social and economic problems, however, 
and have come to be treated as a separate fourth category for administrative and budget 
purposes" ( Dushkin 1961 : 1665). 
8. Dushkin 1961 : 1665. 
9. Economic tests are discussed in chap. 8, §G. below. For an example of a "residence" criterion, 
see the Home Ministry's grants for all persons in selected "undeveloped areas" (RCSCST 
1959­60: I, 323). See chap. 8. §H, below, on geographic criteria. 
­43­ 
lief, and restricting land transfers attempt to protect SC and ST from the economic oppression of 
their more sophisticated neighbors. Antiuntouchability propaganda and the Protection of Civil 
Rights Act attempt to relieve untouchables from the social disabilities under which they have 
suffered.  11.   This legislation is not "compensatory discrimination" in the formal sense of departing 
from equal treatment to favor these groups; it enjoins equal treatment rather than confers preferential 
treatment. But in substance it is a special undertaking to remedy the disadvantaged position of the 
untouchables. Unlike the measures discussed above, these are regulative rather than distributive; 
they involve different claimants in different settings; they are implemented by different sets of 
officials ­­ e.g., by police and magistrates rather than by administrators. There are important 
connections and perhaps important parallels. But the "protections" involve a whole world of 
different problems and richly deserve detailed study in their own right. We shall take up these 
protections here only as they intersect with compensatory programs of the first two types. 
To make the present survey manageable we shall concentrate on several of the most prominent 
programs­­reserved seats in legislatures and preferential treatment in education (including 
reservations in professional colleges). Other programs and some general features of their 
administration will be noted briefly, and the problem of assessing their performance will be 
broached. The following chapter presents a more fully developed "case study" of preferential 
treatment in the area of government employment. 
A. LEGISLATIVE RESERVATIONS 
The most prominent of all preferential policies is the reservation of seats in elective legislative 
bodies. The Constitution specifically provides reserved seats in proportion to their numbers for the 
Scheduled Castes (SC) and the Scheduled Tribes (ST) in the Lok Sabha (lower house of Parliament) 
 12.   and the Vidhan Sabhas (lower houses of the state legislatures).  13.   No seats are reserved in the 
upper houses, central or state. There are no reservations in legislatures for the Other Backward 
Classes or for any minority goups.  14.   
____________________ 
11. The Protection of Civil Rights Act is the new name given to the Untouchability Offences Act (Act 
22 of 1955) upon its amendment by the Untouchability (Offences) Amendment and 
Miscellaneous Provisions Act (Act 106 of 1976). For a review of the background of this 
legislation and its operation into the mid­1960s, see Galanter ( 1972a). 
12. Article 330. 
13. Article 332. 
14. The Constituent Assembly definitively rejected political safeguards for religious and other 
minorities. See Austin 1966 : 146­56. The Constitution does, however, 
­44­ 
All of the other constitutional provisions for preferences are merely authorizations empowering the 
State to make special provision for disadvantaged groups. But legislative reservations are 
specifically provided in the Constitution itself.  15.   Seats are reserved in proportion to the population 
of SC and ST to the total population of each state. Thus in 1976, 78 (14.4%) of the 542 seats in the 
Lok Sabha were reserved for SC, and 38 seats (7%) for ST. Of the 3,997 seats in the Vidhan Sabhas, 
540 (13.5%) were reserved for SC, and 282 (7%) for ST.  16.   
These reservations do not involve "separate electorates" ­­ i.e., the representation of a particular 
group by legislators chosen by an electorate composed solely of members of that group.  17.   The 
seats are "reserved" in the sense that candidates who stand for them must belong to the specified 
groups, but the entire electorate participates in choosing among candidates so qualified. Separate 
electorates for Parliament and the state legislatures are specifically outlawed by Article 325, which 
provides that no person shall be excluded from any electoral roll on grounds of religion, race, or 
caste.  18.   
The constitutional provision of reserved seats is complemented by statutory provisions to enhance 
political participation by SC and ST. Smaller election deposits are required from members of these 
groups, and candidates for reserved seats enjoy more permissive residence requirements.  19.   
____________________ 
provide for nominated representatives of the Anglo­Indian community in the Lok Sabha (Art. 
331) and in the Legislative Assemblies (Art. 333). On the implementation of these provisions, 
see RCSCST 1975­77: I, 37. 
15. The Constitution contains no similar provision for representation in other political bodies. 
However, such political safeguards may lie within the scope of Art. 15(4), which is apparently 
broad enough to authorize the reservation of seats in elective bodies on a local level. 
16. Election Commission 1976a: 3­6. 
17. By separate electorates is meant the representation of religious (or other) minorities by 
legislators chosen by an electorate composed only of members of that minority. The provision 
and extent of such representation was an extremely troublesome political issue in India during 
the forty years preceding Independence. The Government of India Act, 1909, gave separate 
electorates to Muslims; the Government of India Acts of 1919 and 1935 provided separate 
electorates for Muslims, Sikhs, Indian Christians, and other groups. Proposals to give separate 
electorates to the Scheduled Castes under the 1935 Act were withdrawn after adamant resistance 
by Mr. Gandhi. See chap. 2 above. 
18. The Supreme Court has indicated that Article 15 prohibits communal electorates in local bodies. 
Nain Sukh Das v. State of U.P., A.I.R. 1953 S.C. 384. 
19. Representation of the People Act, 1951 (Act 43 of 1951). The smaller deposits apply in contesting 
non­reserved as well as reserved seats. A proposal by Dr. Punjabrao Deshmukh, leading 
spokesman for the Other Backward Classes, to extend the smaller deposits concession to that 
category elicited a gruff dismissal from Dr. Ambedkar, then the Law Minister: "this is the first 
time that I hear that such a concession should be extended to the Backward Classes. Hitherto the 
concessions that have been spoken of as 
­45­ 
Unlike the general authorization of special treatment and the provision for reserved posts in 
government service, the reserved seats in legislatures are subject to a constitutional time limit. It 
was originally provided that such reservations should expire ten years after the commencement of 
the Constitution.  20.   In 1959 these provisions were extended for another ten years,  21.   in 1969 for 
another ten,  22.   and in 1980 for yet another ten.  23.   
The demarcation of constituencies and designation of those reserved for SC and ST are entrusted to 
a Delimitation Commission.  24.   The courts are constitutionally barred from intervening in the 
process of delimitation or allotment of seats.  25.   Generally, constituencies are formed on the basis of 
contiguity, communications, absence of natural barriers, and cultural and ethnic homogeneity. After 
they are formed, certain constituencies are selected as reserved.  26.   In the case of the 
____________________ 
being necessary for the upliftment of the Backward Classes are educational concessions and 
concessions in the services of the country. . . . [S]o far as the Members representing the 
Backward Classes in this House . . . they are capable of paying not only their own election 
deposits, but also the election deposits of many others." The government refused to accept the 
amendment and it was withdrawn. Parliamentary Debates, Vol. XII­XIII (Part II), Cols. 9141­49 
( 19 May 1951). 
20. Article 334. 
21. Constitution (Eighth Amendment) Act, 1959. 
22. Constitution (Twenty­third Amendment) Act, 1969. 
23. Constitution (Forty­fifth Amendment) Act, 1980. 
24. The Delimitation Commission consists of a retired Supreme Court judge, a sitting High Court 
judge, and the Chief Election Commissioner ( Delimitation Act, 1972 [Act 76 of 1972]). In each 
state, the Commission is assisted by ten non­voting associate members (5 Members of Parliament 
and 5 Members of the Legislative Assembly) nominated by their respective Speakers (§5). The 
associated members from each state are consulted after tentative delimitation. After changes at 
this stage, the proposed delimitations are published and objections invited. Where there are 
objections, hearings are held. The location of reserved seats is the most frequent source of 
objections, but they are rarely shifted. 
25. Article 329. 
26. During the 1950s, most of the reserved seats were in double­member constituencies, in which 
one of the seats could be filled only by a member of the "privileged" group and the other was 
open to general competition. Thus in 1952, all of the Scheduled Caste seats and half of the 
Scheduled Tribe seats in the Lok Sabha were double­member constituencies. In 1957, 467 out of 
470 SC seats in the Vidhan Sabhas and 115 of the 221 seats reserved for ST were also in double­
member constituencies. (On the functioning of the double­member constituencies and their 
abolition, see Dushkin 1972 : 189 ff.) 

Double­member constituencies were abolished in 1961 (the Two Member Constituency 
[Abolition] Act, 1961 [Act 1 of 1961]). This aboliton seems to have been inspired principally by 
the concern of politicians with the greater difficulty and expense of campaigning in a double­
sized district. This was compounded by some resentment of the infrequent but well­publicized 
instances in which the unreserved "general" seat as well as the reserved seat was won by SC or 
ST candidates, who were able to compete for both seats, while their opponents could compete for 
only the unreserved seat. This did not happen very often. In nearly two hundred double­member 
constituency contests for Lok Sabha seats in 1952 and 1957, the general seat was won by a 
reserved­seat 

­46­ 
Scheduled Tribes seats there is a single standard for selection ­concentration of the ST population. 
 27.   Since typically the tribal population lives in areas of concentration, the constituencies reserved 
for them contain a high proportion of ST.  28.   In 1962, the ST constituted more than 50% of the 
population in 18 of the 31 Lok Sabha constituencies reserved for them.  29.   Conversely, most of the 
major concentrations of ST population are located in reserved constituencies. Of the 29 
constituencies in which ST make up over 40% of the population, 23 are reserved. Thus the reserved 
seats encompass the bulk of the ST population ­­ perhaps close to 70%. These constituencies are, 
not surprisingly, more isolated and much less urban than are the general constituencies. 
In the case of seats reserved for Scheduled Castes, the Delimitation Commission is instructed to 
apply two standards­­ concentration of SC population and dispersal of reservations. The 
Delimitation Act instructs the Commission to locate seats reserved for SC "in different parts of the 
State and . . . as far as practicable, in those areas where the proportion of their population to the total 
is comparatively large."  30.   The Delimitation Commission achieves this by first reserving the 
constituencies with the highest concentration of SC ­­ usually not much more than 20% ­and then 
selecting the others by the dispersal criterion. Since the SC population, unlike the ST, is rather 
evenly dispersed, the constituencies reserved for them contain far fewer SC members. Among the 76 
Lok Sabha constituencies reserved for SC in the 1962 elections, there was 
____________________ 
candidate in one in 1952 and in five in 1957. It was apparently even less frequent in the legislative 
elections, as indicated by the very small number of non­reserved seats occupied by SC and ST. A 
well­known 1959 Supreme Court decision upheld the right of an ST candidate to take the 
unreserved seat, sealing the defeat of the Labour Minister ( V.V. Giri v. D. Sura Dora, A.I.R. 
1959 S.C. 1318, discussed below in chap. 13,§ B). This decision triggered the move for abolition 
( Dushkin 1972 : 192). New single­member constituencies were designated for the 1962 
elections. In most cases this was done simply by dividing the double­member constituency and 
allotting the reserved seat to the one with the greatest proportion of SC or ST in its population 
( Election Commission 1966 : I, 8). 
27. §9(1) (d). 
28. In the following pages, comparisons of the reserved­seat holders in the Lok Sabha with the 
occupants of general seats are based on data made available to me by Professor Henry C. Hart of 
the University of Wisconsin. These data consisted of (1) coded cards on all 497 members of the 
Third Lok Sabha from 1962 to 1965 and (2) coded cards based on his interviews with 189 
members of the Lok Sabha in 1963­64. Of the sample, 108 were a random sample (30%) of 
Congress members. The remaining 79 were 60% of the opposition and independent members. 
The sample of the latter is not random, but is weighted slightly toward those active in the affairs 
of the Lok Sabha. In the sequel a source given simply as "Hart Data" refers to the larger set of 
cards on all members; the smaller set is referred to as the " Interview Sample." The Hart data is 
analyzed in more detail in Galanter 1979. Other data from the Hart study are presented in Hart 
1971. 
29. Krishna ( 1966 : 28) reports 19. 
30. § 9 (1) (c). 
­47­ 
none in which they were a majority of the population and only 13 in which they were as much as 
30%.  31   The bulk of SC constituencies contain be en 10% and 30% of Scheduled Castes. Thus, the 
vast majority of SC ­­ perhaps 75% ­­ live outside the constituencies reserved for SC. Of the 111 
Lok Sabha constituencies where SC constituted more than 20% of the population in 1961, only 46 
were reserved, whereas 29 constituencies with less than 20% were reserved. On the whole, 
constituencies reserved for SC tend to be political backwaters ­­ slightly less urban, with less 
newspaper circulation and a slightly greater percentage of agricultural laborers. 
The same pattern of concentration of Scheduled Tribes and dispersion of Scheduled Castes 
reappears in the reserved seats in the state legislatures. In Andhra Pradesh, for example, SC make up 
only 21% of the population of the constituencies reserved for them (and 13% of the population of 
the general constituencies). But ST make up 43% of the population of the constituencies reserved 
for ST and only 2% of the general constituencies.  32.   
Designation of a constituency as reserved usually encounters opposition. 
While everyone expressed anxiety that the Scheduled Castes and the Scheduled Tribes 
should get their full quota of reserved seats, very rarely indeed would anybody (unless 
he himself belonged to any of these castes or tribes) agree to any seat being reserved for 
them in a constituency in which he was personally interested.  33.   
The selection of reserved constituencies gives rise to the bulk of the objections to the proposed 
delimitation. As the Election Commissioner reports, 
ģerally speaking the vocal sections of the public objected to their own areas becoming 
part of a reserved constituency and put forward ingenious suggestions for reshaping two 
or more constituencies so that some particular constituency would be a general 
constituency.  34.   
However, these objections rarely bring about a change. The courts will not provide any relief so long 
as the Delimitation Commission has taken into account the factors specified by the Deliimitation 
Act.  35.   
____________________ 
31 My figures are taken from the Hart data. Krishna ( 1966: 28) gives slightly different figures, but 
the overall picture is not very different. 
32. Election Commission 1976b: II­A, 265. Inspection of the Election Commission tables reveals 
that Andhra is not atypical in this regard. 
33. Election Commission 1959: I, 57. For an instance of alleged manipulation in the opposite 
direction, see Saberwal 1972 : 74. 
34. Election Commission 1968: I, 12. 
35. Mastanaiah v. Delimitation Commissioner, A.I.R. 1969 A.P. 1. The Court rejected the argument 
that the Delimitation Commission was required to locate a reserved constituency in an adjoining 
district with a slightly higher percentage of Scheduled Castes. 
­48­ 
Proposals for rotation of reserved constituencies have been rejected, and their location has remained 
relatively stable.  36.   
Of course, Scheduled Caste and Tribe candidates may stand for non­reserved seats. However, they 
have been notably unsuccessful in winning elections to them. In the first six Lok Sabhas, only a 
handful of candidates from SC or ST filled unreserved seats. Five were elected in 1971 and three in 
1977.  37.   In the lower houses of the state legislatures, the record is no more heartening, even though 
the smaller constituencies mean that local concentrations of population and resources should offer 
more opportunities for political success. For example, in 1970­71, there were only three SC and two 
ST representatives sitting in the 2,853 non­reserved seats in the Vidhan Sabhas.  38.   
There is no reservation of seats in the indirectly elected upper houses at both Centre and states. The 
Rajya Sabha at the Centre is elected by proportional representation by the members of the state 
Legislative Assemblies.  39.   The upper houses (Vidhan Parishads) in those state legislatures that have 
them are chosen by a more complicated formula, combining election by the lower house, special 
qualified constituencies, and governmental nomination.  40.   Here, too, the number of SC and ST has 
remained fairly small. In 1969­70, there were 13 SC and 5 ST among the 228 members of the Rajya 
Sabha, and 18 SC and 9 ST among some 703 members of Vidhan Parishads.  41.   
There is no constitutional requirement or statutory provision for reservations of political 
appointments within the legislatures or the government, such as cabinet ministerships  42.   or 
membership on standing committees. However, it early became a convention to have at least one SC 
cabinet minister at the Centre and in each of the states. The number of SC ministers at the Centre 
and the number of ST ministers in the states has slowly risen. 
____________________ 
36. The Election Commissioner characterized a proposal for rotation (on the ground that no area 
should be permanently burdened with a reserved seat) as "mere lip service to the principle of 
reservation" and concluded that the "real purpose of such reservations would be largely defeated 
if the proposal be adopted" ( Election Commission 1959:I, 75). In the 1969 debate preceding 
extension for the third ten­year period, an amendment proposing rotation and supported by a 
disingenuous plea to afford opportunities for election to Harijans in other localities was negatived 
without a division of the House (LSD (4th Series), Vol. 34, No. 17, Cols. 361, 379 [9 Dec. 
1969]). 
37. RCSCST 1971­73:244; 1975­77:I, 20­21. 
38. RCSCST 1970­71:122. 
39. Article 80. 
40. Article 171. It should be noted that in only seven of the larger states is there a second chamber 
(Art. 168). 
41. RCSCST 1969­70:92. 
42. Article 164(1) provides that the states of Bihar, Madhya Pradesh, and Orissa shall have a minister 
in charge of tribal welfare who may in addition be in charge of Scheduled Castes and Backward 
Classes or any other work. 
­49­ 
Reservation of seats on elective bodies at the local level is not specifically provided by the 
Constitution, but such arrangements are allowed by Article 15(4), and nearly all the states provide 
such reservations by election, co­optation, or nomination in various local bodies and panchayats. 
Similarly, most of the states have made some provision for representation of SC and ST in panchayat 
bodies. There is little data available on the extent to which these reservations have been 
implemented, much less on how effective they are. 
We may safely conclude that the presence of the Scheduled Castes and Tribes in legislative bodies is 
accounted for largely by the provisions for reserved seats. Very few members of these groups have 
been successful in open competition. Even their presence in bodies where there is no reservation 
(upper houses, cabinets) is due largely to the political base provided by reservations. Thus 
reservations do provide for a substantial quantitative presence that would otherwise be lacking. 
We cannot infer that without reservations, representation would be reduced to the present level of 
success in non­reserved seats. Their geographical concentration would assure the Scheduled Tribes 
at least a modicum of representation even in the absence of reservations. The more dispersed 
Scheduled Castes would undoubtedly obtain far fewer seats than they now enjoy, but we should not 
conclude that they would have none at all. If there were no reservations, some of the political talent 
among the SC now competing for reserved seats would compete for non­reserved seats. Parties 
would then need appealing candidates in areas with heavy concentrations of SC; perhaps slate­
makers would be less moved by the sense of "they already have their share" that now seems to affect 
the giving of tickets. (And of course, some representation would be provided by legislators from 
other groups, who would be accountable to constituencies with sizable numbers of SC voters.) 
Perhaps the best indication of the scale of representation that might be expected in the absence of 
reserved seats is provided by some data on elections to local bodies where no reservations were in 
force. In the early 1960s local bodies in West Bengal with a total of 1,081 members contained only 
41 SC members (3.8%) and 16 ST members (1.5%).  43.   Among the 66 presidents and chairmen there 
were 3 SC members and I ST. (According to the 1960 census, 19.84% of the population of West 
Bengal were SC and 5.91% were ST.)  44.   A similar picture is presented by some fragmentary data 
from Gujarat where only 35 (0.5%) of 6,863 
____________________ 
43. RCSCST 1963­64: I, 163. 
44. The population figures here are somewhat overstated, since SC and ST are less urban than the 
general population. But offsetting this, the small territorial constituencies ought to give these 
groups an advantage. Also, Bengal is reputed to be less resistant to the mobility of these groups 
than many other areas. 
­50­ 
Sarpanches were from the SC.  45.   The electoral experience of India's other large and (mostly) low­
status minority ­­ the Muslims ­­ does not offer additional reassurance.  46.   
The effect of reservation of seats on the quality of representation is more difficult to assess. How 
well served are the residents of the reserved constituencies? And, apart from them, how well served 
are the members of Scheduled Castes and Scheduled Tribes ­­ a great majority of the former ­­ who 
reside outside the reserved on constituencies? In what sense do the reserved seat legislators 
represent them? Does the presence of their fellows in reserved seats assure that they are better 
served by their own representatives? Finally, is the nation as a whole well served by these 
legislators? What effect do they have on the working of the legislatures? 
In assessing the quality of representation afforded by these representatives, we must recall the 
structural constraints imposed by the arrangements for reserved seats. The reservations are designed 
in a way which filters and muffles the representation of the interests of the Scheduled Castes and 
Tribes insofar as the diverge from the interests of other groups. The reserved­seat legislator ­­ 
especially from the SC ­­ is responsible to and dependent upon a constituency made up 
overwhelmingly of non­members of his group. To the extent that these reserved­seat legislators are 
assumed to be responsible and prudent politicians, this arrangement serves as a filter to keep the 
divergent interests of these groups from being expressed too directly or forcefully. This is especially 
the case with the SC representatives, whose constituencies are populated overwhelmingly by others. 
From a handful of studies we may draw a composite portrait of the Members of Parliament who 
occupy reserved seats.  47.   They are younger and more rural in residence than their fellow legislators; 
fewer have professions (other than landholding), and they have less formal education. The median 
age of reserved seat M.P.'s is in the early forties; that of their fellows in general seats is in the early 
fifties. The disparity in age has remained constant over successive Lok Sabhas, but the gap in 
educational attainments has narrowed appreciably. Those in reserved 
____________________ 
45. RCSCST 1969­70:51 
46. The percentage of Muslims (who make up about 11% of the population of India) in legislative 
bodies is small and declining. Muslim membership declined from 7.2% of the Lok Sabha in 1952 
to 4.6% in 1971 ( Gauba 1973:245). Muslim participation in the State Assemblies is at roughly 
the same level and has been undergoing a similar decline ( Gauba 1973:262­65; Krishna 
1967:187). 
47. My account is drawn from the comparisons of the occupants of reserved and general seats found 
in Chauhan and Chopra 1969 ; Dutta 1969 ; Narayana 1978; Mohopatra 1978. These accounts 
comport with the profile of Scheduled Caste M.P.'s in Parvathamma and Jangam 1969. 
­51­ 
seats are younger when they are first elected to Parliament, and, unsurprisingly, come there with less 
legislative experience at the state level.  48.   We may surmise that similar disparities might be found 
in state legislatures.  49.   
It is widely believed that reserved­seat legislators are less articulate, less assertive, and less 
independent than their fellows. Data on the Third Lok Sabha collected by Professor Henry Hart 
supplies some evidence for the view that reserved­seat holders were as a group less active and less 
influential parliamentarians than those elected to general seats.  50.   They were less likely to hold 
party office, to serve on select or standing committees, to hold ministerships, to contribute to 
debate, and to talk to ministers. Few, compared with holders of general seats, felt that they were 
effective advocates of their views.  51.   One may, of course, wonder whether the most important sorts 
of representation are captured by these measures of participation.  52.   
Obviously, one must be extremely cautious in extrapolating from this data about the Third Lok 
Sabha to more recent performance under changed conditions, much less to conditions that might 
prevail in the future. Since the reserved­seat legislators have come to resemble their fellows in terms 
of education, occupation, and urban residence, there may also be a convergence in legislative 
behavior. Or it may be that changes in political alignments have accorded reserved­seat legislators a 
greater voice and elicited new patterns of participation. 
Low rates of participation do not imply a proportionate absence of effective representation of the 
interests of Scheduled Castes and Tribes. Presumably the activities of reserved­seat legislators are 
concentrated 
____________________ 
48. Mohopatra 1978. 
49. Mohopatra ( 1973) useful study of the Orissa Legislative study in 1969 (n=85 of 139) found that 
SC and ST legislators were younger and less educated, came from more agricultural 
backgrounds, and that fewer of them had independent professions. 
50. The data referred to here is that described in n. 28 above. It would of corse be illegitimate to take 
this picture as reflective of current conditions. And it should also be emphasized that we would 
be on shaky ground generalizing from data on Parliament to conclusions about the working of 
reserved seats in the state legislatures. Nevertheless, the data does enable us to see the legislative 
reservation device in action, and it may provide a useful base line for comparison with later or 
state data. 
51. The data on which these conclusions are based are presented in Galanter 1979. 
52. They provide no measure, for example, of the vision or sustained energy whose absence is 
marked by Saberwal ( 1972: 7) when he notes that the Harijans in the reserved constituency that 
he studied had "no leaders who understood the larger institutional mechanisms which underlay 
and renewed the current inequalities, nor any who had the stamina for changing the larger system 
so that its proceses would generate effective . . . equality." He does not argue that this absence is 
produced by the reservation of seats, but such an argument is made elsewhere in connection with 
criticism of the dependence of reserved­seat legislators. 
­52­ 
on matters that affect these groups,  53.   and their effectiveness in representing the interests of these 
groups should be most evident in their dealing with matters that contribute directly to their welfare: 
(1) establishing programs of which they are the beneficiaries; (2) insuring deployment of resources 
to these programs; (3) pressing for better administration of these programs; (4) resisting cut­backs 
and premature termination of such programs. Although I have no systematic data, there is some 
evidence to suggest that, in the Lok Sabha at least, reserved­seat legislators have been prominent 
and sometimes influential in matters concerning the working of the preference policy.  54.   
The presence of the reserved­seat legislators, if not in itself likely to raise the level of benefits, does 
seem to serve as a warrant of the continuation of these programs at their existing levels. 
Furthermore, their presence in such numbers provides a quantitative basis for Scheduled Caste and 
Tribe participation in leadership at the cabinet level; it is unlikely that there would be many persons 
from these groups in cabinets were it not for the reserved seats. It appears that the initiation and 
expansion of programs for these groups is associated with the presence of group members in cabinet 
posts. The first reservations in government service were established by the efforts of the first 
Scheduled Caste member of the Viceroy's cabinet. Job reservations were first extended from initial 
appointments to the promotion stage under the auspices of the next SC politician to become a 
minister. Similarly, at the state level, the first extension of reservations to the 
____________________ 
53. Kochanek ( 1968:48) found Harijan M.P.'s in 1960 among the strong supporters of land ceilings 
and urban income ceilings and generally giving "higher than average support for any process of 
leveling." But cf. Mohapatra ( 1973:22) finding that reserved­seat holders in Orissa displayed no 
overwhelming concern for these groups. The pattern in state legislatures may differ sharply from 
that in the Lok Sabha, as suggested by Dushkin 1972:210. 
54. Narayana ( 1980) describes the heavy participation on such issues by Scheduled Caste 
representatives in the Lok Sabha from 1962 to 1971 and their success in securing the 
appointment of the Elayaperumal Committee and the standing Joint Committee on Scheduled 
Castes and Scheduled Tribes. The reserved­seat legislators demonstrated their political prowess 
by killing the 1965 Lokur Committee report, which called for the de­scheduling of some tribes 
and castes (see chap. 5, § B, below). The dissatisfaction of the reserved­seat members was 
instrumental in bringing about the Government's first defeat in the Lok Sabha in 20 years, a 
stunning snap vote on a 1967 motion that the constitutional safeguards for SC and ST were not 
being fully implemented. See Dushkin 1972:206­8, who notes that this new assertiveness was 
followed by government's repudiation of the Lokur recommendations, appointment of new 
deputy ministers from among the Scheduled Castes, liberalization of the rules for reserving 
promotion posts in government services, increasing the reservations in direct recruitment, and 
(after years of promises) making converts to Buddhism and Christianity eligible for 
postmatriculation scholarships. 
­53­ 
promotion stage occurred when a Scheduled Caste politician became chief minister of Andhra; the 
ascension of Scheduled Caste Home Minister in Madras was followed by a dramatic increase in the 
vigor of prosecution under the Untouchability (Offences) Act.  55.   
The potential leverage enjoyed by the occupants of reserved seats varies with the balance of party 
power in the legislatures. The more evenly balanced the party alignment, the greater the potential 
impact of the reserved seats.  56.   
Of course, representation is not the only virtue claimed for reserved seats ­­ and misrepresentation is 
not the only sin attributed to them. At the end of this chapter I shall sketch out the range of claims 
about costs and benefits that might be made in connection with compensatory discrimination 
programs. For example, do these schemes produce acceptance of Scheduled Castes or do they lead 
to rejection of them? Do they promote or retard the self­sufficiency and self­regard of the 
beneficiaries? Optimally, we should like to be able to answer such questions for each preferential 
program.  57.   The data to do this are not available. In any event, it is difficult to separate the effects of 
reserved seats from the other components of the compensatory discrimination policy. We shall 
postpone further assessment untill we have surveyed the whole preference policy. But before leaving 
legislative reservations, we should note their special relationship to the whole of the compensatory 
discrimination policy. 
Reservations of government posts or university admissions often operate as guaranteed minimums 
rather than as allotments of places over and above those gained on merit. Therefore the net amount 
of preference they confer automatically declines as members of the preferred groups gain places on 
merit. Reservations in the legislatures do not have this automatic self­liquidating feature. However, 
unlike other preferences, they alone are subject to a constitutional time limit ­­ this limit has been 
thrice extended. Since the presence of the reserved­seat legislators undoubtedly contributed to this 
result, these reservations are in a literal sense self­perpetuating in a way that others are not. 
Furthermore, their presence helps to perpetuate other forms of preference; indeed, that is their 
purpose. They can only perpetuate preferences to the extent that they are influential ­­ so, self­
perpetuation is a mark of some success in fulfilling their function of representation. Thus self­
perpetuation is not to be accounted a negative feature unless the whole system of com­ 
____________________ 
55. Galanter 1972a. 
56. E.g., Krishna ( 1966:32) notes that in 1962, SC legislators in reserved seats supplied the entire 
Congress plurality in 8 of 13 state legislatures, Dushkin ( 1972:205) calculated that in March 
1967, at least 9 state governments were dependent on, reserved­seat legislators. 
57. Galanter ( 1979) attempts to do this for the reserved legislative seats. 
­54­ 
pensatory preference is thought to be unjustified or to have outlived its justification. 
Although there is no time limit on the constitutional authorization of other forms of special 
treatment for Scheduled Castes and Tribes, one of the curious and significant things about the 
debate preceding each extension of reserved seats has been the extent to which the public, the press, 
and many parliamentarians proceeded as if an end to reserved seats was the end of all benefits and 
concessions to these groups. There is no constitutional or legal foundation for such an equation. The 
political reservations seemed in many minds to encompass the entire system of compensatory 
preference, and their extension was tantamount to heading off expiration of all benefits. This 
symbolic fusion may tell us something important about the way in which political reservations 
function to lock into place all the other programs for Scheduled Castes and Tribes.  58.   
When we focus (in Parts Two and Three) on the judicial encounter with compensatory 
discrimination, these reserved seats retreat into the background. Election disputes about reserved 
seats have been the source of almost all the litigation about group membership (taken up in chap. 9) 
because the greatest tangible prize for asserting membership is combined there with the greatest 
incentive for rivals to disprove it. But with that major exception,  59.   legislative reservations have 
remained outside the ambit of the courts. Courts do not get involved in the allocation of reserved 
seats or the delimitation of constituencies. They have no occasion to assess the working of reserved 
seats. Our discussion of' the working of' legislative reservations should remind us that the courts are 
not the only and perhaps not the most crucial shapers of compensatory discrimination policy. If the 
courts, as we shall see, act to brake and confine the preference policy, the reserved legislative seats 
may serve to lend it motive power. 
B. EDUCATION AND OTHER SCHEMES 
The total number of persons directly benefitted by the reservation of legislative bodies and of' 
government jobs constitutes only a small 
____________________ 
58. Of course that insurance function might be performed by other means ­­ e.g., by a constitutional 
guarantee of continued preferential treatment until SC and ST enjoyed specified levels of well­
being relative to the rest of the population. It might thus be possible to de­link the assurance of 
continued special treatment from the reserved­seat device, permitting reserved seats to be 
evaluated in terms of their specific benefits and costs. 
59. Legislative reservations were also the arena for important litigation about the problem of the 
double­member constituency. After it was settled by the courts, statutory change made this 
doctrine moot in the legislative representation area, although it has general implications in the 
administration of preferences. See chap. 13, § B, below. 
­55­ 
fraction of the backward classes. They are far outnumbered by those who benefit from a vast array 
of special schemes for education and economic uplift. Among educational schemes are scholarship 
programs, provision of meals, supplies, and hostels for Scheduled Caste (SC) and Scheduled Tribe 
(ST) students; and the construction and maintenance of schools intended primarily for these groups. 
Joined to these are a lengthy catalog of measures to improve the economic status of the SC and ST: 
allotments of (and subsidies to purchase) agricultural land; installation of irrigation wells; provision 
of livestock, seeds, and tools; encouragement of cottage industries by the provision of training, 
supplies, and equipment; support for agricultural and craft cooperatives. Finally, there are schemes 
for providing housing, roads (in tribal areas), drinking­water wells, health clinics, legal aid, and 
nursery schools. 
The "Backward Classes Sector" of the successive Five Year Plans provides a rough measure of 
expenditures on special measures for the disadvantaged. Planned expenditures for all categories for 
backward classes rose from Rs.300 million in the First Plan ( 1951­56) to Rs. 794 million in the 
Second Plan ( 1956­61) to Rs. 1,139 million in the Third Plan ( 1961­66) to Rs. 1,403 million in the 
Fourth Plan ( 1969­74) to Rs. 3,269 million in the Fifth Plan ( 1974­79).  60.   A breakdown of plan 
expenditure  61.   reveals a continuing emphasis on education as the major undertaking for SC. For ST, 
educational expenditures, now the predominant category, were earlier overshadowed by 
communications and more recently by economic uplift for isolated tribal areas. The plan provisions 
for Other Backward Classes (OBC) have followed a more irregular course owing to the controversy 
about the membership of that category.  62.   Expenditure on OBC was 9.84 crores in the First Plan, 2. 
 63.   in the Second, and 17.2 in the Third Plan. The projected expenditure on OBC in the Fourth Plan 
was 8 crores on education, 2 crores on economic development, and 1 crore on health, housing, 
etc.63 There was no provision under this heading in the Fifth Plan. 
These special undertakings for the backward form but a small fraction of plan expenditures. India's 
planners have been convinced that the improvement of the conditions of these "weaker sections" 
would flow from overall development. The special schemes for them represent a redistributive sub­
theme to narrow the gap that separates them from the rest of the population. It is a distinctly minor 
theme. In spite of the 
____________________ 
60. Planning Commission 1961: 701; 1970: 415. The scheme for post­matriculation scholarships is 
included, except for the Fourth Plan. 
61. "Planned expenditure" refers to new developmental spending and does not include spending to 
maintain established "regular" government programs. 
62. See chap. 6, below. 
63. Department of Social Welfare 1969: 162. 
­56­ 
rise in absolute expenditures, the backward classes sector (including the post­matriculation 
scholarship expenditures of the Education Ministry) accounted for 1.63% of the planned expenditure 
in the First Plan period; 1.87% in the Second Plan; 1.52% in the Third Plan; 1.12% in the Fourth 
Plan; and 0.83% in the Fifth Plan.  64.   
The amounts in question area a tiny fraction of the planned expenditure, and when we consider the 
vast population of the groups to be benefitted, we see how thin this special treatment is spread. 
Table 1 shows the per capita planned expenditure for SC and ST in the successive Plan periods. 
TABLE 1 ANNUAL PER CAPITA EXPENDITURE ON SCHEDULED CASTES AND 
SCHEDULED TRIBES THE FIVE FIVE­YEAR PLANS [IN RUPEES] 
Scheduled Castes 
1951­56  1956­61  1961­66  1968­74  1974­79 
1st  2nd  3rd  4th  5th 
Education  0.20  0.66  0.98  0.85  1.44 
Economic Uplift  0.01  0.15  0.17  0.20  0.51 
0.52 
Health, Housing, etc.  0.10  0.39  0.25  0.33 

TOTALS  0.31  1.20  1.40  1.38  2.59 


Scheduled Tribes 
1st  2nd  3rd  4th  5th 
Education  0.48  0.78  0.98  1.21  1.83 
Economic Uplift  0.41  1.78  1.82  2.20  1.45 
0.75 
Health. Housing, etc.  0.90  1.34  0.78  0.29 

TOTALS  1.79  3.90  3.58  3.72  4.05 


RCSCST 1965­66: 118; RCSCST 1970­71: 116­17; RCSCST 1975­77: I, 
Sources: 
16; II, 
18­20. 
The exchange value of the rupee decreased from $ 0.21 in 1951 to $ 0.12 
Note: 
in 1979. 
The multitude of schemes, the paucity of useful; information about their working, and limitations of 
space and energy make it impossible to undertake even a rough sketch of these programs. To avoid 
being swallowed up in a sea of detail, we shall concentrate on the largest single item among these 
distributive programs­­education. It entails the highest expenditures and directly affects the largest 
number of indi­ 
____________________ 
64. Department of Social Welfare 1969: 162; Planning Commission 1976: 115, 159. 
­57­ 
viduals. Education is widely regarded as the prime instrumentality for improving the condition of 
the backward; and the elevation of the backward is seen as one of the most important social 
products of education.  65.   
Independent India has undergone a remarkable educational explosion. Expenditures and enrolments 
have increased steadily and rapidly. Government expenditures increased from Rs. 1,143 million in 
1950­51 to Rs. 6,000 million in 1965­66 to an estimated Rs. 14,500 million in 1974­75. 66. 
Enrolments increased correspondingly from a total of 18.9 million in 1950­51 to over 58.2 million in 
1965­66 to over 85 million in 1973­74.  67.   In spite of rapid population growth, the growth of 
educational facilities has resulted in a great increase in the proportion of the population receiving 
instruction at every level. For example, enrolment in the elementary grades rose from 37.8% of the 
corresponding age group in 1950­51 to 83.5% in 1973­74.  68.   Enrolment in higher education had 
risen from 0.7% of the 18­23 age group in 1950­51 to 2.1% of vastly larger group by 1965­66.  69.   
Included in this general increase is an even more pronounced increase in the portion of Scheduled 
Castes and Scheduled Tribes children who go to school. These groups are still far below the general 
population in educational accomplishments, but this must be seen against a starting point of almost 
total exclusion from educational opportunities until fairly recently. 
By the 1920s, there were thousands of Depressed Class (as they were then called) students in the 
primary schools ­­ many or most of them in segregated inferior schools .  70.   Few went beyond the 
primary level. In 1927, the Hartog Committee found that enrolment of Depressed Classes children in 
secondary schools was 2.6% of enrolment in primary schools.  71.   Only a handful of students from 
these groups managed to get higher education. In 1931, only 1.9% of the "Exterior Castes" and 0.8% 
of the "Primitive Tribes" were literate, compared with 8% of the total population.  72.   In 1961 7.4% 
of Scheduled Castes and 6.28% of 
____________________ 
65. "One of the most important social objects of education is to equalize opportunity, enabling the 
backward or underprivileged classes and individuals to use their education as a lever for the 
improvement of their condition" ( Ministry of Education 1966: 108). 
66. Planning Commission 1976: 75. 
67. RCSCST 1975­77: II, 32ff. The figure of 85 million does not include higher education, which 
enrolled over one million by 1965­66 ( Ministry of Education 1966: 300). 
68. Ministry of Education 1966: 161; RCSCST 1975­77: II, 32. 
69. Education Commission 1966: 300. 
70. Indian Statutory Commission (Hartog Committee) 1929: 226. 
71. Indian Statutory Commission (Hartog Committee) 1929: 220. 
72. Natarajan 1972: 81, 93, 95. The 1931 and 1961 Census Reports are the only ones that give 
information useful in making these comparisons. "Exterior Castes," as they 
­58­ 
Scheduled Tribes were literate, compared with 24.02% of the total population.  73.   Thus the recent 
figures reflect a vast increase in educational attainments for the lowest groups, though they still lag 
far behind the rest of the population.  74.   
The expansion of educational opportunity for the general population during the decades following 
Independence was accompanied by a special commitment to raise the educational level of the 
Scheduled Castes and Tribes. While free primary education was gradually extended to larger 
portions of the whole population, education through the secondary level was made free to SC and 
ST pupils. By 1955 most states allowed full fee concessions to SC and ST students at all levels. 75. 
Mere exemption from fees was found insufficient. Economic distress, apathy, and reliance on 
school­age children to earn or look after younger children disinclined many parents from enrolling 
their children in schools and encouraging them to stay in school.  76.   Some states have encouraged 
enrolment in the first years of primary school by providing free mid­day meals, free books and 
supplies, and a clothing allowance.  77.   And for a fortunate minority, the few who reach the upper 
primary grades, there were pre­matriculation scholarships.  78.   Even among these, the vast majority 
do not recieve such scholarship help. In some states pre­matriculation scholarships are awarded on 
merit rather than as of right. And state governments have sometimes failed to spend 
____________________ 
were called, do not correspond exactly to the later­designated Scheduled Castes, but they are 
close enough for this crude comparison. Similarly, "Primitive Tribes" are not precisely equivalent 
to Scheduled Tribes. The 1931 tribe figures are for those "20 years and older" and may thus 
understate recent improvement. 
73. Nataraian 1972: 118. The gap between SC and ST and the rest of the population is understated by 
comparing them with a "total" of which they make up almost one­quarter. 
74. It should be noted that there was a change in the operational definition of literacy. In 1931 it 
meant those who had completed grade 4 in primary education; in 1961 it was based on a simple 
test of literal for those who claim to be literate without having completed primary school. 
75. RCSCST 1955: II, 106. 
76. RCSCST 1958­91 I, 48 (on the need for incentives for attendance); RCSCST 1975­77: I, 89­90. 
77. RCSCST 1960­61: 161­68; RCSCST 1971­73: 55­56. 
78. In about 1956, a scheme of pre­matriculation scholarships was introduced for the SC, ST, and 
OBC. Payments under these scholarships start at grade 5 or 6 (age 11) and range from 23 to 72 
rupees a year, increasing with each standard so that by standard 11 they range from 32 to 150 a 
year ( Department of Social Welfare 1969: 191­92). During the Third Plan period a total of 1.7 
million pre­matriculation scholarships or stipends were awarded to Scheduled Castes and over 
one million to Scheduled Tribes ( Planning Commission 1970: 416). This is less than half a 
million a year ­­ a small fraction of population of these groups. Most states have extended the 
schemes for prematriculation scholarships to OBC. Incomplete figures for ten states show over 
88.000 pre­matriculation scholarships for OBC in 1960­61. 
­59­ 
the amounts allocated to pre­matriculation scholarships. The Elayaperumal Committee found that 
the governments of A. P., Bihar, Punjab, Rajasthan, and West Bengal have lapsed the 
amounts allocated for pre­matriculation scholarships, while they deny the scholarships 
[to applicants] on the ground that the fund is not available. . . . This is due no doubt to 
lack of interest on the part of these Governments.  79.   
Lowered barriers, a changing atmosphere, availability of schools, and government aid have 
contributed to a great spurt in the enrolment of SC and ST students. In the lowest grades the 
percentage of children from these groups trailed behind the rest of the population, and in the higher 
grades there were still massive disparities ( table 2 ). 
TABLE2 SCHOOL ENROLMENT AS A PERCENTAGE OF CORRESPONDING AGE 
GROUP, BY COMMUNITIES 1973­1974 

Classes  Others  SC  ST 

I­V  83.5  68.9  59.3 


VI­VIII  35.6  22.1  15.7 
IX+  21.2  11.0  7.7 

RCSCST 1975­77: II, 32­34. For many purposes such figures should be regarded 
with
utmost suspicion (suspicion not allayed by the fact that a number of states report
enrolments of over 100% of the relevant age groups). But for the purpose at 
hand there is
Source: 
no reason to think they do not give a rough idea of the disparities among SC, 
ST, and the
rest of the population. A slightly different set of figures is found elsewhere in 
the same
report: RCSCST 1975­77: I, 88. 
With expanding enrolments, wastage (dropping out) and stagnation (repeating a grade) were 
accentuated as major problems in Indian education generally. Of 100 pupils enrolled in grade I only 
half reached grade 3 in the prescribed 2 years, and only half of that half finish grade 6 on time. The 
rate of attrition is even more pronounced among Scheduled Castes and Tribes.  80.   There is a strong 
correlation between social status and the proclivity to stay in school.  81.   In one revealing study, 
comparing the influence of a multitude of factors, caste was the aspect of social structure with the 
strongest effect on wastage.  82.   
____________________ 
79. Department of Social Welfare 1969: 193. 
80. Sharma and Sapra 1969: 1. 
81. Sharma and Sapra 1969: 81. 
82. Sharma and Sapra 1969: 96 ff. The more influential variables were behavioral (e.g. steady 
attendance) and attitudinal (e.g. motivation for learning from home). It 
­60­ 
For those who manage to remain in school there are serious problems of the suitability and quality 
of the training they receive. The former policy of creating special schools for these groups has been 
severely criticized for perpetuating their segregation. There is some indication, too, that the 
standards in schools intended primarily for Scheduled Castes and Scheduled Tribes are much lower 
than in general schools. It is now generally agreed that mixed schools are preferable wherever 
possible.  83.   
Even where they are located in areas without special problems and enrolled in mixed schools, 
children from low­status groups tend to attend schools with the least­adequate facilities, the least­
qualified teachers, and the least­advantaged fellow students.  84.   Combined with lower resources and 
motivation in the family setting, these disadvantages cumulate as they move up the educational 
pyramid. Of those who survive the course, most emerge with academic records that are not 
prepossessing.  85.   
Notwithstanding these difficulties, the number of Scheduled Caste students who reach higher 
education increases steadily. The contrast with the situation before. Independence is stunning. In 
1927, out of a total of almost 55,000 college students in all of India (excluding Bengal), there were 
82 students from the Depressed Classes­­less than 15/100 of one percent of total enrolment.  86.   By 
the 1961 Census, SC and ST were still only a tiny portion of the nation's graduates­­a very generous 
estimate would be about 1.6%, a more cautious one closer to 1%.  87.   
Access to higher education has remained a matter of prime concern 
____________________ 
might be thought that these things are associated with the educational attainments of the parents. 
But of course the parents with the lowest educational status are most likely to belong to the lower 
castes. D'Souza 1969 : 91. 
83. The number of pupils in special schools declined from 829,000 in 1951­52 to 465,000 in 1960­61 
( Ministry of Education 1955: 111; 1968 : 135). One of the causes of low enrolments is lack of 
school in depressed areas, and as schools are established to meet this need, more children will in 
fact be receiving separate education. This is especially so for Tribals where the policy of 
bringing schooling to remote areas and making the first years available in tribal languages 
effectively separates them. 
84. In a study of 21 schools in Chandigarh the schools were stratified on the basis of fees: in the 
higher­grade schools the pupils were exclusively of the "Advanced Castes," in third­grade 
schools there were 25% "other castes" ( D'Sounza 1969 : 100­5). 
85. Shah and Patel 1977 : 59. 
86. Indian Statutory Commission (Hartog Committee) 1929 : 128, 220. 
87. There were, according to the 1961 Census, 1.147 million graduates. There were 6,307 SC and 
1,087 ST graduates counted in urban areas. The rural SC and ST were listed only as 
"matriculation and above." If we make the unlikely assumption that the same proportion of rural 
as urban matriculates were graduates, we arrive at an estimated total of 16,223 SC and 2,938 ST 
graduates­­about 1.6% of the total. 
­61­ 
to leaders of all backward class groups. A centrally funded scheme of post­matriculation 
scholarships for Scheduled Castes was founded in 1944. It was extended to Scheduled Tribes in 
1948 and Other Backward Classes in 1949. There was an early commitment to provide a scholarship 
for every qualified applicant from the SC and ST. Total scholarships for these groups grew from 731 
in 1948 to 37,372 in 1958, to 156,834 in 1968, and to 350,000 in 1975. Expenditures rose from Rs. 
500,000 to Rs. 136 million.  88.   Virtually every applicant from these groups received a scholarship. 
Scholarships for Other Backward Classes were more highly competitive.  89.   The number of 
scholarships for OBC increased rapidly in the early years, but more slowly after 1955 as the whole 
question of special treatment for this group became controverted. Until 1961 the OBC for purposes 
of this scheme were lists of castes (and communities) supplied by each state. In 1961 the Center 
abandoned the earlier notion of drawing up an authoritative list of OBC and recommended to the 
states a changeover to an economic basis for determining backwardness.  90.   from 1963­64, post­
matriculation scholarships were awarded on an income basis. The applicants were divided into two 
slabs: those with annual family incomes of less than Rs. 1,500 and those with income between Rs. 
1,500 and Rs. 2,000 (except in the case of technical courses, where the upper limit is Rs. 2,400). 
Students from the first slab had first priority for scholarships; any remaining funds were used for 
those in the higher slab. Within the slabs, students from families with lower income were to be 
given priority, but where income was the same and there were insufficient funds available, selction 
was to be on merit.  91.   
For SC and ST applicants, there was a much higher income ceiling: scholarships were available to 
those whose monthly income is below Rs. 500.  92.   Several states pay far less than the stipends 
recommended by the central government­­Rs. 40 per month to hostellers and Rs. 27 to day scholars 
for graduate and undergraduate courses, and Rs. 75 for hostellers and Rs. 60 for day scholars in 
Medicine and Engineering.  93.   
____________________ 
88. RCSCST 1975­77: I. 93; RCSCST 1971­73 : 274­75. 
89. E.g., in 1958­59, there were 37.654 applications but only 12.590 scholarships awarded. 
90. See chap. 6, below. 
91. Government of Maharastra, Education and Social Welfare Dept., Resol. No. CSP 1065­M(SG) 
Statement C, pp. 15­16 ( 1966). The recipients under this new dispensation were called 
"Economically Backward Classes" from 1963 to 1965, when their name was changed to "Lower 
Income Group Students." 
92. There had originally been an income ceiling, which was withdrawn in 1957 and re­imposed in 
1961­62. 
93. E.g., Punjab pays Rs. 15 to intermediate students and Rs. 20 to B.A. and B.Sc. students. 
­62­ 
The central government's standards were set in 1954­55 and have not been raised, in spite of great 
increases in the cost of living. 
As one might expect, SC and ST students are concentrated in the less prestigious, less demanding, 
and less potentially renumerative subjets. Students from the OBC have a much different profile of 
enrolment­­a much larger portion are students in the highly competitive engineering and medical 
fields. In 1958­59, 37.8% OBC students at the postmatriculate level were studying medicine or 
engineering; in contrast, only 6.4% of post­matriculate SC and ST students were enrolled in these 
subjects.  94.   
These students tend on the whole to be less well prepared than their fellows. As we have noted, 
many of them come from schools with poor facilities, as well as from families with little educational 
background. They tend, as might be expected, to be poor academic performers. Although little data 
is available, it seems that the wastage (drop­out) rate among SC and ST college students is 
staggeringly high. For example, a survey conducted in Maharashtra found that 25% to 40% failed in 
their first­year exams, about 70% failed in intermediate, 10% in their junior year, and between 40% 
and 60% in their final exams. Only 8% received their degrees within the prescribed four years, and 
about 85% left college without a degree.  95.   Those who do get through tend to get low grades.  96.   
The policy of reservations has been widely adopted in higher education, but the general expansion 
of facilities has meant that there is room enough for everyone except in a few highly coveted fields. 
Reservations are of consequence only where opportunities are scarce­­notably in institutions for 
professional and technical training. The Ministry of Education requested that all states and 
universities reserve 15% of places in these institutions for SC and 5% for ST; that the maximum age 
be relaxed by three years and the minimum marks for admission be reduced by 5%. Most states have 
provided for such reservations in at least some institutions. But although accepted in theory, they 
have not been fully implemented. A rough indication of the actual extent of such reservation and of 
its effectiveness is provided by the results of a survey of medical and engineering colleges 
conducted by the Elayaperumal Committee. In the 42 colleges that replied, less than 11% of seats 
were reserved for SC and far less than half of these were actually filled by SC candidates.  97.   The 
general level of inclusion reflected in these figures comports with several later readings. From data 
available on 67 of 97 medical colleges in 1972­73, Dushkin computed that 6.6% of medical 
____________________ 
94. RCSCST 1958­59 : 176. 
95. Department of Social Welfare 1969 : 203. 
96. RCSCST 1968­69 : 65. 
97. Department of Social Welfare 1969 : 237­44. 
­63­ 
students in 1972­73 were Scheduled Castes.  98.   Available data show that 157 of 2,301 (6.8%) 
admitted to 22 medical colleges in 1975­76 were Scheduled Castes.  99.   None of these estimates tells 
us about the percentage of SC students who graduate. 
Reservations in professional­school admissions have been of great concern to OBC. Such 
reservations are in force in at least nine states.  100.   Indeed, reservation of these highly prized 
medical admissions can be taken as a crude measure of state commitment to divert resources to the 
groups denominated as OBC. The composition of these groups differs greatly from state to state. As 
reservations for OBC are sometimes larger than for SC and ST and they never go unfilled, they 
represent a major claim on a much­coveted and scarce opportunity. In a number of states 
reservations for OBC in professional colleges have become a focus of conflict between putative 
OBC and other groups.  101.   We shall encounter these reservations frequently in the sequel, for 
virtually all of the litigation about the compensatory discrimination policies in education has been 
about reservations in medical and engineering colleges for the Backward Classes. 
Although it remains unclear just how much of the increase in education for the backward classes 
represents "special treatment" over and above what is provided for the whole population, it is clear 
that compensatory discrimination provisions can be credited with securing the inclusion of these 
groups as beneficiaries of the educational explosion. And it has enabled them to "catch up" to some 
extent. If "catching up" is taken to mean greater percentages gains over a base­line period for the 
backward than for others, there clearly has been catching up in this sense. But "catching up" may 
also mean narrowing the gap (expressed, e.g., in percentages) between the backward and the rest of 
the population. Here, it seems, there has been substantial catching up at the lower educational levels, 
but not at the higher levels. 
C. ADMINISTRATION 
Limitations of space and energy do not allow even a cursory review of protective measures, of the 
development programs in the tribal areas, and of the many distributive schemes outside the field of 
education. These latter include provision of free house­sites, subsidies for the 
____________________ 
100. See table 20, in chap. 6, below. 
101. The politicization of medical admissions is apotheosized by the arrangement in Jammu and 
Kashmir in which the admissions list was issued directly by the Chief Minister of the State. See 
Subhash Mohan Jalili v. Principal, A.I.R. 1967 J. & K. 106. 
98. Dushkin 1979 : 663. 
99. RCSCST 1975­77 : II, 36. 
­64­ 
construction of houses, allotments of acreage to the landless, improvement of methods of agriculture 
and animal husbandry; health measures such as provision of wells, mobile dispensaries, and 
maternity centers; encouragement of cooperatives and cottage industries. Such schemes have been 
mainly for the Scheduled Castes (SC) and Scheduled Tribes (ST). There has been only a small 
expenditure on welfare measures for Other Backward Classes (OBC).  102.   A reading of the official 
reports suggests that such distributive schemes have on the whole been less intensive and less 
successful than efforts in education. Since any assessment of compensatory discrimination policies 
requires that we consider not only the announced schemes but the level of performance, we 
conclude with a very brief characterization of the administration of compensatory discrimination 
schemes. Our gross characterization will undoubtedly do some injustice to detail. Hidden in the 
dismal aggregates are some gems of devoted and effective implementation. But all accounts point in 
the same direction as far as the overall character of implementation is concerned. 
There are persistent "shortfalls" in expenditures. That is, the allocation of government moneys for 
these programs has been greater than the ability of the various agencies to spend it. Over and over 
again there are reports of failure to spend more than a fraction of the money allocated to programs 
for the backward classess.  103.   Shortfalls in expenditure reflect lack of vigorous and imaginative 
leadership, inadequacy of planning, absence of coordination between agencies, and deficient 
arrangements for collecting data on the progress of schemes. There are delays at the Centre in 
making grants; states are disinclined to make sufficient matching grants where that is required. 
Skill, enthusiasm, and political push are in short supply.  104.   
How much of this spending actually represents preferential treatment for the backward classes? 
There is some evidence that administrators sometimes feel that it is not necessary to include 
Scheduled 
____________________ 
102. The annual reports of the Commissioner for Scheduled Castes and Scheduled Tribes provide 
data on central and state schemes for Scheduled Castes and Scheduled Tribes. On schemes for 
Backward Classes it is necessary to consult reports issued by the various states. 
103. Department of Social Welfare 1969 : 130, 144­45. RCSCST 1963­64 : 27, 33. 110; RCSCST 
1959­60 : I, 17. (The Commissiner, notes that part of the spending is only on paper. He estimates 
that fully one­third of the spending reported represented transfers to other accounts to prevent 
lapse of appropriations.) SASTC : I, 100, observes that shortfalls are greater in economic and 
health programs than in education. This suggests that the money is more readily spent where the 
administration's job is to pass it on to interested and alert recipients to spend in existing 
institutions than where the administration is called upon to erect and operate programs of its 
own. 
104. RCSCST 1959­60 : I, 13­14; RCSCST 1960­61 : 66; SASTC : I, 78, 102. 
­65­ 
Castes within general schemes, so that in part the budget for backward classes replaces rather than 
enlarges the benefits to which they are entitled under general welfare measures.  105.   
This is most clearly documented in the case of the Scheduled Tribes. In 1960 the Dhebar 
Commission pointed out that although the Centre's special provision for Scheduled Tribes were 
meant to supplement the general welfare programs of the states, many states had not pursued normal 
development programs for the ST and instead used the central funds as a substitute rather than a 
supplement.  106.   A decade later a Study Team noted "with regret": 
Though the attention of State Governments has been repeatedly drawn to the fact that 
the special provision in the plan for S.T. and O.B.C. is not in lieu of the provision in the 
general development programme but is intended to supplement it . . . no conscious 
attempt has been made by any of the State Governments with the solitary exception of 
A.P. to ensure that the tribals receive a reasonable share of benefits of the general 
development programme to which they . . . are legitimately entitled.  107.   
But persistent complaints suggest that it is also true of the various distributive programs (other than 
education) for the Scheduled Castes. This point emerges starkly from Aggarwal and Ashraf's study 
of Scheduled Castes in Haryana, where we see these policies through the eyes of their beneficiaries. 
A majority of the 500 household heads we interviewed for our study complained that 
they were excluded from the general aid programme of the government. Many asserted 
that their applications were either not accepted or they were rejected mainly on the plea 
that there were special provisions for them. Some said that they were clearly told to go 
to the "Harijan department." We learned that the general impression among the people, 
including the government officials, was that "the Harijans were lucky to have all those 
special provisions." 
Most of our high caste informants, too, shared the same view. They said that Harijans 
should feel satisfied with the special privileges. In fact there was complete ignorance 
about the smallness of the budget and the myriad administrative problems. People who 
normally evince sympathy for the weaker sections and support the idea of special 
privileges know very little about the programme and feel complacent. 
Even the government officials were rather poorly informed on the details of the social 
welfare programme and therefore assumed that enough was being done by the 
govenment for the Scheduled Castes. . . . 
____________________ 
105. Dushkin 1961 : 1702. The Commissioner pointed to the need to make benefits additional rather 
than substitutes (RCSCST 1960­61 : 75; Department of Social Welfare 1969 : 166). 
106. SASTC : I, 73. 
107. STTDP : 13. 
­66­ 
In order to get some idea of what this exclusion means for the Harijans, we may 
compare just one set of figures . . . only Rs. 200,000 was provided in the 1973­74 budget 
of the Welfare Department for the Scheduled Castes and Backward Classes, for business 
loans. Compared to this the Haryana Government allocated over 220 million rupees in 
the year 1972­73 as takavi and crop loans to the farmers. As we can see, this figure is a 
thousand times the amount provided for the Harijans. Had there been no special 
provisions for the Harijans and they were given a proportionate share of the bigger 
amount they would have got 44 million rupees. That would have been 240 times the 
amount that they now get.  108.   
Sometimes the schemes turn out to be special in the sense of separate rather than additional. 
Programs may be administered in such a way as to perpetuate the segregation they are intended to 
overcome. Particularly in housing, which tends to be located away from higher­caste localities­­and 
consequently in schools, wells and meeting halls­­the provision of special facilities has tended to 
perpetuate isolation. (Similarly, officials faced with "controversial" claims by Scheduled Castes for 
the use of facilities, are tempted to achieve a settlement by diverting development funds into 
provision of what are in effect separate wells, etc.) 
Although the total expenditure on these schemes is small, there is a great plethora of schemes. The 
proliferation of small schemes consumes a large portion of expenditures in their administration. 109. 
Benefits can be made available to only a small fraction of those who are nominally eligible under 
any given program. Locating programs and establishing eligibility is a demanding undertaking for 
would­be beneficiaries, made more demanding by passive and torpid administration. Those most in 
need of help are least likely to have the literacy, information, contacts, and skills to navigate the 
tortuous channels leading to inclusion.  110.   
These programs are administered in a routine, mechanical fashion. There is no attempt to generate 
feedback from those affected. Recordkeeping about actual results is fragmentary and incoherent. 
There is no serious attempt to monitor the effectiveness of programs. One officer in a state office 
that had been administering a number of schemes for fifteen years reported that the office had never 
discontinued any scheme on grounds of ineffectiveness. Many programs not only are adminis­ 
____________________ 
108. Aggarwal and Ashraf 1976 : 171­72. 
109. Thus out of a sum of Rs. 100,000 sanctioned for a Tribal Development Block, nearly three­
quarters was spent on staff salaries (SSTDP : 29). 
110. On the problem of illiterate tribals whose applications for land allotments are rejected as 
improperly filled in, see STTDP : 21. Cf. the problem of securing approvals for schemes from 
standing committees of Panchayat Samitis and Zilla Parishads, which meet infrequently and 
irregularly (RCSCST 1962­63 : 5). 
­67­ 
tered in a lackadaisical fashion but are of doubtful suitability to local conditions.  111.   Lower officers 
enjoy little discretion to modify schemes to suit local conditions. Nor is there much enthusiasm, 
initiative, or inventiveness in the higher echelons. Departments dealing with backward classes 
attract little talent. Experienced staff tend to flee these departments, which offer few channels for 
promotion. Well­qualified officers move on to better prospects elsewhere.  112.   The ranks are filled 
with rejects from other departments.  113.   
The listlessness of administering the means is matched by thoughtlessness about goals. The 
mindless, going­through­the­motions quality of administration is nicely captured by Aggarwal and 
Ashraf: 
In our discussions with the Haryana government officials, from the State to the tehsil 
level, we found that no one talked of a goal to be achieved through the special privileges 
programme. Outside the government, too, neither the Scheduled Caste leaders nor the 
high caste persons were aware of any goal. In other words nobody seemed to know what 
was to be achieved and by what date. 
The Haryana Welfare Department publishes a report each year . . . These reports are 
built around the previous year's budget which contains all the details about the schemes, 
the sources of funds, the amount, allocation for each district on the basis of its 
population, and even the number of cases to be served. The reports then simply mention 
the amount spent and the number of grants and loans given. 
There is no projection of what is to be achieved and by what date. If equality is the 
objective, which is implied, no mention is made with whom and by when it is to be 
achieved. 
All of this clearly indicates that this programme is activity centered and not goal 
centered. Such a goalless programme suffers from many deficiencies. One is that it is 
impossible to evaluate. Only activities can be listed and expenditure accounted for. 
Beyond that no one can say whether the programme is achieving anything because 
achievement can be measured only against a goal. Furthermore, it induces a false sense 
of accomplishment in the administrator. He feels satisfied as long as he is able to spend 
the money and cater to a specified number of clients. A feeling of complacency is 
generated which in turn kills initiative for greater efficiency and achievement. With an 
easily achievable target of activities already set; only an extraordinary official will strive 
on his own.  114   
Generally, administration of these programs is a matter of low priority, which commands little 
attention from ministries or from legisla­ 
____________________ 
111. On the failure of housing colonization schemes for tribals, see STTDP: 8, 16­17 
112. RCSCST 1962­63 : 2. 
113. RCSCST 1960­61:13; STTDP : I, 49. 
114 Aggarwal and Ashraf 1976 : 176­77. 
­68­ 
tors.  115.   Thus legislators have paid little attention to the reports of the Commissioner for Scheduled 
Castes and Scheduled Tribes. State officials have been less than assiduous in responding to requests 
for information. There seems to be little genuine curiosity about how the schemes do work in fact. 
In spite of the vast quantity of paper that is generated by the administrators, there is remarkably little 
data about how the schemes work. 
For the most part the departments of social welfare or Harijan welfare found in each state 
Government are not independent administrative departments, but have only advising and 
coordinating functions in relation to the regular governmental departments which administer these 
programs.  116.   Administration of special benefits is by the departments concerned with the general 
welfare programs of housing, education, agricultural loans and so forth. 
Government departments whose major task is dealing with the disadvantaged tend to have little 
prestige or influence within the government. As one M.P. observed of the Central Department of 
Social Welfare: 
Nobody takes the decisions of this department into consideration and no other 
department cares for this department. No department complies with the directions and 
promises of this department. The conditions prevalant in the State level [are] worse.  117.   
This is reflected in the surprised observation of the Elayaperumal Committee that 
although State Ministers in charge of Welfare of B.C.s had unanimously decided that the 
principle of earmarking funds in the general sector for the Welfare of the SCs should be 
implemented this decision was not carried out.  118.   
One well­placed officer offered the suggestive generalization that the minister in charge of 
backward classes or social welfare is usually the weakest in the state cabinet. 
____________________ 
115. See Schermerhorn ( 1969) for an interesting attempt to measure the priority levels of Scheduled 
caste programs in various states. 
116. At the Centre these functions were housed in the Home Ministry until 1964, when all aspects 
other than government employment and the Scheduled Areas were transferred to a new 
Department of Social Security, later renamed the Department of Social Welfare. In a subsequent 
reorganization they were returned to the Ministry of Home Affairs in 1973 (RCSCST 1964­65 : I, 
1; 1972­73 : I, 251). In 1978 for the first time backward classes matters became the subject of a 
separate wing within the Home Ministry. 
117. Shri Baidhar Behara at LSD (4th Series), Vol. 16, No. 50: Col. 3152 ( 25 April 1968). This was 
said of the Social Welfare Department. See n. 116 above. 
118. Department of Social Welfare 1969 : 166. 
­69­ 
Problems in carrying out constitutional policies for these groups were not unanticipated. The 
Constitution established a special "watchdog" agency to monitor the working of the "safeguards" for 
Scheduled Castes and Scheduled Tribes. Article 338 provides: 
1.  There shall be a Special Officer for the Scheduled Castes and Scheduled Tribes to be appointed 
by the President. 
2.  It shall be the duty of the Special Officer to investigate all matters relating to the safeguards 
provided for the Scheduled Castes and Scheduled Tribes under this Constitution and report to 
the President upon the working of those safeguards at such intervals as the President may 
direct, and the President shall cause all such reports to be laid before each House of Parliament. 
 119.   
The Office of the Commissioner for Scheduled Castes and Scheduled Tribes was established 
pursuant to this provision. The original conception that the Commissioner would serve as an 
independent critic of government was gradually effaced as the office took on tasks of oversight and 
evaluation of programs.  120.   It proved no match for the problems of resistance, low priority, poor 
planning, and lack of coordination that beset these programs. Its investigative potential is limited by 
lack of funds, inadequate staffing, and the absence of power to require testimony and records. It has 
no capacity to generate data of its own, and no power to induce other agencies to provide it with 
information. The Commissioner's annual reports are filled with accounts of the unwillingness of 
other agencies to supply information or even to answer specific inquiries.  121.   
The original aspiration to establish the Commissioner as an independent body (like the Election 
Commission or the Union Public Service Commission) was forgotten over the years. In 1967 the 
Commissioner's field organization was dismantled and replaced by a regional organization 
subordinate to the Central Department of Social Welfare. To the Elayaperumal Committee this was 
the culmination of a process in which 
the Commissioner's organisation was gradually . . . reduced to the position of a 
subordinate office of the Government of India with the result that he became powerless 
not only in respect of the various State Governments but even with the various 
Departments/Offices of the Central Government. He 
____________________ 
119. This provision refers only to the SC and ST. Article 338(3) says that those references "shall be 
construed as including references to such other backward classes as the President may, on receipt 
of the Commission appointed under [Art. 340(1)] . . . by order specify." The abandonment of the 
expectation that there would be a Presidential Order specifying the Other Backward Classes is 
recounted in chap, 6, §3, below. 
120. CWSCST (4th) 1 (1969). 
121. On the working of this officer, see Fisher 1970. 
­70­ 
has been powerless even to collect the basic and requisite data for his assessment in 
fulfillment of his constitutional duties. His annual reports are a record of his bewailings 
and beseachings [sic] year after year.  122.   
Without any investigative capacity of his own the Commissioner was reduced, in the words of 
another Study Team, to 
[discharging] his constitutional functions through those very agencies on whose 
performance in certain spheres he may be called upon, by virtue of his office, to sit in 
judgment.  123.   
In 1977 the Commissioner reported that his office had "only two Research Officers and two 
Investigators."  124.   
In 1968, Parliament established its own "watchdog" institution, the Joint Parliamentary Committee 
on the Welfare of Scheduled Castes and Scheduled Tribes. This is a thirty­member body elected by 
the two houses of Parliament (20 from Lok Sabha, 10 from the Rajya Sabha) for two­year terms. 
The Committee is mandated to recommend measures to implement the Commissioner's Reports, 
and to monitor central government programs and practices. Unlike the Commissioner's Office, the 
Joint Committee has investigative powers. The Committee has issued a series of detailed reports 
(there were 65 by early 1976), each concerned with a single agency or problem. A number of these 
are follow­up reports, tracing the response to its earlier recommendations. The Committee has 
focussed heavily on extending and implementing reservations in central government jobs and in 
public corporations, and in higher­echelon educational opportunities.  125.   Comparable oversight 
committees were established in some states in the 1970s. 
Rising concern over atrocities against Scheduled Castes and eagerness to allay charges of neglect, 
combined with dissatisfaction with the Commissioner's Office, led to the establishment of yet a 
third central "watchdog" institution in 1978. A five­member Commission for the Scheduled Castes 
and Scheduled Tribes was appointed and empowered to investigate the implementation of 
safeguards (particularly job reservations) and anti­disabilities laws. The functions of the new 
Commission were described as corresponding broadly to those entrusted to the Commissioner, and 
the government suggested that the Constitution would be amended to reflect this replacement.  126.   
____________________ 
122. Department of Social Welfare 1969: 399. 
123. STTDP: 89. 
124. RCSCST 1975­77: I, 65. 
125. On the establishment of the Joint Committee and its performance during its first years, see 
Narayana 1980. 
126. LSD (6th Series), Vol. 16, No.5: Col. 262­64 ( 21 Jul. 1978). For an interpretation of this 
puzzling event, see Economic and Political Weekly 13: 1159 ( July 28, 1978). 
­71­ 
On a day­to­day basis the function of prodding the administration is performed not by these formal 
"watchdog" institutions but by the M.P.'s and M.L.A.'s ­­ usually the occupants of reserved seats. 
They intervene with the administration to champion beneficiaries' grievances, urge priority for 
programs, overcome inertia, and move things through channels. As one backward classes leader 
observed, "If reservations in the legislature are taken away, educational schemes will not work, for 
there will be no one to press the ministers."  127.   
D. CRITICS, COSTS, AND BENEFITS 
Few in independent India have voiced disagreement with the proposition that the disadvantaged 
sections of the population deserve and need "special help." But there has been considerable 
disagreement about exactly who is deserving of such help, about the form this help ought to take, 
and about the efficacy and propriety of what the government has done under this head. The major 
lines of criticism may conveniently be sorted into objections to what the government has not done 
and objections to what it has done. Critics of the first (omission) type may in turn be divided into 
those who criticize the Government for doing not enough and those who criticize it for doing not 
the right thing. 
Not enough criticisms include insufficiency of appropriations and expenditures, inadequacy in the 
extent and scope of reservation, failure to implement programs with sufficient zeal. Not the right 
thing criticisms point to the government's failure to do things it ought to have done ­e.g., to aim its 
educational efforts at preparing backward class students for offices rather than trades (or vice versa). 
Those who make these criticisms of non­performance or omission have generally been favourable to 
the main lines of the government's policy of special treatment and seek to see it augmented or 
redirected.  128.   
A much more widely articulated set of criticisms has come from those who question the main lines 
of the Government's policy of special treatment. These (commission) critics, too, may be 
conveniently di­ 
____________________ 
127. Interview with G. Latchanna. Hyderabad, 1965. 
128. A great deal of criticism along these lines is found in government sources, especially the topical 
reports of the Committee on the Welfare of the Scheduled Castes and Scheduled Tribes, and the 
annual reports of the Commissioner for Scheduled Castes and Scheduled Tribes. A massive 
compilation of such objections to government policy may be found in the report of the 
Elayaperumal Committee ( Department of Social Welfare 1969). All of the above deal with 
Scheduled Castes and Tribes. For assessment of Schemes for Backward Classes, see Government 
of Tamil Nadu 1971; Government of Karnataka 1975. Views along this line are also forthcoming 
from organizations of beneficiary groups. 
­72­ 
vided into those who think the government has done too much and those who think it has done the 
wrong thing ­­ although many of these critics link the two. The too much view asserts that 
preferential treatment, especially reservations of various kinds, has been overdone and costly in 
terms of efficiency, utilization of talent, morale, and basic fairness. But the great focus of 
dissatisfaction has been the use of caste and community groups in selecting the beneficiaries. In 
what has become the conventional wisdom on the matter, the use of caste units has been blamed for 
perpetuating the caste system, accentuating caste consciousness, alienating and demoralizing the 
groups who are benefitted, augmenting political abuse of caste by promoting "vested interests in 
backwardness," impeding the development of secularism, and other sins as well. Thus it has been 
asserted by two sociologists that 
reservation on the basis of caste . . . has affected the morals, the administration and the 
society adversely. The attempt to atone for the injustice perpetuated on the lower caste 
[sic] in times past has in effect become a tool of aggrandisement in the hands of the 
numerically large and hence politically dominant caste groups in majority of the states. 
 129.   
A political scientist has observed that governmental definition of backwardness in terms of caste has 
perpetuated and accentuated the caste consciousness. The judiciary by upholding 
classifications based on caste, has also helped the growth of the virus of casteism. This 
has undermined the foundations of the secular state and has been detrimental to the 
development of secularism.  130.   
Another sociologist has observed that reservations for backward groups are 
bound to prove harmful from a broader point of view, since it will perpetuate, more or 
less, the existence of the so­called depressed classes. . . .  131.   
____________________ 
129. Bamabas and Mehta 1965: 76­77. 
130. B. A. V. Sharma 1968: 70. 
131. R. N. Saksena 1965: 25. A scholar who could fairly be called the dean of Indian sociology 
expressed dismay and anger at "the process of proliferation and entrenchment of caste under the 
specious plea of social justice and upgrading­levelling of the so­called Backward Classes. . . ." 
( Ghurye 1969: 430­31.) He concludes a lengthy polemical analysis with the prediction that if the 
reservation of legislative seats is extended, "[w]e will thus have a few caste and sub­caste groups 
turned into political parties in miniature and vested interests for preserving their special spoil of 
the social cake of the Government jobs. From there accretion of other economic and political 
interests would not fail to begin and thus to solidify them as exclusivist and separatist 
communities," ( 1969: 441). 
­73­ 
This kind of sweeping and drastic condemnation has not been confined to academic observers. 132. 
Such views have been expressed by administrators and official committees,  133.   and by a wide array 
of intellectuals and publicists. Whether complacent about the progress that has been made or 
despairing at the lack of progress, participants in public discussion of preferences overwhelmingly 
depict them as pernicious. Two recent examples convey some of the flavor of this discourse. The 
first, an article entitled "Reservation: Blessing or Curse" that appeared in the Sunday Statesman  
Magazine in February 1978, begins with this assertion: 
Constitutional reservations have reduced the mass of untouchables into beggars 
endlessly waiting for the caste­Hindu crumbs of bread, thrown at their whims and 
fancies. . . . [T]he majority of caste Hindus grow increasingly jealous and curse the 
recipients as "Government Brahmins." . . . Except producing some elites among Dalits, 
the system has done the greatest disservice to the Dalits as a whole. . . . The alleged 
rising number of atrocities against Harijans is an off­shoot of this upper­caste jealousy 
and prejudice against these reservations . . . . [T]he major portion [of the benefits] has 
gone to create . . . professional elite . . . [which] has turned into a set of new­Brahmins 
living the life of leaches [sic], sucking the blood of their own brothers and sisters. . . . 
The system of reservations . . . . kills the beneficiaries' initiative, drive and capacity. 
Everybody admits that the concessions extended to Dalits have not reached them. . . . 
[T]wenty­six years of reservations have not made any dent on untouchablity ­­ which is 
the root cause of all discriminations against Dalits. . . . 
The author concludes by recommending that reservations be scrapped in the hope that "abrupt 
withdrawal of reservation, which have pampered them so long will rouse them to action." Wounding 
the sleeping tiger (the author's image) will hasten social change.  134.   
About two weeks later an article appeared in the Indian Express entitled "Emancipation or 
Exploitation." The author finds that "there 
____________________ 
132. A number of foreign students of India have also voiced negative assessments. Thus the leading 
study of Indian secularism ( Smith 1963: 316) found "a profound contradiction between the 
objective of a casteless society and the method of evaluating the backward on a caste basis." 
Similarly, a renowned study of Indian politics ( Rudolph and Rudolph 1967: 150) describes the 
caste basis as the "critical contradiction" in the government's efforts to eradicate the effects of 
untouchability ­­ "a kind of blackmail in reverse," which "is not only profoundly disturbing but 
also an important source of alienation and rebellion." 
133. See, e.g., the covering letter of the Chairman of the Backward Classes Commission, the 
Memorandum of the Home Minister on the Commission's Report (1956), and the various reports 
cited in chap. 6, §C, below. 
134. V. T. Rajshekar Shetty in Sunday Stateman Magazine, Feb. 19, 1978. 
­74­ 
is no untouchability today in the manner which existed before Gandhiji except in rural areas. . . ." If 
the cheris (untouchables' hamlets) in the villages are broken up "untouchability will cease to exist as 
a fact. The Brahmins, kshatriyas and vysyas are getting thoroughly mixed up through business, 
marriage and their non­recognition for any special representation, job or privilege." It is only the 
Harijans that remain a distinctive element, and their perpetuation for the purpose of political 
exploitation is the result of "defining the Scheduled Castes by birth."  135.   The solution, again, is to 
phase out preferences on an ascriptive basis and substitute economic and educational measures of 
backwardness. 
Without taking up the task of assembling the evidence relevant to the sweeping assertions in these 
arguments, I present them as specimens of a style of discourse that remains relatively constant. They 
display a notable unanimity: whether you take the Statesman view that untouchability is undented or 
the Indian Express view that it is vanishing, the conclusion is that preferential treatment for low 
castes should be abolished. 
There is no comprehensive study from which we might gauge the extent of public approval or 
disapproval of preferential treatment, but some scattered survey data enable us to sketch a very 
rough profile of public views. Anant's useful 1968 survey of three localities in North India reveals a 
consistent pattern of sharp division about preferential treatment. For example, just over half (52%) 
of his respondents favor reservations of legislative seats for Scheduled Castes, but 62% thought they 
should be terminated (in 1970, when they were to expire). Seventythree percent favored special 
educational facilities for Scheduled Castes.  136.   Anant conducted a follow­up study four years later 
after the flowering of Mrs. Gandhi's egalitarian rhetoric and after the 1969 renewal of reserved seats 
for another ten years. In 1972 he found that an even higher percentage (86%) of residents of these 
localities favored special educational facilities, and there had been a striking increase in 
____________________ 
135. K. Santhanam in Indian Express, March 6, 1978. The view that only compensatory treatment 
prevents the disappearance of caste inequality had greater currency in the 1960s, before the 
upsurge of rural violence against untouchables. For example: "There is at present no problem [of 
untouchability] and all that need to be done has been done and more. . . . It is the perpetuation of 
propaganda on behalf of 'untouchables' in this country that serves as a constant reminder to those 
born in the 'untouchable' castes that they are 'untouchables.' The sooner the propaganda is 
stopped, the better it will be for all concerned, and for the country" (letter from B. G. 
Balakrishnan in the Economic Weekly 17: 624 [ 1965]). 
136. Anant 1972: 92, 97. Cf. Mohopatra ( 1974) finding that of a 1973 sample of political science 
majors (N = 328) in Orissa, 45% favored abolition of reservation of seats in the state legislature, 
and 38% opposed. 
­75­ 
support for reservation of seats in legislative bodies, which was now approved by two­thirds of the 
respondents.  137.   
Rural, lower caste, and less­educated people were consistently more supportive of preferential 
treatment than urban, higher caste, and moreeducated respondents.  138.   This pattern is strikingly 
illustrated (in table 3 ) in the 1968 responses to the question of whether reservations for Scheduled 
Castes in government employment were "necessary" or "ruinous." 
One recent study reports that "the so­called middle range backward castes offer a stiffer resistance to 
the rising status of the Harijan than the 
TABLE 3 VIEWS OF RESERVATION IN GOVERNMENT EMPLOYMENT FOR 
SCHEDULED CASTES: THREE NORTH INDIAN LOCALITIES IN 1968 
Responses 

No Response
Necessary Ruinous
N  % 
%  % 
Groups 
By Location 
Urban  371  43.1  56.6  0.3 
Rural  374  65.5  33.7  0.8 
By Caste 
Brahmin  140  33.6  66.4  0 
Kshatriya  115  46.1  53.9  0 
Vaishya  209  46.4  53.1  0.5 
Sudra  113  70.8  27.4  1.8 
Harijan  168  76.2  23.2  0.6 
By Educational Level 
Illiterate  121  86.8  12.4  0.8 
Less than H.S.  235  68.1  30.2  1.7 
High School  151  49.0  51.0  0 
College, but less than 
91  33.0  67.0  0 
B.A. 
B.A.  95  27.4  72.6  0 
Postgraduate  51  19.6  78.4  2.0 

TOTAL GROUP  749  54.3  45.0  0.7 

Source: Anant 1972: 100. Calculations of significance will be found there. 
____________________ 
137. Anant 1978. 
138. The pattern of higher caste hostility to government help for Scheduled Castes reappears in the 
responses to Anant's open­ended inquiry about "what else should the Government do to help the 
Scheduled Castes." See Anant 1972, chap. 10 
­76­ 
very highest castes. . . . [T]he former . . . feel that their position is threatened by the growing power 
of the scheduled castes. They are also jealous of the benefits and advantages available to the 
scheduled castes under the constitution."  139.   The survey data concur that educated and urban people 
are personally more accepting of mobile untouchables and more opposed in theory to the 
continuation of the caste system.  140.   But even though less­advanced segments of the population 
may have more to fear from the direct competition of Scheduled Castes, it is the higher castes, the 
educated, and the urban who seem most opposed to preferential treatment. 
The articulate criticism of the policies by intellectuals taps a broad vein of dissatisfaction among 
elite groups. A sizable minority of highranking government officials disapprove these policies. 141. 
There is intense resentment of them among government servants, epitomized in the mocking 
reference to the Scheduled Caste officer as a "son­in­law of the government" or a "Government 
Brahmin."  142.   
Although the educated elite have found little to say in its favor, preferential treatment has proved 
durable if not popular among politicians.  143.   And members of the beneficiary groups remain 
generally favorable. While not unaware of some of the costs, Scheduled Castes and Tribes are 
convinced that preferential treatment is benefi­ 
____________________ 
139. Sachchidananda 1977: 172. 
140. See Anant 1972, chaps. 4, 5, 6. In a survey of seven selected Rajasthan villages on readiness to 
accept social change, Oommen found that admission of lower castes to educational institutions 
engendered much less resistance than special development programs and economic concessions. 
He notes that his more change­prone, educated respondents shared a "general disapproval . . . in 
regard to accepting caste as the criterion for extending economic benefits. . . . [T]he educated 
and the literate are likely to feel that special concessions based on caste deprive them of their 
legitimate opportunities" ( 1968: 935). 
141. Taub ( 1969: 183) asked 55 high­ranking officials in Orissa whether they approved or 
disapproved of reserved places in legislatures and administration: 18 (32.7%) disapproved. These 
results are remarkably close to those of Subramaniam ( 1971: 118), who surveyed 98 young 
I.A.S. recruits and 46 older I.A.S. officers in 1963. Thirty­seven percent of the veterans and 34% 
of the recruits favored abolition of special concessions for SC and ST in recruitment for public 
services; only 8% of the veterans and 13% of the recruits thought preferential treatment should 
be expanded; the others chose, "somewhat unenthusiastically," to retain it. 
142. The term "son­in­law" is used, Chitnis ( 1977: 210) observes, to connote "a person whose every 
whim is indulged . . . discriminately." Taub ( 1969: 187) reports the "explosive hostility" that 
inquiry about reservations elicited from "sophisticated and highly educated" I.A.S. officers. On 
resistance to reservation orders in government offices, see chap. 4, below. 
143. Some scattered survey data suggest that elites are arrayed along a scale from politicians, who are 
most favorable, to intellectuals, who are most opposed. In 1964 and 1965 surveys of attitudes 
toward community development programs, various groups of 
­77­ 
cial. In Anant 1968 survey of three areas of North India, Harijan respondents supported reserved 
legislative seats by four to one.  144.   While 72% of other respondents would have terminated 
reservations in 1970, only 26% of Harijans would, far less than any other segment of the population. 
 145   Unlike all other segments, a majority of Harijans disagreed with the assertion that reservation of 
seats tended to separate them.  146.   There was among the Harijans a significant minority: 26% would 
have abolished reservations in 1970, and 23% thought reserved posts "ruinous." But on every 
measure there was more support for preferential treatment from Harijans than from higher caste 
groups. Four years later the support of Harijans for preferential policies was even higher.  147.   
Similar overwhelming support for preferential treatment is shown in Aggarwal and Ashraf's survey 
of 500 Scheduled Castes household heads in Haryana. Only 2.8% thought that guaranteed political 
representation was not a good thing.  148.   Although there was less agreement about the usefulness of 
job reservations, sentiment to continue them for a long time was almost unanimous.  149.   
____________________ 
respondents were asked whether they agreed that "proper implementation of the development 
programmes would automatically contribute to the welfare of all including poorer sections; no 
special programmes for them are necessary." The percentage who chose this statement over an 
alternative which stated or implied that special treatment was essential is as follows: 

Opposition to Special Treatment Among Various Elite Groups ( 1964­65) 

Respondents 
(n) 
Percent agreeing that no
special programmes are
necessary 
Opposition M.P.'s ( 1965) 
(23) 

Congress M.P.'s ( 1965) 
(77) 
20 
Senior Press and Public Figures ( 1965) 
(63) 
30 
Officials ( 1964) 
(360) 
34 
Pradhans, Sarpanches, Panches ( 1964) 
(1,440) 
38 
Executives of cooperative societies ( 1964) 
(900) 
40 
Educationists ( 1965) 
(37) 
46 
Sources: Indian Institute of Public Opinion, Monthly Public Opinion Surveys No.
109­10: 59: No. 124: 19. 
Note: 
Although the statement refers to the "poorer sections," it was prefaced in the
interview by a reference to the "weaker sections." The alternative: "Benefits of
the development programmes go normally to the relatively stronger [and
economically powerful] sections in the village." The alternative given to the three
large groups of respondents omitted the words "and economically powerful" and
included an additional clause: "a special programme for the poor sections is,
therefore, essential." 
144. Computed from Anant 1972: 97. 
145 Anant 1972: 92. 
146. Anant 1972:96. 
147. Anant 1978. 
148. Aggarwal and Ashraf 1976: 145. 
149. Aggarwal and Ashraf 1976: 132­33. 
­78­ 
There is similar endorsement of preferential treatment by the next generation of educated 
untouchables. Thus Chitnis finds that Scheduled Caste college students and high school students 
believe that the programs are beneficial.  150.   Almost two­thirds of the college students disagreed 
with the statement that "the Scheduled Castes have advanced so much that they can stand on their 
own in open competition and do not need reservation any more."  151.   
That beneath their overwhelming approval of special treatment there is considerable unease and 
awareness of the costs is suggested by responses to three of Chitnis's items: the large sample of 
school and college students were asked whether preferences damage self­respect, foster dependence, 
and destroy initiative. In each case, almost threequarters of all the respondents end up in the "can't 
say/don't know" column ­­ a striking departure from the 5% to 30% "can't say" responses on other 
items.  152.   
There is no open public defense of the ancien régime. Everyone is against untouchability and 
against caste. Public debate takes the form of argument among competing views of what is really 
good for the lowest castes and for the country. These views involve a host of assertions about the 
effects ­­ beneficial and deleterious ­­ of compensatory discrimination policies. 
We shall address some of these contentions in the course of this study. For now, a summary of the 
main arguments for and against preferential treatment may help to sensitize us to the issues and alert 
us to be on the look­out for relevant evidence as we examine the programs, their implementation, 
and their treatment by the courts. As we proceed to discuss specific preferential arrangements, we 
shall from time to time attempt to sort out the claimed benefits and costs. Here, it may be 
appropriate to sketch out in the most general terms the full range of claims that are made as to 
benefits and costs ­­ the various ways in which the policy of compensatory discrimination allegedly 
helps or hurts the protected groups, others, and India as a whole. 
Rough and redundant as it is, this anthology of claims will provide us with a checklist that will help 
in devising appropriate standards for evaluating specific schemes. For convenience each claimed 
benefit is paired with the opposite claim of cost. It would simplify matters if each of these pairs 
represented points on a single dimension that could be unambiguously measured. Unfortunately, the 
most that can be claimed for them is that each is a composite of sometimes reinforcing but 
____________________ 
150. Chitnis, 1977: 231. A much more primitive study almost 20 years earlier found similar results: of 
286 Backward Classes students receiving government scholarships, 88% favored reservations in 
admissions and 79% favored reservations in government services ( Ayyar 1965: 60­61). 
151. Chitnis 1977: 233. 
152. Chitnis 1977: 239. 
­79­ 
occasionally conflicting qualities that may conveniently be grouped together. Since the lines 
between the claimed effects are not always distinct, some overlap and redundancy is unavoidable. It 
was necessary, for convenience in labelling, to devise some rubrics that are not found in ordinary 
talk of these matters, imparting a somewhat stilted quality to the list. These claims are listed here (in 
table 4 ) in an order that proceeds roughly from the most focussed and immediate to the most 
capacious and remote, from those which speak of impacts primarily on the beneficiary groups to 
those which concern the shape and career of the whole society. 
This catalog is clearly not a set of explanations of why these policies were adopted, although such 
goals undoubtedly played a part. It is a set of standards for judging these policies. But whose 
standards? Obviously they are the author's but I would claim that they are more than the author's ­­ a 
claim supported by their provenance, for they are refinements and generalizations of arguments 
found in current Indian discourse about these policies. I would claim further that the list 
encompasses most of the standards that would occur to a disinterested policy­maker. (By this I mean 
a policy­maker concerned with these policies per se rather than with their implications for his 
political fortunes.) In this accounting of costs and benefits I have put to one side those benefits (and 
costs) that accrue to individual actors from supporting or opposing a particular program (apart from 
some that enter incidentally under the heading of diversion). This is not because I suppose that 
policy­makers omit consideration of the personal and political gains and losses that such positions 
entail. The course that policy takes is very much shaped by this second level of costs and benefits. 
But although any given actor may have his own schedule of priorities and his own admixture of 
second­level objectives, almost all would share at least some part of the goals implied by these 
standards. And virtually all would appeal for support in terms of these standards. That actors differ 
in their priorities and goals as well as in their estimates of fact does not reduce the usefulness of 
exposing and articulating standards for judging these policies. 
The evaluation of these compensatory programs involves a two­stage inquiry. First, there is what we 
might call the problem of performance. Does the program actually deliver the goods (more jobs or 
housing or better performance in schools or whatever)? In making such judgments we must be wary 
of all the pitfalls of measuring program effects. Having satisfied ourselves that the program has the 
projected effect, we then face what we might call the problem of achievement. Has the program 
produced the results that it is supposed to achieve ­­ do more jobs for Scheduled Castes produce 
considerate treatment by officials, or stimu­ 
­80­ 
TABLE 4 ALLEGED BENEFITS AND COSTS OF POLICY OF COMPENSATORY 
DISCRIMINATION 
DIVERSION 
REDISTRIBUTION  vs. 

These resources are enjoyed by a small
Preferences provide a direct flow of segment of the intended beneficiaries
valuable resources to the
and do not benefit the group as a
beneficiaries in larger measure than whole. 
they would otherwise enjoy. 
MISREPRESENTATION 
REPRESENTATION  vs. 

Preferences provide for participation By creating new interests which di­
decision­making by those who effec­ verge from those of the beneficiaries,
tively represent the interests of the preferences obstruct accurate repre­
beneficiaries, interests that would sentation of their interests. 
otherwise be under­represented or
neglected. 
ALIENATION 
INTEGRATION  vs. 

By affording opportunities for partici­
pation and well being, preferences pro­ By emphasizing the separateness of
mote feelings of belonging and loyalty these groups, preferences reduce their
among the beneficiaries, thereby opportunities (and feelings of) com­
promoting the social and political in­ mon participation. 
tegration of these groups into Indian
society. 
REJECTION 
ACCEPTANCE  vs. 

Preferences frustrate others by what
Preferences induce in others an aware­ they consider unfair favoritism and
ness that the beneficiary groups are educate them to regard the be­
participants in Indian life whose in­ neficiaries as separate dements who
terests and views have to be taken into enjoy their own facilities and have no
account and adjusted to.  claim on general public facilities. 

MANIPULATION 
INTEGRITY  vs. 

Preferences permit forms of action that Preferences subject these groups to
promote pride, self­respect, a sense of manipulation by others, aggravate
achievement and personal efficacy their dependency, and undermine
that enable the beneficiaries to contri­ their sense of dignity, pride, self­
bute to national development as wil­ sufficiency and personal efficacy. 
ling partners. 
OVER PROTECTION 
INCUBATION  vs. 

By broadening opportunities, prefer­ Preferences provide artificial protec­
ences stimulate the acquisition of skills tion, which blunts the development of
and resources needed to compete suc­ the skills and resources needed to suc­
cessfully in open competition.  ceed without them. 
­81­ 

MOBILIZATION  VS.  ENERVATION 


By cultivating talents, providing op­ By making them dependent, blunting
portunities and incentives and pro­ the development of talent, undermin­
moting their awareness and self­ ing self­respect, preferences lessen the
consciousness, preferences enhance capacity for organized effort in their
the capacity of the beneficiary, groups own behalf. 
to undertake organized collective
action. 
SEDATION 
STIMULATION  VS. 

By increasing the visibility of the be­ By projecting an image of comprehen­
neficiary, groups, promoting their sive governmental protection and pref­
placement in strategic locations, and erment, preferences stir the resent­
emphasizing the national commitment ment of others, allaying their concern
to remedy their condition, preferences and undermining initiatives for mea­
serve as a stimulus and catalyst of en­ sures on behalf of the beneficiary
larged efforts for their uplift and groups. 
inclusion. 
SELF­PERPETUATION 
SELF­LIQUIDATION  VS. 

These arrangements create vested in­
terest in their continuation, while dis­
The benefits of preferential treatment couraging the development of skills,
are mutually reinforcing and will
resources, and attitudes that would en­
eventually render unnecessary any able the beneficiaries to prosper with­
special treatment. 
out special treatment. 

UNFAIRNESS 
FAIRNESS  VS. 

Preferences compensate for and help Preferences place an unfair handicap
to offset the accumulated disablements on individuals who are deprived of op­
resulting from past deprivation of ad­ portunities they deserve on merit. 
vantages and opportunities. 
COMMUNALISM 
SECULARISM  VS. 

By reducing tangible disparities By recognizing and stimulating group
among groups and directing attention identity, preferences perpetuate in­
to mundane rather than ritual stand­ vidious distinctions, thereby under­
ing, preferences promote the develop­ mining secularism. 
ment of a secular society. 
STAGNATION 
DEVELOPMENT  vs. 
Preferences contribute to national de­ Preferences impede development by
velopment by providing incentive, op­ misallocation of resources, lowering of
portunity, and resources to utilize morale and incentive, and waste of
neglected talent.  talent. 
­82­ 
late educational accomplishment, or produce social integration? To what extent does delivering the 
jobs entail the costs alleged by critics of preferential treatment ­­ such as stigmatizing the 
beneficiaries, fomenting group resentments, lowering self­esteem? 
This is not a list of the direct effects of compensatory programs ­­ e.g., more jobs, higher literacy, or 
better nutrition. It is a list of the good and. bad effects that are alleged to flow from the performance 
promised by the program, including the effects attributable to the specifically preferential aspects of 
the program. Even if all of these dimensions enter into estimation of the overall working of the 
compensatory discrimination policy, it is not implied that all of them are involved in every specific 
scheme. Nor is it implied that they are to be accorded equal weight in making such evaluations. 
Presumably, specific schemes in different fields (education, housing, etc.) or for different groups 
(SC, ST, OBC) have a different mix of intended effects. And, the relative weights to be assigned to 
those effects (and to unintended by­products) will differ among various participants and observers. 
­83­ 
4 Reservations in Government Employment: A Closer Look at the Paradigm Program 

A MORE DETAILED account of one policy area may serve to give us a firmer sense of how such 
programs operate and an appreciation of the complexities of evaluating them. If only one policy 
area is to be developed at length, government employment is an obvious choice, for in a number of 
ways it is the paradigm case of the Indian policy of compensatory discrimination. It relies on the 
reservation device as the core of the program; it is promoted by both state and central government; 
policy in this area is made with due deliberation by informed agencies; the various groups of 
interested participants are attentive and responsive; the actors have recurrently invoked judicial 
intervention. It is an area that in the eyes of proponents and opponents occupies a central symbolic 
position in envisioning the compensatory discrimination policy. To all of this must be added one 
more factor: the data that make assessment possible, while insufficient in many respects, are more 
abundant, accessible, and readily comparable than in any other field of policy. 
Government in India is regarded, not as just another employer, but as one that affords a degree of 
security, prestige, and authority not obtainable elsewhere. For a broad sector of the India population, 
including both established middle classes and many aspiring groups, government employment is the 
1. (Perhaps it would not be unfair to say that the social role of government 
prime focus of ambition.    
employ­ 
____________________ 
1. The overwhelming preference for government over private employment is confirmed by some 
survey evidence. Eldersveld, Jagannadham, and Barnabas ( 1968: 31­32) report that in Delhi 
State 76% of urban and 89% of rural respondents preferred government over private­firm 
employment "if the pay were the same." Similar percentages accorded greater prestige to doctors, 
clerks, and watchmen in government employ than to those privately employed. 
­84­ 
2. As one Indian 
ment in India is comparable to that of corporate employment in the United States.)    
observer put it: 
With us the Government, be it liked or not, be it respected or not, is surrounded with an 
aura of glamour. It is the fount of power, of prestige, of patronage, and last but not least, 
of employment. If we are on good terms with it, and its representatives, life is all the 
3. 
more easy for us.    
Because government has been their benefactor and because their access to employment in the 
private sector has been severely limited, dependence upon government for employment 
opportunities is especially pronounced among Scheduled Castes (and to some extent among other 
4. Government employment is a matter not only of private ambition but of group 
backward groups).    
advancement. Employment in the upper reaches of government service is often used as a measure of 
5. 
advancement and security.    
The presence of its members in the administration had been regarded by many groups in India as an 
assurance of accessibility and sympathetic response to their interests and a warrant of fair dealing 
on the part of the government. 
Having men belonging to their own fold at different levels, gives the people a sense of 
confidence that they will get a fair deal, they will be treated with consideration and no 
officer will scorn them as people of no consequence.     6. 

Thus, from the outset "untouchable" leaders have seen in government employment not only career 
opportunites, but relief from unsymphathetic and oppressive administation, facilitation of the 
utilization of opportunities, and readier implementation of programs for their benefit. 7. 
Furthermore, government work is regarded as a source of prestige for both the individual and the 
group. A leading Scheduled Caste politician observes: 
____________________ 
2. Blaug, Layard, and Woodhall ( 1969:94 ff.) estimate that over two­thirds of employed graduates 
and nearly two­thirds of matriculates work in the public sector, which makes up over 58% of the 
"organized sector" (defined as public sector plus private firms employing at least 25 persons). 
3. Bonarji 1956. 
4. Cf. Dahl ( 1961 : 294) observations that American Blacks have had little choice but to seek 
employment in the public sector because of their exclusion from the private sector of the 
economy and that upwardly mobile Blacks have regarded government as warm and receptive. 
5. Cf. Dushkin 1972 : 179. This measuring rod is used not only by untouchables but by others as 
well. 
6. Tamil Nadu Backward Classes Commission 1971 : I, 84. 
7. M. C. Rajah 1925 ; R. Srinivasan at Indian Central Committee 1929 : 384; Ambedkar cited at 
Austin 1966:20. 
­85­ 
If today scheduled castes and scheduled tribes are anxious to get proper representation 
in the Government service it is not that it presents a solution for their economic 
8. 
difficulties but it adds to their prestige and gives them status in society.    
Over the years spokesmen for the backward classes have expended more concern on government 
employment than on any other scheme of preferences. 
A. THE SCOPE OF RESERVATIONS 
The central government has provided reservations in government employment for Scheduled Castes 
9. and for Scheduled Tribes (ST) since 1950.  10.   Since 1947, of posts recruited 
(SC) since 1943,    
directly on an all­India basis by open competitive examination, 12½% are reserved for SC and 5% 
for ST. (These percentages were raised to 15% and 7½% in 1970.)  11.   Of posts filled on an all­India 
basis other than by open competition, 16⅔% are reserved for SC and 5% (raised to 7½% in 1970) 
for ST.  12.   In the case of lower posts likely to attract candidates 
____________________ 
10. In 1947 the Government decided that no useful purpose would be served by reservations for 
Scheduled Tribes, as too few were available for appointments. In December 1947 instructions 
were issued that appointing authorities should keep in view the desirability of recruiting suitable 
tribal candidates to vacancies reserved for other minorities (RCSCST 1951 : 29). 
11. Times of India, 4 April 1970. 
12. The basic policy document is the Ministry of Home Affairs Resolution of 13 September 1950, 
reprinted at RCSCST 1951 : 120­22. The stated percentages apply 
8. Jagjivan Ram at Planning Commission 1965:21. Similarly, the Tamil Nadu Backward Classes 
Commission observed that government service is primarily a status symbol and that "[t]he 
symbolic and prestige value of Government Service is in almost direct proportion to the extent of 
the castes' social backwardness" ( Tamil Nadu Backward Classes Commission 1971:I, 39, 88). 
They find the craze for government service part of an "aversion to follow the traditional pursuits . 
. . due to social stigma and low ritual status" even where wages are satisfactory (id., at 129). 
9. Instructions were issued in July 1934 that duly qualified candidates belonging to the Depressed 
Classes should not be deprived of fair opportunities for appointment because they could not 
succeed in open competition, but no definite percentage was specified. In 1942 the government 
concluded that the slow progress was due mainly to the unavailability of suitable candidates, but 
that reservation of a definite percentage might provide a needed stimulus. In August 1943 orders 
were issued reserving 8⅓% of posts filled by direct recruitment, raising the age limit and 
lowering the examination fee for members of these castes. In June 1946 the percentage of 
reserved vacancies was raised to 12½%, corresponding to their percentage in the population 
(123/4% in undivided India). In August 1947 reservation for other minorities in posts filled by 
open recruitment was withdrawn, though it was continued in respect to posts filled otherwise 
than by open competition. Reserved for Scheduled Castes were 162/3% of posts filled on other 
than open competition, a percentage apportioned to their presence in the population of divided 
India (15½%) (RCSCST 1951:27 ff.). 
­86­ 
only from the locality, percentages are fixed on the basis of the population of SC and ST in the area. 
The central government has not at any time provided any reservation in government employment for 
Other Backward Classes. 
At the state level there are similar reservations for SC and ST. Reservations range from 5% to 25% 
of posts for SC and 3% to 80% for ST. In a few places, both groups share a common reservation. 
The combined reservation in higher post  13.   for both groups ranges from a low of 10% in Kerala to a 
high of 40% in Orissa. Several states reserve a slightly larger percentage of lower­grad posts.  14.   All 
the states except Orissa, Madhya Pradesh, and Jammu and Kashmir had reservations in effect by 
1951.  15.   In several states these reservations were part of wider schemes for promoting entry of 
backward classes into government service. Over the years, the reservations for SC and ST have 
remained largely unchanged, although they have increased by a few percentage points in erstwhile 
Madras, West Bengal, and Punjab, and almost doubled in Uttar Pradesh, from 10 to 18%. In several 
states where these groups had been treated together with Other Backward Classes, reservations for 
them were separated out at the insistence of the central government. 
Although there are no reservations for Other Backward Classes the Centre, such reservations are not 
uncommon at the state level.  16.   In 1978, at least thirteen states reserved posts for backward classes 
other than the Scheduled Castes and Tribes. Reservations for Other Backward Classes were found 
throughout southern India, in Maharashtra and Gujarat, and across the north from Bihar to Kashmir. 
The proportion of reservations for these groups was heaviest in the south. 
Reservations are accompanied by an array of other special provisions, designed to enhance the 
ability of the preferred groups to compete successfully for government posts. These include age 
concessions­­i.e., relaxation of the maximum age for entry into the service; fee concessions­­i.e., the 
waiver or reduction of examination fees; reduction of the minimum qualifying marks on 
examinations and 
____________________ 
primarily to posts in Classes I and II. The Resolution provides that for lower posts likely to attract 
candidates from the locality only, the Government shall fix percentages of reservations upon 
taking into account the population of SC and ST in that area. 
13. Central government jobs are divided into four grades: Class I (senior administrative), Class II 
(other administrative), Class III (clerical), and Class IV (attendants, peons). On the classification, 
see Pai Panandiker 1964. The states and public sector employers employ similar classifications. 
14. Planning Commission 1965:285 ff. 
15. For a survey of reservations in employment dating from 1925, see BCC I : 127 ff. 
16. See table 6 for current figures, and chap. 6 for details. For a profile of the situation in the early 
1960s, see Planning Commission 1965:285 ff. 
­87­ 
the waiver of a passing mark on viva voce examinations; specific safeguards against retrenchment; 
authorization for Public Service Commissions and appointing authorities to relax minimum 
standards; freedom for Scheduled Caste and Scheduled Tribe jobholders to register for better jobs 
while working for the government; travel allowance (to reach the interviewing center), pre­
examination training for Scheduled Castes and Tribe aspirants for high­echelon posts, etc. 
Thus preferences not only exclude others from competing for some portion of the available posts 
(reservations per se), but let into the competition those who would not otherwise be eligible (e.g., 
age and minimum­qualification waivers), make it easier for them to enter the competition (e.g., fee 
waivers, travelling allowance), and in a few cases enhance their capacity to compete (i.e., through 
coaching). It should be added that not all of these devices are in effect in all places or for all posts. 
There has been, during the period that reservations have been in effect, a striking increase in the 
numbers of Scheduled Caste and Scheduled Tribe government servants. In 1953 there were 20 SC 
(0.35%) in Class I jobs. In 1975 there were 1,201 ­­ 3.4% of a much expanded work force. Similarly 
there was a jump from 1.3% to 5.0% of Class II posts and 4.5% to 10.7% of Class III. Scheduled 
Tribes increased from 0.14% to 0.6% of Class I; from 0.24% to 0.6% of Class II; and from 0.47% to 
2.3% of Class III. The small percentage changes conceal a great increase in absolute numbers of 
these groups. For example, during this period the total number of employees in Classes I and II 
more than sextupled (from 14,500 to 89,200). But the number of SC and ST in these classes 
increased almost 25 times, from 160 to 4,436. In spite of this increase these groups, making up over 
21% of the population, held less than 5% of these important and coveted positions. The increase at 
lower levels was less spectacular in percentage terms but more substantial in absolute numbers. 17. 
Table 5 summarizes the position in the central government in 1953 an 1975. 
During this period, the number of Scheduled Castes and Scheduled Tribes in upper grades increased 
manyfold over the very low starting point. The upward movement in their proportion of total 
employment in these grades was modest but steady. At the same time there was a very 
____________________ 
17. Available data does not indicate how much of the intake of Scheduled Castes and Tribes can be 
attributed directly to reservations­­i.e., how many of the successful candidates would have failed 
to obtain the post on merit. In the upper echelons at least the portion is high. Of the 59 SC and 
ST recruited to the I.A.S. in the years 1959, 1961, and 1962, only three would have succeeded 
without the reservation ( Planning Commission 1965:279). Presumably the ratio of candidates 
who succeed on merit is higher in less­exalted posts. 
­88­ 
TABLE 5 SCHEDULED CASTE AND SCHEDULED TRIBE EMPLOYMENT IN 
CENTRAL GOVERNMENT SERVICE, 1953 AND 1975 
Scheduled Tribes 
Scheduled Castes 

1953  1975  1953  1975 

No.  No.  No.  No. 


(%) 
(%)  (%)  (%) 

CLASS I  20  1,201  6  218 


(0.6%) 
(Higher Administrative)  (0.35%)  (3.4%)  (0.10%) 

CLASS II  113  2,695  21  322 


(0.6%) 
(Lower Administrative)  (1.29%)  (5.0%)  (0.24%) 

CLASS III  24,819  174,119  2,548  36,893 


(2.3%) 
(Clerical)  (4.52%)  (10.7%)  (0.47%) 

CLASS IV  161,958a  230,864b  14,512  49,464 


(Menial)  (20.52%)  (18.6%)  (1.80%)  (4.0%) 
Sources: I BCC 134; RCSCST 1974­75: 82. 
Notes:  a Includes Sweepers. 
b Excludes Sweepers 

substantial rise in the numbers of SC and ST in the inferior grades and a sizable increase in their 
proportion of the overall figures. 
The situation in the states, though somewhat uneven, is roughly similar. There have been small gains 
in higher services from practically no members of these groups to at least token representation. In 
lower services there are many of these groups in menial grades. Table 6 summarizes the situation in 
the states, comparing the reserved percentages with percentages of places actually held. 
Data on places actually filled by Other Backward Classes (OBC) is difficult to come by. The census 
gathers no statistics on these groups; the central government no longer recognizes them as a 
category; the states that maintain such reservations tend not to be forthcoming with data about 
them. The general impression is that there is no lack of candidates from the OBC to fill all of the 
posts reserved for them.  18.   This 
____________________ 
18. Thus OBC made up well over half of the "qualified applicants" for every grade of posts in 
Andhra Pradesh for the years 1962­63 and 1963­64 ( Andhra Pradesh Back­ 
­89­ 
TABLE 6 SCHEDULED CASTLES, SCHEDULED TRIBES, AND OTHER BACKWARD 
CLASSES IN GOVERNMENT SERVICE: POSTS RESERVED AND POSTS HELD, 
ABOUT 1976 

GRADE OF POSTS 

CLASS I  CLASS II  CLASS III  CLASS IV 

Gover­ Percentage % % % % % % % %
nment  Group  Year  Population  Reserved  Held  Reserved  Held  Reserved  Held  Reserved  Held  Sou
CENTRE  SC  1975  14.59  3.4  5.0  10.7  18.6 

ST  1975  6.93  0.6  0.6  2.3  4.0 

ANDHRA  SC  1974  13.27  14  5.5  14  3.3  14  5.5  14  14.0 
ST  1974  3.79  4  1.0  4  0.2  4  1.2  4  2.5 

OBC  1975  38.0  25 ...  25 ...  25  NI  25  NI 

ASSAM  SC  1976  6.10  7  NI  7  NI  7  NI  7  NI 


ST  1976  12.84  Hills 12  NI  12  NI  12  NI  12  NI 

Plains 10  NI  10  NI  10  NI  10  NI 

BIHAR  SC  1976  14.11  14  NI  14  NI  14  NI  25  NI 
ST  1976  8.74  10  NI  10  NI  10  NI  20  NI 

OBC  1978  NI  26  NI  26  NI  26  NI  26  NI 

GUJARAT  SC  1975  6.84  5  NI  5  NI  5  NI  5  NI 


ST  1975  13.97  10  NI  10  NI  10  NI  10  NI 
OBC  1972  NI  NO RESERVATIONS AFTER 1959 
­90­ 


HARYANA  SC  1975  18.89  20  3.6  20  2.6  20  8.2  20  27.6 

HIMACHAL 
PRADESH  SC  1967  22.24  22.5  1.3  22.5  7.6  22.5  8.5  22.5  15.3  A 

ST  1967  4.09  5  1.7  5  8.9  5  2.7  5  5.1 

JAMMU & 
KASHMIR  SC  1976  8.26  8  NI  8  NI  8  NI  8  NI  A 
OBC  1973 ...  42  NI  42  NI  42  NI  42  NI  D 
KARNATA­ 
KA  SC  1973  13.14  15  4.1  15  4.1  15  4.1  15  4.1  A 
ST  1973 .78  3  0.4  3  0.6  3  0.8  3  1.9  A 
E.F 
OBC  1978  [45]  40  NI  40  NI  40  NI  40  NI 

KERALA  SC  1974  8.30  8  2.0  8  2.7  8  7.0  8  2.5  A 


ST  1974  1.26  2  0.1  2  0.1  2  3.0  2  1.5  A 

OBC  1970  NI  40  NI  40  NI  40  NI  40  NI 

MADHYA  SC  1976  13.09  15  2.0  15  1.9  16  6.0  16  14.4  A 

PRADESH  ST  1976  20.14  18  0.6  18  0.8  20  2.8  20  5.0 

MAHA­  SC  1975  6.00  13  2.9  13  4.2  13  11.4  13  25.2  A 
RASHTRA  ST  1975  5.85  7  0.6  7  0.8  7  2.8  7  5.0  A 
OBC  1966  12  14  NI  14  NI  14  NI  14  NI  H 
­91­ 

GRADE OF POSTS 

CLASS I  CLASS II  CLASS III  CLASS IV 

Govern­ Percentage % % % % % % % %
ment  Group  Year  Population  Reserved  Held  Reserved  Held  Reserved  Held  Reserved  Held 
ORISSA  SC  1976  15.09  16  0.4  16  1.6  16  6.6  16  14.7 

ST  1976  23.11  24  1.1  24  1.2  24  22.3  24  7.7 

PUNJAB  SC  1976  24.71  14(R&P)  6.3  14(R&P)  4.7  20(P)  12.0  20(P)  37.9 

OBC  1964  NI  1  NI  1  NI  1  NI  1  NI 

RAJASTHAN  SC  1976  15.82  16  3.4  16  14.3  16  12.1  16  18.4 

ST  1976  12.11  12  1.7  12  7.2  12  8.2  12  10.0 

TAMIL  SC  1976  17.76 


NADU  ST  1976 .75  16  4.0  16  11.0  16  10.0  16  11.0 

OBC  1972 ...  31  31  31  31 

UTTAR  SC  1976  21.00  18  3.1  18  3.6  18  8.0  18  12.8 
GRADE OF POSTS 

CLASS I  CLASS II  CLASS III  CLASS IV 

Govern­ Percentage % % % % % % % %
ment  Group  Year  Population  Reserved  Held  Reserved  Held  Reserved  Held  Reserved  Held 
PRADESH  ST  1976 .21  2  ­­  2  0.1  2  0.2  2  0.4 
OBC  1978  NI  16  (R)  NI  16  (R)  NI  16  (R)  NI  16  (R)  NI 
15  (P)  15  (P)  15  (P)  15  (P) 

WEST  SC  1969  19.90  15  2.2  15  3.2  15  3.2  15  13.7 
BENGAL  ST  1969  5.70  5  0.1  5  0.5  5  1.0  5  2.4 
A: RCSCST 1975­77: II, 15; B: Andhra Pradesh Department of Social Welfare, Performance Bud
1975­76 : 1­3; C: Overseas Hindustan
Times, 6 April 1978; d: Janki Prasad v. State of Jammu and Kashmir, A.I.R. 1973 S.C. 930, 933,
E: The Statesman, 18 March 1978 : 4; F:
Sources:  Government of Karnataka 1975: I, 316­317; G: Jayasree v. State of Kerala, A. I. R. 1976 S.C 2381
H : Maharashtra Backward Class
Welfare Wing 1966­67; I: Tamil Nadu Directorate of Backward Classes 1972­3: 1; J: The Statesm
March 1978; 8 June 1978; K:
RCSCST 1974­75: 82; L: State of Punjab v. Hira Lal, A.I.R. 1971 S.C. 1777, 1778. 
Population figures for SC and ST from 1971 Census. State service Classes A, B, C, D, are shown
Note:  Classes I, II, III, IV, respectively.
R = Recruitment. P = Promotion. NI = no information available. 
­92­ 
is particularly the case where the OBC category is broad and includes some prosperous and 
politically potent groups. In some cases the actual posts obtained by OBC are augmented beyond the 
stated reservation by the carryover of unfilled reservations from the SC and ST. (Thus the Punjab 
Evaluation Committee found that where reservations were made interchangeable "backward classes 
fed at the cost of scheduled caste vacancies out of all proportion.")  19.   
B. THE EFFECTIVENESS OF RESERVATIONS 
In estimating the effectiveness of reservations, we must remember that reservations are not of the 
full complement of posts but only of current appointments.  20.   Therefore, if the size of a service 
remained constant and the reservations were filled every year, it would take the length of an average 
service career (approximately 33 years) for the percentage of posts held to reach the percentage of 
appointments reserved.  21.   At the time our most recent data was collected, reservations had been in 
effect at the Centre (and in most states) for a period roughly three­quarters the length of a service 
career. We might then expect that if reservations were fully utilized, the percentage of posts filled by 
these groups should now be approaching three­quarters of the 
____________________ 
ward Classes Commission 1970: 120). See also Department of Social Welfare 1969: 290. R. D. 
Bhandare ( Government of Maharashtra 1964: 152 ff) gives the following figures for posts held 
by OBC in Maharashtra in 1961: 
Class of Posts 
Reservation 
% Held 

12½% 
0.4 
II 
(shared with SC and ST) 
2.1 
III 
9% 
9.3 
IV 
11% 
7.7 
19. Punjab, Welfare Department 1966: 91. 
20. There have been instances of states reserving much higher portions of current intake, the 
enhanced figure to apply until the reserved percentage was reached. E.g., in the early 1950s, 50% 
of all vacancies were reserved in Madhya Bharat (RCSCST 1952: 69) and 100% in Saurashtra 
( Backward Classes Commission 1955: 132). In 1960 Bihar announced a policy of hiring only SC 
in Class IV posts until the quota was reached ( Department of Social Welfare 1969: 295). This 
policy was withdrawn after Devadasan v. Union of India, A.I.R. 1964 S.C. 179 (discussed in 
chap. 12,§B, below) established an upper limit of 50% on such temporarily enhanced 
reservations. In the late 1960s the Punjab was reported to have adopted a policy of reserving 50% 
of the intake of constables ( Department of Social Welfare 1969: 372); U.P. had a policy of hiring 
45% SC in inferior services until its 18% quota was reached (id., at 389). 
21. With a policy of massive reservation for all communities other than Brahmins (i.e., for 95% of its 
population), "it took [ Mysore] a period of 30 years to reduce the percentage of Brahmins from 
69% to about 39%" in 1952 (I BCC 138). 
­93­ 
reserved percentage of appointments. But this would be a considerable understatement, for during 
this period the size of the government's work force has more than doubled. Far more than three­
quarters of present government employees have entered the service during the period since 
reservations came into effect. So we might expect, very conservatively, that if reservations had been 
fully utilized each year, Scheduled Castes and Scheduled Tribes would be present in government 
service at about three­quarters of the percentage reserved for them. Actual representation falls short 
of the expected. Thus we would expect, conservatively,  22.   to find over 9% of Scheduled Caste 
employees in Class I central government jobs, but by 1975 there were only 3.4%; we would expect 
3.75% Scheduled Tribes, but there were only 0.6%. 
However, at the very peak of the system, in the elite Central Government Services, and down in the 
more populous clerk grades that make up the bulk of the pyramid, actual employment comes much 
closer to the expected. We would expect to find over 9% SC and 3.75% ST in the Indian 
Administrative Service (IAS) and in fact we find 8.56% and 4.08%, respectively.  23.   In Class III 
posts, we find 10.7% SC and 2.3% ST.  24.   
This general profile is confirmed by a somewhat more cumbrous and slightly more refined way of 
measuring the effectiveness of reservations.  25.   By comparing employment for the years 1953 and 
1975 (the earliest and latest for which complete figures are available) we can estimate the number of 
new jobs. To these are added the estimated 
____________________ 
22. Our expectation is that the levels of reservation used to compute the expected presence of SC, 
and ST are 12½% and 5% respectively, the levels that obtained until 1970, when they were raised 
to 15% and 7%. 
23. RCSCST 1975­77: I, 51 (this is as of Jan. 1, 1976). The representation of SC and ST in the 
various other elite service corps (Indian Police Service, Indian Forestry Service, etc.) lags behind 
that in the I.A.S., so the figures here should not be taken as applicable to the whole of this 
stratum of jobs (id.). 
24. The rate of effectiveness for Class III is overstated by our use of the 12½% figure. Many of these 
posts are no doubt covered by the 16⅔ reservation for posts not recruited directly on an all­India 
basis. 
25. All of these measures of "effectiveness" are subject to several distorting factors. First, they 
overestimate the "expected" total (by assuming that the coverage of reservations is universal) and 
thus underestimate the effectiveness of reservations. Second, the), overestimate effectiveness by 
underestimating the reservation in the lower grades and in the later years by using the 12½% and 
5% figures throughout. Third, they overestimate the effect of reservations by attributing all of the 
increases of these groups to the reservation even though some portion of that increase would 
have been attained on "merit" without any reservation. But this seems fair in so far as the 
reservation is intended not only to secure places to those who would otherwise not get them, but 
also to motivate aspirants from these groups and to insure that deserving candidates are not 
eliminated by prejudice or misjudgement. 
­94­ 
number of jobholders who were replaced (owing to retirement or death) during that period, and a 
total of the number of jobs to be filled is arrived at. Then we can calculate the number of Scheduled 
Caste employees who would have been hired were the reservation universally applicable and 
completely filled. This is the number by which SC employees should have increased by 1975 
(assuming that none of them were among the deceased or retired, an assumption that introduces 
little distortion since there were so few SC among older employees). Then the actual increase in SC 
employees is compared with the expected increase. 
Table 7 confirms that actual increases most closely approximate the projected increase at the very 
peak and in the clerical grades. In the wide ranges of administrative grades, the increase falls far 
short of that projected. In Class I posts, the actual increase is a bit more than a quarter of the 
expected for the Scheduled Castes and a bit more than a eighth for Scheduled Tribes. In Class II it is 
four­tenths for SC but less than an eighth for ST. The gap is considerably greater for ST at every 
level except at the very top. Both calculations depict this U­shaped distribution. Twenty­five years of 
reservations have produced solid representation in the lower ranks and a strong and growing 
presence at the very pinnacle, but there remains great under­representation throughout the middle to 
higher ranges. 
By dividing our period into three periods of approximately equal length and making the same 
computation of effectiveness for each period, we can gauge changes in the effectiveness of 
reservations over time. There has been a great increase in the effectiveness of reservations at the 
peak. In the latest period Scheduled Castes and Tribes have been coming into the all­India services 
in numbers close to the reserved portion. In Class I and Class II there have been substantial 
increases in effectiveness for SC, but reservations for ST in Class I and Class II have shown much 
smaller improvements. This reflects the siphoning effect of the availability of the highest positions 
and the limited number who can meet the basic educational qualifications.  26.   In Class III, there has 
been improvement in the recruitment of ST, while hiring of SC remained stable. 
In explaining the profile of effectiveness and ineffectiveness, we must take into account the 
somewhat deceptive character of the announced reservations. A very large number of positions are 
effectively out­ 
____________________ 
26. There was not a complete count of SC: and ST graduates in the 1961 census. There is a count of 
urban graduates (6,307 SC, 1,087 ST). If we assume (undoubtedly an overestimation) that a 
comparable percentage of rural matriculates from these groups had earned a university degree, 
we derive an estimated total of about 16,000 SC graduates and 3,000 ST graduates in 1961 (based 
on census [ 1966]). 
­95­ 
TABLE 7 APPROXIMATE EFFECTIVENESS OF RESRVATIONS FOR SCHEDULED 
CASTES AND SCHEDULED TRIBES IN CENTRAL GOVERNMENT SERVICES OVER A 
21½ YEAR PERIOD (1953­1975) 

1  2  3  4  5  6 
RATIO
PROJECTED  ACTUAL
NO NEW  ESTIMATED  INCREASE  PROJEC
POSTS  REPLACEMENT  ESTIMATED  IF 12.5%/5%  PERCENTAGE  INCREA
GRADE  CREATED  OF  TOTAL  RESERVATI ACTUAL  OF EST.  A. Sched
OF  SINCE  INDIVIDUALS  HIRING  ON FOR SC/ INCREASE  TOTAL  Castes
POSTS  1953  a
SINCE 1953     SINCE 1953 b  ST  OF SC/ST  HIRING 

I.A.S     2,136  734  2,870  359  359  9.0%  72
I  29,310  3,834  33,144  4,143  1,181  4.0%  28
II  45.373  5,837  51,210  6,401  2,582  5.0%  40
82
B. Sched
III  1,083,097  366,200  1,448,297  181,162  149,300  10.3% 
Tribes


I.A.S     2,136  734  2,870  143.5  131  4.5%  91
I  29,310  3,834  33,144  1,657.0  212  0.64%  12
II  45,373  5,837  51,210  2,560.0  301  0.58%  11
III  1,083,097  366,218  1,449,315  72,465.0  34,345  2.30%  47
Source:  Derived figures at I BCC 134; RGSCST 1974­75:82. 
Note:  aBased on estimate of 33 years for average length of service career (i.e., .667 × number of posts in 195
bSum of new posts and replacements (col. 1 + col. 2). 
cPeriod for Indian. Administrative Service runs from January 1954 to 1 January 1976. 

­96­ 
side the scope of the announced reservations. Some departments are not covered; posts filled by 
promotion (over one­quarter of all posts) are not covered (except in the few instances indicated 
above); posts filled at the creation of a new office are not covered; nor, in some cases, are temporary 
posts. A very considerable proportion of government posts are thus exempt from the coverage of 
reservation­­just how many is hard to say.  27.   
Nonetheless, a very large number of reservations have gone unfilled. The poor showing is attributed 
by the appointing authorities to the lack of qualified candidates among the Scheduled Castes and 
Scheduled Tribes. This is undoubtedly a factor in some cases, but the existence of large numbers of 
persons in these groups with the requisite educational credentials lends some credence to the 
frequent complaints that lack of qualification is not the whole reason.  28   Whatever the shortcomings 
of the candidates, they are compounded by the shortcomings that characterize the administrative 
implementation of reservations. Indifference or hostility on the part of the appointing authorities, 
insufficient publicization of vacancies, the sheer expense of application, all contribute to the under­
utilization of reservations. Although there is no data from which it is possible to estimate their 
relative contribution, it is clear that inflexibility and apathy are endemic in the administation of 
reservation policy. A committee investigating reservations in government service in Maharashtra 
observed: 
Generally speaking, there exists no thoughtful awareness of' the intentions and purpose 
of Government in reserving posts for the Backward Classes and consequently, the 
obvious and immediate reasons for the unavailability of candidates are accepted as a 
matter of course. . . . The picture that arises is . . . one of stolid apathy, lack of sympathy, 
understanding and flexibility. . . . Years of routine approaches and looking to the formal 
interpretation of orders and rules have contributed to the elimination of initiative, 
enthusiasm and thinking which is particularly essential when dealing with problems 
arising from a social malaise.  29.   
____________________ 
27. Some figures from the working of the Employment Exchanges give a rough indication of the 
shrinkage between announced and actual reservation coverage. In 1966, only 3.8% of the 
vacancies notified to the Employment Exchanges by the central government were reserved for SC 
and only 2.37% for ST. Among vacancies notified by the state governments, only 0.7% were 
reserved for SC and only 0.14% for Scheduled Tribes, (RCSCST 1966­67: 101­3). Of course, the 
disparity might also result from failure to notify the Employment Exchanges of reservations. 
28 At the end of 1968 there were about 94,000 "Matriculates and above" SC on the live register of 
employment exchanges, including over 7,000 graduates. There were about 14,000 matriculates 
and above ST, about 1,000 of them graduates, on the Live Register (RCSCST 1968­69: 211). 
29. Government of Maharashtra 1964: 15. 
­97­ 
Implementation of reservation policy is almost invariably a matter of very low priority for 
administrators.  30.   The intricate maze of provisions for implementing reservations often simply fails 
to secure adequate attention in offices overburdened with other matters. Thus the Maharashtra 
committee reported that "the directions concerning the carrying over of vacancies and to make up 
the existing shortfalls are either not understood or not followed."  31.   A more recent investigation 
revealed the same state of affairs. On the basis of spot checks in twenty government offices the 
Commissioner for Scheduled Castes and Scheduled Tribes concluded that "reservation orders and 
various other relaxations concessions were being ignored by almost all the organizations." 32. 
Officers and Staff charged with implementing concessions were "not fully conversant" with the 
relevant rules and instructions; there was an inclination to apply "whatever interpretation suits a 
particular officer."  33.   
This administrative obduracy is compounded by a genuine concern for efficiency. Many 
administrators believe that Scheduled Caste and Tribe candidates are of inferior abilities. As one 
recent account puts it, the "attitude . . . [of] a major section management, is one of 'protect your 
organization against SC/ST.'"  34.   Even if one regards the formal requirements with deserved 
skepticism, there is very likely some factual basis for such reactions. Reservations, as we have seen, 
have a siphoning effect: they draw in candidates with lower performance on various indicators. In 
many cases, lower performance undoubedtly reflects inexperience, lack of exposure, or other 
shortcomings that are temporary and readily remedied on the job. But to some extent the scores 
reflect limitations of ability or training that are not so readily overcome­­the less readily since 
others, expecting poor performance, expend less effort in on­the­job training. So any original 
deficiencies may in many cases be amplified by the reaction to them­­and the expectation of poor 
performance becomes a self­fulfilling prophecy! 
The Centre and some state governments have adopted various devices to offset the lack of zeal of 
appointing authorities: requirements that de­reservation be authorized by (or notified to) high 
ranking officers; requirements of periodic returns by appointing authorities and departments; 
creation of special grievance cells.  35.   
____________________ 
30. Nor is it a matter that engages those involved in devising policy for reform of the administration. 
For example, the Administrative Reforms Commission does not discuss reservations anywhere in 
its 1969 Report on Personnel Administration. 
31. Government of Maharashtra 1964: 15. 
32. RCSCST 1975­77: I, 65. 
33. RCSCST 1975­77: I, 45. 
34. Kirpal and Kelkar 1976: 7. 
35. Department of Social Welfare 1969:295­96; Punjab Welfare Department 1966: 96. 
­98­ 
The most visible and dramatic attempt to improve chances of Scheduled Caste and Scheduled Tribe 
applicants has been the establishment of pre­examination training centers to prepare them for 
highlevel examinations. The first of these, funded by the Home Ministry, was established in 
Allahabad in 1958. Of 411 candidates who prepared at this Centre between 1959 and 1966, 153 
(37.2%) were selected for the Indian Administrative Service, Indian Police Service or the other 
central services.  36.   By the late 1970s there were six such establishments operated by the central 
government, fifteen at the state level, and four new "coaching cum guidance centres."  37.   The 
performance of these centers has declined as their numbers have increased. From 1970 to 1975 the 
three longest established centers ( Allahabad, Madras, and Patiala) trained 652 SC candidates and 
202 ST candidates. Of these, only 49 of the SC (7.5%) and 43 of the ST (21.2%) were finally 
selected­­just 10.8% of the total.  38.   That such a program could go into its twentieth year without 
any evident urgency about upgrading this deteriorated performance is a telling indicator of energy 
and critical intelligence expended on the design and administration of preferential treatment. (It may 
also indicate that a large and inefficient program is preferred to a more select and effective one­­in 
spite of the administrative and human costs.) 
It is often contended that SC and ST candidates are unfairly eliminated at the later stages of the 
selection process, particularly where there is scope for personal evaluation in the form of an 
inerview or personality test.  39.   Scattered data on the selection process at the state level indicate very 
severe attrition at the interview stage. For example, over a two­year period in Madhya Pradesh only 
21% of SC candidates (who had fulfilled all of the educational and tested qualifications) survived 
the interview, compared with 41% of other candidates.  40.   On the other hand, figures on the work of 
the Union Public Service Commission suggest that the higher the qualifications for the post in 
question, the less the elimination of Scheduled Castes and Tribes in the later stages of the 
appointing process.  41.   
Preferences may extend not only to initial recruitment into government service but to advancement 
within it. Promotion depends in part on favourable evaluation by superior officers who rate each 
subordinate 
____________________ 
36. Planning Commission 1970: 416. 
37. RCSCST 1975­77: I, 53. 
38. RCSCST 1975­77: II, 11. 
39. E.g., RCSCST 1959­60: I, 143; Government of Maharashtra 1964: 154 ff. (Separate Note of Prof. 
R. D. Bhandare); Department of Social Welfare 1969: 280 (disqualification by "personality 
tests"). 
40. Department of Social Welfare 1969: 314. 
41. Planning Commission 1965: 266 ff. 
­99­ 
in a "Confidential Report" as Outstanding, Very Good, Good, etc.  42.   It is widely believed by 
Scheduled Caste employees that superior officers deliberately obstruct their promotion by giving 
them unfavorable ratings.  43.   The Commissioner for Scheduled Castes and Scheduled Tribes concurs 
that such reports are "generally biased."  44.   The difficulties that Scheduled Castes and Tribes have 
experienced in securing promotions have combined with the demand to be represented at higher 
levels to generate an intense concern to insure their access to promotion posts. Government has 
responded to this concern in various ways. 
Standards for promotion may be relaxed, as they are in some states, or there may be a definite 
reservation of posts to be filled by promotion. The central government's policy on reservation in 
promotions has vacillated. Somewhere in the range of one­quarter to one­third of all central 
government posts are filled by promotion, rather than by direct recruitment,  45.   including virtually 
all of the upper ranges of Class II and Class III. Promotions are made on three different bases and 
by different procedures: (1) by competitive examination open only to candidates within the 
department ("departmental examination"); (2) by selection ("merit"); (3) on grounds of "seniority­
cum­fitness." 
Before 1957, there was no reservation of any promotion posts. In 1957 the Home Ministry 
recommended that existing reservations be extended to those promotion posts filled by departmental 
examination­­a suggestion implemented by some departments and later held to be constitutional by 
the Supreme Court.  46.   In 1963 the Ministry of Home Affairs severely curtailed reservations in 
promotion. It extended them to selection posts as well as those filled by departmental examination, 
but it confined them to Classes III and IV and to only those posts to which there was no direct 
recruitment. This virtually ended them. Two and a half years later an organization of Scheduled 
Castes and Scheduled Tribes Central Government Employees declared: 
Since the withdrawal of the concessions in November, 1963 . . . not a single Scheduled 
Castes/Tribes candidate has been promoted either on the basis of departmental 
promotion examination or on the basis of ad hoc selections.  47.   
____________________ 
42. Dave 1966: 536. 
43. Department of Social Welfare 1969: 278. 
44. RCSCST 1975­77: I, 43­47. Bad confidential reports can lead not only to loss of promotions, but 
to enforced early retirement. The incidence of such enforced retirements of SC employees had led 
to special procedures for review at high levels. 
45. Dave 1966: 534. 
46. In General Manager v. Rangachari, A.I.R. 1962 S.C. 36, discussed in chap. 11. §B, below. 
47. Central Government Scheduled Castes and Scheduled Tribes Employees Welfare Association 
1966: 11. 
­100­ 
In 1968 the orders were further refined, and somewhat more scope was allowed for reservation in 
promotions in posts within grades II, III, and IV filled by departmental examination and for which 
direct recruitment did not exceed 50%. In promotion by selection from grade III to II, within grade 
II, and from grade II to the lowest rung of grade I, there was provision for weightage of individual 
SC and ST candidates in competing for up to 25% of the available vacancies.  48.   
Some state governments have followed the central government in extending reservations to the 
promotion stage.  49.   It is promotions rather than initial recruitment that have been the focus of the 
most controversy and the most adamant resistance from administrators and from other employees. 
The Commissioner complains, for example, of "the tendency of administrative authorities to 
continue ad hoc promotions for long periods to disallow the SC and ST their due share."  50.   Another 
measure of the sensitivity of this issue is provided by our litigation data: an overwhelming majority 
of the cases about reserved posts concern promotion rather than initial recruitment.  51.   
We have discussed state and central government services in the narrow sense: our figures do not 
include "government employment" in local bodies or in public sector undertakings. Reservations are 
applic­ 
____________________ 
48. Department of Social Welfare 1969: 252 ff. The weightage takes the form of giving SC and ST 
employees who are not considered unfit for promotion "one grading higher than the grading 
otherwise assignable to them on the basis of' their record of service­­i.e., if any Scheduled Caste . 
. . employee has been categorized . . ., on the basis of his record of service as 'good', he should be 
recategorized as 'Very Good,' (MHA [O.M. No. 1/12/67], dated 11th July 1968, reprinted at 
RCSCST 196768: 120). 
49. There is little information available on the working of the short­lived ( 1961­64) promotion 
reservations in Andhra. The Andhra Pradesh Backward Classes Commission ( 1970: 116­17) 
recommended against re­instituting promotion reservations because of the "demoralizing effect" 
on other employees. Punjab adopted a policy of reservation in promotions in 1963, reserving 
10% of all promotion posts for members of SC, ST, and OBC "provided [they] . . . possess the 
minimum necessary qualifications and have at least a satisfactory record of service," ( Punjab, 
Welfare Department 1966: 94). In 1966, the state government revised its policy, aligning it with 
the then central government policy­­expanded promotion reservations to 20% for SC and 2% for 
BC­­but confined these reservations to those Class III and IV posts filled by selection or 
departmental examination. An Evaluation Committee found the change "demoralizing, 
retrograde and injurious to the interests of the Scheduled Castes. It has practically denied all 
chances of promotion for Harijans to higher posts. In actual practice, ever since reservation in 
promotions was laid down, there have hardly been any benefits to the Harijans in Class II and 
higher posts, but with this change in policy, whatever semblance of hope there was, has been 
completely ruled out" (id., at 93). 
50. RCSCST 1975­77: I, 45. And see the Commissioner's Complaint about the unwillingness of the 
personnel authorities to modify a zoning scheme, designed to even out promotion prospects 
among departments, which was felt to undercut the reservation policy (id., at 50). 
51. See chap. 14, §C below. 
­101­ 
able to employment by local bodies in some states but not in others, but application of them has 
been rather lackadaisical.  52.   The percentage of SC in these jobs is less than in state services. 
Although extension of reservations to public sector undertakings was agreed to "in principle" as far 
back as 1954, such a policy had not been implemented by the mid­1960s. A comparison of 
government and public sector employment at that point and subsequently is instructive. The 
Elayaperumal Committee collected data on Scheduled Caste employment in 159 public 
undertakings whose total employment seems to be about equal to that of the entire central 
government. It was found that in Classes I, II, and III combined, these undertakings employed only 
one­eighth as many Scheduled Castes as did the Centre­­and this in spite of the fact that most of 
them were new enough to have done virtually all of their hiring after reservation had become 
accepted public policy. The comparison is given in table 8. A similar comparison for Scheduled 
Tribes is in table 9. The situation in public sector undertakings by state and local governments 
seemed not very different.  53.   
The sequel offers dramatic testimony to the efficacy of reservations in securing the inclusion of the 
beneficiary groups. Reservations in public 
TABLE 8 COMPARISON OF SCHEDULED CASTE AND SCHEDULED TRIBE 
EMPLOYMENT IN GOVERNMENT SERVICE AND PUBLIC SECTOR 
UNDERTAKINGS, MID­1960s 
Public Sector Undertakings (1965) 
Class 
Total 

SC 
Central Government 
% SC 
Posts ( 1966) 
Total 

SC 

%SC 

0.20 
20,379  361  1.77  35,512  73  II 

1.07 
30,001  974  3.25  15,820  274  III 

1,117,554  99,017  8.86  1,261,166  11,608  0.92 


Total 

ST 

%ST 
Total 
Public Sector Undertakings (1965) 
Class 
Total 

SC 
Central Government 
% SC 
Posts ( 1966) 
Total 

SC 

%SC 

ST 

% ST 

0.023 
20,379  103  0.52  a 
17,521     4  a  II 

0.034 
30,001  80  0.27  a 
23,730     8 a  III 

a  1,516  a 
1,117,554  12,356  1.10  246,654     0.65 
Source: Department of Social Welfare 1969: 257, 275; STTDP 257, 261. 
Note: a=This information is based on the returns from 56 Public Sector undertakings,
whereas the information on Scheduled Castes is based on 159 undertakings. 
____________________ 
52. See Department of Social Welfare 1969: 292, 298, 321, 370. 
53. E.g., Department of Social Welfare 1969: 298, 353­54. 
­102­ 
sector employment were introduced in the late 1960s.  54.   By 1975 the numbers of Scheduled Castes 
and Scheduled Tribes in the administrative grades were still below those found in government, but 
they were beginning to be more than token. And in Class III the totals were comparable to ­­ 
perhaps beyond ­­ those in government service. Incomplete returns show an addition of some 2,000 
Scheduled Tribes and Castes in administrative posts ­­ and almost 150,000 in clerical jobs. 
No attempt is made to extend the principle of' reservation to the private sector. There is no provision 
for reservation by those enterprises which are aided by the government or do business with it. 
(Some states have made requests along this line to their contractors.) Nor is there any requirement 
that employers in the private sector hire members of these groups. Until recently it was not required 
that private employers refrain from discriminating against them. The Untouchability (Offences) Act, 
the central government's anti­discrimination law, did not reach private employment until its 
amendment in 1976.  55.   Nor has any, state attempted to enforce its own anti­discrimination policies 
in private sector employment. It is not clear whether or how much there has 
TABLE 9 COMPARISON OF SCHEDULED CASTE AND SCHEDULED TRIBE 
EMPLOYMENT IN GOVERNMENT SERVICE AND PUBLIC SECTOR 
UNDERTAKINGS, 1975 
Scheduled Castes  Scheduled Tribes 
Public Sector 
Public  a 
   
Class  Government  a  Government 
Sector     I 

0.3% 
3.4%  1.4%  0.6%  II 

0.4% 
5.0%  3.0%  0.6%  III 

6.0% 
10.7%  13.7%  2.3%  IVb 

18.6%  26.2%  4.0%  11.9% 


Source: RCSCST 1974­75: 82, 93. 
aBased on returns from 120 undertakings. 
bSweepers omitted from Class IV. 

____________________ 
54. The CWSCST (4th) 2 ( 1969) traces the process through more than a decade of urgings, 
admonitions, and occasional compliance to establishment of a system of implementation of 
reservations in 1969. 
55. The original Untouchability (Offences) Act 1955 (Act 22 of 1955) was amended by the 
Untouchability (Offences) Amendment and Miscellaneous Provision Act, 1976, (Act 106 of 1976) 
and has become the Protection of Civil Rights Act. An addition to §4 (iii) makes it an offense to 
discriminate on grounds of untouchability in connection with "employment in any job." The 1976 
amendments strengthen the old act in many ways. For a discussion of the weaknesses of the old 
act and the tendency of the courts to construe its provisions narrowly, see Galanter 1972a. 
­103­ 
been any improvement in the employment position of Scheduled Castes and Tribes in the private 
sector. Given the enhanced opportunities in government service and the difficulties in the private 
sector, it is doubtful if many ambitious Scheduled Castes and Scheduled Tribes have attempted to 
make careers there. 
The greater presence of Scheduled Castes and Tribes in government employment than in the private 
sector reflects greater desire for government work, greater assurance of acceptance upon applying, 
the existence of personal networks of information about openings, and a selection process that 
makes some allowances for their deficiencies and lack of resources. It would seem to be the absence 
of these, not the presence of greater "prejudice" that is responsible for the poorer showing in posts 
where there are no reservations. For example, let us compare government jobs with Khadi and 
Village Industries Boards, establishments which we may assume would harbor a goodly share of 
devoted Gandhians who would score low in prejudice against untouchables. Yet we find that their 
performance in hiring Scheduled Castes is far worse than that of the government in general. For 
example, in Maharashtra the KVI Board had only 27 SC employees out of 653 (4.1%) ­­ and all of 
them in lower grades.  56.   In Orissa, the KVI Board had 29 SC (4.9%) among its 538 employees ­­ 
again all of them in lower grades.  57.   
If the data suggest a picture of slow but steady overall progress, it is progress against resistance that 
is seldom articulated but nonetheless dogged. It is resistance compounded of ignorance, apathy, 
prejudice, and resentment mixed with genuine concern for efficiency and fairness. The data on 
litigation allows us to locate the strain and tension associated with implementation of these 
preferences.  58.   Reservations in government employment gave rise to some 27 reported instances of 
litigation during the years 1950­77 ­­ not an impressively large total in a country where government 
servants are not loathe to litigate about seniority, discipline, and other matters. In particular, there 
has been surprisingly little litigation in regard to preferences in initial recruitment into government 
service. Of the 27 cases, 19 were concerned with reservations in promotions. (Of the remaining 8, 5 
were brought by lawyers seeking to obtain posts of munsif [a minor judge]). 
These cases involved, not rival contenders fighting over entrance into government service, but 
government servants (and lawyers) fighting over opportunities for advancement. The battle focusses 
not at the top, where reservations have relatively little impact on career chances, nor at the bottom, 
where then, are hardly needed (and those involved might 
____________________ 
56. Department of Social Welfare 1969:329. 
57. Id., 370. 
58. The universe of cases that is the basis of these computations is described at chap. 14, n. 45. 
­104­ 
not be able to afford litigation), but in the middle: particularly in the upper ranges of Class III and 
the jump from Class III to Class II. Naturally, such litigation clusters around those departments and 
posts where provision for reservations exist and are implemented with some vigor. For example, of 
the 13 cases involving the central government, 11 involved three departments (Railways, Posts and 
Telegraphs, Central Secretariat). As we shall see, the courts have provided resisting government 
servants with a resource which may be used to curtail and restrict the operation of compensatory 
discrimination.This incomplete and spotty data enables us to make an initial quantitative sketch of 
the performance of reservations of government posts: 
1.  reservations succeed, where exhortation and good will do not, in getting members of the 
beneficiary groups into government service; 
2.  generally, the reservations fall short of announced goals; 
3.  this is more so the higher the posts, except at the very peak; 
4.  reservations are far fewer in number than the announced level conveys; 
5.  reservations tend to be clustered in certain services, departments and grades; 
6.  the process of achieving substantial representation is slower than indicated by the announced 
level and scope of reservation; 
7.  reservations are resented and resisted; resentment and resistance is more articulate and 
focussed at the middle levels and at the promotion stage than at the initial recruitment stage. 
C. ASSESSING THE COSTS AND BENEFITS 
Our crude assessment of the sheer quantitative impact of the reservation policy leads us to the 
question of its overall effects. What are the consequences of propelling members of backward 
groups into government posts by a policy of preferential treatment? There has been no systematic 
study of these consequences and none can be attempted here. Here we can only try to formulate the 
issues and sort out some of the scattered and inconclusive but suggestive bits of evidence that are 
available (and indicate areas where there is no evidence at all). The assessment that results is 
necessarily incomplete and tentative. It is meant to launch inquiry, not to conclude it. 
We may, conveniently, reformulate our general list of alleged costs and benefits with specific 
reference to compensatory discrimination in government employment ( table 10 ). 
­105­ 
TABLE 10 ALLEGED BENEFITS AND COSTS OF PREFERENCES IN GOVERNMENT 
EMPLOYMENT 
REDISTRIBUTION  VS.  DIVERSION 
Reservations provide a  Reservations benefit directly only a
direct flow of small segment within these groups,
valuable resources (e.g.,  and the benefits do not spread widely. 
salaries, REPRESENTATION 
status, information,  VS. 
contacts) to the MISREPRESENTATION 
backward classes. 
Preferential treatment in government
The presence of backward 
jobs creates new interests which
classes in
diverge from those of the backward
government jobs 
classes. If there is effective
facilitates the
representation, it is of some smaller
sympathetic consideration 
segment within these groups. 
of their
INCUBATION 
needs in policy 
VS. 
formulation and
OVERPROTECTION 
administration. 

Reservations provide artificial
Reservation in  protection, remove the necessity for
government jobs competition, blunt the incentive to
stimulates the educational  acquire skills needed to compete on a
and other "merit" basis. 
accomplishments  INTEGRATION 
necessary to utilize VS. 
them.  ALIENATION 

By affording opportunities 
for
constructive and  By emphasizing the separateness of
prestigious these groups, preferences reduce their
participation, preferences  sense of common participation. 
promote ACCEPTANCE 
feelings of belonging and  VS. 
loyalty REJECTION 
among the employees (and 
others)
from these groups. 
REDISTRIBUTION  VS.  DIVERSION 
Preferences frustrate others by what
they view as unfair favoritism, educate
The presence of backward 
them to regard the beneficiaries as
classes
separate elements who enjoy their own
afforded by preferences 
facilities and have no claims on general
induces others
public opportunities. 
to accept them as fully 
INTEGRITY 
entitled
VS. 
participants in Indian life. 
MANIPULATION 

Holding of government 
jobs promotes
pride, self­respect, and a  The separate treatment involved in
sense of preference in government jobs
achievement and personal  undermines their sense of dignity, self­
efficacy sufficiency, and personal efficacy, and
that enable these groups to  aggravates the dependency of these
contribute groups by depriving them of potential
to national development as  leaders. 
willing
partners. 
­106­ 

MOBILIZATION  VS.  ENERVATION 


By imparting skills and  By siphoning the most talented into
providing government service, reservations
incentive for cultivating  deplete the pool of organizational skills
talents, available to backward communities
reserved posts increase the  and direct attention away from
capacity of collective action to individual
the beneficiary groups to  advancement through reservations. 
undertake STIMULATION 
organized action on behalf of  VS. 
their SEDATION 
collective interests. 
By portraying the inclusion of the
By providing a model for the 
backward classes as a responsibility of
presence
government, reserved posts make
of backward classes in 
others feel it is unnecessary to take any
responsible
measure to include them where it is not
positions, reserved posts 
required by government. 
stimulate
EFFICIENCY 
their inclusion even in 
VS. 
settings where it
INEFFICIENCY 
is not mandated by law. 

Preferences utilize neglected  Preferences lead to employment of
MOBILIZATION  VS.  ENERVATION 
less­qualified persons and lower the
and morale and efficiency of other
under­utilized talents and  employees. 
provide FAIRNESS 
administrators sensitive to. VS. 
problems of these groups.  UNFAIRNESS 

Preferences compensate for, 
Preferences place an unfair handicap
and help
on individuals who are deprived of
to offset, the accumulated
opportunities they deserve on merit. 
disablements resulting from 
SELF­LIQUIDATION 
past
VS. 
deprivations of advantages 
SELF­PERPETUATION 
and
opportunities. 
By increasing confidence 
through
stimulating development of 
talent and
acquisition of skills, and by
Reservations, by lowering incentives,
encouraging acceptance, 
encouraging alienation and rejection,
reservations
creating vested interests in separate
will render themselves 
and preferential treatment, lead to
unnecessary by
perpetuation of reservations. 
increasing the number of 
backward
classes who can enter 
without
preference. 
Note: This table is an adaptation of table 4. "Development" there is specified here as "Effi­
ciency." The "Secularism vs. Communalism" item, omitted here because it does not raise issues
that are distinctive to the employment area, is discussed in chap. 16. 
­107­ 
Redistribution vs. Diversion; Representation vs. Misrepresentation; Incubation vs. 
Overprotection 

Reservations in government services clearly entail some substantial redistribution of resources to 
the Scheduled Castes and Scheduled Tribes ­­ not only of salaries and perquisites, but of 
information, contacts, prestige, and patronage. Although the percentage increases in the 
representation of these groups are small, and although the fortunate beneficiaries of reserved seats 
remain a small portion of the total population of these groups, we should not overlook the 
significance of the absolute increase. In the early 1970s Aggarwal and Ashraf found that every 
eighth Scheduled Caste household in their survey of Haryana State had one member or more in a 
reserved job.  59.   Although the vast majority are not directly benefited, reserved jobs bring to the 
beneficiary groups a manyfold increase in a number of families liberated from circumscribing 
subservient roles, able to utilize expanding opportunities and support high educational attainments. 
Although such families are only a tiny fraction ­­ and optimistic guess might be 6%  60.   ­of all 
Scheduled Caste families they provide the critical leaven from which effective leadership might 
emerge.  61.   It is not clear how widely these benefits have been spread among the various parts of the 
Scheduled Castes and Tribes population or how much they have entailed any spillover from the 
direct recipients to others in these groups.  62.   Scattered evidence suggests that the mobile minority 
utilizing the reservations in government employment is by no means evenly spread through the SC 
and ST. There is a severe clustering of op­ 
____________________ 
59. Aggarwal and Ashraf 1976:125. 
60. Isaacs ( 1965: 111) estimates that perhaps as many as 10% of the SC population is "coming up," 
largely through reserved posts. Our figures suggest that his estimate is probably high. By 1975, 
about 180,000 SC were in Class III or higher service with the central government. Let us assume 
(generously) about an equal number of Class III or higher in public sector enterprises and the 
same number in state services. That gives us about 500,000. If we make the very optimistic 
assumption that for every such person there is another person similarly situated in local 
government or private sector employment, we come to a total of 1 million. If we assume that 
each of these persons is the head of a family of five, we come to a total of 5 million persons in 
such mobile families­­just over 6% of the entire SC population. If we were willing to include 
Class IV employment as having similar potential for mobility, we would more than double our 
figures and come out with something close to Isaacs's ten percent. 
61. Cf. Zelliot ( 1969:4, 53) discussion of the way in which army service in the 19th century 
provided the most important source of leadership for modernization and mobility among the 
Mahars. 
62. In his study of the Agra Lynch Jatavs ( 1969:108) suggests the absence of such spillover: 
"Individual education and advancement do not necessarily rebound to the benefit of the rest of 
the caste." 
­108­ 
portunities in some of the larger and more advanced groups among them.  63.   
In some cases the presence of Scheduled Castes and Scheduled Tribes does seem to produce more 
sympathetic administration and increased responsiveness to the needs of these groups. At least one 
would assume that their presence might inhibit the overt mistreatment of members of these groups 
by other officers. But few SC and ST officers have risen to policy­making positions. Whatever the 
benefits, they are at the cost of the channelling of the most talented and qualified members of these 
communities into government service, rendering them unable to contribute directly to community 
leadership (and perhaps at the cost of luring many into dead­end white­collar employment in which 
their capacity to advance either themselves or their community remains limited). 
Knowledge that jobs are accessible to them through reservations acts as a stimulus to educational 
accomplishment. It is often asserted that feeling assured of jobs, members of these groups feel no 
need to apply themselves strenuously to education. Aggarwal, on the basis of interviews with 
successful Harijans, portrays preferential treatment in a much more positive light. He concludes that 
for those starting in illiterate families "the availability of privileges was the major facilitator."  64.   
Integration vs. Alienation; Acceptance vs. Rejection; Integrity vs. Manipulation 
There is clearly some integration and acceptance in the office setting. The presence of Scheduled 
Castes and Scheduled Tribes on the job creates relations of interdependence with both fellow­
officers and clientele. To some extent this is offset by the alienation experienced by the SC and ST 
employees which they attribute to rejection on the part of their fellow­officers.  65.   
____________________ 
63. There is evidence of severe clustering in the distribution of educational benefits (see above, chap. 
3, §B, and below, chap. 13, §C) and presumably this is reflected in differential capacity to utilize 
reservations in employment. 
64. Aggarwal 1980:3. 
65. A measure of SC and ST employees' perceptions of resistance to their presence is provided by the 
number of complaints forwarded to the Commissioner for Scheduled Castes and Scheduled 
Tribes about service matters. In 1976 the Commissioner's office received 17,000 complaints, 
more than ten times the total of five years earlier (RCSCST 1975­77: I, 66; RCSCST 
1973­74:167). If we estimate that the total number of SC and ST employees in all grades of 
government or public sector jobs was less than one million, we conclude that in a single year 
almost 2% of these made written representations of mistreatment to a single remote (and 
powerless) officer. This reveals not only much (perhaps increasing) unhappiness on the job but 
also an absence of effective channels of redress. 
­109­ 
The Scheduled Castes/Tribes employees are treated as trespassers, coming into the 
services at the expense of really deserving hands and recruited on the basis of "birth" 
and not "worth." They thus start with an initial disadvantage in the sense that others are 
instinctively prejudiced against them. They are harshly judged by their superiors. . . . 
[They] find it very difficult to establish a personal rapport with their bosses and 
colleagues. . . .  66.   
As this plaint suggests, even the beneficiaries assign much of this resentment to the preference 
policy per se. The antagonism that exists toward the beneficiary groups seems to be accentuated by 
antagonism toward the policy that produces their presence. Educated untouchables, Aggarwal 
observes, are "taunted and slighted and constantly reminded of the unfairness of their privileges." 67. 
So while preferences do get them into the office setting, they may augment other barriers to 
integration. Unfortunately, we have no studies which trace the interplay between antagonism to 
reserved­seat holders and the momentum of office activities. To assess the effect of preferences on 
integration and acceptance, one would want to compare situations in which Scheduled Castes and 
Tribes were present on a "merit" basis. This may be complicated by one of the by­products of 
reservation: the widespread overestimation of their scope and the consequent perception of all SC 
and ST as recipients of preferential treatment. 
Reservations can probably be credited with promoting feelings of integrity and self­respect 
associated with prestigious and responsible government employment. However, this is undermined 
by feelings of receiving separate and second­class treatment, by resentment of features of preference 
schemes seen as demeaning, and by aggravation of a sense of dependency.  68.   
Mobilization vs. Enervation; Stimulation vs. Sedation 
Reserved posts provide a powerful incentive for undergoing the sacrifices required to cultivate 
talents by obtaining education. The growing number of government servants includes many who 
devote their talents to group betterment through educational and social work. But government 
servants are not permitted to participate in political activity, so backward groups are deprived of the 
skills of those who might be the most effective organizers. By rendering inactive many of the most 
____________________ 
66. Central Government Scheduled Castes and Scheduled Tribes Employees Welfare Association 
1966: 11­12. 
67. Aggarwal 1980: 10. 
68. Isaacs 1965: 116ff. Cf. the "no response" data discussed in chap. 3, §D, above. 
­110­ 
talented and qualified, reserved seats deprive the community of leadership and render it more 
manipulable, less assertive, and less selfreliant.  69.   
The record of private sector employment (and public sector employment before the introduction of 
reservations) undermines the stimulation argument. It seems hard to deny that reservations have 
been accompanied by lack of concern to include these groups in settings where such inclusion is not 
mandated. But against a long history of the absence of such concern, it is difficult to conclude that 
its current absence is attributable to reservations. 
Efficiency vs. Inefficiency 
It is widely believed that compensatory discrimination in government employment is responsible for 
a decline in the efficiency of government operations. It is clear that reservations bring about the 
presence of government servants with less qualification according to existing methods of measuring 
it. We should hesitate to assume that there is a direct connection between the tested 
accomplishments of entrants and the effectiveness of government administration. In any event, the 
effect on "quality" of Scheduled Caste and Scheduled Tribe recruitment is overshadowed by rapid 
expansion of government service and a general lowering of standards. A Study Team of the 
Administrative Reforms Commission found that in the All­India non­technical services (Indian 
Administrative Service, Indian Forest Service, Indian Police Service, Central Government Class I), 
the percentage of first­division entrants dropped from 42% in 1950­55 to 25% in 1960­65, while 
third­class degrees rose from 4% to 19%.  70.   Since this shift in educational profile has been cited as 
evidence of the deleterious effects of reservation,  71.   it deserves close examination. The falling off of 
first­division entrants can hardly be attributed to reservations.  72.   It reflects a general decline in the 
____________________ 
69. In his study of the Jatavs of Lynch Agra ( 1968: 107) concludes that reservations deprive the 
community of leadership from its best­qualified members. 
70. Administrative Reforms Commission, Study Team 1969: 5. 
71. Ghurye ( 1969:440) cites a newspaper account of these figures as conng "common knowledge" 
and infers that "[t]o a fair­minded person and patriotic Indian it must become clear that the 
proviso of the maintenance . . . of efficiency of administration of Article 335 requiring the 
Government to take the claims of the Scheduled Castes to be represented in the services has been 
set at naught." 
72. Except as the declining appeal of government service may reflect some apprehension that 
opportunities to rise in government service might be constricted by the presence of reservations. 
I know of no study which has attempted to isolate the contribution of this factor. I would surmise 
that such an effect has been produced by the expansive reservations for Backward Classes in 
several states, where others' op­ 
­111­ 
appeal of government service to the academically most qualified.  73.   The rise in entrants with third­
class degrees is attributable in part to the intake of Scheduled Castes and Tribes under reservations. 
But the Study Team's own figures show that only a minority (160 out of 377) of the third­class 
degree holders selected in 1960­65 were Scheduled Caste and Tribes.  74.   By rearranging the Study 
Team's data we can compare the educational performance of Scheduled Castes and Tribes with that 
of the other successful candidates ( table 11 ). 
In estimating the impact of the reservations on the academic profile of the intake, it is important to 
isolate their net effect. Let us assume that reservations were responsible for the selection of all of 
the third­class and some large portion of the second­class degree holders among the Scheduled 
Castes and Scheduled Tribes. If there were no reservations, they would be replaced by "others." 
Presumably, it was the less qualified others who were eliminated by the reservation. So the 
academic profile of these eliminated others is bound to be lower than the others 
TABLE 11 PRIOR EDUCATIONAL PERFORMANCE OF SUCCESSFUL CANDIDATES 
FOR ALL­INDIA HIGHER NON­TECHNICAL SERVICES, 1960­1965 
Scheduled  Others 
Castes and Tribes 


Type of Degree  n  %  n  I 

29.5% 
18  5%  484  II 

57.2% 
178  50%  936  III 

160  45%  217  13.3% 

100% 
Total  356  1,637  100% 

Source: Derived from Administrative Reforms Commission 1969:5, 6. 
____________________ 
portunities for entrance and advancement are severely restricted. But it seems less likely that 
comparable discouragement would be produced by the modest reservations for SC and ST in 
central services. There is also a question of whether such discouragement is a bad thing: do 
talented individuals make more of a contribution to development outside government service? 
73. In 1950­55, an average of 1 in every 3.5 holders of first­class degrees took the Combined 
Competitive Examination for higher central services: there was an average of 3 firsts competing 
for each vacancy. In 1960­65, only 1 in every 13 first­class degree holders took the Examination, 
and there were only 1.8 First Classes competing for each vacancy ( Administrative Reforms 
Commission, Study Team 1969:4). 
74. By subtracting the figures given by the Study Team on p.6 from the figures it gives on p.5, we 
can determine that 160 of the 356 (44.9%) SC and ST recommended in 1960­65 held third­class 
degrees, as did 217 of the 1,637 (13.3%) of the other candidates. 
­112­ 
who were successful in spite of the reservation. This successful group included over 13% third­class 
degrees. We must assume that the next two or three hundred down the list would include an even 
higher percentage of third­class degrees. Thus the net difference between the two groups is 
overstated by our table (for it compares the SC and ST only with those competitors who weathered 
the "bumping" caused by the reservations). Thus it overstates the effect of the reservation on the 
academic profile of the intake. 
The translation of lower academic accomplishment into inefficiency in administration is difficult to 
trace. It is not clear how well academic performance correlates with administrative talent. Nor is it 
clear that differences in the level of such talents are directly reflected in efficiency or inefficiency of 
administration. In the light of the many forces which affect administrative inefficiency, it seems 
appropriate to confront assertions about the effect of reservations with some skepticism, if only to 
guard against the widespread tendency to attribute any inefficiency in governmental operations to 
the presence of Scheduled Castes.  75.   
Another "efficiency" argument is that reservations injure the morale of other workers, lowering the 
general level of administration. Low morale might stem from (1) inferior performance by SC and ST 
(which in turn might be caused by rejection­­i.e., abrasive social relations, absence of informal on­
the­job training, etc.); (2) fear for career prospects ­­ especially where reservations in promotions 
give rise to apprehension that one's co­worker (and competitor) of today is destined to become 
undeservedly one's superior tomorrow;  76.   (3) displeasure at the enforced contact with these groups. 
Since there are plenty of government offices with no Scheduled Castes or Scheduled Tribes, and 
many where no reservations in promotions are in effect, a comparative study of effects of these 
arrangements on morale is conceivable, but has not been done. For the moment we may accept that 
reservations entail some cost in morale­­but it is not clear how much is due to the reservation device 
per se and how much is just rejection of Scheduled Castes and Scheduled Tribes and would be there 
even if they attained their posts on "merit" alone. 
The arguments about efficiency lead back to the "siphoning effect" that was mentioned earlier  77.   ­­a 
phenomenon with implications for the values that we have summed up under the headings of 
acceptance and integrity. Reservations, particularly those for high­echelon posts, 
____________________ 
75. For some example, see Isaacs 1965: 133. 
76. Thus Kirpal and Kelkar ( 1976:8) convey widespread reports that "[c]aste Hindus . . . tend to 
withhold information at the time of training their SC/ST subordinates, who, they fear, may 
supercede them by virtue of seats reserved for promotion." 
77. See text at n. 26 above. 
­113­ 
pull in candidates with lower performance on various indicators. The pool of candidates from the 
beneficiary groups who have passed the threshold educational qualifications to aspire beyond the 
clerk's grade is rather small. Once the threshold is passed, members of these groups tend to shoot 
for the very top. Thus in 1961 Scheduled Castes candidates made up 8.9% of applicants for I.A.S. 
and Class I posts, but only 4.5% of applicants for Class II posts. Among Scheduled Tribes, the 
siphoning is even more pronounced: they made up 1.5% of the applicants for I.A.S., etc., but only 
0.06% of applicants for Class II posts.  78.   Reservations place these candidates in positions alongside 
peers with higher performance on various indicators (examination scores, etc.). In part, the higher 
scores of others may reflect cultural advantages which are irrelevant to the business at hand; in part, 
the lower scores of beneficiaries may reflect a remediable lack of polish and experience rather than 
lack of native ability. But with all due allowance for the irrelevance of the indicators, performance 
on them probably reveals something about ability to deploy one's talents. And with all due 
allowance for "diamond in the rough" qualities, it is clear that not everyone can surmount greater 
challenges while still suffering the effects of earlier disadvantages. Reservations propel SC and ST 
candidates into settings in which their ability to perform competently and comfortably is placed 
under considerable strain­­strain compounded by the hostility and resentment of those around them. 
If this leads to some instances of heroic triumph over challenge, it must in many others be damaging 
both to performance on the job and personal well­being. The forced­draft aspect of reservations at 
the top may carry heavy personal costs for beneficiaries as well as for others. 
Fairness vs. Unfairness 
The fairness argument is of course based on the principle of historical compensation and more 
immediately on the notion that members of these groups on the whole have less opportunity to 
cultivate their talents so as to show up well in recruitment testing. The charge of unfairness is based 
on the notion that it is unfair to individuals that their life chances be impaired to provide such 
compensation. There is no reason to suppose that the individuals who are "bumped" from valued 
opportunities were disproportionately benefitted by invidious discrimination in the past. If current 
ability to perform meritoriously is thought a product of accumulated past advantage, those who are 
bumped seem to enjoy less rather than more of this advantage. And since reserved places are 
distributed among those eligible in order of merit, the ben­ 
____________________ 
78. Planning Commission 1965:266 ff. 
­114­ 
eficiaries will be those who are least disadvantaged among the eligible group. 
Because reservation policies tend to vary from department to department in their incidence and in 
the effectiveness with which they are implemented, there is some clustering­­so that the "unfair 
competition" impinges heavily on some government servants while others are virtually untouched. 
We may distinguish two separate claims of unfairness: (1) that which rejects the principle of 
collective recompense for historical disabilities; and (2) that which objects to the adminstration of 
the principle so that the burden is borne by a few who are no more responsible for the historical 
deprivations that underlie it than the many who escape. The administration of compensatory 
discrimination measures does seem to involve considerable "unfairness" of the latter kind. If some is 
inseparable from compensatory discrimination, it is clear that more could be done to distribute the 
burden on nonbeneficiaries more widely and more evenly. 
Self­liquidation vs. Self­perpetuation 
Reservations in government posts are not self­perpetuating in a literal sense (as are reservations of 
legislative seats). They set up expectations of and pressures for continuation and expansion ("vested 
interests"):­­but in the field of Government service the pressures seem more directed to enhanced 
benefits for existing recipients (e.g., extension of reservation to promotions) than to increasing the 
beneficiaries. And, in an important sense, reservations in government service are self­liquidating: 
many reservations operate as guaranteed minima, not as fixed quotas.  79.   Thus they are potentially 
self­liquidating to the extent that there is adequate incubation of talent. Thus, if programs for 
delivering education to these groups were fully effective, the reserved posts would become largely 
redundant. 
Considering the readiness with which many Indian intellectuals, including some eminent social 
scientists, have passed severe judgments on compensatory discrimination, it is surprising how little 
systematic study of its presumed effects we have. Our cursory sketch suggests several considerations 
that seem crucial if the heightened interest in these questions is to equip us to arrive at a credible 
assessment. First, there is the necessity of isolating the net effect of compensatory discrimination 
from effects of the presence (apart from compensatory discrimination) of Scheduled Castes and 
Tribes. For example, to the extent that 
____________________ 
79. See chap. 13, §B, below. 
­115­ 
rejection, alienation, lowered morale, etc., are associated with the presence of Scheduled Castes and 
Tribes, apart from preferential treatment, they cannot be accounted as costs of compensatory 
discrimination. Reservations may provide the occasion rather than the source of the negative effects. 
Second, there is the necessity of considering not only the design of the scheme in assessing its costs 
and benefits, but also the level of performance at which it is operating. As the government 
employment area suggests, programs of compensatory discrimination may operate at a level of 
performance considerably below their announced promise. One consequence is to project to 
interested (and resentful) publics an image of ubiquitous and unqualified preference with immense 
power to affect them adversely, while to the beneficiaries it appears a program of narrow and 
contingent preference with limited power to benefit them. The output of those values we have 
labelled costs and benefits may be augmented or diminished by changing levels of performance as 
well as by changes in authoritative policy. 
Third, perceptions of costs and effects may have a powerful influence quite apart from the presence 
of objective indicators of such costs, etc. Compensatory discrimination policies are symbolic as well 
as administrative realities. Their consequences are not reducible to their perceived consequences, 
but the perceived consequences of these policies are important components of any assessment. 
Notwithstanding all the problems of ascertaining their performance and evaluating their 
achievements, the compensatory discrimination policies mark the inclusion of the beneficiary 
groups into the visible public life of Indian society to a degree that would have been unimaginable a 
few decades ago.  80.   Although the beneficiary groups still are significantly lower on virtually all 
measures, they are no longer completely excluded from the mainstream politically, occupationally, 
or educationally. 
Yet it is evident that actual performance of these policies falls far short of announced aspirations. 
We have seen that reservations go unfilled in good measure; most of those receiving educational 
benefits drop out early; protective legislation is unenforced and circumvented; allotted funds go 
unspent; and programs may give substitute services instead of additional ones. The general pattern 
is that normative com­ 
____________________ 
80. Our preliminary assessment of performance has been concerned almost entirely with the 
preferences for Scheduled Castes and Scheduled Tribes. For the OBC, there is less data available, 
there is more variation state by state, and the goals are less clear, so it is difficult to make any 
generalizations about OBC preferences. But one important by­product should be noted­­the 
operation of OBC preferences has done much to obscure and confound public understanding of 
the preferences for SC and ST. 
­116­ 
mitments outrun institutional capacities (and political will) to implement them. Weak and 
unaggressive implementation combined with a concentration of attention on "higher echelon" 
benefits (i.e., those which require some degree of previous success) mean that utilization is by those 
among the potential beneficiaries better supplied with other advantages. Thus the total redistributive 
effect is smaller than is projected and tends to be concentrated disproportionately in the upper layers 
of the beneficiary groups. 
These programs, with their pervasive gap between promise and performance, provide the setting 
within which we shall examine the work of the courts as critics, checks, energizers, definers, and 
legitimators of this complex of policies. But first we must take up the process by which the 
beneficiaries of these policies have been selected and identified. 
­117­ 
Part Two IDENTIFYING THE BENEFICIARIES 

Analytically, the first question that arises about India's policy of compensatory discrimination is 
"who are the beneficiaries?" But historically the process is more complex: the identification of some 
groups as requiring and deserving special treatment gives rise to a concept of compensatory 
treatment, and this in turn produces controversies about those who should be included within its 
scope. Even after compensatory preference is firmly established, who is to benefit continues to be a 
central question in designing and administering programs­­and crucially important for the 
continuing justification of these programs to others. Public and professional discourse about the 
identity of the beneficiaries is thus an integral part of the operation of these programs. These 
disputes also reveal how (some) Indians see their society and how they relate their aspirations for it 
to their perceptions of its present state. We can observe how those perceptions and aspirations are 
filtered through the institutional style of the Indian legal culture. 
Chapters 5 to 8 address the general question of the selection of the beneficiary groups. Chapter 5 
takes up the groups designated earliest and at the national level. Section A recounts the emergence 
of the notion that there were a distinct layer of "untouchables" at the bottom of Indian society, the 
controversies about drawing the line, resolved by the formation of the Scheduled Castes category in 
the 1930s. Section B describes the institutionalization and stabilization of that category in post­
Independence India. Section C (of chapter 5) sketches the comparable development of the 
Scheduled Tribe category. 
­119­ 
Chapters 6 to 8 take up the more recently controverted questions of the other beneficiaries of 
preferential treatment, the groups that have come to be called the "Other Backward Classes." 
Chapter 6 parallels the account in Chapter 5: it traces governmental action and public controversy in 
reference to the designation of the Other Backward Classes, a story that is complicated by great 
diversity among the states and by tensions between central and state governments. There is another 
dimension to the selection of Other Backward Classes: it is an area that has been the scene of steady 
and significant judicial intervention; it is an area permeated by resort to the courts and by ideas 
worked out in the courts. 
Chapters 7 and 8 analyze the judicial response to the Other Backward Classes. Chapter 7 takes up 
the controverted question of what are the social units that government may designate as backward 
classes (here the controversy over the "caste basis" of selection). Chapter 8 takes up the question of 
the standards and the methods by which the backwardness of these groups is to be measured. 
Chapter 9 moves from the question of which groups to the question of who are members of those 
groups. This includes disputes concerning both individuals whose personal history makes their 
membership problematic and groups (e.g., neo­Buddhists) whose inclusion in a larger category is 
controverted. These controversies reproduce in a more precise form and on a more personal scale 
some of the conflicts discussed in the earlier chapters. This chapter analyzes the jurisprudence of 
personal and group identity which the courts have developed in their encounter with these 
controversies. 
Chapter 10 summarizes the perplexities of governmental superimposition of official categories of 
beneficiaries on the complex processes of group and personal identity. It concludes with a 
discussion of the implications of these perplexities for the role of courts in implementing the 
compensatory discrimination policy. 
­120­ 
5 The Designation of the Scheduled Castes and Scheduled Tribes 

THE CONSTITUTION DOES not contemplate reservations and special treatment as a general 
principle of operation. Such treatment is authorized only for the limited purpose of advancing the 
backward. Who, then, are the legitimate recipients of such preferential treatment? 
1. ­­i.e., 
The Constitution permits preferences for three categories of groups: (a) Scheduled Castes    
2. and (c) other (socially and educationally) backward classes. 
"untouchables", (b) Sheduled Tribes;    
 3.   The term "backward classes" is commonly used in two senses: (1) as a generic term including the 
Scheduled Castes and Scheduled Tribes as well as the so­called Other Backward Classes;     4. or (2) as 
a designation of those backward groups not included in either of the first two categories (i.e., as 
equivalent to "Other Backward Classes.")    5. 

The Constitution itself does not define these groups, nor does it provide detailed standards by which 
6. In the case of the Scheduled Castes and Tribes, it does prescribe an 
they may be determined.    
agency and a method for designating them. Not only are Backward Classes left undefined in the 
Constitution, but no such method or agency for their determination is provided. Before taking up 
(chaps. 6, 7, and 8) the 
____________________ 
1. Arts. 15, 16, 46, 341, 335. Cf. Arts. 330 and 332. 
2. Arts. 15, 16, 335, 342. Art 244, Vth and VIth Schedules. Cf. Arts. 330 and 332. 
3. Arts. 15, 16, 46. 
4. See Art. 16(4). In General Manager v. Rangachari, A.I.R. 1962 S.C. 36, it was held that 
"Backward Classes"' in Art. 16(4) includes the Scheduled Castes and Scheduled Tribes. 
5. See Art. 15(4), as well as Art. 340 and Art. 164 where it is apparently used in this restricted 
sense. Cf Sudhir Kumar Ghose v. Ank Kameya, XXVII Cuttack Law Times (Short Notes) 5 
(1961). (Election set aside on ground that candidate was a member of a Backward Class, not of a 
Scheduled Caste.) 
6. Section (24) of Art. 366 defines Scheduled Castes as those so designated under Art. 341. § (25) 
of Art. 366 defines Scheduled Tribes as those designated under Art. 342. No definition of 
"Backward Classes" appears at any point in the Constitution. 
­121­ 
more complex issues that have arisen in the designation of the Backward Classes, I shall trace the 
problems of identifying the Scheduled Castes and Scheduled Tribes. 
The Scheduled Castes and Scheduled Tribes constitute the bulk of those receiving preferential 
treatment, and they receive a larger quantum of preferences than Other Backward Classes. The 
Scheduled Caste and Tribe categories are intended to comprise those who were at the bottom or the 
margins of the Indian social order­­those groups who because of their low ritual status in the 
traditional Hindu hierarchy or their spatial and cultural isolation were subjected to imposition of 
disabilities and lack of opportunity. The Constitution prescribes that the Scheduled Castes and 
Tribes shall be designated by Presidential Order in the first instance, with subsequent modification 
7. It further provides that the President shall appoint a special officer to 
only by Act of Parliament.    
investigate all matters relating to the safeguards for the Scheduled Castes and Tribes and report to 
Parliament on their working.    8. Accordingly, the office of the Commissioner of Scheduled Castes 
and Scheduled Tribes was established in 1950 as a continuing body with coordinating and reporting 
but not administrative functions. 
A. THE INVENTION OF THE SCHEDULED CASTES 
"Scheduled Castes" is the most recent of a long line of official euphemisms for "untouchables." The 
Scheduled Caste category is intended to comprise those groups isolated and disadvantaged by their 
"untouchability"­­i.e., their low status in the traditional Hindu caste hierarchy which exposed them 
to invidious treatment, severe disabilities, and deprivation of economic, social, cultural, and political 
opportunities. In the early years of the century the "depressed classes" (as they were then called) 
became an important focus of concern among reformers. After 1901, fears of diminished Hindu 
majorities and proposals for special legislative representation for these classes propelled 
"untouchability" from the realm of philanthropy into the political arena. As concern about 
untouchables grew and political demands were made in their behalf, there was sharp disagreement 
about the number of persons who belonged in this category. 
From 1908 on, it was claimed by numerous writers that untouchables numbered 50 millions or 
more­­about 24% of the Hindu population or just under 16% of the total population of India 
9. Although Madras compiled a list of Depressed Classes during 
according to the 1911 Census.    
____________________ 
7. Arts. 341, 342. 
8. Art. 338. 
9. Risley ( 1908) counted 50.6 million Hindu untouchables (in addition to 17.2 million animists). 
This estimate was omitted from the ­1915 edition of his book, as noted by 
­122­ 
the 1920s,  10.   the Government of India refrained from any official classification (until 1936) on the 
ground that it would be unfair to stigmatize these groups by official acknowledgement of their low 
status.  11.   But various governmental bodies attempted to assess their numbers. The first official 
estimates in 1917 and 1918 arrived at the figure of about 30 millions for British India.  12.   The 
Franchise (Southborough) Committee, which was established to advise on the franchise under the 
1919 reforms, found 42.2 millions in British India.  13.   Since there were presumably one­quarter to 
one­third as many more in the princely states, this did not depart far from the 50 million figure. The 
1921 census and the Reforms Enquiry (Muddiman) Committee of 1924 arrived at comparable 
figures.  14.   However, in 1928, official estimates reverted to the lower range: two officials identified 
the number of untouchables as 28.5 millions and 29 millions,  15.   and the Hartog Com­ 
____________________ 
Weitbrecht­Stanton 1920:175. Cf. H. Madras 1909: 1022 ("the outcastes of Hindu society form all 
over India a distinct section of the population, numbering about 50 millions"); Holderness 
1911:102 (50 to 60 millions); S. N. Singh 1913:376 ("over 50,000,000"); Fisher 1919:97 (53 
million). Somewhat higher estimates are given by Ketkar 1911:78 (one­fourth of Indian 
population) and Brown 1914:548 (one third of Hindu population). A lower estimate was given by 
Baines 1912 (34.8 million). 
10. They were scheduled by notification under the Madras Local Boards Act of 1920 ( Franchise 
Committee 1932:I, 112). 
11. "Though officers of Government have from time to time attempted to estimate the total number 
of the depressed classes in the country, the Government of India have consistently avoided 
themselves making any precise classification of the groups forming what are called the depressed 
classes if only on the ground that owing to the social disabilities to which members of the 
depressed classes are exposed it would be in the highest degree undersirable that any official 
authorization might appear to extend to such classification. The fluidity of social distinctions and 
the efforts of the classes lowest in the scale, aided by social reformers, to improve their status 
make it the more desirable that Government should abstain from doing any thing which would 
tend to give rigidity to these distinctions" ( Indian Statutory Commission 1930:V, 1341). 
12. The estimate of Sir Henry Sharp, Educational Commissioner in 1917, was 31.5 millions ( Indian 
Statutory Commission 1930:V, 1350 ff.) In response to the Government of India's inquiry, the 
provincial governments (except Assam and Punjab) provided lists in 1917­18 of depressed classes 
whose total population was 27.4 millions, ( Indian Statutory Commission 1930:V, 1341 ff.). 
Weitbrecht­Stanton 1920 comes to a similar figure. 
13. Indian Franchise Committee 1919:I. The figures appear in the appendices under each province. 
The figures are compiled and discussed, id.367, in an appended despatch from the Government 
of India, Home Department, signed by the Chelmsford Committee. 
14. Census 1921:I, Pt. 1, p. 193; Reforms Enquiry Committee 1925:56. The Committee noted that its 
figures were "a rough estimate of the minimum numbers given in the last census report" (id.55). 
15. In the Indian Legislative Assembly on February 23, 1928, Mr. G. S. Bajpai, Secretary in the 
Department of Education, Health and Lands, stated that according to his calculation based upon 
available material the depressed or untouchable classes in 
­123­ 
mittee set the number at 29.76 millions.  16.   
A new round of counting had begun with the impending of further constitutional changes. In 1929, 
the Indian Central (Nair) Committee, set up to advise the Simon Commission, arrived at a figure of 
44.5 millions,  17.   and the Simon Commission itself estimated the number at 43.6 millions.  18.   (See 
table 12.) 
The Indian Franchise (Lothian) Committee, established to advise on the franchise under the new 
reforms, received conflicting estimates from a number of quarters. The 1931 Census Commissioner 
estimated untouchables at 31.2 millions excluding Bengal;  19.   the Provincial Governments supplied 
somewhat lower figures, totalling 37.45 millions.  20.   The Provincial Committees, set up to advise the 
Franchise Committee, 
TABLE 12 PERCENTAGE OF UNTOUCHABLES IN BRITISH INDIA (According to Indian 
Statutory [Simon] Commission, 1930) 
Number of % of % of
Province  Untouchables Hindu Pop.  Total Pop. 
(Millions) 
MADRAS  6.5  18  15.5 
11 
BOMBAY  1.5  8.0 

BENGAL  11.5  a  57  24.5 


UTTAR PRADESH  12.0  a  31  26.5 
PUNJAB  2.8  42  13.5 
BIHAR & ORISSA  5.0  a  20  14.5 
CENTRAL PROVINCES  3.3  33  24.5 
ASSAM 
24  13.0 
1.0 

28½ 
TOTALS  43.6  19 

Source: Indian Statutory Commission 1930a: 40. 
Note:a Noted as "subject to . . . warning". 
____________________ 
the Governor's Provinces (excluding Burma and Assam) numbered 28½ millions. Later that year, 
in Parliament on the 23rd of April the Secretary of State for India estimated the total number of 
untouchables in British India at 29 millions "on information provided in Provincial Educational 
Reports" ( Indian Statutory Commission 1930:218). 
16. The Hartog figures were based on the figures given to the Legislative Assembly ( Indian 
Statutory Commission [Hartog Committee] 1929:217­18). 
17. Indian Central Committee 1929:44, 86. 
18. Indian Statutory Commission 1930:I, 40. 
19. Hutton's tentative figures to Indian Franchise Committee 1932:I, 119. 
20. Indian Franchise Committee 1930:I, 119. The Memoranda from the provincial governments are 
in vol. III. 
­124­ 
produced figures that totalled less than half of the then­current estimates (18.62 millions).  21.   The 
dissenting members of the Franchise Committee minimized the number even further to 14.42 
millions.  22.   The Franchise Committee itself was unable to arrive at a firm figure.  23.   The various 
estimates are presented in table 13. 
As table 13 indicates, there was little dispute about who were the untouchables in southern and 
western India­­Madras and Bombay Presidencies and the Central Provinces. The dispute was about 
northern and eastern India. Which groups there ought to be treated on a par with the untouchables 
of the south and west? This, in turn, depended on how the depressed classes were to be defined­­
whether strictly in terms of specific indicia of ritual and social exclusion or more broadly in terms 
of low status. 
The earlier investigations had tended to proceed on the assumption that untouchables throughout 
India were set off by some uniform and distinctive pattern of practices and that a test could be 
devised that would identify them on an all­India basis. Thus, the Government of India's 1919 
Despatch, criticizing the inadequacy of the Southborough Committee recommendations for 
representation for the Depressed Classes, observes that, 
though they are defined in varying terms, [they] are broadly speaking all the same kind 
of people. Except for the differences in the rigidity of their exclusion they are all more 
or less in the position of the Madras Panchamas, definitely outside that part of the 
Hindu community which is allowed access to their temples.  24.   
But this assumption proved difficult to apply in northern and eastern India. Touch and distance 
pollution extended throughout the caste system in the south, while in the north touch and distance 
pollution were not expressed uniformly enough. The Simon Commission, a decade later, conceded 
that in Bengal, the United Provinces, and Bihar and Orissa, there was a wide margin of possible 
error. In these provinces 
the connection between theoretical untouchability and practical disabilities is less close, 
and a special investigation might show that the number of those who are denied equal 
rights in the matter of schools, water and the like is less than the total given for the 
depressed classes in these areas.  25.   
____________________ 
21. Indian Franchise Committee 1932:I, 119. The Provincial Franchise Committees' reports are 
contained in Vol. III. 
22. Minutes of dissent of Messrs. S. B. Tainbe, C.Y. Chintamani, and R. R. Bakhale. Indian 
Franchise Committee 1932:I, 219. 
23. Indian Franchise Committee 1932:I, 118. 
24. Fifth Despatch in Indian Constitutional Reforms 1919:366, 367. 
25. Indian Statutory Commission 1930:I, 41. 
­125­ 
TABLE 13 ESTIMATES OF UNTOUCHABLE POPULATION IN BRITISH INDIA BY 
VARIOUS OFFICIAL BODIES, 1917­1941 
Reforms 
Franchise  India
Sir  Enquiry Indian 
Provincial  Committee 1921  Hartog  Statut
Henry  Committee  Central 
Governments  (Southborough Census  Committee  Commis
Sharp  (Muddiman Committee 
)  (Simo

YEAR 1917 1917­18 1919 1921 1924 1928 1929 1930


SOURCE  A  B  C  D  E  F  D  G 

MADRAS  5.7  6.48  6.4  6.4  6.4  6.53  6.50  6.5 


BOMBAY  1.6  1.63  0.6  2.8  2.8  1.46  1.47  1.5 
BENGAL  6.7  6.74  9.9  9.0  9.0  6.64  11.50  11.5
UTTAR 
8.4  8.37  10.1  9.0  9.0  7.89  13.00  12.0
PRADESH 
PUNJAB  2.1 ...  1.7  2.8  2.8  1.70  2.80  2.8 
BIHAR & 
1.2  1.10  9.4  8.0  8.0  2.53  5.00  5.0 
ORISSA 
CENTRAL 
3.0  3.06  3.8  3.3  3.3  3.01  2.67  3.3 
PROVINCES 
ASSAM 
2.7 
...  0.3  2.0  2.0  0  e  1.55  1.0  1.9 

TOTALS  31.5 ...  42.2  43.3  43.3  29.76  44.50  43.6


Reforms 
Franchise  India
Sir  Enquiry Indian 
Provincial  Committee 1921  Hartog  Statut
Henry  Committee  Central 
Governments  (Southborough Census  Committee  Commis
Sharp  (Muddiman Committee 
)  (Simo

­126­ 
The very different nature of the matter in the north was brought out nicely in the testimony before 
the Indian Franchise Commission of Mr. G. S. Pal, representing the United Provinces Hindu 
Backward Classes League, which had submitted a list of 115 castes (estimated to make up over 60% 
of the population) for inclusion as Depressed Classes in the United Provinces.  26.   

Dr. Ambedkar. My point is this. Is this a list of what might be called backward 
communities, economically backward and educationally backward? Or is this a list of 
castes which are called untouchables, whose touch causes pollution? Is this a list of 
castes which are touchable but which are educationally and economically backward or 
which are untouchable and also educationally and economically backward? 
Mr. Pal. I respectfully submit that this is a list of all the castes which in the opinion of 
the League are depressed classes; untouchability in the sense in which it is understood 
in Madras or Bombay does not exist here. Bhangi is the only caste that is universally 
untouchable, pollution by whose touch is regarded as something very undesirable. Here 
untouchability exists, of course, but in a very mild form or exists in some form or 
another in all the castes I have enumerated. 
. . . . 
Sir Sunder Singh Majithia. How would you define depressed class people? 
Mr. Pal. Not in the sense of the definition in Bombay and Madras but apparently here 
the Hindu class is divided into high and low castes. The problem is very acute, as acute 
as untouchability is in Bombay and Madras. I apply three tests, the first is we are 
socially despised; then there is economic backwardness; and educational bankruptcy.  27.   
In the most elaborate attempt to specify criteria by which to identify untouchable groups, J .H. 
Hutton, the 1931 Census Commissioner, proposed a series of tests which revolved around the 
incidence of disabilities: 
1.  Whether the caste or class in question can be served by Brahmans or not. 
2.  Whether the caste or class in question can be served by the barbers, water­carriers, tailors, etc., 
who serve the caste Hindus. 
3.  Whether the caste in question pollutes a high­caste Hindu by contact or proximity. 
____________________ 
26. The list and estimate are at Indian Franchise Committee 1932:IV, 825­26. 
27. Id.,827­28. Cf. another definition of "untouchables" in the U.P. setting, suggested in a 
memorandum of the United Provinces Adi­Hindu (Depressed Classes) Association: refusal of 
caste Hindus to take food or water which they touched; refusal of entry into eating­houses; 
refusal of accommodation in dharamsalas (id.,833). This memorandum contained a list of 64 
castes said to meet these tests. 
­127­ 

4.  Whether the caste or class in question is one from whose hands a caste Hindu can take water. 
5.  Whether the caste or class in question is debarred from using public conveniences, such as 
roads, ferries, wells or schools. 
6.  Whether the cȧste or class in question is debarred from the use of Hindu temples. 
7.  Whether in ordinary social intercourse a well­educated member of the caste or class in 
question will be treated as an equal by the high­caste men of the same educational 
qualifications. 
8.  Whether the caste or class in question is merely depressed on account of its own ignorance, 
illiteracy or poverty and but for that, would be subject to no social disability. 
9.  Whether it is depressed on account of the occupation followed, and whether but for that 
occupation it would be subject to no social disability.  28.   
These tests are not altogether free of difficulty. They involve terms like "clean Brahmin" and "high­
caste Hindu" which have merely a local and comparative reference. And whether a caste is 
"polluting" or "debarred" from public facilities or temples may admit of no unequivocal answer, 
since pollution and debarment may exist in grades or degrees. Some may be forbidden to approach a 
temple, while others are only excluded from the inner sanctuary. Debarment from wells might imply 
use of adequate but separate water supply or onerous exclusion from any decent water supply. Tests 
7 and 8, as Lelah Dushkin points out, "seem a bit academic, since the [Census] Superintendents 
were often in no position to find out whether they applied or not. . . ."  29.   Whether a caste "can be 
served" is not necessarily the same as whether it is habitually served unless one assumes 
correspondence between local learning on the matter and local practice that does not invariably 
obtain. 
However, the tests clearly point to the incidence of disabilities as the crucial test. "[I]t is at least 
clear that occupation, 'ignorance, illiteracy or poverty' or any other such forms of 'backwardness' 
were not intended to enter into the definition process."  30.   
Hutton suggested: 
From the point of view of the State the important test is the right to use public 
conveniences­­roads, wells and schools­­and if this be taken as the primary test, 
religious disabilities and the social disabilities involved by them may be regarded as 
contributory only. Some importance must be attached to them, since obviously if the 
general public regards the persons of cer­ 
____________________ 
28. Census of India, 1931, Vol. 1, Part 1, App. 1, p.472. (Reprinted in Hutton 1961:194.) 
29. Dushkin 1957:88. 
30.  Ibid.   
­128­ 
tain groups as so distasteful that concerted action is resorted to in order to keep them 
away, persons of those groups do suffer a serious disability.  31.   
But the "right to use public conveniences" might be one of the most difficult and fruitless tests to 
apply, since roads were generally open to all, practices regarding schools and wells were subject to 
much local variation and came in a number of gradations, and finally, because practice might 
diverge sharply from "rights." 
The existence of multiple (and ambiguous) criteria permitted regional variation by giving the 
Census superintendents scope for discretion. Each provincial superintendent was instructed to draw 
up his own list, taking these tests into account, "and to reckon as depressed only those castes who 
definitely suffer from serious social and political handicap on account of their degraded position in 
the Hindu social scheme."  32.   The final reckoning, then, did not apply Hutton's tests pure and 
simple. Excluded were some castes which were theoretically untouchable but who, by virtue of 
prosperity and vigorous organization, "have built up for themselves a strong position which obviates 
the need of any special measures for their social, political or religious protection."  33.   Conversely, 
some groups were included even though theoretically they were not untouchables.  34.   The total of 
Depressed Classes reported by the census was 39.1 millions in British India and 11.1 millions in the 
States and Agencies­­a total of 50.2 millions.  35.   
The insufficiency of any single test to identify the untouchables provoked a variety of responses on 
the Franchise (Lothian) Committee. Dr. Ambedkar, the leading untouchable spokesman, points out 
that the aim is to demarcate those who suffer from the contempt and aversion of higher caste 
Hindus: "It is a fatal mistake to suppose that differences in tests of untouchability indicate 
differences in the conditions of the untouchables."  36.   The crucial common element is the odium 
and avoidance of the higher castes. Once this element is present "the whole of the class of 
untouchables so ascertained must be taken into account for the purpose of representation without 
any further distinction between rich and poor, advanced and backward, educated and uneducated . . . 
."  37.   Both of these points were rejected by the dissenters, S. B. 
____________________ 
31. Hutton 1961:195. 
32.  Ibid., 196­97
   . 
33.  Ibid., 197
   . 
34. Dushkin ( 1957:91­95) observes that in Orissa, Bihar, and Bengal local census superintendents 
utilized "ad hoc" criteria, in addition to those given by Hutton, to include Hinduized tribals, 
nomadic and criminal tribes. Their numbers were negligible except in Bengal where these "ad 
hoc" castes made up over a quarter of the population on the final list. 
35. Hutton 1961:199. 
36. Indian Franchise Committee 1932:I, 203. 
37. Id.,210. 
­129­ 
Tambe, C. Y. Chintamani, and R. R. Bakhale. They find that outside of Madras, Bombay, and the 
Central provinces, untouchables "do not form a distinct and separate element in the population." In 
the other provinces "untouchability . . . is the adjunct not of a person but of the occupation he 
occupies."  38.   The majority of the Committee, adhering to denial of access to temples and causing 
pollution by touch or approach as "generally accepted tests of untouchability"  39.   reiterated its 
conviction that the depressed classes should include only untouchables and not "those Hindus who 
are only economically poor and in other ways backward but are not regarded as untouchables." 40. 
But in the face of conflicting estimates it confessed its inability to reach any conclusions about the 
untouchable population in the north and east.  41.   
The groups in this category were finally listed (i.e. scheduled) in 1936 for purposes of giving effect 
to the provisions for special electoral representation in the Government of India Act, 1935.  42.   The 
list finally promulgated did not correspond exactly with any of the earlier estimates: it hewed closely 
to the low figures for Assam, the high figures for the United Provinces and Bihar and Orissa, and 
took a middle course in Bengal.  43.   In the event, the list reflected definitions of untouchability with 
an admixture of economic and educational tests and considerations of local politics.  44.   Lelah 
Dushkin observes that 
the authorities who made up the final list made quite a departure from the original strict 
criterion of untouchability . . . many castes listed in the North would not have been 
included in the South. However, whatever the criteria were, all of the castes included in 
the final list were socially the lowest Hindu castes in their particular provinces and they 
were listed partly, at least, on that basis.  45.   
The list contained over 40 million persons. The 1941 Census enumerated 48.8 million members of 
the listed groups (very close to the 50 million figure current at the inception of the debate a quarter­
century earlier). They constituted over 19% of the Hindu population and 12.6% of the total 
population of undivided India. (See table 14.) 
____________________ 
38. Id.,220 
39. Id.,110. 
40. Id.,109. 
41. E.g., for U.P. the Committee had estimates of 0.58 million, 6.8 million, 9.2 million, and 11.2 
million (the latter accompanied by 1.4 million "untouchable" but depressed (id.,114­16). 
42. The Government of India (Scheduled Castes) Order, 1936. 
43. For an analysis of the discrepancies between Hutton's list and the Order's for north and eastern 
India, see M. Singh 1947. 
44. M. Singh 1947 gives an estimate of the groups included in the Order which were not 
"untouchable" according to the Census: U.P., 75,936 (0.7%); Bihar, 146,397 (3.2%); Bengal, 
2,386,515 (26.2%). 
45. Dushkin 1957:95. 
­130­ 
TABLE 14 POPULATION OF SCHEDULED CASTES IN 1941 
SC Pop.  % Total Pop. 
% Hindu Pop. 
(mils.) 

BRITISH INDIA  39.92  20.90  13.49 


SC Pop.  % Total Pop. 
% Hindu Pop. 
(mils.) 

MADRAS  8.068  18.85  16.35 


BOMBAY  1.855  11.20  8.89 
BENGAL  7.378  29.44  12.23 
UTTAR PRADESH  11.717  25.57  21.39 
PUNJAB  1.248  16.53  4.38 
BIHAR  4.340  16.37  11.49 
5.578  16.74  12.37 
ORISSA  1.238  18.12  14.18 
CENTRAL PROVINCES  3.051  23.57  41.34 
ASSAM  0.676  16.04  6.62 
N.W.F.P. 
SIND  0.191  15.58  4.22 
AJMER­MERWARA 

TOTALS  39.80  20.87  13.45 


STATES AND AGENCIES  a  8.892  13.86  9.78 
COCHIN  0.141  15.73  9.92 
HYDERABAD  2.928  21.99  17.92 
MYSORE  1.405  21.03  19.16 
RAJPUTANA . . .  . . .  . . . 
TRAVANCORE  0.395  11.15  6.50 

TOTALS (throughout

India)     48.81  19.15  12.62 

Source:  Census of India, 1941, I. Part I:98. 
Notes:  aGreater than sub­totals, owing to omission of minor units. 
bSome units omitted. 

B. SCHEDULED CASTES UNDER THE CONSTITUTION 
The Constitution makers eschewed any attempt at a connotative definition of untouchability and 
adopted the same sort of solution as had the British. Lists were to be compiled for each area of those 
castes which were regarded as needful and deserving of such treatment. Rather than provide any 
standard of selection for these groups beyond the general understanding that they were the 
"untouchables," the 
­131­ 
Constitution provided only a procedure for designating them. The President is empowered to 
specify, after consulting with the Governor of a state, those "castes, races or tribes or parts of or 
groups within castes, races and tribes which shall for purposes of this Constitution be deemed to be 
Scheduled Castes in relation to that State. . . ."  46.   Once promulgated, this list can be changed only 
by Act of Parliament.  47.   
A Scheduled Castes Order was promulgated by the President in 1950 which basically re­enacted the 
1936 list.  48.   The major additions were four Sikh castes (discussed below in chapter 9, §E) and the 
provision for the first time of lists for areas which had none previously, (notably Rajasthan, Gwalior, 
and Madhya Bharat). 
There were a few changes in 1951, but these were in the nature of adjustments and elimination of 
anomalies, rather than any basic change in policy.  49.   The 1951 census revealed that the Scheduled 
Castes contained over 52 million members.  50.   More discrepancies, errors, and oversights came to 
light, and in 1956 the lists were revised again.  51.   The principal changes were the addition of about a 
million persons each in Rajasthan and in U.P. and the inclusion now of all Sikh untouchables. The 
revised list contained 55.3 millions (using 1951 census figures). In the 1971 census, the Scheduled 
Castes contained 80 millions (14.6% of the population).  52.   (See table 15.) 
The changes in the list have been more in the nature of removing anomalies and correcting 
oversights: moving groups from the Scheduled Caste to Tribe lists, sorting out discrepancies and 
errors, rationalizing administration.  53.   There have been no important policy departures (other than 
inclusion of the Sikhs). Indeed, there has been no further attempt to formulate any criterion for 
inclusion. In his first report the 
____________________ 
46. Art. 341(2). 
47. Art. 342(2). 
48. Constitution (Scheduled Castes) Order, 1950. 
49. Constitution (Scheduled Castes) (Part C States) Order, 1951. 
50. The original 1951 Census figure of 51.3 millions understated the total for a number of reasons, 
including the omission of untouchables who had returned themselves under generic names 
Scheduled Caste, etc.) or under "no caste" (Ad­Dharmis, etc.). In 1953, the census authorities 
made adjustments which brought the total to 52.1 million (RCSCST, 1954:154­55, 332­33). 
51. The Scheduled Castes and Scheduled Tribes Order (Amendment) Act, 1956 (Act 63 of 1956). 
The Backward Classes Commission, which was formed in 1953 and reported in 1955, had been 
asked to recommend alterations in the Scheduled Castes and Scheduled Tribes lists. 
52. These numbers increased slightly when the list was amended in 1976. See n. 74, below. 
53. For some examples of administrative confusion see, e.g. RCSCST 1952: 27 (overlapping separate 
lists of "Depressed Classes" and "Scheduled Castes" in Mysore); RCSCST 1958­59:I, 9 (separate 
and different Scheduled Caste Lists issued by two government departments in Kerala). 
­132­ 
TABLE 15 SCHEDULED CASTES POPULATION (IN MILLIONS) AND PERCENTAGES, 
1950­1951, 1956, 1961, 1971 
According to 
According to 1950
1956 1961 Census  1971 Cens
and 1951 Orders 
Order 

SC SC %  SC SC %  SC SC % 


Total Popula­ of Popu­ of Total Popula­ of Total
Popula­ tion  Total lation  Total Popula­ tion  Total Popu­ SC
tion Popula­ Popula­ tion Popula­ lation Popula­
( 1951)  tion  tion  ( 1961)  tion  ( 1971) tion 

Source  A  A  A  A  A  B  B  B  C  C 

ANDHRA 
31.3  4.406  14.09  4.415  14.3  36.0  4.973  13.50  43.50  5.774 
PRADESH 
ASSAM  9.0  0.424  4.68  0.424  4.68  11.9  0.732  6.16  14.96  0.912 
BIHAR  38.8  4.913  12.66  4.913  12.67  46.5  6.563  14.04  56.35  7.950 
BOMBAY  48.3  5.130  10.62  5.202  10.78 
GUJARAT  20.6  1.367  6.63  26.70  1.825 
MAHARASHTRA  39.6  2.226  5.61  50.41  3.025 
JAMMU & 
4.4  N.A . . .  0.156  3.54  3.6  0.263  7.55  4.62  0.381 
KASHMIR 
KERALA  13.5  1.254  9.25  1.207  8.81  16.9  1.422  8.40  21.35  1.772 
MADHYA 
26.1  3.502  13.43  3.912  15.01  32.4  4.245  13.12  41.65  5.453 
PRADESH 
MADRAS [ TAMIL 
30.0  5.391  17.98  5.381  17.95  33.7  6.072  18.00  41.20  7.315 
NADU] 
MYSORE 
19.4  2.589  13.33  2.583  13.31  23.6  3.117  13.20  29.30  3.850 
[ KARNATAKA] 
ORISSA  14.6  2.639  17.95  2.629  17.95  17.5  2.763  15.71  21.94  3.310 
PUNJAB  16.1  3.062  18.97  3.490  21.64  20.3  4.139  20.32 
PUNJAB  13.55  3.348 
HARYANA  10.04  1.895 
RAJASTHAN  11.0  1.584  9.91  2.502  15.67  20.2  3.359  16.62  25.77  4.075 
UTTAR PRADESH  63.2  1.855  18.75  13.100  20.72  73.7  15.417  20.90  88.34  18.548 
WEST BENGAL  26.3  4.855  18.56  4.743  18.04  35.0  6.950  19.84  44.31  8.816 
UNION 
4.1  0.57  13.83  0.664  16.23  7.8  0.916  11.74  6.34  0.750 
TERRITORIES 

TOTALS  61.2  2.2  14.5  55.3  15.32  39.1  54.504  14.64  47.95  79.995 

Source: A: RCSCST 1956­57: II, 15: B: RCSCST 1961­62: II, 74; C: Census of India. 1971, Series I, 
Part II­A (ii): xxvii­xxix. 
Note: Totals include minor units not given here. 
­133­ 
Commissioner for Scheduled Castes and Tribes observed that Hutton's 1931 census criteria "appear 
to hold good for purposes of specifying the Scheduled Castes."  54.   Although there was no indication 
that they were applied any more literally than in 1931 or 1936, no one seemed troubled. In the late 
1950s the Commissioner remarked that the criteria for classifying a caste as Scheduled were "fairly 
definite."  55.   
Formally, the power to make and change the lists lay with the President (and later with Parliament), 
but it appears that the dominant role was played by the executive, center and state. The lists 
remained in charge of the Ministry of Home Affairs except for an interval from 1964 to 1973 when 
this task was housed in the erstwhile Department of Social Security (later Social Welfare). 56. 
Although the role of the state is formally consultive, designation seems in fact to be a two­stage 
process in which center or state proposes and attempts to get consent of the other. Where the state is 
unwilling, listings favored by the center can be held up for years. On the other hand, where states 
have proposed changes in the list, especially inclusions, these have been rejected for lack of any 
supporting data.  57.   
The selection of Scheduled Castes, then, has proceeded without benefit of a single connotative 
definition. The selection has proceeded primarily on the basis of "untouchability"­­measured by the 
incidence of social disabilities­­but this criterion has been combined in varying degrees with 
economic, occupational, educational, and (as we shall see) residential and religious tests. 
The resulting list, then, designates all of those groups who in the view of Parliament require the 
special protections provided by the Constitution: it defines who may stand for reserved seats and 
enjoy benefits and reservations for the Scheduled Castes. But it does not necessarily include every 
person or group that might be considered "untouchables" by any conceivable definition. It omits 
some groups which historically suffered disabilities (e.g. Ezhuvas) or which would be untouchables 
in terms of the 1931 census tests. And it excludes nonHindus (other than Sikhs) who would clearly 
seem to be untouchables within the judicial test of "origin in a group considered beyond the pale of 
the caste system."  58.   There is, thus, no single inclusive list of all groups considered untouchable, 
just as there is no single criterion for identifying them. 
____________________ 
54. RCSCST 1951: 10­11. 
55. RCSCST 1957­58: I, 7. Cf. Ministry of Home Affairs 1956: 1. 
56. RCSCST 1963­64: I, 13; 1972­73: I, 251. 
57. RCSCST 1960­61: I, 22. Department of Social Security 1965:55 if. 
58. See Devarajiah v. Padmanna, A.I.R. 1958. Mys. 84 at 85, discussed below at nn. 105 and 109. 
­134­ 
In spite of the absence of a definitive criterion (or perhaps because of it), the lists have remained 
stable for more than thirty years. Some groups continue to petition for inclusion. The Commissioner 
observes that "there has been a tendency of some castes and communities to use every opportunity 
for pressing their claims for inclusion in the list of Scheduled Castes/Scheduled Tribes."  59.   But the 
process of getting on is so cumbersome that few have succeeded. If the pre­Independence figures of 
12.3% are adjusted to allow for the failure of some areas to have lists at all and for the removal of 
predominantly non­Hindu areas by partition, the 1971 figure of 14.6% (the same as in 1961) 
represents very little change in overall proportion. There have been no significant inclusions (other 
than the Sikhs). The changes that have occurred have been rationalizations and extensions of 
existing criteria to overlooked groups and areas previously omitted, rather than any change in 
criteria. Nor have there been any significant exclusions with the exceptions of the "voluntary" 
departure of the converts to Buddhism. There has been no litigation about inclusion or exclusion of 
any group challenging the "untouchability" criterion or its application­­but there has been litigation 
on the religious and residence tests. 
With whatever admixture of other factors, the Scheduled Castes have been chosen primarily by the 
criterion of low social and ritual standing. In the selection of Other Backward Classes, the use of 
social and ritual status as a criterion of the backwardness of the group has been severely restricted 
by the courts.  60.   But such criteria have been predominant in the selection of Scheduled Castes, who 
have been chosen precisely on the ground of their low social and ritual status in the traditional social 
hierarchy. The justification for employing low status and attendant disabilities as a criterion for 
preferences is that these groups are generally lowest in income, education, health, cultural resources, 
and economic opportunities. Generally, more than a mere correlation is implied. These castes are, in 
the prevailing view, backward because they were subject to disabilities which impeded advancement 
by individual effort and denied them opportunities for group social mobility.  61.   
This emphasis has endured over the years. In the debate preceding passage of the 1976 amendments 
to the Scheduled Castes Order the 
____________________ 
59. RCSCST 1957­58: I, 7. Petitions for inclusion are still made. See the representations of South 
Indian barbers and washermen to the Joint Parliamentary Committee, reported in The Mail, 16 
September 1968. In a few instances, castes have petitioned for removal from the list on the 
ground that inclusion was degrading. See e.g., RCSCST 1952:66. 
60. See chap. 8, § B, below. 
61. M. N. Srinivas ( 1956:377) observes that "the barrier of untouchability" cuts off untouchables 
from the opportunities for group social mobility that do exist within the caste system. 
­135­ 
view surfaced that "it is basically the economic and social backwardness of a caste which should 
entitle it to be categorized as a Scheduled Caste."  62.   But the Home Minister was emphatic in 
reiterating that the concept of Scheduled Castes is one of backwardness stemming from 
untouchability. "[I]t is not neglect, it is not mere poverty, it is not mere backwardness that entitles a 
man to come under the Scheduled Castes."  63.   
However, to the extent that the legislative prohibition of disabilities and changing social practice 
result in a lessening of disabilities, and to the extent that preferential treatment increases educational 
attainments and economic opportunities, the correlation between low ritual standing and economic 
and social backwardness will be weakened. If "untouchability" is a useful criterion for identifying 
those groups which are at the bottom in terms of economic, educational, and cultural resources and 
opportunities, the very success of redistributive measures might lead to a discrepancy between ritual 
standing and the other indicia of backwardness. It would seem that Parliament's power to designate 
Scheduled Castes is sufficiently broad to permit the use of standards which would decrease the role 
of ritual standing and disabilities and increase the emphasis on educational, economic, and cultural 
criteria. There is room for eventually changing the mix of ritual and other factors in the selection of 
Scheduled Castes. There has been some discussion about changing the mix­­at least to the extent of 
applying economic and educational tests to de­schedule those groups which have allegedly been so 
benefitted that the correlation no longer holds. 
In June 1965 a committee was appointed by the Government to advise on revision of the existing 
lists of Scheduled Castes and Scheduled Tribes, under the chairmanship of B.N. Lokur, the Law 
Secretary. The Lokur Committee reported promptly on August 25, 1965.  64.   It found that in view of 
the constitutional abolition of untouchability, 
[I]t would indeed be inappropriate . . . to apply the sole test of untouchability in 
preparing the list of Scheduled Castes . . . Nevertheless, having regard to the historical 
background, we have, in revising the list, adopted the test of extreme social, educational 
and economic backwardness of castes, arising out of the traditional custom of 
untouchability.  65.   
____________________ 
62. S hri Santosh Kumar, RSD Vol. 97, No. 18: Col. 35 ( 3 Sep. 1976). 
63. Shri K. Bhrahmananda Reddy, LSD (5th Series), Vol. 64, No. 16: Col. 125 ( 2 Sep. 1976). 
64. Department of Social Security 1965. The CSCST suggests that the Basavalingappa case's 
requirement that all synonyms and subgroups be listed (see chap. 9, § A, below) was the 
occasion for the appointment of the Lokur Committee (RCSCST 1965­66:65). 
65. Department of Social Security 1965:6. 
­136­ 
The committee noted with satisfaction that untouchability was "fast disappearing, particularly in 
cities and towns, and, even where it is practised, it is considerably diluted in form."  66.   They 
observed that "[t]he line of demarcation between high castes and low castes, which was fairly clear 
in the past, has tended to become blurred."  67.   Asserting that "a lion's share" of the various benefits 
and preferences were "appropriated by the numerically larger and politically well organized 
communities" on the list,  68.   and noting the deleterious effects of creation of "vested interests" in the 
listing, they concluded that "the time has come when the question of descheduling of relatively 
advanced communities should receive serious and urgent consideration."  69.   
The Committee devoted most of' its attention to many technical changes in the lists, including a 
number of minor exclusions and inclusions. On major de­scheduling, though, they curiously omitted 
to utilize any of the fresh and abundant 1961 census data. Instead, they cautiously put forward a list 
of communities adjudged to be "relatively forward" by "several persons who appeared before us, 
including eminent social workers."  70.   In addition to 14 tribes, this list included 28 Scheduled Castes 
listings, ranging from the insignificant to such numerous groups as the Chamars ( Bihar, U.P., and 
Punjab) and the related Jatav ( Madhya Pradesh); the Dhobis ( West Bengal; Orissa, Bihar, and 
U.P.); the Mahars ( Maharashtra and Madhya Pradesh); the Malas ( Andhra); and the Namasudras 
and Rajbanshis of West Bengal. In a kind of dim reflection of the disputes of the early 1930s, the 
proposals would have left the southern Scheduled Castes virtually intact, but practically halved the 
Scheduled Caste population in the north. In addition, it would have eliminated about a fifth of the 
Scheduled Tribes.  71.   
____________________ 
66. Id., at 6. 
67. Id., at 6. 
68. Id., at 8. 
69. Id., at 10. 
70. Id., at 11. Dushkin ( 1972:220 ff.) exposes the thinness of the Committee's assertions about the 
advancement of the groups it proposes to de­schedule. Only a handful had literacy rates higher 
than the average for Scheduled Castes in their states; over half had literacy rates lower than the 
all­India rate for Scheduled Castes. 
71. Estimated percentage of Scheduled Castes and Scheduled Tribes in each state who would have 
been de­scheduled if the provisions of the Lokur Commission had been implemented 
Scheduled Castes 
Scheduled Tribes 
Andhra Pradesh 
34.8 
­­ 
Assam 
32.01 
40.48 
Bihar 
34.07 
­­ 
Gujarat 
11.35 
­­ 
Kerala 
4.51 
­­ 
Madhya Pradesh 
66.69 
57.98 
Madras 
0.04 
­­ 
Maharashtra 
35.12 
20.32 
(Table continued overleaf) 
­137­ 
The report was received with angry resistance by Scheduled Caste spokesmen. At a tumultuous 
meeting with the Law Minister (then in charge of the Social Security portfolio), Scheduled Caste 
Members of Parliament agreed with the few inclusions but would not countenance any exclusions 
from the list. The Minister was reported to have retreated. Scheduled Caste leaders and government 
officials concurred that the report was dead.  72.   
After the 1967 elections, the question was raised again ­­ now with the added dimension of the 
impending debate on whether to extend the reserved seats, due to expire in 1970. In the meantime, 
Congress (the principal political beneficiary of reserved seats) had become considerably more 
dependent on the adherence of the occupants of the reserved seats, who now supplied its much 
narrower plurality in the Lok Sabha. The Congress plurality was 38 seats (279/520); Congress held 
72 of 114 reserved seats.  73.   
On August 12, 1967, a bill to amend the lists was introduced by the Government; the bill's proposed 
list included every one of the groups whose possible deletion had been suggested by the Lokur 
Committee. A Joint Parliamentary Committee established to review the lists submitted its report in 
November 1969. Among the amendments proposed by the Joint Committee were the inclusion in 
Scheduled Castes of women who married Scheduled Caste men, the exclusion from Scheduled 
Tribes of converts to Christianity and Islam, and the elimination of all area restrictions on 
Scheduled Caste status. The conversion proposal provoked great controversy, and neither it nor the 
removal of area restrictions were accepted by the Government, which proposed a host 
____________________ 
71 
(Table continued from 
previous page) 
Mysore 
8.60 
­­ 
Nagaland 
­­ 
98.91 
Orissa 
9.6 
4.95 

Punjab 
38.56 
­­ 
Uttar Pradesh 
62.22 
­­ 
West Bengal 
31.81 
­­ 
All India 
33.88 
20.29 
Percentage of SC 
recommended to be de­
scheduled in four 
southern states: 13.34. 
Percentage of SC 
recommended to be de­
scheduled in the rest of 
India: 40.44. 
Source: Computed from 
Census 1966. 

72. One cynical and well­placed observer concluded that the Lokur Report 
was never meant to be taken seriously, since the Government would not 
remove electorally potent communities like Chamars, Mahars, and 
Namasudras from the list a year before the general elections, but found it 
convenient to remind them that there was such a possibility and how 
much they had to be grateful for. Interview, B. P. Maurya, December 
1965. Whatever the motivation of the political leaders, there is no doubt 
that the committee members were serious and well­intentioned. 
73. Chandidas, et. al 1968:553­643. 
­138­ 
of amendments to the Joint Committee's version when the bill was taken up in November 1970. 
Mired in controversy, the bill lapsed with the dissolution of the Fourth Lok Sabha in December 
1970. 
A new version, brought forward during Emergency Rule and finally passed in August 1976, 
eliminated most intra­state area restrictions, adding about 2.5 million to the Scheduled Castes 
population and over 3 million to the Scheduled Tribes.  74.   Apart from this, the Act maintained the 
status quo: there were no significant inclusions or exclusions. It did not address any of the touchy 
issues of personal status (of wives, converts, or migrants) which had come to surround the 
application of the lists. 
Caste remained the primary basis for designating the Scheduled Castes. Caste plays a dual role in 
the selection of Scheduled Castes: castes in the sense of social units are typically the groups which 
are listed, and it is caste in the sense of rank or status in the social­religious hierarchy which is the 
criterion for choosing these groups. But caste is not the only factor in the selection: two other factors 
are explicitly employed­­territory and religion. Scheduled Castes are designated by state and 
sometimes by districts or regions within states. The same caste may be scheduled in one state but 
not in an adjoining state, or one district but not in an adjoining district. 
The power of the President to specify Scheduled Castes by districts 
____________________ 
74. Scheduled Castes and Scheduled Tribes (Amendment) Act, 1976 (Act 108 of 1976). Only 64 of 1, 
126 intra­state area specifications were retained. The additions to the Scheduled Caste and Tribes 
population can be seen in the following table: 
Increases (and Decreases) in Scheduled Caste and
Tribe Population due to 1976 Act (in lakhs) 
State 
Scheduled Castes 
Scheduled Tribes 
Andhra Pradesh 
+0.41 
+5.68 
Bihar 
+4.35 
Gujarat 
+0.64 
+0.23 
Himachal Pradesh 
+0.38 
Karnataka 
+4.27 
+0.31 
Kerala 
+2.30 
­0.76 
Madhya Pradesh 
+2.98 
+14.28 
Maharashtra 
+1.51 
+8.87 
Orissa 
­0.04 
+0.03 
Rajasthan 
+1.40 
+0.09 
Tamil Nadu 
+0.22 
+1.38 
Uttar Pradesh 
+5.46 
West Bengal 
+0.84 
+0.70 

Total 
+24.76 
+31.57 
+0.04 
+0.76 
Source: RCSCST 1975­77: I, 16. 
­139­ 
has been upheld on the ground that the social and educational backwardness of a group may vary in 
degree in different areas.  75.   Specification by states has been held to be constitutionally required. A 
postal clerk residing and working in Orissa was a member of the Konda Kapus, a group listed as a 
Scheduled Tribe in neighboring Andhra Pradesh but not in Orissa. After being appointed to a higher 
post against a reservation for Scheduled Tribes, he was reverted on the ground that Konda Kapu was 
not a Scheduled Tribe in Orissa where he was a permanent resident. He argued that he should be 
accounted a Scheduled tribe anywhere for purposes of central government employment, since 
central legislation clearly outlawed state residence requirements for central government jobs.  76.   The 
High Court, turning to the constitutional provisions for designating Scheduled tribes (Art. 342[1]), 
finds that the phrase "shall be deemed to be Scheduled Tribes in relation to that State" means that in 
order to get the benefit of being a member of a Scheduled Caste or a Scheduled Tribe in the matter 
of public employment, the person claiming it should be a member of such caste or tribe in relation 
to the particular area or state where he is residing and where he seeks employment.  77.   Thus the 
court takes a restrictive reading of residence rules to be a constitutional requirement. On numerous 
occasions, officials and political leaders have taken the view that only a constitutional amendment 
could abate these restrictions.  78.   
However, the area limits have been troublesome. There is some doubt as to whether they correspond 
with any exactness to the different levels of deprivation and need. (Indeed, the decision not to 
collect caste­wise census data for any groups other than those already listed as Scheduled Castes or 
Tribes makes this exceedingly difficult to determine.) More troubling still are the difficulties 
presented by cases of spatially mobile members of these groups. The Scheduled Castes Order says 
rather obscurely that castes shall "be deemed to be Scheduled Castes so far as regards members 
thereof resident in the localities specified in relation to them respectively." . . .  79.   In his first report 
in 1951 the Commissioner of Scheduled Castes and Scheduled Tribes 
____________________ 
75. Bhaiya Lal v. Harikrishnan Singh, A.I.R. 1965 S.C. 1557, 1560. 
76. Public Employment (Requirement as to Residence) Act, 1957 (Act 44 of 1957). 
77. K. Appa Rao v. Director of Posts and Telegraphs, A.I.R. 1969 Or. 220, at 221. 
78. E.g., the Lokur Committee ( Departmet of Social Security 1965:14); the Commissioner for 
Scheduled Castes and Scheduled Tribes (RCSCST 1971­2/1972­3:242­43) and the Home 
Minister [LSD (5th Series), Vol. 64 No. 16: Col. 166 ( 3 Sep. 1976)]. 
79. Para. 2. Like much else, this was carried over from the 1936 order, evidently without any thought 
about its purpose. Since the principal "benefit" envisaged in 1936 was special electoral 
provisions, the residence requirement was in the nature of provision for determining electorates, 
but wholly inappropriate now that the lists are used for a variety of welfare schemes. 
­140­ 
observed that the provision was ambiguous.  80.   Taking the example of the Khatik caste, which is a 
Scheduled Caste in Punjab but not in neighboring Uttar Pradesh, he propounded the following 
questions: When, if at all, does a U.P. Khatik who moves to Punjab become a member of a 
Scheduled Caste in relation to Punjab? Is a Punjab Khatik who moves to U.P. no longer a member of 
a Scheduled Caste in relation to Punjab or in relation to concessions granted by the central 
government? In a strikingly unhelpful interpretation the Ministry of Law opined that "each case will 
have to be decided on its facts." But since "residence at the time in question" was the "determining 
factor," they inclined to the view that the Punjab Khatik would not be a Scheduled Caste while 
residing in U.P.  81.   
The Commissioner's example contains a series of puzzles which we might usefully sort out and 
label. First, there is the out­migration problem. Suppose an X leaves a state where Xs are a 
Scheduled Caste and goes to a state where they are not so listed. Such persons are excluded from 
any benefits by the prevailing reading of the Order. Thus a Dusadh who moved his residence from 
Bihar (where Dusadh was a Scheduled Caste) to Madhya Pradesh (where it was not) was not a 
member of the Scheduled Castes for purposes of making a lower deposit when filing a nomination 
paper.  82.   The same rule applies to intra­state territorial limitations. A railway worker resident 
outside the taluks listed for his tribe was held properly denied eligibility for reservation in 
promotion.  83.   
There is the converse problem of in­migration. Suppose a member of group X migrates from an area 
in which the Xs are not scheduled to an area in which they are. Here the residence test stressed by 
courts and officials would seem to argue for inclusion. But the one court that has addressed it 
decided the other way. "Suryabashi" was listed as a Scheduled Caste in Jabalpur District. But a 
Suryabashi migrant to Jabalpur who hailed from an area where Suryabashis were not Scheduled was 
held ineligible for benefits on the ground that "only Suryabashis residing in Jabalpur district . . . 
were declared to be Scheduled Castes under the Presidential order."  84.   Apparently the court is 
looking beyond residence to some notion of origin in the designated groups, a test that proved 
insufficient in the out­migration cases. 
These issues are decided entirely by textual interpretation without any consideration of policies 
which might provide some guidance here. 
____________________ 
80. RCSCST 1951:8. 
81. Id., at 8­9. 
82. Jyoti Bhushan v. Bodh Ram Muritram, 1974 M.P.L.J. 565. 
83. Wasudeo v. Union of India, 1973 Mh. L.J. 994. Residence is equated with "permanent abode" 
rather than place of origin or ancestral home (id., at 998). 
84. Lalwani v. Collector, A.I.R. 1976 M.P. 44, 45. 
­141­ 
For example, in­migration involves the enlargement of the number of beneficiaries, whereas out­
migration does not. Again, extension of benefits to in­migrants might induce mobility that was not 
otherwise economically justifiable; but extension of benefits to out­migrants might be necessary in 
order to prevent the inhibiting of economically justified moves. 
The logic of these cases points to an even more grotesque result in the case of lateral movement. 
Suppose a member of group X migrates from state A where the group is scheduled to state B where 
it is also scheduled. According to the residence argument of the out­migration cases, he is no longer 
entitled to any benefits as an X of state A. Nor, according to the origins argument of the in­
migration case, is he entitled to any benefits as an X of state B. One suspects that administrative 
failure to give full effect to the logic of the cases, perhaps augmented by some dissimulation on the 
part of the migrants, has tempered the effect of these rules and accounts for the infrequency with 
which these issues have come to court. 
That migrant members would lose their Scheduled Caste benefits remained troublesome. In 1959, 
the Commissioner had raised the matter again,  85.   this time in connection with the case of a Dhobi 
whose family origins were in U.P. (where Dhobi is a Scheduled Caste) but who was born, brought 
up, and employed in Bombay (where Dhobi is not a Scheduled Caste). He claimed that his ancestral 
property, house, etc., were in U.P. and that he had no property or house in Bombay, but lived there 
only for the sake of his employment. This case was referred to the Law Minister, who again read the 
order so that he was excluded­­a result that the Commissioner found anomalous, in view of the fact 
that 
[s]uch persons may actually continue to suffer from all the disabilities resultant from the 
practice of untouchability as they may have to maintain all ties with their relations in the 
home State.  86.   
He recommended that migrant Scheduled Castes and Tribes should be eligible for benefits from the 
central government and their home state at least for a generation.  87.   No relief for inter­state migrants 
has been forthcoming. 
In 1965 the Lokur Committee recommended the removal of interstate area limitations on the ground 
that they inhibit mobility.  88.   The Scheduled Castes and Scheduled Tribes Orders (Amendment) 
Bill,1967
____________________ 
85. RCSCST 1958­59: I11 
86. Id. 
87. Id.; the Backward Classes Commission had earlier recommended that migrating members should 
be treated as Other Backward Classes ( I: 155). In 1963 the Commissioner noted that the 
Government of India had still not taken any decision on this issue, ( 1962­63: I, 12). 
88. Department of Social Security, 1965: 13. 
­142­ 
1967, proposed that when a caste is Scheduled for one area of a state, members from that area shall 
continue to be included so long as they reside anywhere in that state  89.   The 1976 amendment of the 
lists did not contain this provision, but the need for it was largely dispelled by the abandonment of 
almost all intra­state area restrictions. 
The conferring of scheduled status on a group is a monopoly of the Centre. The 1950 Scheduled 
Tribes Order list for Maharashtra included an item for Halba from six specified taluks in three 
districts. In 1967 the Government of Maharashtra ordered that the Halba Koshtis of the whole of 
Vidarbha region be treated as belonging to the Scheduled Tribes. A Halba railway worker from 
outside the listed taluks was deleted from a promotion panel on the ground that he was not a 
member of Scheduled Tribe. His challenge was repulsed by the Bombay High Court, which held 
that the state's pronouncement could have no effect on his status as a Scheduled Tribe in connection 
with any of the affairs of the central government. The court left open the question whether the state 
might treat him as a Scheduled Tribe in regard to state services.  90.   
This issue has never been fully addressed. While it might be desirable to allow the state to employ 
its better­informed judgement about local conditions, there is of course the danger that state use of 
this power could dilute benefits to those deemed deserving beneficiaries by national policy. The 
most prominent instance of this­­Maharashtra's treatment of the Buddhists on a par with Scheduled 
Castes­­has not been challenged in court. 
Religion was introduced as a qualification into the first Scheduled Castes Order in 1936, which 
provided that no Indian Christian (nor, in Bengal, those professing Buddhism or a tribal religion) 
should be deemed a member of a Scheduled Caste.  91.   Earlier, it was often recognized that there 
were comparable depressed groups among Christians and Muslims. But in the disputes leading up to 
the listing of Scheduled Castes, it was agreed that Muslims and Christians should be excluded. 92. 
This execution was readily understandable, for the major purpose of the list was to provide for 
electoral representation, and Christians and Muslims were the beneficiaries of special electoral 
treatment as minorities.  93   In spite of the constitutional ban on religious 
____________________ 
89. Lok Sabha Bill No. 119 of 1967, p. 7. 
90. Wasudeo v. Union of India, 1973 Mh. L.J. 994. 
91. The Government of India (Scheduled Castes) Order, 1936, §3. 
92. Thus Hutton's note to the Indian Franchise Committee observes that it was decided to omit 
Christians and Muslims ( Indian Franchise Committee 1932: I, 111). 
93 For the composition of the legislative bodies under the Government of India Act, 1935, see the 
First and Fifth Schedules to that Act. 
­143­ 
discrimination, the elimination of separate representation for religious minorities and the change in 
purpose of the list from electoral to administration of welfare, the religious qualification (or, more 
properly, disqualification) was retained after Independence. The President's 1950 Order provides 
that "no person professing a religion different from Hinduism shall be deemed a member of a 
Scheduled Caste."  94.   An exception was made for Sikh members of four castes. In 1956 it was 
broadened to include all Sikh untouchables.  95.   
The religious test for Scheduled Castes is employed, not as a positive test for selecting appropriate 
groups for inclusion, but as a disqualification of individuals and groups who otherwise meet the 
criteria, thereby inevitably discouraging conversion. There is reason to think that this was at least 
part of its purpose.  96.   It does not operate as an encouragement of Hindu orthodoxy, for the legal 
definition of Hinduism is so broad that few individuals are likely to have difficulty with this test 
other than those who explicitly convert to a non­Hindu religion. In Punjabrao v. Meshram, the 
Supreme Court rejected the argument that Buddhists were included in the meaning of Hindu.  97.   The 
courts have upheld the exclusion of non­Hindus without reaching the broader question of religious 
discrimination. Nor have the courts addressed the factual question of the effect of conversion in 
dissipating the conditions that lead groups to be listed as Scheduled Castes. 
Persistent legislative attempts to include the Buddhists, either by dropping the religious qualification 
or by specifying that "Hindu" be read to include them, have been unsuccessful­­as has a recent 
attempt to extend the religious qualification to the Scheduled Tribes. The interplay of religion with 
caste and tribe membership is discussed in chapter 9. 
Since we shall be examining the impact of the courts on the working of the policy of "compensatory 
discrimination," it is worth noting that the role of the courts in the process of designating the 
Scheduled Castes has been a very minor one. The search for a uniform criterion that could be 
employed throughout India to distinguish untouchables from other Hindus proceeded without any 
assistance from the courts. Although Scheduled Castes were selected on the basis of the disabilities 
they suffered, the jurisprudence of disabilities was too spotty and too enmeshed with local 
variations.  98.   Although untouchables could be read­ 
____________________ 
94. Constitution (Scheduled Castes) Order, 1950, §2. 
95. Scheduled Castes and Scheduled Tribes Order (Amendment) Act, 1956 (Act 63 of 1956). 
96. The problem of converts in taken up in chap.9. 
97. A.I.R. 1965 S.C. 1179 (discussed in detail below, chap. 9, §E.) 
98. See Galanter 1972. 
­144­ 
ily identified with the lower end of the varna scale, the traditional jurisprudence of varna standing 
was of little assistance in identifying the groups which were supposedly untouchable. Modern 
"untouchability" bore little correspondence to the chandala category of classical law. In reference to 
determinations of customary rights, courts had sometimes employed varna categories. Untouchables 
had sometimes, particularly in South India, been referred to as a fifth varna, below the Sudras. 99. 
But in other places they were regarded as Sudras, albeit "unclean" ones.  100.   For purposes of 
applying Hindu personal law, the courts had never attempted to distinguish untouchables from 
Sudras; all Hindus other than the twice­born were lumped together as Sudras.  101.   Even where 
untouchables were popularly regarded as Sudras, they could not be equated with them, since there 
were non­untouchable groups belonging to this category. Thus the tests used for distinguishing 
Sudras from the twice­born could not be used as a satisfactory measure of untouchability. 
Once the lists were promulgated, the courts contributed little to their refinement.  102.   Although on at 
least one occasion, judicial demands for precision helped to provide an opportunity for re­
examination,  103.   courts have contributed nothing to the critical understanding of the lists for the 
criteria that underlie them. 
Although it might seem that identification of "untouchable" groups would proceed from the 
definition of "untouchability" in order to select those who suffer from it, in fact the opposite process 
has taken place. The contemporary legal treatment of untouchability has derived from the listing of 
these groups (which in turn are chosen, as we saw, with at least one eye to an ill­defined notion of 
untouchability). The term "untouchability" had no technical legal meaning before the Constitution 
abolished it.  104.   The few judicial encounters with the concept have not succeeded in defining it in a 
way that would be useful in identifying untouchable groups. 
The courts have indicated that untouchability does not include all instances in which a person is 
treated as ritually unclean and a source of pollution. It does not include such temporary and expiable 
states of 
____________________ 
100. See e.g., Atmaram v. King­Emperor, A.I.R. 1924 Nag. 121. 
101. See e.g., Muthusami v. Masilmani, I.L.R. 33 Mad. 342 (1909); Manickam v. Poongavanammal, 
A.I.R. 1934 Mad. 323. 
102. A somewhat exaggerated expression of judicial impotence in this area is found in the assertion of 
then Chief Justice Ray in State of Kerala v. N. M. Thomas, A.I.R. 1976 S.C. 490, 501, that "no 
court can come to a finding that any caste or any tribe is a Scheduled Caste or a Scheduled 
Tribe." Cf. the cases discussed in chap. 9, §A, below. 
103. See ch. 9, n. 16, below. 
104. On the history of the terminology, see chap. 2, nn. 21 and 46, above. 
99. See e.g., Sankaralinga Nadan v. Raja Rajeshwari Doral, 35 I.A.C. 176 (1908). 
­145­ 
pollution as those suffered by, e.g., women at childbirth, menstruating women, mourners, persons 
with contagious diseases, persons who eat forbidden food or violate prescribed states of cleanliness. 
 105.   Nor does it include that "untouchability" which arises from incidents of personal history. 106. 
Nor does it refer to situational or relative impurity, such as that between ordinary worshipper and 
priest or temple attendant.  107.   It does not include every instance in which a person is stigmatized as 
unclean, polluting, or inferior because of his origin or membership in a particular group­­i.e., where 
he is subjected to invidious treatment because of difference in religion or membership of a lower or 
different caste.  108.   The "untouchability" with which Article 17 is concerned is that which "refers to 
those regarded as untouchables in the course of historical development" and which is related to the 
relegation of persons "beyond the pale of the caste system . . . on the ground of birth in certain 
classes."  109.   
Thus untouchability is confined to disabilities imposed upon groups commonly regarded as 
"untouchable," and its meaning is to be determined by reference to those who have traditionally 
been considered untouchables. But it is, as we have seen, no easier to define untouchables than to 
define untouchability. "Beyond the pale of the caste system" 
____________________ 
105. The Mysore High Court, the first to address itself to the problem, pointed out that the word 
"untouchability" appears in Article 17 between inverted commas and inferred that "the subject 
matter of that Article is not untouchability in a literal or grammatical sense but the practice of it 
as it had developed historically in this country" ( Devarajiah v. Padmanna, A.I.R. 1958 Mys. 84 
at 85). Cf. the remarks of Professor K. T. Shah , VII CAD 668, and Mr. Nazirrudin Ahmad, VII 
CAD 665. 
106. Thus the ban on untouchability does not make excommunication from caste unconstitutional, 
although it makes a person "for all practical purposes untouchable." Hadibandhu v. Banamali, 
A.I.R. 1960 Or. 33. (But cf. the argument of Sinha, C.J., dissenting in Saifuddin Saheb v. State of 
Bombay, A.I.R. 1962 S.C. 853 at 866: forbidding excommunication was only carrying out "the 
strict injunction of Art. 17." The majority disagreed, finding that the power to excommunicate 
may in some circumstances be the constitutionally protected right of a religious sect.) 
107. Parameswaran Moothathu v. Vasudeva Kurup, I.L.R. 1960 Ker. 73. Consonant with this 
understanding, temple­entry legislation puts untouchables on a parity with other worshippers, 
but does not attempt to erase distinctions between worshippers and priests or attendants. All 
temple­entry legislation confers rights to enter and worship "in the same manner and to the same 
extent" as is permissible for other persons professing the same religion, etc. On several occasions 
the Supreme Court has implied that the ambit of the temple­entry power would not extend to 
erasing the distinction between ordinary worshippers and temple functionaries ( Sastri 
Yagnapurushadasji v. Muldas Bhundardas Vaishya, A.I.R. 1966 S.C. 1119) or sectarian initiates ( 
Sri Venkataramana Devaru v. State of Mysore, A.I.R. 1958 S.C. 255). 
108. An attempt to explicitly attach this meaning by amending Art. 17 to read "no one shall, on 
account of his religion or caste, be treated or regarded as all 'untouchable'" was defeated in the 
Constituent Assembly ( VII CAD 665, 669). 
109. Devarajiah v. Padmanna, A.I.R. 1958 Mys. 84 at 85. 
­146­ 
is a misleading and unworkable formulation. Even the lowest castes are within the traditional system 
of reciprocal rights and duties, though they had more of the latter and fewer of the former. Their 
disabilities and prerogatives are articulated to those of other castes. Presumably, the court that 
devised this standard meant outside the four varnas of the classical lawbooks. But, as we have seen, 
the varna test is not workable either­­although it continues to crop up in discussions of 
untouchability.  110.   Thus the courts have not developed any learning on "untouchability" that could 
be employed in refining the criteria of selection of the Scheduled Castes. 
However, there is the possibility that as the correlation between low ritual status and economic and 
educational condition breaks down, some of the learning developed by the courts in addressing the 
problems of identifying the Other Backward Classes (see chapter 8 below) will be utilized in the 
Scheduled Castes and Tribes area. 
C. THE SCHEDULED TRIBES 
There has been much less dispute about the identification of the Scheduled Tribes (ST) than about 
the Scheduled Castes or the Other Backward Classes. The general notion is that this category should 
include those groups distinguished by "tribal characteristics" and by their spatial and cultural 
isolation from the bulk of the population.  111.   The British attempted to protect these "aboriginal" 
people by placing areas in which they were concentrated outside of ordinary administration to 
permit a policy of insulating them from exploitative or demoralizing contact with more 
sophisticated outsiders. These enclaves, (called Excluded Areas and Partially Excluded Areas in the 
Government of India Act, 1935) were placed outside the operation of general law and were the 
object of broad executive powers to provide special protective regulation.  112.   Some provinces had 
undertaken a policy of protective 
____________________ 
110. Thus Justice Krishna Iyer in his passionate defense of special protection for Harijans identifies 
them as "this mixed bag of tribes, races, groups, communities and non­castes outside the four­
fold Hindu division." ( State of Kerala v. N. M. Thomas, A.I.R. 1976 S.C. 490 at 535). 
111. In the early 20th century tribals were sometimes included within the Depressed Classes, but by 
1930 they were carefully distinguished. Weitbrecht­Stanton ( 1920) counts 18 million aboriginals 
(excluding Burma). The first serious attempt to identify the "primitive tribes" was in the 1931 
Census, which counted 24, 613, 848 ( Census of India 1931, Vol. I, Pt. 1, p. 503). 
112. Government of India Act, 1935, §91, 92. These enclaves were called "Backward Areas" in the 
Government of India Act, 1919 §52­A (2), and before that Scheduled Districts (see Scheduled 
Districts Act, Act XIV of 1874). For a review of earlier British Policy, see SASTC I: 21­31; 
Ghurye 1963, chap. 4. 
­147­ 
treatment of tribals outside these areas.  113.   In the 1935 Act, for the first time, provision was made 
from some representation for "Backward Tribes" in the reformed Provincial Legislatures,  114.   and a 
list of Backward Tribes was promulgated in 1936 for all of the provinces except Punjab and Bengal. 
 115.   The 1941 census counted 25.4 million tribal people (6.58% of the total population).  116.   
The Constitution carried forward this dual treatment of tribal areas and tribal people in a set of 
provisions for their protection and advancement far more elaborate and detailed than the provisions 
for the other categories of backward classes. Along with reservations in legislatures and services 
comparable to those for the Scheduled Castes, there are provisions for direct central control over 
administration,  117.   for direct central financial responsibility,  118.   for annual reports by the Governors 
to the President,  119.   for extensive executive power over Scheduled Areas to insulate them from the 
application of inappropriate law and to fashion protective regulation (especially regarding allotment 
and transfer of land and money­lending),  120.   and for the formation of Tribal Advisory Councils to 
participate in the formulation of policy.  121.   
The formal mechanism of designating the Scheduled tribes is identical with that for Scheduled 
Castes. The President may specify, after consulting with the governors, the "tribes and tribal 
communities or parts of or groups within tribes or tribal communities" to be the Scheduled tribes 
for each state.  122.   Once promulgated, the list can be varied only by Act of Parliament.  123.   The 
Scheduled Areas are to be ordered by the President, who retains the power to alter or de­schedule 
them.  124.   
____________________ 
113. See. e.g., Bombay 1930 [ Starte Committee]: 3­7. 
114. The quantity was only token: a total of 24 of 1,535 seats in the provincial legislatures were 
reserved. 
115. The Government of India (Provincial Legislative Assemblies) Order, 1936, Thirteenth Schedule. 
Reprinted in Rajagopala Aiyanger 1937. The identity of the list for Bombay with the Starte 
Committee's list suggests that these lists corresponded to already existing state lists, but I have 
been unable to check this for the other provinces. 
116. Census 1943: 101. The total included 19.7 millions in British India (5.65% of the population) 
and 8.7 millions in the states and agencies (9.61% of the population). 
117. Art. 339 (2) (Scheduled Tribes); Fifth Schedule §3 (Scheduled Areas). 
118. Art. 275 (1). 
119. Fifth Schedule, §3. 
120. Fifth Schedule, §4. 
121. Fifth Schedule, §5. The discussion here pertains to the Scheduled Areas outside Assam. The 
Scheduled Areas in Assam are dealt with separately in the Sixth Schedule, which provides for 
autonomous districts and regions with elected councils which enjoy extensive law­making 
powers over land, forests, agriculture, social and political life, administration of justice, 
education, revenue, licensing, and moneylending. 
122. Art. 341 (1). 
123. Art. 342 (2). 
124. Art. 244, Fifth Schedule, §6. 
­148­ 
In 1950, the President promulgated the list of Scheduled Areas  125.   and a list of Scheduled Tribes, 
 126.   apparently by making some additions to the 1935 list of Backward Tribes.  127.   (Some list of the 
ST must have been in use earlier, for the central government's scheme of postmatriculation 
scholarship included ST beginning with the academic year 1948­49.)  128.   Some groups that had 
been listed as Scheduled Castes in Bengal were now moved to the ST list.  129.   The 1951 census. 
counted 19.1 millions in these groups (5.30% of the total population). Some tribes had been 
omitted; about a million members of these "left out" tribes were temporarily accommodated in the 
Other Backward Classes category  130.   until 1956, when the Scheduled Tribes Order was amended, 
 131.   adding 2.5 million tribals, mostly in Rajasthan and Madhya Pradesh, bringing the total to 22.5 
millions (6.93% of the total population). In 1971 these groups numbered over 38 millions (6.93% of 
the total population). In 1976 over 3 million persons were added, bringing the total to almost 40 
millions.  132.   (See table 16.) 
Although the lists contain groups throughout the country, the tribal population is heavily 
concentrated in central, eastern, and northeastern India.  133.   Unlike the relatively dispersed 
Scheduled Castes population, about 90% of the tribals are found in nine states. Three states had less 
than 1% Scheduled Tribes in 1961, while Orissa and Madhya Pradesh had more than 20%. More 
than one­third (10.6 millions) of the Scheduled Tribes population resided in the Scheduled Areas 
(where they formed less than two­thirds of the population). (See table 17.) More 
____________________ 
125. Scheduled Areas (Part A) Order, 1950; Scheduled Areas (Part B) Order, 1950. 
126. Constitution (Scheduled Tribes) Order, 1950; Constitution (Scheduled Tribes) (Part C States) 
Order, 1951. 
127. See Department of Social Security 1965:6. 
128. Ministry of Education and Scientific Research 1957:2. This report suggests that the decision to 
extend the scheme to Scheduled Tribes was made in 1947. 
129. Dushkin 1957:159. 
130. RCSCST 1954:238 ff. 
131. Scheduled Castes and Scheduled Tribes Order (Amendment) Act, 1956 (Act 63 of 1956). 
132. Virtually all the sizable groups with a plausible claim to inclusion in the Scheduled Castes list 
are included. There are no prospects of any major additions to the list. However, there is a 
controversy about a possible further sizable addition to the Scheduled Tribes list. For years 
preceding the 1976 revision of the lists there was an effort to secure the inclusion of some 2 to 4 
million tea­garden laborers of tribal origin in Assam, a project that threatened unwelcome 
political consequences in that state and attracted the intense opposition of the state government ( 
LSD [Fifth Series], Vol. 64, No. 17: Col. 58 and passim [2 Sep. 1976]; Department of Social 
Security 1965: 18 ff.). 
133. For two decades the Government of Uttar Pradesh refused to countenance a Scheduled Tribes 
category in spite of the presence of groups that appeared to have all the requisite characteristics. 
The refusal of U.P. to classify any groups as Scheduled Tribes "even though the conditions of life 
and surroundings of these tribals are almost 
­149­ 
TABLE 16 SCHEDULED TRIBES POPULATION 
1950 & 1951 1956 Order  1961 Census  1971 Census 
Orders 


Source  A  A  A  A  B  B  C 

State  (mils.)  %  (mils.)  %  (mils.)  %  (mils.)  % 


ANDHRA  0.766  2.45  1.149  3.68  1.324  3.68  1.657  3.79 
ASSAM  1.735  19.18  1.761  19.48  2.068  17.42  1.920  12.84 
NAGALAND .343  93.03 .457  88.46 
BIHAR  3.827  9.86  3.880  10.00  4.204  9.02  4.932  8.74 
BOMBAY  3.671  7.60  3.743  7.76 
GUJARAT  2.754  13.34  3.734  13.97 
MAHARASHTRA  2.397  6.03  2.954  5.85 
JAMMU & KASHMIR  No Scheduled Tribes 
KERALA .074  0.54  0.134  0.99  0.207  1.23  0.270  1.26 
MADHYA PRADESH  3.859  14.79  4.844  18.58  6.678  20.62  8.387  20.14 
MADRAS  0.060  0.20  0.136  0.45  0.252  0.74  0.311  0.75 
MYSORE  0.145  0.23  0.080  0.41  0.192  0.81  0.231  0.78 
ORISSA  2.967  20.25  3.009  20.55  4.223  24.02  5.071  23.11 
PUNJAB  0.002  0.01 .002  0.02  0.014  0.06 ...  ... 
HARYANA ...  ...  ... 
RAJASTHAN  0.348  2.18  1.744  11.11  2.309  11.02  3.125  12.11 
UTTAR PRADESH  No Scheduled Tribes .198 .21 
WEST BENGAL  1.386  5.27  1.566  5.96  2.063  5.91  2.532  5.70 
1950 & 1951 1956 Order  1961 Census  1971 Census 
Orders 


Source  A  A  A  A  B  B  C 

State  (mils.)  %  (mils.)  %  (mils.)  %  (mils.)  % 

TOTALS  19.147  5.30  22.511  6.23  29.846  6.80  38.015  6.93 


Sources
A: RCSCST 1956­57 : II, 15; B: RCSCST 1961­62: II, 74; C: Census of India, 1971, 

Series I, Part II­A ( ii): xxx­xxxi. 
Note:  The totals may be slightly higher than the columns indicate, owing to the inclusion of 
Scheduled Tribes in the Union Territories. 
than half of the Scheduled Tribes lived in parliamentary constituencies where they formed a 
majority of the population.  134.   
The Scheduled Tribes were defined partly by habitat and geographic isolation, but even more on, the 
basis of social, religious, linguistic, and cultural distinctiveness­­their "tribal characteristics." just 
where the line between "tribals" and "non­tribals" should be drawn has not always been free from 
doubt.  135.   We shall see that there are still 
____________________ 
identical with tribals in neighboring states" was described as "unusual" and "unsound" by the 
Backward Classes Commission ( I BCC155­56). The matter engaged the Government of India 
for a long time without any action by the state government ( RCSCST 1965­66: 66), but finally in 
June 1967 a notification was issued by the President declaring five communities as Scheduled 
Tribes in U.P. ( RCSCST 196667:27). 
134. See chap. 3, text at notes 28 ff. 
135. This has been entwined with a controversy over the extent to which tribals were "Hindus." For a 
taste of the long­standing academic controversy over this, compare 
­150­ 
TABLE 17 SCHEDULED AREAS POPULATION 
1961 Total 1961 ST 1961 ST in SA 1961 ST in SA
STATE  SA Pop.  Pop. in SA  as % Total SA  as % Total ST 
(mils.)  (mils.)  %  % 
ANDHRA PRADESH . . .  0.438 . . .  33.08 
BIHAR  5.257  2.929  55.71  69.67 
GUJARAT  1.987  1.306  65.60  47.42 
MAHARASHTRA  1.165  0.856  63.26  35.71 
JAMMU & KASHMIR  No Scheduled Areas 
KERALA  No Scheduled Areas 
MADHYA PRADESH  3.395  2.147  63.52  32.15 
MADRAS  No Scheduled Areas 
MYSORE  No Scheduled Areas 
ORISSA  3.950  2.392  60.50  56.64 
1961 Total 1961 ST 1961 ST in SA 1961 ST in SA
STATE  SA Pop.  Pop. in SA  as % Total SA  as % Total ST 
PUNJAB  0.020  0.014  68.90  100.00 
RAJASTHAN  0.877  0.564  64.31  24.42 
UTTAR PRADESH  No Scheduled Tribes 
WEST BENGAL  No Scheduled Areas 

TOTALS  16.659  10.647  61.28a  35.70 


Source: RCSCST 1965­66 : 186. 
Note: a Andhra Pradesh not included in total. 
problems of overlap with the caste order (discussed in chapter 9, §§ C,D, below) and controversy 
about whether a specific group is more approporiately classed as a Scheduled Tribe or a Scheduled 
Caste.  136.   In any event, not all tribal peoples were listed, but only those 
____________________ 
Ghurye 1963: 19 ("the only proper description of these people is that they are the imperfectly 
integrated classes of Hindu society. . . . [T]hey are in reality Backward Hindus") with von Furer­
Haimendorf 1967. On the theoretical complexity of the tribe­caste distinction, see Bailey 1961. 
136. A recent and fascinating instance surfaced in the Lok Sabha debate over the revision of the 
Scheduled Castes and Tribes Order in 1976. Shri Dasaratha Deb proposed that the Laskar 
community in Tripura be treated as a Schedules Caste rather than as a Scheduled Tribe because 
of their lack of affinity with the tribes of Tripura and especially because their treatment as tribals 
has enabled them to play a "dirty game" of using their marginal status to circumvent the 
restrictions on transfer of tribal lands to non­tribals. According to his account, as tribals, tribal 
land could be transferred to them without prior sanction of the government; they then used their 
Bengali surnames to transfer the land to non­tribals without arousing the suspicion of the 
registration authorities. Apparently this group are the descendants of migrants from Bengal who 
were accorded tribal status by the former Maharaja. They are not separately listed in the Tripura 
List as either a Scheduled Caste or a Tribe (nor were they previously). Shri 
­151­ 
whose primitiveness, backwardness, or isolation made them deserving of special treatment.  137.   No 
religious test was specified. A recent attempt to exclude converts failed to gain acceptance.  138.   
In 1951 the Commissioner noted that no uniform test for distinguishing the Scheduled Tribes had 
been formulated.  139.   He circularized the state governments for their views. After surveying the 
divergent responses, which included racial, religious, linguistic, and cultural tests as well as tests of 
political, economic, and social organization, and degree of contact, isolation, or 
assimilation/distinctiveness, the Commissioner proposed as the common elements: tribal origin, 
primitive way of life, remote habitation, and general backwardness in all respects.  140.   
Whatever theoretical imperfections may underlie the lists, nothing takes on fixity so readily as a 
rough approximation made to solve an immediate and intractable problem. A few years later the 
criteria for Scheduled Tribes seemed wholly unproblematic. The Home Minister in 1956 referred to 
the "well­defined and easily ascertained" criteria for distinguishing them.  141.   By the time the 
Dhebar Commission reported in 1961, it felt no need to devote any attention to the problem of 
identifying the tribes.  142.   The Lokur Committee in 1965 reported that, in revising the lists, 
we have looked for indications of primitive traits, distinctive culture, geographic 
isolation, shyness of contact with the community at large and backwardness; we have 
considered that tribes whose members have by and large mixed up with the general 
population are not eligible to be in the list of Scheduled Tribes.  143.   
The provisions for Scheduled Tribes are distinct not only in their scope but in their intent, for 
preservation of their separate integrity, rather than complete assimilation, is one element in policy 
toward 
____________________ 
Dasaratha Deb's Amendment to exclude them from the item for "Tripura" by adding "except 
Laskar termed as Deshi Tripura by ex­ruler's Order" was defeated ( LSD [Fifth Series], Vol. 64, 
No. 17: Cols. 45, 156, 166 [ 2 Sep. 1976]) 
137. Department of Social Security 1965: 47. 
138. See discussion in chap. 9, § E. 
139. RCSCST 1951 : 11. 
140. RCSCST 1951 : 109­11. 
141. Ministry of Home Affairs 1956: 1. Cf. Majumdar ( 1960: 219) 1954 reference to the relative ease 
of their classification by "self­evident and homogeneous tribal organization." 
142. SASTC. 
143. Department of Social Security 1965: 7. The first part of this language is utilized almost verbatim 
by the Home Minister more than 10 years later in the debate about revision of the lists ( LSD 
[Fifth Series], Vol. 64, No. 17: col. 124 [ 2 Sep. 1976]). 
­152­ 
them. In the case of the Scheduled Castes, the Other Backward Classes, and the Denotified Tribes; 
it is hoped that the disabilities and disadvantages that separate them from their compatriots will 
eventually be overcome by the preferential treatment. That is, the aim of the policy is to eliminate 
their distinctiveness by dispelling all of the differences that set them apart from other Indians. The 
aim of Scheduled Tribes' policy is more complex­­to balance improvement of their condition and a 
degree of assimilation with preservation of their distinctiveness and a measure of autonomy.  144.   
Special treatment of the Tribes will, it is hoped, foster development "along their own lines." Because 
of this difference in purpose (and because of their geographical separation), the objections that have 
been made against the use of communal criteria in other categories of backward classes have not 
been made with the same force in regard to the Scheduled Tribes. 
Although Scheduled Tribes share with Scheduled Castes a variety of reservations­­in legislatures, 
government service, and educational institutions­­the principal schemes for the Tribes have been 
protective and developmental (economic, educational, health, and communications). Tribal aspirants 
have not pressed hard on the reservations; many have gone unfilled. With the exception of a few 
election cases raising problems of tribal membership,  145.   tribes have been little involved in 
litigation about compensatory discrimination. They have had little effect on the law, which is being 
developed in cases involving the Scheduled Castes and the Other Backward Classes. 
____________________ 
144. Just what the mix of assimilation and distinctiveness should be has long been a matter of 
dispute. Cf., e.g., the Dhebar Report [SASTC] and Ghurye ( 1963) critique. 
145. Discussed below in chap. 9, §C, D. 
­153­ 
6 The Other Backward Classes: The Process of Designation 
A. THE PRE­HISTORY OF THE BACKWARD CLASSES CATEGORY 

THE OTHER BACKWARD Classes for whom preferential treatment are authorized are not defined 
in the Constitution, nor is any exclusive method or agency for their designation provided. For all the 
uncertainty surrounding the term "Scheduled Castes" (and its predecessor "Depressed Classes"), its 
central purpose is clear: to identify the victims of "untouchability." At the time of Independence, the 
term "Backward Classes" had a less fixed and definite reference. The term had been around for 
1. but it had a variety of referents, it had shifted rapidly in meaning, and it had come to 
some time,    
mean different things in different places. For purposes of tracing these varied meanings in the 
following discussion, we may portray some of the variety of meanings in schematic form in figure 
1. 
In 1917 the Maharaja of Kolhapur recounted to Mr. Montagu that he had "been taking very keen 
2. [Denotation 3?]. 
interest . . . in uplifting the Backward Classes and especially the untouchables"    
Soon after, the term appears in the terms of reference of the Southborough Committee, which was 
to advise on measures to secure representation of "minorities, of special interests, or of backward 
classes."  3.   But the Committee did not mention any such groups in its report, other than Depressed 
Classes [Denotation 1?]. The Joint Select Committee of the British Parliament which reviewed the 
Southborough Report mentioned in passing the importance they attached to "the educational 
advance­ 
____________________ 
1. The Fort St. George Gazette No. 40 of November 5, 1895, mentions grants in aid to schools for a 
list of "Backward Classes" which includes most of the "untouchable" castes of Madras 
Presidency. Irschick ( 1977) notes that the term was used as far back as 1880 to describe a list of 
groups, also called illiterate or indigent classes, entitled to allowances for study in elementary 
schools. See Saraswathi 1974: 107 ff. 
2. Latthe 1924: II, 574. 
3. Indian Constitutional Reforms 1919: 1. 
­154­ 
FIGURE 1 

VARIOUS DENOTATIONS OF THE "BACKWARD CLASSES" 1. As a synonym for Depressed 
Classes, untouchables, Scheduled Castes. 2. As comprising the untouchable, aboriginal and hill 
tribes, criminal tribes, etc. 3. As comprising all the communities deserving special treatment, 
namely those included in (2) above and in addition the lower strata of non­touchable communities. 
4. As comprising all non­tribal communities deserving special treatment. 5. As comprising all 
communities deserving special treatment except the untouchables. 6. As comprising the lower strata 
of non­untouchable communities. 7. As comprising all communities above the untouchables but 
below the most "advanced" communities. 8. As comprising the non­touchable communities who 
were "backward" in comparison to the highest castes. 9. As comprising all communities other than 
the highest or most advanced. 10. As comprising all persons who meet given non­communal tests of 
backwardness (e.g., low income). Note: this figure portrays varying usages, of which examples are 
given in the text. Exact specifications may differ along other dimensions as well­­e.g., whether non­
Hindus are included or whether prosperous families within these groups are excluded. 

­155­ 
4. There is no indication that the term was meant to 
ment of the depressed and backward classes."    
include anyone besides the depressed classes [Denotation 1? 2?]. 
"Backward Classes" first acquired a technical meaning in the Princely State of Mysore. In 1918 the 
Mysore Government appointed a committee to enquire into the question of encouraging members of 
5. In 1921 preferential recruitment of "backward 
the "backward communities" in public service.    
communities" was instituted, and they were defined as "all communities other than Brahmins, who 
6. [Denotation 9]. 
are not now adequately represented in the public service."    
The Reforms Enquiry Commission ( 1924) did not find occasion to use the term, but the Ministry 
Report refers to its use as a synonym for the Depressed Classes (untouchables) and in 
7. [Denotation 1]. 
contradistinction to "non­Brahmins."    
Although "Non­Brahmin" and "Maratha" were much more frequently used in the setting of the 
Bombay "non­Brahmin" movement of the 1920s, the Backward Classes rubric was employed both 
8. In 1925 a 
popularly and officially in a broad meaning, somewhat akin to that in Mysore.    
Government Resolution defined Backward Classes as all except Brahmins, Prabhus, Marwaris, 
9. [Denotation 9?]. Reservations in government service were provided 
Parsis, Banias, and Christians    
for this group. 
The Hartog Committee ( 1928) defined Backward Classes in their glossary: 
Castes or classes which are educationally backward. They include the depressed classes, 
aboriginals, hill tribes, and criminal tribes.  10.   [Denotation 2] 
In 1929 the Indian Central Committee distinguished the problem of the "'backward classes,' among 
whom may be counted aboriginals, criminal tribes and others among the less advanced of the 
inhabitants of British India"  11.   [Denotation 5]. Apparently the Committee not only excludes the 
Depressed Classes but includes, in addition to tribals, some strata of the caste population, for they 
mention an estimate of 16 
____________________ 
10. Indian Statutory Commission (Hartog Committee) 1929: 399. 
11. Indian Central Committee 1929: 45. 
4. Reprinted in P. Mukerji 1920: 528. 
5. A 1921 G.O. referring to the 1911 G.O. is reprinted in Mysore Backward Classes Committee 
1961: 57­61. 
6. Id., at 58. An attached education memorandum employs the term "backward classes" (id., at 61), 
but it is not clear that it was exactly synonymous. 
7. Reforms Enquiry Committee 1925: 147. 
8. Omvedt 1976: 184, 186, 188. 
9. Government of Bombay Finance Department Resolution No. 2610 of Feb. 5, 1925, cited by 
Omvedt 1976: 343. 
­156­ 
million backward classes in U.P. (excluding the untouchables).  12.   But the category is not used as 
expansively as the Mysore or earlier Bombay usage, for it did not include the Madras non­Brahmins 
or Mahrattas in Bombay, who were discussed separately.  13.   The term could not have been a familiar 
one, for the Report finds it necessary to distinguish Backward Classes from untouchables several 
times. In a separate note, Mr. M. C. Rajah, an untouchable spokesman, mentions Backward Classes 
only as a confusing synonym for Depressed Classes  14.   [Denotation 1]. 
In 1930 the Starte Committee in Bombay devoted careful consideration to the question of 
nomenclature. They noted that in 1924 the term Depressed Classes had been accorded a wider 
meaning to include 
aboriginal tribes and the Criminal Tribes and some other wandering and backward 
Castes . . . [which had] resulted in much confusion of thought in this Presidency, as in 
ordinary usage the phrase Depressed Classes is taken as meaning the untouchables . . . 
whereas they do not form half of this new and enlarged grouping of Depressed Classes. 
 15.   
The Committee recommended that "Depressed Classes" should be used in the sense of 
untouchables, a usage which "will coincide with existing common practice."  16.   They proposed that 
the wider group should be called "Backward Classes" [Denotation 3], which should be subdivided 
into Depressed Classes (i.e., untouchables); Aboriginals and Hill Tribes; Other Backward Classes 
(including wandering tribes). They note that the groups then currently called Backward Classes 
[Denotation 8] should be renamed "intermediate classes."  17.   In addition to 36 Depressed Classes 
(approximate 1921 population 1.475 millions) and 24 Aboriginal and Hill Tribes (approximate 1921 
population 1.323 millions), they list 95 Other Backward Classes (approximate 1921 population 1.041 
millions)  18.   [Denotation 6]. 
The Simon Commission, though it refers to "intermediate castes" and takes note of the non­Brahmin 
movement, makes no mention of Backward Classes.  19.   But shortly after, in the early 1930s, we find 
a flowering of the term. In the hearings held by the Indian Franchise Committee in the United 
Provinces, the term recurs in the testimony in a number of significations. Thus a note by Dr. S. S. 
Nehru, I.C.S., presents a list of Depressed Classes of which portions are designated 
____________________ 
12. Id. 
13. Id., at 47. 
14. Id., at 365. 
15. Bombay 1930 [ Starte Committee]: 8. 
16. Id., at 8. 
17. Id., at 9. 
18. Id., appendix II. The total Backward Classes population was 3.840 millions, or 14.4% of the 1921 
population of the presidency. 
19. Indian Statutory Commission 1930: I35­37. 
­157­ 
"Vagrant Tribes" and "Backward Classes"  20.   [Denotation 6]. Presumably these are the groups who 
least fit the "untouchability" criterion. However, the memorandum of Pandit Sheokaran Nath Misra 
(Deputy Collector of Fyzabad) suggests that "depressed classes should include untouchables as well 
as backward classes." He equated Depressed Classes roughly with all Sudras excepting Kayasthas 21. 
[Denotation 4]. The United Provinces Hindu Backward Classes League (founded in 1929) submitted 
a memorandum which suggested that the term "Depressed" carried a connotation "of untouchability 
in the sense of causing pollution by touch as in the case of Madras and Bombay" and that many 
communities were reluctant to identify themselves as depressed. The League suggested the term 
"'Hindu' 'Backward'" as a more suitable nomenclature.  22.   The list of 115 castes submitted included 
all candidates for the untouchable category as well as a stratum above [Denotation 4]. "All of the 
[listed] communities belong to nonDwijas or degenerate or Sudra classes of the Hindus." They were 
described as low socially, educationally, and economically and were said to number over 60% of the 
population.  23.   
This inclusive usage was adopted elsewhere. Travancore in 1937 abandoned the Depressed Classes 
nomenclature and substituted the term "Backward Communities" to include all educationally and 
economically backward communities  24.   [Denotation 3]. However, in Madras and elsewhere the 
term "Backward Classes" was used to refer to the strata above the untouchables.  25.   The Madras 
Provincial Backward Classes League, consisting of the less forward non­Brahmin communities, was 
founded in 1934 for the purpose of securing separate treatment from "the forward non­Brahmin 
communities." According 
____________________ 
20. Indian Franchise Committee 1932: IV, 734. 
21. Id., 738. 
22. Id., 826. In an interview in 1966, Mr. S.D. Singh Chaurasia of Lucknow, who was associated 
with the League and subsequently became a leading spokesman of Backward Classes and served 
on the Backward Classes Commission, recalled that he coined the term in 1930 or 1931. Since 
1974 Mr. Singh Chaurasia has been a member of the Rajya Sabha. 
23. Id., 826. 
24. Anonymous 1935­37: 364. The official announcement was that communities not then classified 
as depressed would be included and that advanced groups among the depressed would be 
reclassified. For a similar inclusive usage, note the announcement in December 1936 that the 
Government of the Central Provinces constituted a committee to promote the education of the 
Backward Classes. Anonymous 1935­37:273. 
25. A writer in the Times of India ( July 1936) claimed that the estimates of Depressed Classes 
population were highly inflated by the inclusion of "backward class Hindus, such as dhobis, 
barbers, potters and others who are not untouchables." See response of S. R. Venkataraman, 
reprinted in Anonymous 1935­37: 161. Cf. the March 1937 manifesto on the Christian Duty to 
Depressed Classes and Backward Classes, reported in Anonymous 1935­37: 354. 
­158­ 
to the list they presented to the Madras Government in 1944, the Backward Classes comprised more 
than a hundred communities with about 25 million people (more than two­thirds of the non­
Brahmin category and 50% of the total population of the Presidency)  26.   [Denotation 6]. In 
November 1947, separate reservations in the Madras services were provided for these "Backward 
Hindus."  27.   
In the meantime, the Objectives Resolution of the Constituent Assembly, moved by Jawaharlal 
Nehru on December 13, 1946, had resolved to provide adequate safeguards for "minorities, 
backward and tribal areas, and depressed and other backward classes"  28.   [Denotation 5? 6?]. 
Thus, the term had never acquired a definite meaning at the all­ India level. There had been no 
attempt to define it or employ it on the national level, and there were no nationwide backward 
classes organizations or spokesmen. The term had definite meanings in local contexts, although 
these differed somewhat. After the listing of Scheduled Castes, the usage as a synonym for 
untouchables [Denotation 1] drops away. Two major species of usage emerge: (1) as the more 
inclusive group of all those who need special treatment [Denotations 3, 9]; (2) as a stratum higher 
than the untouchables but nonetheless depressed [Denotations 6, 7]. This double usage continues 
today: the former in the usage of backward classes in the wide sense (including Scheduled Castes 
and Scheduled Tribes); the latter in the usage as equivalent to "Other Backward Classes." It is with 
the latter that we are now concerned here. 
B. THE CONSTITUTIONAL DESIGN 
By the time of the Constituent Assembly the usage of the term "Backward Classes" to refer to some 
larger or smaller portion of the population deserving of special treatment was familiar in many parts 
of the country. In the Assembly, delegates from the north expressed puzzlement at the provisions for 
"Backward Classes," (in what came to be Article 16 (4). It struck them as vague, and some thought 
it was meant merely as a synonym for the Scheduled Castes,  29.   while others were 
____________________ 
26. G.O. 190 Public Department (Services) (Jan. 20, 1944). Saraswathi 1974: 121­22. 
27. G.O. 3437 Public Department (Services) (Nov. 21. 1947). Saraswati 1974: 124. This Backward 
Hindus list at this point included 145 communities, who formed nearly one­half the population of 
the Presidency. 
28. Quoted in Banerjee 1948: I, 313. 
29. Pandit Hirday Nath Kunzru (U.P.) at VII CAD 680; Shri Ari Bahadur Gurung (West Bengal) at 
VII CAD 685; Shri R.M. Nalavade (Bombay) at VII CAD 686. Dr. Dharam Prakash (U.P.) 
proposed that Depressed Classes or Scheduled Castes should be used in place of Backward 
Classes, since the former terms had a definite meaning (VII CAD 687). 
­159­ 
concerned that it might mean more.  30.   But representatives from Madras, Mysore, and Bombay 
assured their colleagues that Backward Classes was a distinct term with a technical meaning. 
Examples were given of Mysore where Backward Classes included all but Brahmins,  31.   of Madras 
where it referred to a stratum of non­untouchable Hindu castes,  32.   and of Bombay where it included 
not only Scheduled Castes and Tribes but others who are economically, educationally, and socially 
backward.  33.   A representative from Bihar (which had the most active Backward Classes movement 
in the north) explained that Backward Classes were a section of society between the highest castes 
and the Scheduled Castes: "the third [section] occupying the middle position . . . and consisting of a 
large portion of our people is what may be termed the Backward Class. . . . No doubt they are not 
treated as untouchables."  34.   
Mr. K. M. Munshi assured the house that the term would include the Scheduled Castes, when 
doubts were expressed, and explained that it was "necessary to find a generic term." "It signifies 
people­­touchable or untouchable, belonging to this community or that, who are so backward that 
special protection is required in the services."  35.   
Although one speaker pointed out that classes were not necessarily castes and that literacy might be 
the test of backwardness,  36.   it was generally envisioned that the backward classes would be 
communities. When asked what is a backward community, Dr. Ambedkar, defending the draft, 
explained: "we have left it to be determined by each local Government. A backward community is a 
community which is backward in the opinion of the Government."  37.   However, the matter was not 
as clear as this suggests. The most prophetic statement was Mr. T. T. Krishnamachari's prediction 
that the clause would be "a paradise for lawyers."  38.   
It was anticipated, then, that the backward classes other than the 
____________________ 
30. Shri H. J. Khandekar (C.P. & Berar) preferred the use of Scheduled Castes lest "the people of 
other castes would claim to be backward . . . ." (VII CAD 692.) Sardar Hukum Singh (E.P.) 
expressed uncertainty as to whether the term covered the Scheduled Castes and religious 
minorities (VII CAD 694). 
31. Shri T. Channiah (Mysore) at VII PAD 689­90. 
32. Mhd. Ismail Sahib (Madras) found the Madras usage objectionable in that it did not extend to 
"the backward classes of minority communities" (VII CAD 692). 
33. Dr. K. M. Munshi at VII CAD 697. 
34. Shri Chandrika Ram (Bihar) VII CAD 687. 
35. VII CAD 697. 
36. T. T. Krishnamachari at VII CAD 699. 
37. VII CAD 702. In the light of the later analysis and the text itself, this seems rather misleading, 
for the "opinion" of the government is determinative only of the group's "under­representation," 
not of its backwardness. See chap. 8, §D, below. 
38. VII CAD 699. 
­160­ 
Scheduled Castes and Tribes were to be designated at the local level. The delegation to local 
authorities undoubtedly reflected an acceptance (at least temporarily) of the divergence of existing 
practices, a desire to preserve flexibility, and an awareness of the difficulties of prescribing 
universally applicable tests of backwardness in view of the varying local conditions. It may also 
have been presumed that "Backward Classes" were sufficiently potent politically to look out for 
their own interests on a local level, and unlike Scheduled Castes and Tribes, central control of their 
designation was not required to insure the inclusion of the deserving. However, the central 
government was not entirely excluded from the process. The President is instructed to appoint a 
Backward Classes Commission 
to investigate the conditions of socially and educationally backward classes . . . and the 
difficulties under which they labour and to make recommendations as to the steps that 
should be taken by the Union or any State to remove such difficulties and to improve 
their condition . . . .  39.   
No special central machinery for supervision of programs for backward classes is provided in the 
Constitution, but Article 338 (3) provides that the Commissioner of Scheduled Castes and Tribes 
shall include in his duties such other groups as the President may specify on receipt of the report of 
the Backward Classes Commission. It is not clear from the text that this presidential specification 
was meant to be exclusive. 
Even before the Constitution came into force there was a great expansion in the employment of the 
Backward Classes category. Several states created such a category for the first time,  40.   and of those 
who 
____________________ 
39. Art. 340. While the Constituent Assembly was at work, A.V. Thakkar, a member of that body 
and one of India's leading social workers, explained that the Draft visualized a commission "to 
go through the whole of the country . . . and find out which castes of Hindus, Muslims, etc., are 
really backward according to certain standards, educational, social, economic etc." ( Hindustan 
Times, Dec. 26, 1948, reprinted in Jagadisan and Shyamlal 1949: 338). That what was 
contemplated was reservation for backward "communities" is clear from the recommendations of 
the 1948­49 University Education Commission 1950­51: I, 53 that "the needs of justice to the 
members of the Scheduled Castes and the communities declared to be backward by the 
government . . . can be met by reserving a certain proportion of the seats . . . for qualified 
students of these communities. . . ." 
40. E.g., the Bihar Government made provisions in 1947 for Other Backward Classes in post­
matriculation studies and, on the basis of the Ministry of Education's list, propounded its list of 
Backward Classes in 1951. (Interview with Deva Charan Singh, one of the founders of the All­
India Backward Classes Federation, and later Chairman of the Bihar Legislative Council, on 30 
March 1966.) It was estimated that these groups made up 60% of the state's population. 
Educational concessions for Other Backward Classes began in U.P. in 1948. The list of 56 castes 
was estimated to make up 65% of the population (interview with Mr. S. D. Singh Chaurasia on 
17 March 1966). 
­161­ 
already had it, several expanded the benefits conferred upon the Backward Classes. The central 
government was pressed to extend its scheme of post­matriculation scholarships to the Other 
Backward Classes,  41.   and when it did so, it compiled lists of Backward Class in each state for this 
purpose.  42.   Backward Classes organizations emerged;  43.   a national federation was formed.  44.   
Apart from the central scholarships, provisions for Other Backward Classes roughly followed a 
regional pattern as indicated by table 18. In South India and in Bombay, provisions for Other 
Backward Classes flourished. In the north, there were mostly a scatter of educational concessions, 
and many states did not have any lists at all (other than those compiled for purposes of the central 
post­matriculation scholarship scheme). 
____________________ 
41. Deshmukh (see n. 44 below) reports that his "first concrete attempt" took the form of personal 
letters to Prime Minister Nehru and Maulana Azad in March, 1948, "requesting that some 
amount should be set apart for the award of scholarship to the Other Backward Classes also as 
was done for the Scheduled Castes and Scheduled Tribes. This request . . . was accepted in 
principle by Government and the Other Backward Classes were added to the list for award of 
scholarships for the first time in 1948­49. The State Governments were then asked to prepare 
schedules of these castes. . . . As soon as the Central Government took this decision . . . I was 
nominated as a member of the Scheduled Castes and Scheduled Tribes' Scholarships Board . . . ." 
( AllIndia Backward Classes Federation 1955: 2). 

Post­matriculation scholarship were provided for Other Backward Classes beginning in 1949­50, 
with an expenditure of 2.46 lakhs, less than half of that for the Scheduled Castes. By 1953­54, 
the expenditure had increased tenfold to 26.51 lakhs, approximately equal to that for Scheduled 
Castes ( RCSCST 1954 : 366). 

42. The central government's lists may be found at RCSCST 1951: 163 ff. Although I have not made 
an exhaustive comparison, it seems that these lists follow very closely the lists of those states 
which then had Backward Classes lists. Thus, for example, the Mysore list includes not only 
Muslims, Christians, and Jains, but all Hindu groups other than Brahmins. The Madras and 
Bombay lists, however, are more restricted and do not include the higher non­Brahmin castes. 
43. The Bihar State Backward Classes Federation was founded in 1947 (interview with Deva Charan 
Singh, as above in n. 40). 
44. Majumdar ( 1955 : 474­75) in 1954 counted 88 organizations working for the Backward Classes 
in 15 states, of which 74 represented individual communities, and 14, Backward Classes in 
general on a local or state basis. To some extent, at least, the national Backward Classes 
movement represented a coming together of remnants of the non­Brahmin movements of the 
South and Bombay with the less successful Backward Class movement of the north (especially 
U.P. and Bihar). Their relative success is reflected in the disparate benefits afforded by the states 
in table 18. The chief spokesman for the Other Backward Classes at the national level was Dr. 
Punjabrao Deshmukh of Madhya Pradesh ( Minister of Agriculture from 1952 to 1957 and 1958 
to 1952). He reported that when he came into the Constituent Assembly he "mustered courage to 
take up the problem [of Backward Classes] and place it on an all­India plane. . . ." The All­
Indian Backward Classes Federation came into existence on January 26, 1950, the day that the 
Constitution came into force ( All­India Backward Classes Federation 1955: 2). 
­162­ 
TABLE 18 CONCESSSIONS FOR OTHER BACKWARD CLASSES, 1951­1952 
Education: 
fee 
concessions, 
stipends,  Reser­
training,  vations in  Recruitment  Expenditures, 
facilities,  Educational  Other  to Govern­ 1951­52  Expenditures, 
STATE  etc.  Insts.  Welfare  ment Posts  (lakhs)  1953­54 
ANDHRA  ­  ­  ­  ­  ­  4.36 
Education: 
fee 
concessions, 
stipends,  Reser­
training,  vations in  Recruitment  Expenditures, 
facilities,  Educational  Other  to Govern­ 1951­52  Expenditures, 
STATE  etc.  Insts.  Welfare  ment Posts  (lakhs)  1953­54 
ASSAM  X  O  O  O  18.04  2.49 
BIHAR  X  O  O  O  6.22  10.26 
BOMBAY  X  X  X  X  24.95  50.15 
MADHYA 
O  O  O  O  0  0.77 
PRADESH 
MADRAS  X  X  X  X  13.04  41.25 
ORISSA  O  O  O  O  0  0 
PUNJAB  O  O  O  O  0  0 
UTTAR 
X  O  O  O  4.75  6.29 
PRADESH 
WEST BENGAL  X  O  O  O ...  5.86 
HYDERABAD  O  O  X  O  36.27  1.70 
MADHYA 

BHARAT 
MYSHORE  X ?  X  X ?  9.14 
P.E.P.S.U.  O  O  O  O  0  0 
RAJASTHAN  X  O  O  O  0.29 ... 
SAURASHTRA  O  O  O  Xa  0  6.44 
TRAVANCORE­ 
COCHIN  X ?  ?  X ?  8.71 
AJMER  X  O  O  O  0  0 
BHOPAL  O  O  O  O  0.05  0.05 
BILASPUR  O  O  O  O ?  ... 
COORG  O  X  O  O ...  0.06 
DELHI  X  O  O  O ...  ... 
HIMACHAL 
O  O  O  O  3.47 ... 
PRADESH 
KUTCH  O  O  O  O ...  3.7 
MANIPUR  O  O  O  O ...  ... 
TRIPURA  0 ... 
VINDHYA 
O  O  O  O ...  ... 
PRADESH 
CENTRAL 
GOVERNMENT  X  O  X?  O 
Source: This table is a complete of information found in RCSCST 1952:202­5 with introductory 
chapters X, XI. The 1953­54 expenditures are taken from I BCC146. 
Education: 
fee 
concessions, 
stipends,  Reser­
training,  vations in  Recruitment  Expenditures, 
facilities,  Educational  Other  to Govern­ 1951­52  Expenditures, 
STATE  etc.  Insts.  Welfare  ment Posts  (lakhs)  1953­54 
Note: aOBC lumped together with SC and ST. 
­163­ 
It remained unclear just how many people were included in the Backward Classes category.  45.   The 
never­completed 1951 census enumeration, based in part on the existing state lists, counted 67 
millions (18.9% of the total population).  46.   The Planning Commission in July 1951 estimated that 
the Other Backward Classes were approximately 20% of the population,  47.   but in late 1952 noted 
that "these other backward classes . . . are believed to number about [54.6 millions]."  48.   Other 
observers predicted a smaller list,  49.   but the President in his address to the inaugural meeting of the 
Backward Classes Commission is reported to have mentioned the figure of 70 millions.  50.   
A pair of Supreme Court decisions in April 1951 shattered the legal foundation of the system of 
communal quotas which prevailed in South India. State of Madras v. Champakam Dorairajan struck 
down the Madras reservations in educational institutions and by implication barred all preferential 
treatment outside the area of government employment.  51.   
____________________ 
45. The Backward Classes category was found to provide a convenient administrative depository for 
groups which were felt to deserve some special treatment, but which were not included or 
includable under other headings. Two and a half million "left out" tribal peoples, who had 
mistakenly been omitted from the list of Scheduled Tribes were included until the Schedule was 
amended in 1956 ( RCSCST 1953: 196). The denotified tribes (formerly Criminal Tribes) were at 
first listed separately for budget purposes but were later assimilated to the Other Backward 
Classes for budgetary purposes. Backward Classes lists typically included converts from 
Scheduled Castes to non­Hindu religions, and several states used this category to provide some 
concessions to sections of their Muslim population. This is especially true of Bihar and 
Hyderabad. (See the lists in RCSCST 1951.) 
46. The 1951 Census, unlike previous Censuses, had decided not to collect caste data, except for 
"Special Groups," including not only Scheduled Castes and Tribes, but Backward Classes 
specified by the state governments. However, the statistics on Backward Classes were not fully 
tabulated and were never published. The Census authorities provided the Backward Classes 
Commission with two figures for "Backward Classes: their enumerated population" (67 million, 
or 18.9% of the total population) and their estimated 1951 population of each caste (73 million, 
or 20.5% of the total population) (III BCC9). The similarity of totals conceals great 
discrepancies at the state level ((gee table 19 ). Although Professor Shah (III BCC9) indicates 
that the census totals are based on the existing state lists, there seem to be some cases in which 
this is not so­­e.g., Mysore, where the census enumeration is 19.3% of the population, while the 
state list was closer to the Commission's figure of 65%. 
47. Planning Commission 1951: 231. 
48. Planning Commission 1953: 243. 
49. E.g., a writer in the Times of India anticipated that the Backward Classes Commission, soon to 
be formed, was expected to enumerate about 20 million in the Backward Classes­­that is, less 
than one­third of the census total ( Times of India, December 3, 1952). 
50. Dr. Deshmukh at All­India Backward Classes Federation 1954: 5. 
51. A.I.R. 1951 S.C. 226. Both judgments were handed down on April 9, 1951; but they could not 
have been wholly unanticipated, for the Madras High Court decided similarly in the Champakam 
case on July 27, 1950 (A.I.R. 1951. Mad. 120). A select 
­164­ 
Venkataramana v. State of Madras struck down the Madras quotas in government posts for all 
groups other than the Scheduled Castes and "Backward Hindus," confining potential recipients to 
those who could qualify as the "backward classes" mentioned in Article 16(4).  52.   These decisions 
caused a political furore in South India and occasioned the prompt addition of Article 15(4) to the 
Constitution.  53.   
The debate over the amendment centered on the desirability of providing educational preferences to 
the Backward Classes, and it revolved in part around the question of who were the backward classes. 
 54.   Although the Scheduled Castes and Tribes were equally discomfited by Champakam Dorairajan 
and presumably equally concerned with educational preferences, there was little mention of them in 
Parliament. The question worrying the house was the identity of Backward Classes and the possible 
abuse of the new provision by advanced groups. 
In an attempt to confine the possible meaning of Backward Classes, the wording of Article 15 (4) 
was keyed to that of Article 340,  55.   which provided that the not yet established Backward Classes 
Commission should list the "socially and educationally backward classes of citizens." Some 
speakers drew assurance from the notion that the listing by 
____________________ 
Committee to report on proposed amendments was established on 16 May 1951. With the 
exceptions noted below, the debate on the Amendment took place between 29 May 1951 and 2 
June 1951. 
52. A.I.R. 1951 S.C. 229. 
53. In the debate on the amendment, Prime Minister Nehru remarked: "The House knows very well 
and there is no need for trying to hush it up, that this particular matter in this particular shape 
arose because of certain happenings in Madras" (Parliamentary Debates, Vol. XII­XIII (Part II), 
col. 9615). While others concurred in locating the thrust for the amendment in Madras, one 
member pointed out: "it is not only the Madras Government that is concerned with this but the 
whole of South India­­the States of Mysore, Travancore­Cochin and even Bombay" ( Shri 
Shankaraiya, id., at col. 9000, 17 May 1951). Dr. Deshmukh saw the regional origin as a question 
of time­lag: the problem was not confined to Madras, but was bound to arise elsewhere as soon 
as the Backward Classes became more aware and assertive (id., at 9775). 
54. The bill was referred to a select committee after some discussion on May 16, 1951. Further 
debate on Article 15(4) took place on May 18, 29, 30, 31, and June 1 and 2. The clause inserting 
Art. 15(4) was passed on June 1; the entire bill on June 2. This amendment was one of the three 
major changes made by the Constitution (First Amendment) Act, 1951. The amendment both of 
Art. 19 (freedom of speech) and of Art. 31 (acquisition of property) received far more of 
Parliament's time and attention. Of the three major provisions, Art. 15(4) encountered the least 
opposition. The clause was passed 243 to 5. (The Art. 19 amendment passed 228 to 19; Article 
31, 239 to 7.) 
55. The original draft of Art. 15(4) would have added to Art. 15(3), which authorized special 
provision for women and children, the words "or for the educational, economic or social 
advancement of any backward class of citizens" (id., at 8929, 17 May 1951). Prime Minister 
Nehru explained that the Select Committee chose the final wording "because they [socially and 
educationally] occur in Article 340 and we wanted to bring them bodily from there" (id., at 
9830). 
­165­ 
the Commission and subsequent presidential specification would be determinative.  56.   Others 
assumed that the identity of the Backward Classes would remain a question for the state 
governments and counselled trust in their good faith.  57.   
Examination of the debate leaves it abundantly clear that the Backward Classes, by whomever 
designated and according to whatever tests they were chosen, were expected to be a list of castes or 
communities. Dr. Ambedkar, then Law Minister, forthrightly observed that the amendment was 
needed precisely because "what are called backward classes are . . . nothing else but a collection of 
certain castes."  58.   There was considerable concern that the provision should not permit communal 
quotas to be enjoyed by more advanced groups. While there was discussion of the economic 
backwardness of the groups who deserved preferences, it was not merely the poor that the drafters 
and speakers had in mind.  59.   (Indeed, if they had, an amendment would hardly have been 
necessary.) Some speakers argued that preferences should be directed to the economically poor, but 
the predominant concern was to provide some special treatment to offset and remedy specifically 
those social inequalities of caste and community which were seen as underlying and compounding 
economic differences.  60.   
____________________ 
56. While some members liked the final phrasing because they thought it confined Backward 
Classes to those to be specified by the President under Art. 340 ( Pandit Thakur Das Bhargava , 
id., at 9719; Shri M. A. Ayyangar, id., at 9817), others objected that they were not so confined 
( Sardar Hukum Singh, id., at 9823; Dr. S. P. Mookerjee, id., at 9824). An amendment to make 
explicit this limitation to the groups specified under Art. 340 was not accepted by the 
government and was defeated by the House. (id., at 9832­33). 
57. Seth Govind Das, id., at 10051; Shri Venkataraman, id., at 10081. 
58. Id., at 9006 ( 18 May 1951). 
59. Prime Mininister Nehru observed that "we have to deal with the situation where for a variety of 
causes for which the present generation is not to blame, the past has the responsibility, there are 
groups, classes, individuals, communities, if you like, who are backward. They are backward in 
many ways­­economically, socially, educationally­sometimes they are not backward in one of 
these respects and yet backward in another. The fact is therefore that if we wish to encourage 
them in regard to these matters, we have to do something special for them" (id., 9616). In spite of 
his reluctance to talk about caste, it is clear that what was intended were not measures to erase all 
inequalities, but specifically those which were associated with traditional social structure. "[W]e 
want to put an end to . . . all those infinite divisions that have grown up in our social life . . . we 
may call them by any name you like, the caste system or religious divisions, etc. There are of 
course economic divisions but we realize them and we try to deal with them. . . . But in the 
structure that has grown up . . . with its vast number of fissures or divisions. . . ." 

The Prime Minister's extraordinary reticence about using the word "caste" was not shared by all 
of his fellow members. Cf. the more straightforward remarks of Reverend d'Souza, id., at 
9689­90; Dr. Deshmukh, id., at 9775­76. 

60. Thus, Professor K. T. Shah, the strongest advocate of an individualized 
­166­ 
The Venkataramana case had indicated that the "Backward Classes" included only those caste 
groups of whose backwardness there was some assurance. In spite of some unflattering observations 
about that judgment, there was nothing in the proceedings to indicate that Parliament sought to 
overturn it.  61.   Indeed, it was seen as the foundation for judicial protection against unwarranted use 
of the power to confer preferences.  62.   Upon passage of the amendment, Madras revised its former 
system of communal quotas to accord with the new requirements; the situation in other states 
remained unchanged. There was no further litigation over the Backward Classes for eight years.  63.   
C. THE FAILURE TO CENTRALIZE POLICY 
The Backward Classes Commission was established in 1953 and directed to 
determine the criteria to be adopted in considering whether any sections of the people . . 
. (in addition to the Scheduled Castes and Tribes . . .) should be treated as socially and 
educationally backward classes; and, in accordance with such criteria to prepare a list of 
such classes. . . .  64.   
____________________ 
approach, avowed that the backwardness to be remedied was economic (id., 9821) and proposed 
to eliminate the word "classes" and to add "economically" to the qualifiers of the term "backward 
classes" (id., 9815). Prime Minister Nehru, explaining his unwillingness to accept any of the 
amendments, indicated that he had no objection to adding "economically" but that to do so would 
put the language at variance with that of Art. 340. He then observed: "But if I added 
'economically' I would at the same time not make it a kind of cumulative thing but would say that 
a person who is lacking in any of these things should be helped. 'Socially' is a much wider word 
including many things and certainly including economically" (id., 9830). 
61. Although Dr. Ambedkar attributes the necessity for the amendment to both judgments 
collectively, it is difficult to discern that he has a specific objection to the Venkataramana case 
(id.,9006, 18 May 1951). Shri Venkatraman ( Madras) closed the debate with the observation that 
the present amendment puts the position in regard to education on a par with that in government 
service and expressed satisfaction with the Supreme Court's holding in that area (id., at 10081). 
62. Shri Venkataraman assured the house that there was no need for ministerial assurances, for "there 
is a decision of the Supreme court . . . You have adequate protection, even as the law now stands . 
. . to prevent any abuse of this clause" (id., 10081). 
63. Madras listed 155 communties as the Backward Classes. See e.g., Government of Madras, Public 
(Service) Department G.O. No. 2687 (dated 28 September 1953.) Nearly two decades later, the 
Tamil Nadu Backward Classes Commission estimated that 105 communities that made up the 
then current list were 51.3% of the population of the state ( Tamil Nadu Backward Classes 
Commission 1971: I, 176). 
64. The charge to prepare a list is, of course, not specified in the constitutional provision for such a 
committee (Art. 340), although some listing seems to be anticipated by Art. 338(3). 
­167­ 
The directions express an expectation of centrally­established uniform standards and a central 
master list of Backward Classes.  65.   Accord­ 
TABLE 19 ESTIMATES OF POPULATION OF OTHER BACKWARD CLASSES, EARLY 
1950s 
1951 Census  Backward Classes
Commission 

Total Enumera­ Percent­ Estimated Estimated


Percent­ Percent­
Population ted Pop. of age of Pop. of Pop. of
age of age of
(lakhs)  OBC  Enum.  OBC  OBC 
Est. to Est. to
to
Total  Total 
Total 
Pop.  Pop. 
STATE  Pop. 
ASSAM  90.44  18.84  20.8  26.15a  28.9  28.66  31.7 
BIHAR  402.26  62.76  15.6  66.90  16.6  153.21  38.1 
BOMBAY  359.56  44.90  12.5  16.13  4.5  110.10  30.6 
MADHYA 
212.48  66.48  31.3  56.04a  26.4  79.03  37.2 
PRADESH 
MADRAS (& 
570.16  197.16  34.6  263.95  46.3  126.81  22.2 
ANDHRA) 
ORISSA  146.46  41.73  28.5  15.85  10.8  13.56  9.3 
PUNJAB  126.41  N.A.  N.A.  7.89a  6.2  25.56  20.2 
UTTAR 
632.16  41.51  6.6  41.51  6.6  269.10  42.6 
PRADESH 
WEST BENGAL  248.10  13.59  5.5  1.81  0.7  22.66  9.1 
HYDERABAD  186.55  94.13  50.5  122.85  65.9  137.66  73.8 
MADHYA 
79.54  13.68  17.2  13.25  16.7  19.37  24.4 
BHARAT 
MYSORE  90.75  17.53  19.3  20.62  22.7  59.64  65.7 
P.E.P.S.U.  34.91 ...  ...  3.79a  10.8  4.42  12.7 
RAJASTHAN  152.91  32.16  21.0  34.30a  22.4  34.31  22.4 
SAURASHTRA  41.37  15.78  38.1  11.06  26.7  12.16  29.4 
TRAVANCORE­
92.80  4.16  4.5  2.85  3.1  9.12  9.8 
COCHIN 
AJMER  6.93  0.35  5.1  0.35  5.1  2.98  42.9 
BHOPAL  8.36  1.25  14.4  1.35  13.5  2.98  35.2 
HIMACHAL 
PRADESH
& BILASPUR  11.09 ...  ...  2.38a  21.4  3.15  28.4 
COORG  2.29  0.69  30.3  0.68  29.7  0.64  27.8 
DELHI  17.44  N.A.  N.A.  2.41a  13.8  3.18  18.2 
KUTCH  5.68  0.01  0.2  1.08  19.0  2.01  35.4 
MANIPUR  5.78 ...  ...  ...  ...  0.35  6.1 
TRIPURA  6.39  0.30  4.7  4.12  64.5  0.69  10.1 
VINDHYA 
35.75  7.40  20.7 ...  ...  13.56  38.5 
PRADESH 
ANDAMAS &  .30 ...  ...  ...  ...  ...  ... 
1951 Census  Backward Classes
Commission 

Total Enumera­ Percent­ Estimated Estimated


Percent­ Percent­
Population ted Pop. of age of Pop. of Pop. of
age of age of
(lakhs)  OBC  Enum.  OBC  OBC 
Est. to Est. to
to
Total  Total 
Total 
Pop.  Pop. 
STATE  Pop. 
NICOBARS
INDIA  3568.29  674.39  18.9  730.01  20.5  1135.10  31.8 
Source: III BCC14­15. (All figures in lakhs). 
Note: aRepresents estimated 1951 population of educational backward classes. (The source is not 
entirely
clear regarding Assam.) 
____________________ 
65. The expectation that there would emerge an official central list of Backward Classes was 
expressed by the Commissioner for Scheduled Castes and Scheduled Tribes, who noted in his 
report for 1953 that "at present there are no 'Backward Classes' as such, which may have been 
officially recognized by the Government of India. This would be done after the recommendations 
of the Backward Classes Commission are available" ( RCSCST 1953: 196). See also Planning 
Commission 1953: 243. 
­168­ 
ingly, the Commission, after two years of work, presented a list of 2,399 backward groups and 
recommended various measures for their economic, educational, social, cultural, and political 
advancement.  66.   It was estimated that these groups totalled more than 116 million members (about 
32% of the total population of India).  67.   (See table 19.) This does not include women as a separate 
group, although the Commission recommended that all women in India made up a Backward Class. 
 68.   Nor does it include Scheduled Castes and Scheduled Tribes, who in 1951 made up over 14% and 
6% respectively of the total population. 
It had been generally anticipated,  69.   if not universally approved,  70.   
____________________ 
66. During the Commission's deliberations, estimates of the Backward Classes had risen as high as 
70% of the total population (III BCC9). It was proposed and seriously considered to list 
"advanced" groups with all of the remainder to constitute the Backward Classes (I BCC48). This 
is discussed at length in Shri Chaurasia's minute of dissent (III BCC41 ff). 
67. This population figure is for only 913 of the 2,399 Backward Communities (Ministry of Home 
Affairs 1956: 2). The Commission had no population figures for the others, but it may be 
presumed that few were very numerous. 
68. I BCC31­32. 
69. This expectation was evident in the Parliamentary debate over Art. 15(4), discussed above in nn. 
53­62. It was very clearly expressed in 1954 by N. D. Majumdar, a civil servant, writing in a 
government publication on social welfare: "The backwardness with which the commission is 
expected to deal is much more special. For one thing, the commission cannot recognize 
individual backwardness, however widespread. For its purpose, the backwardness must be 
collective. . . . Even this is not enough. The group must not be a purely economic group or even a 
social group of the modern type. For example, industrial labour. . . . The commission cannot 
concern itself with these groups because it cannot trench on the ground of labour or agrarian 
welfare. In order to come under the scope of the commission's investigations, a social group has 
to be a genetic or hereditary group, and exhibit disabilities and backwardness as a group confined 
to lowliness by birth. . . . [T]he commission is expected to deal with certain undesirable 
consequences of the evolution of Hindu society, and developments of the same kind in the non­
Hindu societies which have been influenced by the Hindu social pattern" ( Planning Commission 
1960: 219). An eminent anthropologist (who had served as Gandhi's secretary and was later to 
serve as Commissioner for Scheduled Castes and Scheduled Tribes) writing in 1954 also 
visualized the specific purpose of the Backward Classes Commission as the devising of 
measures to eliminate not all inequalities but specifically those associated with caste. "It is . . . 
the desire and will of the Indian nation to do away with the hierarchy of caste and of its 
consequent social discrimination, and prepare the ground for full social equality" ( Bose 1967 
[1954]: 182). He suggests that "castes whose water is not acceptable to Brahmins, and who are at 
the same time very poorly represented in school registers or in Income Tax and Union Board 
[tax] lists in proportion to their population, should be considered 'backward' in the sense of the 
Commission's enquiry" ( Bose 1967: 188). 
70. Nehru is reported as saying at the inauguration of the Backward Classes Commission (on March 
18, 1953) that he disliked the term Backward Classes, and that it was basically wrong to label any 
section as backward even if they were so, particularly when 90% of Indians were poor and 
backward (I BCC3). This seems to mark shift from his 1951 position in the debate over Art. 
15(4). (See n. 59 above.) 
­169­ 
that the "classes" designated by the Commission would be castes or communities. While indicating 
its desire to avoid perpetuating the evils of caste and its eagerness to avoid caste, the Commission 
"found it difficult to avoid caste in the present prevailing conditions."  71.   The Commission felt it was 
"not only correct but inevitable" to interpret its terms of reference "as mainly relating to social 
hierarchy based on caste."  72.   As general criteria of backwardness the Commission listed trade and 
occupation, security of employment, educational attainments, representation in government service, 
and most emphatically, position in the social hierarchy.  73.   It used these standards, not directly to 
isolate categories of backward persons, but to isolate backward communities. The units to which 
these tests were applied were for the most part caste and sub­caste groups. In identifying the 
backward, the Commission used caste in two ways: first, it used caste groups as the units or 
"classes" to be classified; and second, it used the position or standing of these groups in the social 
hierarchy as the principal criterion for determining their backwardness. 
In classifying communities, the Commission aimed to take into consideration 
the social position which a community occupies in the caste hierarchy, the percentage of 
literacy and its general educational advancement; and its representation in Government 
service or in the industrial sphere. The economic backwardness also had to be kept in 
view . . . as also the recent trends in its advancement as a result of various [Government] 
measures . . . during the past one or two decades.  74.   
The Commission was deluged by communities claiming to be backward.  75.   But it found that the 
paucity of data fully matched the immensity of its task.  76.   The state governments were found not to 
have relevant statistics, "administrators and census officers pleaded their inability to supply the 
relevant material," and figures supplied by the communities themselves were chiefly guesswork. 77. 
The decision to de­emphasize caste in the 1951 census operations had left the Commission without 
any figures on the literacy, income, or occupation of the various communities.  78.   The Commission 
had no facilities for generat­ 
____________________ 
71. I BCC41. 
72. Id., 42. 
73. Id., 46. 
74. Id., 47. 
75. The Commission received a total of 3,344 memoranda and interviewed a total of 5, 636 persons 
(I BCC217). 
76. In addition to a compilation of a list of Backward Classes, the Commission was subsequently 
charged with recommending revisions of the lists of Scheduled Castes and Scheduled Tribes. 
77. I BCC7. 
78. I BCC11. Apparently the absence of caste data was the deliberate policy of Sardar Patel, the 
Home Minister until 1950, who rejected caste tabulation as a device to 
­170­ 
ing data themselves. In the face of these obstacles, the Commission plowed bravely on: 

In the absence of reliable facts and figures, the only course open to us was to rely on the 
statistics available from the various Governments and the previous census reports, and 
to go by the general impressions of Government officers, leaders of public opinion and 
social workers.  79.   
In some cases there was no data at all and "the decision had to be taken on the strength of the name 
of the community only, on the principle of giving the benefit of doubt."  80.   The Commission did not 
undertake to document the application of its tests to the communities on its list, for which it 
provided only names, traditional occupations, and estimated population.  81.   In addition to listing the 
Backward Classes, the Commission was instructed to 
investigate the conditions of all such socially and educationally backward classes and 
difficulties under which they labour: 
and make recommendations 
i.  as to the steps that should be taken by the union or any state to remove such difficulties or to 
improve their conditions; 
ii.  as to the grants that should be made. . . .  82.   
True to its charge, the Commission recommended a vast array of schemes for the protection and 
advancement of the backward, includ­ 
____________________ 
confirm the British theory that India was a caste­ridden country and as an expedient "to meet the 
needs of administrative measures dependent upon caste division" (from a 1950 address to the 
Census Conference, quoted as I BCC9). CE III BCC18. 
79. Id., at 8. Cf I BCC47, where the Commission notes that the existing backward classes lists of the 
states and the Ministry of Education "formed the basis of our information." 
80. III BCC9. Cf. I BCC49. 
81. Hardgrave 1969: 141ff. provides a revealing account of the Nadars' successful efforts to be listed 
by the Commission as a Backward Class. This numerous Tamil caste, whose traditional 
occupation was toddy­tapping, had produced educated and prosperous strata by dint of strenuous 
efforts at self­improvement. The Nadar Mahajana Sangam submitted to the Commission that 
"barring a few individuals who can be counted in numbers, the community is illiterate, 
economically poor and socially boycotted." A Nadar partisan was informed by the Secretary of 
the Commission that it would be "impossible to include 'Nadars' since many witnesses have told 
the Commission that the 'Nadars' are not all backward." The Commission's solution was to 
include in its list "Shanan," a derogatory appellation for the caste which the Nadars had long 
fought to discourage. (The Madras Government later adopted the same device, and after a period 
of some confusion, finally ordered that "Nadars" be treated as synonymous with "Shanan.") 
82. I BCC2. 
­171­ 
ing a number of major changes in rural life: redistribution of land, protection of tenants, help to the 
small agriculturalists (credit, price supports, irrigation, etc.).  83.   They recommended the creation of 
a separate ministry for Backward Classes' Welfare.  84.   They proposed reservations for Backward 
Classes in government service of at least 25% in Class I, 33⅓% in Class II, and 40% in Classes III 
and IV.  85.   In addition there were various aids to the education of these groups and a reservation of 
70% in medical, scientific, and technical colleges.  86.   
In a last minute volte face, the chairman virtually repudiated the Commission's work, having 
concluded that "it would have been better if we could determine the criteria of backwardness on 
principles other than caste."  87.   He finds the caste test repugnant to democracy and inimical to the 
creation of "a casteless and classless society" (in the then fashionable phrase) by perpetuating and 
encouraging caste division. It is not entirely clear what he would put in its place. He recommends 
that backwardness be measured by residential, economic, educational, and cultural criteria. 88. 
Apparently he repudiates not only caste standing as the test of backwardness but also the use of 
caste units. At several places in his covering letter he suggests that only individuals and families 
should be the units whose backwardness is ascertained.  89.   Elsewhere, however, he recommends 
policies which seem incompatible with this position (e.g., that where an income test in employed, 
members of backward communities should be given priority).  90.   
The chairman's last­minute desertion foreshadowed (and perhaps augmented) the negative reception 
that awaited the report.  91.   The spectacle of numerous groups vying to display their backwardness, 
the feeling that caste classifications were divisive and unfitting, the casual­ 
____________________ 
83. I BCC51 ff. 
84. I BCC143 ff. 
85. I BCC140. The Commission indicated that it was inclined to recommend reservations in 
proportion to population, where educational attainments permitted, but adopted the minimum 
figures to leave sufficient scope for highly qualified candidates. 
86. I BCC125. This recommendation was meant to apply until accommodation was available for all 
students qualified for admission. 
87. I BCCxiv. 
88. I BCCxiv­xv. 
89. I BCCiv, xiv. 
90. I BCCviii. Cf. his suggestion that the most backward communities should be treated separately to 
prevent all benefits from going to the most advanced among the backward (I BCCxxi ff). 
91. In addition to the Chairman's repudiation, three other members, of the Commission, including 
the Secretary, filed minutes of dissent objecting to the caste basis of classification (Minutes of 
Dissent of Dr. Anup Singh, Shri Arunangshu De, Shri P.G. Shah). Two other members dissented 
on the ground that the Commission had not gone far enough: Shri T. Mariappa merely objected 
to the failure to include urban Lingayats and Vokkaligas in Mysore; Shri S. D. Singh Chaurasia 
set forth a detailed proposal for equating Backward Classes with Sudras. The various Minutes of 
Dissent make up Volume III of the Commission's Report. 
­172­ 
ness of the Commission's application of its criteria, the vastness of the number they found 
backward, and the expansiveness of the preferences they proposed exposed their work to widespread 
criticism.  92.   
The Commission's report was laid on the table of both houses of Parliament on September 3, 1956, 
accompanied by a withering critique from the Minister of Home Affairs, expressing disappointment 
with its criteria and its conclusions.  93.   The emphasis on caste, the Minister asserted, displayed the 
"dangers of separatism."  94.   Not only was the caste basis unfair to the backward outside these 
communities, but the caste system was undeniably "the greatest hindrance in the way of our 
progress toward an egalitarian society, and the recognition of specified castes as backward may 
serve to maintain and perpetuate the existing distinctions on the basis of caste."  95.   The 
Commission's standards other than caste were "obviously vague."  96.   The very expansiveness of the 
Commission's list undermined its usefulness, for if everyone "barring a few exceptions, has thus to 
be regarded as backward, the really needy would be swamped by the multitude and hardly receive 
any special attention. . . ."  97.   Thus the Commission failed to find "positive and workable criteria." 
 98.   "[F]urther investigations will have to be undertaken so that the deficiencies that have been 
noticed in the findings of the Commission are made good and the problem is solved with due regard 
to the requirements of Art. 340. . . ."  99.   The state governments were requested to undertake ad hoc 
surveys to determine the numbers of Backward Classes and in the meantime to "give all reasonable 
facilities" to the Backward Classes "in accordance with their existing lists and also to such others 
who in their opinion deserve to be considered as socially and educationally backward in the existing 
circumstances."  100.   Thus the matter went back to the states; the Commission's report remained on 
the table, and in spite of occasional agitations, was not taken up by Parliament until 1965. 
When the replies from the state governments "were not found helpful," the central government 
requested the Office of the Registrar General to conduct ad hoc surveys to determine suitable 
criteria.  101.   It was hoped that the Registrar General could determine occupational tests of 
backwardness. On the basis of a three­state sample survey, an 
____________________ 
100. Id., at 4­5. 
101. RCSCST 1957­58: I, 9. 
92. Opposition to the way in which the Commission proceeded might have been anticipated in view 
of the Government's decision to ignore caste in the 1951 Census (see n. 78 above) and from 
Prime Minister Nehru's remarks at the Commission's inauguration (see n. 70 above). 
93. Ministry of Home Affairs 1956: 3­4. 
94. Id., at 3. 
95. Id. 
96. Id. 
97. Id., at 4. 
98. Id. 
99. Id. 
­173­ 
occupational test with an income ceiling was tentatively proposed.  102.   The Home Ministry's 
suggestion that the states adopt such an occupational criterion enjoyed a mixed reception. 103. 
However, this effort collapsed when the Deputy Registrar General reported that it would be 
impossible to draw up a precise and complete list of occupations whose practitioners were socially 
and economically backward.  104.   
Those who entertained hopes of generous preferential treatment for the Other Backward Classes 
(OBC) found some encouragement in the late 1950s. State expenditures on schemes for the OBC 
increased (particularly in education, where they were substantial),  105.   although the expansion of the 
central government's scheme for postmatriculation scholarships slowed down after 1955­56. 
Apparently at some point in 1957 the Government of India instructed all state governments to extend 
the benefits of various welfare schemes to the OBC, using the Ministry of Education lists, pending 
finalization of a list of OBC. Furthermore, it requested the state governments to fill unused 
vacancies in seats reserved in educational institutions for Scheduled Castes and Tribes with students 
from the OBC,  106.   a measure found very gratifying by the All­India Backward Classes Federation 
(AIBCF), which observed that "at long last the voice of the Federation is being effectively heard." 
 107.   
____________________ 
102. RCSCST 1958­59: I, 11­12. The Census authorities indicated that "it might be possible to draw 
up a list of socially and educationally backward occupations on the basis of: (a) any non­
agricultural occupation in any State . . . in which 50% or more of the persons belong to the 
Scheduled Castes or Scheduled Tribes; or (b) any nonagricultural occupation in which literacy 
percentage of the persons depending thereon is less than 50% of the general literacy in the 
State." It was suggested that this test be used in conjunction with an income limit of Rs. 1,000 
per family. 
103. RCSCST 1958­59: I, 12. 
104. Letter of the Ministry of Home Affairs to Chief Secretaries of all State Government/Union 
Territories, 14 August, 1961, reprinted at RCSCST 1960­61: II366. 
105. A number of states gave fee concessions to Other Backward Classes ( RCSCST 1956­57: II, 65). 
106. Beginning in 1957, the Ministry of Home Affairs and the Education Ministry recommended that 
vacancies in reservations for Scheduled Castes and Scheduled Tribes should be filled with Other 
Backward Classes candidates provided they met minimum qualifications. Letters from Ministry 
of Home Affairs No. 10/41/57SCT(IV), dated 30 July 1957; Letter No. 10/32/57­SCT(IV), dated 
11 June 1958. This provision was withdrawn in letter No. 28/6/61­SCT(I), dated 4 Jan. 1962, 
after the government had decided not to draw up an all­India list of backward classes. The latter 
two letters are reproduced in Planning Commission 1965: 234­35. 
107. All­India Backward Classes Federation 1958: 6­7. In December 1958 the Federation's executive 
committee reiterated its thanks to the Government and its appeal that the same be done for 
unfilled reservations in government posts ( All­India Backward Classes Federation 1959: 
Appendix vi). In an interview with Mr. Jai Narayan Singh Yadava ( 10 April 1966) he estimated 
that under this arrangement Backward Classes candidates obtained roughly 6% of places in the 
affected institutions. 
­174­ 
By the beginning of the 1960s the tide was running strongly against definition of the Backward 
Classes by community. Opposition within the government  108.   was augmented by criticism from 
academics and much of the national press, who voiced a common suspicion of the caste criterion. 
 109.   For the first time since 1951, a court intervened to strike down a scheme for Backward Classes 
in a decision widely acclaimed as a blow at casteism.  110.   Revulsion from communal criteria was 
reinforced by reports of their abuse. In a situation where many thoughtful persons were increasingly 
concerned about the dangerous potentialities of social cleavages,  111.   the alleged divisive tendencies 
of the communal criterion seemed a serious threat to national unity and integration.  112.   
____________________ 
108. In 1959 the Study Team on Social Welfare and Welfare of the Backward Classes recommended 
the abolition of the OBC Category and its replacement with economic criteria for selecting 
beneficiaries for government schemes. The recommendation is contained in the introduction 
(dated July 1959), but the Committee's position is hardly clear in the body of its report ( 1959: 7; 
cf. id., 127). 
109. See e.g., Srinivas 1957 presidential address to the anthropology section of the Indian Science 
Congress. After a dispassionate review of developments, he suggests "it is time to give serious 
thought to evolving 'neutral' indices of backwardness. . . . The criteria of literacy, landownership 
and income in cash or grain should be able to subsume all cases of backwardness." But other less 
disinterested critics perceived the development of a "vested interest in backwardness" and all 
sorts of dire effects on national integration and efficiency. See, e.g., Mehta 1963. (There were 
few to argue the other side. For a rare example, see Subbiah 1963.) For a convenient review with 
many references to the scholarly and polemical literature of this period, see Barnabas and Mehta 
1965. 
110. Ramakrishna Singh v. State of Mysore, A.I.R. 1960 Mys. 338. The Court's judgment (discussed 
below passim in chaps. 7 and 8) on 18 September 1959 was acclaimed as a blow at casteism. 
E.g., the Times of India's Mysore Newsletter for September 23, 1959, began: "The Mysore High 
Court has done what no politician of this State would dare to do. It has put a brake on the race to 
seek educational privileges by all and sundry." The next day the decision was lauded in an 
editorial which observed: "The only rational test for backwardness is a person's economic 
circumstances and not his caste . . . . The Mysore and similar orders elsewhere show how caste­
ridden the Congress is despite the anticasteist protestations of its leaders." 
111. In 1960 Selig Harrison's book appeared with its gloomy reflections on the disintegrative forces 
of language, region, and caste and its serious questioning of whether India could remain united. 
Although he points out that caste tests of backwardness consolidate caste consciousness, he does 
not assign them any major disintegrative role: "[T]he constitutional guarantees only 
institutionalize a group awareness that would in any case exist to a great extent. Since the upper 
castes often refuse to recognize the achievement of an individual member of a low caste who 
may, for example, acquire education, the individual invariably concludes that mobility on any 
significant scale must be a group phenomenon" ( Harrison 1960: 104). 
112. E.g., in 1961 the Ministry of Education's Committee on Emotional Intergration received over 
1,200 replies in a poll of highly­placed educators. Over 70% were convinced that divisive forces 
had increased since Independence. When asked to 
­175­ 
While a "casteless and classless society" remained the avowed aim of the Congress and a wide 
section of the intelligentsia, there had been a subtle shift in notions of how this aim was to be 
pursued. A decade before, it was widely thought that special redistributive measures were required 
specifically to offset inequalities associated with caste, even while general development programs 
addressed other aspects of inequality. The notion of caste differentials as themselves a significant 
form of inequality deserving of special governmental attention to eliminate their effects gave way to 
a notion that the salient differences were economic; specific redistributive measures directed at caste 
differences were not necessary, since overall development would raise the general level. 113. 
Recognition of caste differences in order to offset their effects was replaced by an enhanced 
reluctance to recognize them at all; indeed, any recognition of such differences was seen as itself a 
violation of egalitarian principles and productive inequality. 
In May 1961 the cabinet decided that no national list of Other Backward Classes should be drawn 
up, and the states were informed that in the view of the Government of India "it was better to apply 
economic tests than to go by caste."  114.   At the end of May a Conference of Chief Ministers to 
consider matters relating to National Integration "agreed that economic backwardness rather than 
community or caste would provide an appropriate criterion for giving aid to individuals in matters 
of education including professional and technical training."  115.   In August 1961 the Home Ministry 
informed the state governments of the Centre's decision not to list Backward Classes. In the 
Ministry's view, the very expansiveness of such proposed enumerations as that of the Backward 
Classes Commission' militated against them: "If the bulk of 
____________________ 
identify the "disintegrative forces," the highest number chose "casteism/communalism" (62%)­­
significantly more than chose religious bigotry, etc. (39%) or regionalism, etc. (44%) ( Ministry 
of Education 1962: 189). Cf. the Times of India's view that if the use of preferences were not to 
defeat the establishment of a "casteless and classless society" the government must amend the 
Consitution to eliminate the permissibility of communal units in distributing preferences ( Times  
of India, August 30, 1961). 
113. Thus the Estimates Committee of the Lok Sabha (Forty­eighth Report) proposed that the 
"weaker sections of society" should be defined by economic criteria as well as educational and 
social backwardness. "Progressive emancipation from economic backwardness should help the 
people belonging to Scheduled Castes and Scheduled Tribes to overcome their social 
backwardness" (quoted at Study Group on the Welfare of the Weaker Sections of the Village 
Community 1961: I, 12). The shift in emphasis is apparent in the October 1961 report of the 
Study Group (chaired by Jayaprakash Naravan). The "weaker sections" are identified as the 
economically backward­practically all of the village community. Caste inequality is a 
contributing cause of economic backwardness which is to be addressed by remedial measures 
framed along lines of a means test ( 1961: I, chap. 2). 
114. Ministry of Home Affairs 1962a: 38 
115. Id. 
­176­ 
the country's millions were to be regarded as coming within the category of Backward Classes, no 
useful purpose could be served by separate enumerations of such classes." Furthermore, the caste 
criterion was objectionable: "the remedies suggested on the basis of caste would be worse than the 
evil of backwardness itself."  116.   But subsequent efforts to discover usable criteria on economic lines 
"did not yield any useful results." Whereas in 1956 the Home Ministry had acknowledged an 
obligation to compile a list in accordance with the requirements of Article 340, it now pointed out 
that the Constitution did not require the Centre to draw up a list. Since, even if it were to do so, "it 
will still be open to every state government to draw up its own lists, any all­India list would have no 
practical Utility."  117.   More important, the "crying need" of the day was social cohesion and 
emotional integration. There was the grave danger that different treatment of the "backward" would 
foster divisive tendencies and would undermine efforts for general economic uplift and the 
reduction of disparities between different classes. The states were urged to emphasize the expansion 
of welfare and educational benefits to all of the poor, employing economic rather than communal 
categories."  118.   
The withdrawal of the central government from involvement in preferences for the Other Backward 
Classes was confirmed by the omission of any provision for them in the central sector of the Third 
Five Year Plan.  119.   The amount contributed by the central government for postmatriculation 
scholarships for OBC, which had increased steadily since 1949, was frozen at the 1958­59 level. 120. 
And beginning in 1963, the state lists of OBC used in administering this scheme were abandoned in 
favor of a test of family income.  121.   
____________________ 
116. Letter of Minister of Home Affairs to Chief Secretaries of All State Governments/Union 
Territories, 14 August, 1961, reprinted at RCSCST 1960­61: II, 366. 
117. Id. 
118. Id. These views were endorsed by the Ministry of Education's Committee on Emotional 
Integration ( 1962: 45) which concluded that "the time has now come, in our opinion, when 
increasingly assistance should be based on economic criteria. In some states powerful groups 
have exploited 'backwardness' to their own advantage and to the detriment of society as a 
whole. . . . It also results in the long run in making the backward classes less self­reliant than 
they should be." 
119. The Other Backward Classes category had become, from the Centre's point of not a stratum of 
the population but a catch­all category for groups other than the Scheduled Castes and Tribes felt 
to be deserving of government help. In 1960 the Commissioner for Scheduled Castes and 
Scheduled Tribes noted that Denotified Communities, Nomadic Tribes, and Neo­Buddhists were 
"decidedly, the groups to get Government help as Other Backward Classes during the Third Five 
Year Plan period" ( RCSCST 159­60: I,9). 
120. RCSCST 1959­60:238. 
121. RCSCST 1964­65: 155, 158, 159. 
­177­ 
The central government's campaign for economic criteria in the states was given added impetus by 
the first (since 1951) intervention of the Supreme Court into the matter of who are the Backward 
Classes. In September 1962 the Supreme Court struck down the Mysore Backward Classes list, 
whose defects included exclusive reliance on caste standing as a measure of backwardness, adding 
the onus of constitutional disrepute to the caste criterion.  122.   The Court's judgment, which warmly 
commended economic tests, was widely acclaimed and widely interpreted (mistakenly, as we shall 
see in chapter 7) as outlawing entirely the use of caste tests. This case marked the emergence of the 
judiciary as the institution within which the problem of who are the Backward Classes was most 
carefully and coherently addressed. 
In 1965, when the Report of the Backward Classes Commission was finally discussed in Parliament, 
the central government's spokesman firmly reiterated its opposition to communal criteria. Caste 
criteria were not only administratively unworkable, but were contrary to the "first principle of social 
justice" in their unfairness to the other poor. They were contrary to the Constitution, would 
perpetuate caste, and would create in the recipients both vested interests and a sense of helplessness. 
The Centre endorsed economic criteria, but refrained from enforcing it on the states, preferring "the 
path of persuasion."  123.   Eight states were said to have adopted the economic criterion.  124.   
Backward Classes organizations continued to campaign for greater benefits and for a revival of 
central responsibility and interest in the Backward Classes.  125.   In particular, they petitioned for a 
restoration 
____________________ 
122. Balaji v. State of Mysore, A.I.R. 1963 S.C. 649. 
123. LSD (3rd Series), Vol. 48 No. 16, cols. 3973­3976 ( 25 Nov. 1965). 
124. The list included two states ( Gujarat and Maharashtra) that had retained the caste test outside 
the scholarship area; one state ( Punjab) that not only retained the caste test outside the 
scholarship area, but employed a list of communities in the scholarship area as well, albeit in 
conjunction with an income test; one state that had no schemes of its own for Other Backward 
Classes (Orissa); and two states that had none outside of scholarships ( Assam, West Bengal). 
The only states with economic tests and a high level of benefits for Other Backward Classes 
were Mysore and Andhra Pradesh. The latter reverted to a caste test less than a year later, and the 
former in 1977. 
125. Backward Classes groups themselves differ in their definitions of the Backward Classes and 
their population estimates, but they tend to be rather expansive often including all religious 
minorities as well as Hindu groups. Thus speakers at the 1966 conference of the All­India 
Backward Classes Federation constantly used the figure of 85% of the population­­this was 
intended to include the Scheduled Castes and Tribes, Muslims, Buddhists, and Christians as well 
as the Other Backward Classes who were said to make up 40% of the total population. (Author's 
notes). Welcoming the participants in a 1969 Seminar on Backward Classes and the Fourth Five 
Year Plan (organised by the All­India Backward Classes Federation), Mr. V. P. Singh mentioned 
the figure of 60­70 million Backward Classes. The chief guest, A. N. Jha, Lieutenant Governor 
of Delhi, used the 85% figure while warning the delegates of the futility of attempting to define 
Backward Classes ( Backward Classes Review 1( 3):18). 
­178­ 
of the caste basis, for implementation of the long­ignored report of the Backward Classes 
Commission, and for creation of a ministry to attend to the problems of the Backward Classes. 126. 
Organizations of particular communities and some composite Backward Class organizations at the 
local level (especially in the south and in Bihar) continue to flourish, but many of the local 
composite organizations have become moribund.  127.   After a period of desuetude, the national 
AIBCF had a brief revival in the late 1960s. 
D. THE VARIED STATE PATTERNS 
As the central government's role in designation of Backward Classes moved from an attempt to 
prepare a single nationwide list to suggesting standards to the states, courts involved in litigation 
over Backward Classes lists made it clear that the Constitution did not confer on these 
____________________ 
126. E.g., the resolutions of the All­India Backward Classes Conference held in New Delhi in March 
1966: "The Federation is of the firm opinion that even though ultimately a class of people are to 
be judged by the economic well­being, in the transition period when large sections suffer from 
social disabilities in addition to economic poverty it would not be in the national interests to 
determine backwardness in terms of economic criteria alone. Social backwardness­­as laid down 
in the Constitution­­can only be determined in terms of castes and communities to which the 
stigma applies as a whole and till the society becomes casteless it only injures the interests of the 
'socially backward' to determine backwardness in terms of economic criteria only." The 
Federation supported generous help for the "poorer sections of socially advanced classes" but 
asserted that while economic tests might be appropriate for these, "for the Backward Classes, the 
criterion should be 'Social.'" Cf. Sathi (n.d.) who expresses the view, widely shared in Backward 
Classes circles, that Art. 340 represents a definite (and unfulfilled) commitment to a specific 
stratum of communities. M. S. A. Rao ( 1968: 781) reports that at the Conference of the Yadava 
Mahasabha in 1968, implementation of the BCC Report and the revival of the caste criterion 
were among the most prominent demands. 
127. All of these composite groups are oriented to the "caste" basis, and there is some indication that 
they have declined in direct correspondence to the slackening of the flow of benefits along these 
lines. The desuetude of the All­India Backward Classes Federation can be roughly measured by 
the fact that the last of a series of nicely printed reports begun in 1955 appeared in 1961. (There 
was a quickening of activity after the 1967 elections. The Federation began publishing a 
publication called the Backward Classes Review, whose first issue appeared in December 1968.) 
The Mysore State Backward Classes Welfare Association, established in 1960 (apparently in 
response to initiatives from the Federation) and devoted mainly to organizing and representing 
the "more backward classes" (i.e., not the Vokkaligas and Lingayats), became defunct after 
Mysore abandoned the caste list (interview with Mr. Venkataswamy, Secretary, in Bangalore, 
1966). Organizations of particular communities have proved much more. durable. Thus, 
ironically, the government's policy against communal criteria has discouraged the more broadly 
based composite organizations and left the field to organizations that are communal in the literal 
sense. I encountered no organized groups of recipients of benefits distributed along economic 
lines. 
­179­ 
central proposals any conclusiveness in the identification of the Backward Classes.  128.   While the 
State might make special provision for "any" Backward Class, it is under no constitutional 
obligation to make such provision for every class designated as backward by the Backward Classes 
Comisssion or by any other agency. Indeed, the State may not rely on the findings of the Backward 
Classes Commission to establish conclusively that a given class is backward.  129.   Conversely, 
preferences may be given to a group which does not appear on the Commission's list. State power to 
provide preferential treatment to the backward is not limited to those listed by the Commission or 
any other agency. The original expectation that the Commission's list or some part of it would be 
confirmed by presidential specification has gone unfulfilled. This failure does not invalidate other 
listings of Backward Classes, for the courts have held that presidential specification would not 
exhaustively define the Backward Classes for purposes of preferences under Articles 15(4) and 
16(4).  130.   Central proposals may, of course, carry some persuasive weight with the states (or, as 
guides to "reasonableness," with the courts).  131.   But it is the "State" in the broad sense 
____________________ 
128. But of Professor Tripathi's argument ( 1972: 206) that the way to vindicate rights against caste 
discrimination is to confine the power to make provisions for Backward Classes exclusively to 
parliament. ( Tripathi 1972: 206). Presumably he would then confine the power to designate 
Other Backward Classes to the center as well. On the failure of textual arguments to this effect, 
see n. 130, below. 
129. Jacob Mathew v. State of Kerala, A.I.R. 1964 Ker. 39 at 56. 
130. Art. 338(3) served as the cornerstone of an ingenious argument that the constitutional plan for 
the designation of Backward Classes is analogous to the exclusive central control over 
designation of Scheduled Castes and Scheduled Tribes: that Articles 340 and 338 (3) provide for 
exclusive central designation corresponding to Articles 341 and 342, the only difference being 
the additional step of the Commission Report. This argument was rejected in Ramakrishna 
Singh v. State of Mysore, A.I.R. 1960 Mys. 338 at 342. The court, noting the absence of any 
provision corresponding to Art. 366 (24) or (25), pointed out that there is no indication that the 
presidential specification of Backward Classes for purposes of the operations of the 
Commissioner of Scheduled Castes and Scheduled Tribes was to define Backward Classes 
exhaustively for all constitutional purposes. 

Another variant of this argument for exclusive central control over designation of Backward 
Classes put forward in Balaji v. State of Mysore, A.I.R. 1963 S.C. 649, 657­8, relies on the 
provision in Art. 340 for the appointment of the Backward Classes Commission. It was argued 
that "Backward Classes" could be designated only by the President in pursuance of 
recommendations of the Commission. Noting that Art. 340 (I) refers to measures that might be 
taken by the states and the Union government, the Supreme Court found that exclusive 
presidential power in this area was not contemplated by the Constitution. (This conclusion is 
reinforced by the fact that the Commission is not a continuing body with power to revise its list. 
Nor is there any provision for revision of a presidential listing of Backward Classes.) This 
argument (that reservations can only be made for Backward Classes identified by the 
Commission contemplated by Art. 340) is advanced again in Pradip Tandon v. State of U.P. 
A.I.R. 1975 All. 1,6, and encounters a similar rejection. 

131. Thus, in the Balaji case, the court took careful note of the central government's 
­180­ 
of all governmental organs that retains the power of designation.  132.   Preferences and who is to 
receive them may be provided by executive (as well as by legislative) action ­­ and they almost 
always are.  133.   
With the abandonment of central attempts to define the Backward Classes, and the relinquishment 
of whatever control might accompany central funds for Other Backward Classes, the matter reverted 
to the states. The composition of the OBC, the scope of preferential programs, and the level of 
benefits continued to vary widely from state to state. Some of the varying state practices for 
selecting Backward Classes are summarized in table 20. 
For all its incompleteness and oversimplification, this list gives us a rough profile of who were the 
Backward Classes in the late 1970s. Later chapters will discuss the standards by which such classes 
were selected. But at this point it is worth noting a few of the salient features of this profile. Caste 
and communal units remain the predominant "classes" deemed backward. Caste lists range in 
magnitude from those which include a substantial portion of the state's population to those 
constituting a narrow stratum just above the untouchables. Income tests are also employed in many 
cases, sometimes independently and sometimes in conjunction with communal units. The level of 
benefits ranges from none through scholarships and fee concessions to an array of reservations in 
government posts and medical college admissions. 
There is important regional variation. For convenience we may think of three contiguous groupings. 
First, there is what we might call the peninsular bloc comprising the four Dravidian states ( Andhra 
Pradesh, Karnataka, Kerala, and Tamil Nadu) and Maharashtra. In these states the Backward 
Classes categories have a long history descending from pre­Independence arrangements; there are a 
wide range of benefits; and (except in Maharashtra) a major segment of the population ­from 38% to 
55% ­­ is included, and a major segment of scarce opportunities are reserved for them. 
In stark contrast is what we might call the eastern­middle band, stretching across India from Assam 
in the northeast through West Bengal and Orissa, across Madhya: Pradesh to Rajasthan ( Gujarat 
recently moved from this category to a pattern like that of 
____________________ 
suggestion (incorporating the recommendations of the All­India Council for Technical 
Education) that reservations in higher education should not exceed 25% (or, in exceptional areas, 
35%). A.I.R. 1963 S.C. at 656. 
132. The "State" in Arts. 15 (4) and 16(4) must be interpreted as defined in Art. 12 as including 
governments as well as legislatures and state and local as well as central authorities. 
133. This was early established by Ramakrishna Singh v. State of Mysore, A.I.R. 1960 Mys. 338 at 
343. The latest assertion of this point is in Triloki Nath Tiku v. State of Jammu and Kashmir [II], 
A.I.R. 1969 S.C. 1 at 2. 
­181­ 
TABLE 20 A PROFILE OF OTHE BACKWARD CLASSES IN THE LATE 1970s 

NAME EST. MEDICAL


OF OF COLLEGE
STATE  YEAR  GROUP  CRITERIA  POP.  POSTS  ADMISSION  OTHER  COMMENTS
"Other" 
benefits
only for those 
Scholarship; with
ANDHRA 92 listed  fee  income under
PRADESH  1975  BC  communities  38%  25%  25%  concessions  Rs. 3,600 
NAME EST. MEDICAL
OF OF COLLEGE
STATE  YEAR  GROUP  CRITERIA  POP.  POSTS  ADMISSION  OTHER  COMMENTS
All with 
income
Economically under Rs.  Scholarship
BC  1,500 ...  None  None  hostels 
Members of 
list of 128
communities 
whose 
BIHAR  1979  BC  50%  20%  N.I.  N.I. 
annual
income is 
less than Rs. 
12,000 
All with 
income 
under
Economically Rs. 12,000 
...  3%  N.I.  N.I. 
BC  annually, 
other
than SC/ST/
OBC
1977  BC  N.I.  N.I.  N.I.  10%  N.I. 
82 listed  5% Classes I& II
GUJARAT  1978  BC  N.I.  N.I.  N.I. 
communities  10%  classes III&IV
Income less 
Economically than Scholarships,
1972  N.I.  None  None 
BC  Rs.  etc. 
4,800/year 
Direct 
recruitment
2%
HARYANA  1978  BC  N.I.  N.I.  N.I.  N.I.  Promotion in
2% 
classes III & 
IV 
Direct 
recruitment
HIMACHAL 10%
1978  BC  N.I.  N.I.  N.I.  N.I.  Promotion in
PRADESH  2% 
classes III & 
IV 
1973  BC  N.I.  N.I.  N.I.  2%  N.I.  1 seat of 60 
1. 
Occupations
Socially and N.I.
JAMMU & 2. 19 
1973  educationally N.I. 42%  N.I.  N.I. 
KASHMIR  communities
BC  8 % 
3. Border & 
poor areas 
­182­ 
List of 16 
communities, [45 40%  Direct  G.T.
KARNATAKA  1979  BC  a N.I.  N.I. 
129 castes &  %]      recruitment  [N] 
62 tribes 
All with 
annual 
Economically
income N.I.  15%  N.I.  N.I.  U 
BC 
under Rs. 
4,800 
28
1977  BC ...  N.I.  N.I.  N.I.  C 

Members of 
listed
Socially and
communities  25 C, 
KERALA  1977  educationally N.I.  N.I.  N.I. 
with income %  [H] 
BC 
less than Rs. 
10,000 
Members of 
listed
Socially and
communities  Scholarships,
1972  educationally N.I.  N.I.  N.I.  I 
with income loans 
BC 
less than Rs. 
6,000 
Members of 
560 listed
communities 
1970  BC  N.I.  40%  N.I.  N.I.  O 
with income
less than Rs. 
8,000 
10
MAHARASHTRA  1977  OBC  N.I.  N.I.  N.I.  N.I.  C 

Exclusive 
of reser­
vations for
List of  12 Scholarships. Nav­
1966  OBC  14%  N.I.  J 
communities  %  loans, etc.  Buddhas 
and
some 
tribals 
MEGHALAYA  1978  BC  N.I.  N.I.  5%  N.I.  N.I.  R 
PUNJAB  1978  BC  N.I.  N.I.  5%  N.I.  N.I.  R 
[List of 
1977  BC  N.I.  N.I.  5%  N.I.  C 
communities] 
Backward 
N.I.  N.I.  2%  N.I.  [P] 
Areas 
List of 105 
communities
Hostels;
(plus add'l.  [51 31
TAMIL NADU  1972  BC  31%  special  K, Q 
list for Kanya %]  % 
training 
Kumari 
District) 
List of 58 
UTTAR  Initial 
1978  BC  communities  N.I.  16%  N.I.  N.I.  L 
PRADESH  b  recruitment 
   
15%  Promotion  M 
1977  Hill Areas ...  N.I.  N.I.  3%  N.I.  C 
Notes aIt is probable that this figure is an overstatement in that it fails to distinguish between the listed 
groups and the Economically Backward Classes noted just
below. Of the 40% total, 5% may be a reservation for the Economically Backward Classes and thus a 
partial duplication of the 15% reservation for that group. 
bStruck down as unconstitutional in Chhotey Lal v. State of U.P., A.I.R. 1979 All. 

Sources: A: Andhra Pradesh Department of Social Welfare, Performance Budget 1975­76: 1­3; B: Blair 
1980:65; C: Handbook of Medical Education 1977; D:
Gujarat Directorate of Education 1971­72: 76­77; E: Sukhvinder Kaur v. State of H.P., A.I.R. 1974 H.P. 
35; F: Janki Prasad v. State of Jammu and Kashmir,
A.I.R. 1973 S.C. 930, 933, 939­42; G: The Statesman, 18 March 1978: 4; H: Jayasree v. State of Kerala, 
A.I.R. 1976 S.C. 2381, 2383; I: Kerala Harijan
Welfare Department 1971­72: 10; J: Maharashtra Backward Class Welfare Wing 1966­67; K: Tamil Nadu 
Directorate of Backward Classes 1972­73: l;
L: The Statesman, 15 March 1978; M: The Statesman, 8 June 1978; N: Government of Karnataka 1975: 
1,316­17; O: Kerala (Backward Classes Reservation
Commission) 1971:II, 167­74; P: Gurinder Pal Singh v. State of Punjab, A.I.R. 1975 P. & H. 125: 126; Q: 
Tamil Nadu Backward Classes Commission
1971:I, 176; R: RSD Vol. 107, No. 23, cols. 139­40 ( 21 Dec. 1978); S: National Convention of Christian 
Leaders 1978:23; T: P. Singh 1980; U: The
Hindustan Times, 28 June 1979. Information from source earlier than date in column 2. 
­183­ 
Maharashtra). In these states there is no significant use of the Other Backward Classes category. 
The northern tier of states displays an intermediate pattern. Jammu and Kashmir, with its history of 
communal quotas, resembles the southern pattern. The admixture of geographical criteria there is 
also found in Punjab and Uttar Pradesh along with use of communal categories. Bihar, like Jammu 
and Kashmir, approximates the peninsular pattern. The Backward Classes are selected on a 
communal basis and make up a sizable portion of the population, but the benefits have not been as 
extensive as in the south. Extension of benefits to reservations of government posts in 1978 set off a 
political crisis. 
What the central government tried in 1965 to portray as a trend toward substitution of economic for 
communal criteria was even then largely a rhetorical artifact albeit one built around two substantial 
items ­­ the abandonment of communal units in Andhra and in Mysore.  134.   From the vantage point 
of the late 1970s this "trend" appeared as a transient deflection from the main line of development 
of Backward Classes lists. That line of development has been one of continuing use of communal 
units, but with increasing refinement and restraint. Under pressure from the courts, almost all the 
state governments that made extensive use of the Backward Classes category set up commissions to 
identify the Backward Classes. From 1965 to 1976, such commissions reported in Kerala, Jammu 
and Kashmir, Andhra Pradesh, Tamil Nadu, Kerala again, Karnataka, and Gujarat.  135.   
All of these commissions, and the government orders based on their recommendations, use 
communal units to designate the Backward Classes. But Jammu and Kashmir relies heavily on 
occupational and territorial groups, and Kerala employs an income cut­off. The selection of 
communities is more sophisticated: the commissions attempt to assemble (and sometimes generate) 
evidence about occupation, income, and education as well as status and disabilities. Most of them 
strive to eliminate the well­off. In some cases, the number of groups designated is more modest than 
earlier.  136.   Benefits are recommended 
____________________ 
134. See n. 124 above. 
135. Kerala ( Kumara Pillai Commission) 1966; Government of Jammu & Kashmir 1969; Andhra 
Pradesh Backward Classes Commission 1970; Tamil Nadu Backward Classes Commission 1971; 
Kerala ( Backward Classes Reservation Commission) 1971; Government of Karnataka 1975; 
Government of Gujarat 1971. The Bihar report by the Mungeri Lal Commission is described in 
Blair 1980. Two other reports that are not strictly part of this genre might be mentioned in 
connection with these: Punjab Welfare Department 1966 and the report of the Chhedi Lal Sathi 
Commission on the "most backward" classes in U.P., described in Chhotey Lal v. State of U.P., 
A.I.R. 1979 All. 135 [NIC]. 
136. E.g., in Andhra Pradesh the number of groups on the list fell from 139 in 1963 to 112 in 1966 to 
92 in 1970. See State of A.P. v. Balaram, A.I.R. 1972 S.C. 1375 at 1387 ff. 
­184­ 
for a limited span of time, and there is concern about termination and re­assessment. The best of the 
recent commissions display an industry and a sophistication in gathering and analyzing data that is a 
far cry from the uncritical and suppositious compilations of some of the early reports on backward 
classes.  137.   As we shall see, state governments are not always receptive to the critical assessments 
of their commissions. 
Perhaps some of this state activity was in response to the animadversions of the Centre on 
communal criteria, but more directly and palpably it was a response to the courts. In some cases it 
was a response to courts actually striking down schemes; in others, to courts threatening to do so if 
they were not reformed; in still others it was to the implicit threat of litigation. The states aligned 
their schemes with the prescriptions of the courts rather than with the pronouncements of the 
central executive. 
In part this may be because the Centre offered so little, apart from commending the income test. In 
retrospect it is surprising that the Centre never attempted to formulate a workable quantitative 
standard for selecting backward communities ­­ e.g., those groups whose average income and 
average literacy were less than one­half of the state average. It may have been felt that practically 
such a course was foreclosed by the decision to omit castewise data on the censuses of 1951 and 
1961, but the absence of such attempts suggests that the animus against communal tests eclipsed any 
consideration of making them more workable. 
What emerges from the interaction of state governments, commissions, and courts are lists of 
communal groups, with some admixture of geographic and income factors, chosen on the basis of 
low status, low educational attainments, and poverty. We might generalize very tentatively about the 
magnitude of the Backward Classes category: the lists tend to converge on something like the 
second and third lowest quintiles of the population ­­ in terms of "traditional" status ­­ assuming the 
Scheduled Castes and Tribes make up roughly the lowest quintile. Northern and southern states 
arrive at this position from very different starting points. In the peninsula (and in Jammu and 
Kashmir) the commisson process represents a pruning away of the more prosperous and powerful 
groups from a comprehensive list that approximated a regime of communal quotas. In the northern 
tier, however, provision for this stratum is added slowly, first in education, then in government jobs, 
converging on a somewhat similar position. 
Through the course of this development the term "backward classes" has retained a multiplicity of 
meanings. It is used to describe the totality of groups entitled to preferential treatment on the basis 
of their 
____________________ 
137. The methods of the commissions are discussed in chap. 8, §F, below. 
­185­ 
"backwardness"  138.   ­­ i.e., the Scheduled Castes and Tribes as well as "Other Backward Classes" ­­ 
but not those accorded special treatment because of temporary or situational disadvantage (e.g., 
disaster victims, refugees, defense personnel). The term also refers specifically to those backward 
groups other than, the Scheduled Castes and Tribes. There remain fundamental, if rarely articulated, 
disagreements about who these groups are. Some would confine this category to the lowly ­­ those 
"far below" the mean in welfare and resources, or those whose deprivations are comparable to those 
of the Scheduled Castes and Tribes;  139.   others use the term Backward Classes to describe a wide 
middle stratum of Indian society, who require and deserve special help because they are lagging 
behind the most advanced groups.  140.   There is, as we have seen, further disagreement over whether 
the term refers to the less well­off in all communities or whether it encompasses only those 
communities that suffer "backwardness" as a group. 
As we saw in §A, the question of who were the Scheduled Castes was debated and roughly settled 
before Independence within the executive and without the participation of the courts. But who are 
the Backward Classes is a post­Independence question which the constitutional recognition of the 
category made one of all­India scope. The Constitution left the matter with the executive at the state 
level with an option for the Centre to unify it. When the executive at the centre first failed and then 
declined to provide a resolution, the question reverted to the states. In the wake of the Janata victory 
in the 1977 elections, the Backward Classes returned to the national political agenda. Pursuant to its 
electoral promise,  141.   the Janata Government appointed a new Back­ 
____________________ 
138. In this broad meaning it is more or less equivalent to "weaker sections" in current political 
usage. (Cf. the reference to "weaker sections of the people" in Art. 46.) 
139. See chap. 8, §E., 
140. This view is neatly epitomized in the remarks of Shri K. Hanumanthaiya, a Congress M.P. from 
Karnataka, during the debate over the 1976 revision of the lists of Scheduled Castes and Tribes. 
He observed that although he welcomes reservations for the latter, "[i]t is the middle or the 
backward classes that have been completely ignored and they are the people who are suffering 
the most in this set­up . . .it is the middle classes which suffer most during the time of inflation. 
These most neglected middle classes or the backward classes . . .[should] be helped in the spirit 
of the Kalelkar Commission['s] . . .recommendations." LSD (Fifth Series), Vol. 64, No. 16, cols. 
76­77 ( 2 Sep. 1976). 
141. The 1977 election manifesto of the Janata party called for an end to caste distinctions and 
promised the establishment of an independent and autonomous civil rights commission 
"competent to ensure that the minorities, scheduled castes and tribes, and other backward classes 
do not suffer from discrimination or inequality." It promised a radical reduction of disparities by 
a "policy of special treatment" in favor of the "weaker sections of our society." In connection 
with the provision of "preferential opportunities for education and self­employment to these 
sections" the party promised 
­186­ 
ward Classes Commission in February 1978.  142.   Northern state governments under Janata control 
substantially enlarged the preferences for Other Backward Classes, leading to massive violence in 
Bihar. In its declining days, the Janata caretaker government came close to instituting job 
reservations for Backward Classes in central government service.  143.   
During the hiatus of central involvement, what the states did was increasingly subjected to the 
examination of the courts. It has been the Supreme Court rather than the central government which 
has been the unifying and limiting influence, and presumably any new central policy will be shaped 
in light of two decades of judicial predominance in this area. It is to the judicial encounter with the 
Backward Classes category that we now turn. 
____________________ 
to "reserve between 25 and 33 per cent of all appointments to Government service for the 
backward classes, as recommended by the Kelkar [sic] Commission. . . . The Party will 
formulate a special program within the framework of the five­year plan's for the substantial 
advancement of the scheduled castes and tribes and other backward classes and will provide 
adequate funds for the purpose . . . .Special machinery will be set up to implement the 
programme and assure fulfillment of the relevant Constitutional guarantees" (from Commerce, 
Feb. 19, 1977). 
142. The new five­member commission was established in December 1978 under the chairmanship of 
B. P. Mandal, M.P. and instructed to report by the end of 1979. Its terms of reference, 
reminiscent of those of the 1953­55 Backward Classes Commission (I BCC 2) with the addition 
of a specific inquiry about reservation of government posts, were (i) "to determine the criteria 
for defining the socially and educationally backward classes; (ii) to recommend steps to be taken 
for the advancement of the socially and educationally backward classes of citizen so identified; 
(iii) to examine the desirability or otherwise of making provision for the reservation of 
appointments or posts in favour of such backward classes of citizens which are not inadequately 
represented in the services of both the Central and State Government/Union Territory 
Administrations; and (iv) present a report setting out the facts as found by them and making such 
recommendations as they think proper." When asked under what authority this Commission was 
constituted, the Minister of State for Home Affairs replied that it was "constituted under Article 
340 of the Constitution" (RSD Vol. 107, No. 23, cols. 171­73 [ 21 Dec. 1978]). 
143. It is reported that a few weeks before the 1980 elections, the coalition government headed by 
Charan Singh proposed to reserve 25% of central government jobs for Backward Classes. This 
proposal was dropped after the President objected that it violated an agreement that the caretaker 
government would refrain from taking Policy decisions which might amount to electoral 
initiatives ( Overseas Hindustan Times, Dec. 27, 1979). 
­187­ 
7 The Backward Classes and the Judiciary: The Selection of Beneficiary Classes 

From this point forward the analysis focusses on the work of the courts. Readers who 
are unfamiliar with courts and litigation in India may find it helpful, before proceeding, 
to took ahead to chapter 14. A quick acquaintance with the structures and operations of 
the courts, the system of authority and precedent, litigation patterns, and the legal 
profession may be obtained by reading the opening pages of each of the sections of 
chapter 14.
A. CASTES AS CLASSES 
WE TURN NOW to the emergence of the doctrinal frame which shaped, and was in turn shaped by, 
the policy developments recounted in chapter 6. One of the crucial tasks of this doctrine is to assign 
a meaning to the term "classes" as it appears in the constitutional authorization of preferential 
treatment. To make sense of the course of development it may be helpful to identify three basic and 
recurrent views of the constitutional meaning of "classes," views that reflect more comprehensive 
inclinations about equality and society in India. First, there is what we might label the "modernist" 
view, which would confer preferential treatment on lines of income, occupation, illiteracy, or other 
"neutral" or "secular" characteristics and would eschew the communal groupings whose use is 
otherwise forbidden in the Constitution. Thus "classes" would be collections of individuals selected 
for special treatment because they suffered one or another form of disadvantage. The opposite 
position is taken by proponents of what may be labelled the "historic" view. In this view the 
constitutional provisions for backward classes aim at overcoming the cumulated disparities of 
power, wealth, and culture among historic communities. "Classes" then refers only to 
­188­ 
those historic social formations (caste, and religious groups, perhaps occupations and localities) that 
were the object of constitutional policy. Finally, there are adherents of an "elastic" position, who 
regard "classes" as broad enough to include both historic social formations and ad hoc categories. In 
this view government is neither commanded to use one of these nor excluded from using the other. 
These views are introduced here as a mnemonic convenience. In the following account, which is 
arranged along other lines, we shall encounter some of the arguments of constitutional history, 
textual interpretation, and policy consequences that are advanced on behalf of each. Later we shall 
consider the more comprehensive positions as to the pursuit of equality in Indian society that are 
implicated in these competing readings of the constitutional provisions. 
The Constitution names as the permissible recipients of preferences, not backward individuals or 
families, nor backward castes, religious communities, occupational or regional groups, but 
backward "classes" of citizens. These "classes" are restricted neither to economic classes nor to 
1. The term seems to be used in the broad 
classes in the senses familiar to modern social science.    
connotation of any group of persons having certain common characteristics. In particular it would 
seem to include, though it is not confined to, those classifications otherwise forbidden in Articles 
15, 16, and 29(2)­­e.g., racial, religious, and caste groups. This is so because Articles 15(4) and 
16(4) are exceptions or provisos to these Articles, limiting the operation of their provisions. If the 
proviso that "nothing in this article" shall prevent the State from making preferences for backward 
classes is not broad enough to authorize the use of the forbidden classifications, then it was 
unnecessary to have any proviso at all, since other classifications would be permissible without it. 
Furthermore, the history of Articles 15 (4) and 16 (4) indicates that they were included for this 
purpose. The entire debate on Article 16 (4) in the Constituent Assembly revolved around the 
question of which "communities" were intended to be included.     2. The addition of Article 15(4) by 

amendment was for the specific purpose of permitting preferences for caste groups.     3. 

In a discussion of the use of caste in designating Backward Classes, there are two related and easily 
confused but distinguishable questions. First, may castes or communities be used as the units or 
classes that are designated as backward? Second, may the rank, standing, or prestige of a caste 
group be used as a measure or criterion of its backwardness? The 
____________________ 
1. Cf. Article 23(2), where class is used in what appears to be an economic sense. 
2. VII CAD 673­702, discussed in chap. 6 above. 
3. Discussed in chap. 6, above, and chap. 11, below. 
­189­ 
latter points to "caste" as the measuring rod; the former points to "castes" as the things to be 
measured. 
This distinction, emphasized here for purposes of clarity, introduces an unavoidable element of 
artificiality into our analysis of the cases to the extent that courts, along with the other participants, 
continually discuss the "caste test" or "caste criterion" as a unity. As the passages quoted in the 
ensuing discussion indicate, usually it is either caste as a unit or caste as a status that is meant. But 
there are instances in which both meanings are fused. That the courts have succeeded in clarifying 
the distinction only partially and after a long conceptual struggle points to the significance of this 
fusion;  4.   even a body so detached and devoted to analytical distinctions as the judiciary has found it 
difficult to envision even momentarily a caste group abstracted from its status in a ranked order of 
groups.  5.   
The courts, along with government agencies and the Backward Classes Commission, have accepted 
caste as a permissible unit of classification. In Venkataramana v. State of Madras, the Supreme 
Court upheld reservations for "backward Hindus," a list of castes designated as backward by the 
Madras Govemment.     6. The series of cases arising out of Mysore's expansive list of Backward 
Classes confirms this. Although the reservations were invalidated as arbitrary in Ramakrishna 
7. It 
Singh v. State of Mysore, the court indicated that caste was a permissible basis of classification.    
was argued that "class" in Article 15(4) excluded "caste" by implication, since that Article specified 
"Scheduled Castes" as opposed to "backward classes." Thus,, it was argued, the prohibition on caste 
differentia in Article 15 (1) should not be read as qualified by Article 15(4)in toto but only by that 
8. The Court held that "class" included 
part of Article 15(4) which refers to Scheduled Castes.    
persons 
____________________ 
4. Although the courts have not clarified the distinction, it has not escaped the most perceptive 
students of the Supreme Court's decisions in these areas. Thus L. G. Havanur ( 1965:60) 
explains: "The word 'Caste' can be used in two senses­­one in the concrete and the other in the 
abstract. In the concrete sense 'a caste' means a group of persons or families. In the abstract sense 
'the caste' means status or rank or position" (italics in original). Cf. Radhakrishnan 1966. 
5. The difficulty many Indians experience in separating reference to a caste group from a notion of 
standing or rank became more comprehensible to me when I imagined some American analogs. 
We find it easier to isolate terms describing occupational, religious, and ethnic groups from 
notions of relative rank, but it is difficult for an American to abstract the term "working class" or 
"upper middle class" from its comparative and relative connotations. Similarly, a knowledgeable 
baseball fan might find it difficult to envision a set of teams totally abstracted from their relative 
standing in the league. 
6. A. I. R. 1951 S.C. 229. 
7. A.I.R. 1960 Mys. 338. 
8. It should be noted that the text of Art. 16(4) supplies no ground for such a distinction. 
­190­ 
grouped together on the basis of their castes, though it would certainly be open to the government to 
determine the classes on any other basis. In Partha v. State of Mysore the court, upholding this 
aspect of the state scheme, indicated that the caste "classification will only be open to challenge if it 
can be shown that the criterion adopted for determining their backwardness is useless as a test of 
backwardness, so that the preference would amount to a preference on grounds of caste alone." 9. 
That is, it was permissible to use caste units, but there had to be some criterion of backwardness in 
addition to caste rank or standing. 
The law set forth in these high court cases has remained essentially unchanged. But a series of 
notable Supreme Court cases refined it and widely conveyed a notion of substantial change. In 
Balaji v. State of Mysore the Supreme Court for the first time explicitly dealt with the question of 
the permissibility of the use of caste as a unit.  10.   There are a number of places in its opinion where 
the Court is critical of the use of caste as a criterion of backwardness. But it is quite clear on close 
inspection that the Court puts forward no constitutional objection to the use of caste or community 
as a unit of designation, although it deplores this use on policy grounds. In its examination of the 
state's classification of educationally backward communities, the Court clearly accepts the use of 
"communities"­­including caste groups­­as the units whose backwardness is to be determined. It 
concludes that "only those communities which are well below the State average [of literacy] can 
properly be regarded as educationally backward classes of citizens."  11.   The Court takes strenuous 
objection to the exclusive reliance of Mysore on caste (i.e., in the sense of rank or standing) as a 
criterion of social backwardness and notes that occupations and place of habitation are also relevant 
to determination of the backwardness of "a community of persons."  12.   
It is clear that communities are permissible units: "It is for the attainment of social and economic 
justice that Article 15 (4) authorizes the making of special provisions for the advancement of the 
communities there contemplated, even if such provisions may be inconsistent with the fundamental 
rights guaranteed under Article 15 or 29(2)."  13.   The communities that may be used appear to be 
broader than castes in the narrow sense, but include all communities­­for one of the Court's 
objections to emphasis on caste is its inapplicability to non­Hindu groups.  14.   
____________________ 
10. A.I.R. 1963 S.C. 649. A decade later Balaji was still referred to respectfully as the. locus  
classicus of learning on Backward Classes ( Janki Prasad Parimoo v. State of Jammu and 
Kashmir, A.I.R. 1973 S.C. 930 at 937). 
11. Id., 660. 
12. Id., 659. 
13. Id., 664 (italics added). 
14. See chap. 8, §B, below. 
9. A. I.R. 1961 Mys. 220 at 230. 
­191­ 
In short, Balaji seems to leave intact caste or communal units as the things to be measured, while 
confining the role of caste as the measuring rod.  15.   However, the failure to make clear the 
distinction between castes as units and caste rank as a measure of backwardness put both uses of 
caste in some disrepute and encouraged the notion that "caste" was in all respects eliminated from 
the selection of Backward Classes. 
As a result of the Balaji case, Mysore once more revised its scheme for reservations in technical and 
medical colleges. It deemed as Backward Classes all persons following the occupations of 
cultivator, artisan, petty businessman, inferior service, or other occupations involving manual labor 
and with an annual income of less than Rs. 1,200. It reserved 30% of the seats for these. Upholding 
this new order, the Mysore High Court expressed grave doubts about this classification, which it 
found to be far from satisfactory. It read Balaji to indicate that there could be no satisfactory 
classification of Backward Class Hindus while ignoring the caste basis. "[I]n addition to the 
'occupation' and 'poverty' tests, the State should have adopted the 'caste' test."  16.   Pointing out the 
difficulties of the income/occupation test and its detrimental effect on the most socially backward 
castes, the High Court concluded with the hope that the state would make "a more appropriate 
classification lest its bona fides should be questioned."  17.   
This case went up on appeal on another question, and appellant's counsel urged the Supreme Court 
to repudiate these observations of the High Court as conflicting with the Balaji case. Surprisingly, 
the Court complied in a series of dicta that undermined the limited clarity attained by Balaji. 
In Chitralekha v. State of Mysore, the Supreme Court finds that Balaji established that (1) "the caste 
of a group of citizens may be a relevant circumstance in ascertaining their social backwardness" and 
(2) "it cannot be the sole or dominant test in that behalf."  18.   It then (3) repudiates the High Court's 
view that caste is a necessary test of social backwardness: 
We would hasten to make clear that caste is only a relevant circumstance in ascertaining 
the backwardness of a class and there is nothing in [ Balaji] . . . which precludes the 
authority concerned from determining the social backwardness of a group of citizens if 
it can do so without reference to caste. While this court has not excluded caste from 
ascertaining the backwardness 
____________________ 
15. "What was questioned in [ Balaji] . . .was not . . .the taking of a caste . . .or a section of a 
community as a unit . . . .But the principle . . .of taking the status of a community or a caste in 
the hierarchy of castes as the determining factor of its backwardness . . . ." ( Havanur 1965: 
100­1; italics in original) 
16. Viswanath v. Government of Mysore, A.I.R. 1964 Mys. 132 at 139. 
17. Id., 140 
18. A.I.R. 1964 S.C. 1823 at 1833. 
­192­ 
of a class of citizens, it has not made it one of the compelling circumstances. . . . [T]he 
authority concerned may take caste into consideration . . .but, if it does not, its order 
will not be bad on that account. . . .  19.   
So far the Court has only added to Balaji the notion that caste is a permissible but not a mandatory 
measure of social backwardness. It has not said anything about what are the "classes of citizens" 
whose backwardness is to be measured­­which, as we have seen, Balaji tacitly permitted to be caste 
or communal units. 
Subba Rao, J., then presents some reflections about the use of caste units which compound the 
ambiguities: 
The important factor to be noticed in Article 15(4) is that it does not speak of castes, but 
only speaks of classes. If the makers of the Constitution intended to take castes also as 
units of social and educational backwardness, they would have said so. . . . Though it 
may be suggested that the wider expression "classes" is used . . . as there are 
communities without castes, if the intention was to equate classes with castes, nothing 
prevented the makers of the Constitution to use the expression "Backward Classes or 
castes." The juxtaposition of the expression "Backward Classes" and "Scheduled 
Castes" in Article 15(4) also leads to a reasonable inference that the expression "classes" 
is not synonymous with castes. It may be that for ascertaining whether a particular 
citizen or group of citizens belong to a backward class or not, his or their caste may have 
some relevance, but it cannot be either the sole or the dominant criterion for 
ascertaining the class to which he or they belong.  20.   
He suggests that this restricted use of "caste" is necessary to prevent exploitation of these provisions 
by well­off sections within groups that are largely backward. 
If we interpret the expresson "classes" as "castes," the object of the Constitution will be 
frustrated and the people who do not deserve any adventitious aid may get it to the 
exclusion of those who really deserve it. This anomaly will not arise if, without equating 
caste with class, caste is taken as only one of the consideration to ascertain whether a 
person belongs to a backward class or not. On the other hand, if the entire sub­caste, by 
and large, is backward, it may be included in the Scheduled Castes by following the 
appropriate procedure laid down by the Constitution.  21.   
____________________ 
19. Id., at 1833. 
20. Id., at 1833­34. 
21. Id., at 1834. The argument that if the whole "sub­caste" is backward it should be added to the list 
of Scheduled Castes is elaborated by Tripathi ( 1972 : 204), who portrays the legitimate use of 
caste as exhausted by the provisions for identification of the Scheduled Castes: "castes which 
ought, as such, to be regarded as backward have 
­193­ 
With a bow to the complexity of the problem, the Court concluded: 
[W]hat we intend to emphasize is that under no circumstance a "class" can be equated to 
a "caste" though the caste of an individual or a group of individuals may be considered 
along with other relevant factors in putting him in a particular class.  22.   
The practical policy that the Court tells us it is trying to effect­elimination of the well­off within the 
backward group­­could of course be achieved as readily by requiring income ceilings. But the Court 
seems to go much further than addressing this problem; it comes close to saying), that caste units 
cannot be used at all­­although it does not explain what sort of "group of individuals" it has in mind. 
Caste in the sense of rank or standing may be part of the measuring rod, caste in the sense of unit 
cannot be the thing that is measured.  23.   
The argument is flagrantly unhistorical. Whatever else classes have included in Indian 
jurisprudence, they have clearly included caste and communal units.  24.   Furthermore, the textual 
argument is unconvincing, since it would either undermine the well­established meaning of 
"backward classes" in Article 16(4), which clearly includes the Scheduled Castes, or lead to the 
conclusion that the "backward classes" in Article 16(4) are different than those referred to in Article 
15(4). 
The Chitralekha judgment reinforced the popular reading of Balaji as completely outlawing the 
caste basis. The national press editorialized with satisfaction about the demise of the caste test of 
backwardness. 
____________________ 
already been included in the category of scheduled castes and have been expressly mentioned in . 
. . [Article 15 (4)]; therefore the rule of expressio unius est exclusio ulterious must apply, and the 
expression 'classes' in the sub­clause must be construed as contemplating groups other than 
caste­group." This piece of interpretation not only involves departures from the original (and 
subsequent) understanding of Art. 15 (4), but it leads to an unseemly textual tangle, for 
"backward classes" in Art. 16(4) would then mean something very different from the "backward 
classes" referred to in Art. 15(4). 
22. Id., at 1834. The warning that "caste" is illegitimately used as a "class" because it cannot be 
"equated" with class seems to imply that no term narrower than class may be used, for by being 
narrower it would not be equatable. But the question is really not what is equatable with class, 
but what is included within it! 
23. Mudholkar, J., dissenting at some length on the interview question, joins the majority in 
condemning the observations of the High Court Judge­­but for exactly the opposite reasons. He 
agrees that the presence of Art. 15(4) as an exception to Arts. 15(1) and 29(2) "does not justify 
the inference that castes have any relevance in determining what are socially and educationally 
backward communities" (A.I.R. 1965 S.C. at 1843). Apparently he accepts communities as the 
units, but implies that Arts. 15 and 29 bar any use of caste rank as a measure of backwardness. 
24. Havanur ( 1965) not only points to textual evidence (the heading of Part XVI, the wording of Art. 
16(4)), but provides numerous instances of previous usage from the Constituent Assembly, the 
High Courts and Privy Council, official documents, and political leaders. See also 
Radhakrishnan 1965. 
­194­ 
The impact of these decisions in the courts was more complex. The success of the petitioners in 
Balaji had stimulated litigation in several states where there were extensive reservations for 
Backward Classes­especially Andhra Pradesh and Kerala. 
The first court to which Chitralekha's dicta on caste were cited was the Division Bench of the 
Kerala High Court in State of Kerala v. Jacob Mathew.  25.   A single judge had struck down the state 
government's order specifying Ezhuvas and Muslims as Backward Classes on the round that "the 
predominant if not the sole test [of social and educational backwardness] that has weighed in the 
mind of the State Government is the test of caste or community," and that this is impermissible. 26. 
Here, "no investigation has been made, or data or materials collected, for the purpose of enabling 
the State Government to consider . . . which particular groups or sections . . .have to be treated as 
backward classes. . . ."  27.   On appeal to a Division Bench, respondents urged that Chitralekha 
(which had been decided in the interim) made it impermissable to use caste units in specifying 
Backward Classes. The Division Bench declined to read it that way, and taking Chitralekha with 
Balaji, found nothing which 
precludes the conclusion that if the whole or a substantial portion of a caste is socially 
or educationally backward, then the name of that caste will be a symbol or a synonym 
for a class of citizens who are socially and economically backward and thus within the 
ambit of [Article 15(4)].  28.   
The Division Bench proceeded to restore the reservations for Ezhuvas, Muslims, and Latin 
Catholics. This case was admitted on appeal to the Supreme Court, but was withdrawn in the course 
of oral argument on the understanding that the state was reviewing its policy on educational 
reservations. A Commission established for this purpose reported in late 1965. It recommended, and 
the Government adopted, a scheme of reservations for the same backward communities, but now 
qualified by an income ceiling.  29.   The "means­cum­community" test was subsequently upheld 
against the challenge of an applicant who satisfied the means test but not the community test. 30. 
Communal units without an income ceiling continued to be used in distributing reserved posts in 
____________________ 
25. I.L.R. 1964(2) Ker. 53. 
26. A.I.R. 1964 Ker. 39 at 57. 
27. Id., at 58. 
28. I.L.R. 1964(2) Ker. 53 at 60. Mayakrishnan and Radhakrishnan ( 1965 : 28) observe that the 
Division Bench "failed to appreciate" that Chitralekha was irreconcilable with Balaji on this 
point. 
29. Kerala (Kumara Pillai Commission) 1966­­known as the Kumara Pillai Commission after its 
chairman, a retired High Court Judge. 
30. Laila Chacko v. State of Kerala, A.I.R. 1967 Ker. 124. 
­195­ 
government service. This was challenged and reluctantly upheld in Hariharan Pillai v. State of 
Kerala.  31.   
The Andhra High Court responded to the Balaji and Chitralekha cases in a very different fashion. 
The long­standing castewise list of Backward Classes had gone unchallenged until Balaji, Indeed, 
the government in 1963 added two castes of doubtful backwardness to the list. This was challenged, 
and a single judge struck down the reservations in educational institutions on the ground that the list 
was prepared entirely on the basis of caste.  32.   The state government then dramatically shifted its 
policy on preferences: it withdrew the list of Backward Classes, adopted an economic criterion for 
educational and other benefits, and eliminated reservations in professional colleges and government 
employment. This withdrawal was then challenged by members of the erstwhile Backward Classes, 
who sought the restoration of the list and of the reservations. In Dasa Rayadu v. Andhra Pradesh 
Public Service Commission, the High Court denied this on the ground that reservations were 
permissible but not mandatory.  33.   Adopting the Balaji­Chitralekha position that caste is a relevant 
but not a necessary factor in determining backwardness, the Court then quotes Justice Subba Rao's 
broader dicta on the distinction between caste and class and infers that 
[D]etermination of backward classes on the basis of castes would therefore be not only 
derogatory of Articles 15 and 16 but would also go quite contrary to the avowed 
principles enunciated in the Constitution. . . .  34.   
It is not clear whether the Court really thinks it is impermissible to use caste units if there is some 
way other than caste rank of measuring their respective backwardness. 
In the meantime, political pressure from the erstwhile Backward Classes led to the formation of a 
Cabinet Sub­Committee to re­examine the matter. On the basis of this Committee's finding, the 
reservations were restored and the Backward Classes defined much as before in a list of 119 castes. 
This list, too, was struck down by the High Court, but this time the Court's emphasis was on the 
absence of adequate findings of backwardness, not on the use of communal units, which were, 
implicitly, permissible to use.  35.   
Thus, before the issue returned to the Supreme Court, we see the Kerala and Andhra Pradesh 
judiciaries displaying contrasting styles in handling the litigation stimulated by Balaji.  36.   The 
Kerala Court took a 
____________________ 
31. A.I.R. 1968 Ker. 42. 
32. Sukhdev v. Government of Andhra Pradesh, 1966(1) Andh. W.R. 294 (1963). 
33. A.I.R. 1967 A.P. 353. 
34. Id., at 361. 
35. Sagar v. State of Andhra Pradesh, A.I.R. 1968 A.P. 165. 
36. The only other state in which litigation was stimulated by the Supreme Court's decision in Balaji 
was Bihar, where it was more sparse and intermittent. In Nanda KishorSharma v. State of Bihar 
­196­ 
soft line, cautious and deferential to the state government, applying pressure for rationalization and 
elimination of anomalies while assuming responsibility for the proper working of the system of 
preferences. The Andhra Court, on the other hand, takes a more detached and indifferent stance. 
The difference in judicial response corresponds to the contrasting fashions in which the state 
governments handled the matter. The Kerala Government was consistent and determined in its 
policy; the Andhra policy was subject to sharp vacillations, which exposed its character as a product 
of shifting political forces rather than an expression of basic state policy. In Kerala, the 
government's position was founded on the careful and workmanlike report of the Kumara Pillai 
Commission, marshalling relevant data and arguments;  37.   in Andhra, the government's policy was 
unsupported by data, and the court was left with no choice between swallowing it completely or 
overturning it completely. In spite of the differences in style and result, neither high court relied 
upon the animadversions of Chitralekha against caste units. Building cautiously on Balaji, the 
Kerala Court completely ignored this aspect of Chitralekha. In Andhra the question of caste units 
was relegated to the background, and other objections to the government's scheme brought it down. 
In a series of cases decided in early 1968, the Supreme Court moved somewhere closer to 
articulating the distinction between caste units and caste rank. In P. Rajendran v. State of Madras the 
Court upheld the use of castes as the units by which Madras's Backward Classes were designated: 
[I]t must not be forgotten that a caste is also a class of citizens and if the caste as a 
whole is socially and educationally backward, reservation can be made in favor of such a 
caste on the ground that it is a socially and educationally backward class within the 
meaning of Article 15(4).  38.   
____________________ 
Sharma v. State of Bihar, A.I.R. 1965 Pat. 372, a 10% reservation for Backward Classes in 
medical college admissions was held invalid on the ground that the determination of 
backwardness was, contrary to Balaji, solely on the basis of caste. Reservations for Backward 
Classes remained in force, however. See Umesh Chandra Sinha v. Singh, A.I.R. 1968 Pat. 8. In 
late 1967, a Full Bench had no difficulty in finding the principle of Balaji inapplicable in a case 
where a notoriously "backward community is described conveniently by its caste name." Chait 
Ram v. Sikander, A.I.R. 1968 Pat. 337 at 339; Hridaya Narain Singh v. Mohd. Sharif, A.I.R. 1968 
Pat. 296. 
37. The Kerala Court had the additional advantage that the number of groups involved was small and 
encompassable. The judges themselves had a sense of who the groups were and what their 
situation was. See chap. 8, §F, below. 
38. A.I.R. 1968 S.C. 1012 at 1014­15. Ghouse ( 1969) finds P. Rajendran "irreconcilable" with Balaji 
and Chitralekha and calls for its repudiation. It appears, however, that he confounds the remarks 
of Chitratekha against caste units with the Balaji holding that caste standing is unacceptable as 
an exclusive test of backwardness. From the point of 
­197­ 
It is clear that what the Court is upholding here is the caste as the unit, not caste standing as the 
measure of backwardness. Listing by caste, the Court continues, "does not necessarily mean that 
caste was the sole consideration" in finding these groups backward.  39.   
In State of Andhra Pradesh v. Sagar, the Court confirmed its awareness of the distinction, but 
succeeded in blurring it once more. It found that in the context of Article 15 (4), 
"class" means a homogeneous section of the people grouped together because of certain 
likenesses or common traits and who are identifiable by some common attributes such 
as status, rank, occupation, residence in a locality, race, religion and the like. In 
determining whether a particular section forms a class, caste cannot be excluded 
altogether. But in the determination of a class a test solely based upon the caste or 
community cannot also be accepted.  40.   
Shah, J.,'s qualification that "determination of a class" cannot be "solely based upon the caste" is 
obscure and puzzling. Apparently, the question of what is a class ­­ to which caste as a unit is 
relevant ­­ is not yet severed from the question of the standards to be employed in measuring 
backwardness s ­­ to which caste as a status or rank is relevant. 
The position is clarified by the decision, if not by the language, in Triloki Nath Tiku v. State of 
Jammu and Kashmir, which came before the Supreme Court twice. The state provided reservations 
for "Muslims of Kashmir" and "Hindus of Jammu." In 1966 the Supreme Court remanded the case 
to have the High Court determine whether the communities named were backward and ordered it to 
submit a report on their population, backwardness, and representation in the services.  41.   In 1968, 
when the Supreme Court took up the matter again,  42.   it was careful to specify that Backward 
Classes were not synonymous with castes or communities. Members of an entire caste or 
community 
may in the social, economic, and educational scale of values at a given time be 
backward and may on that account be treated a backward class, but that is not because 
they are members of a caste or community, but because they form a class.  43.   
____________________ 
view that regards such a distinction as unwelcome, P. Rajendran marked a "setback" to the "new 
trend" towards vindicating Art. 15(1) rights ( Tripathi 1972: 202). 
39. Id., at 1015. 
40. A.I.R. 1968 S.C. 1379 at 1382. 
41. A.I.R. 1967 S.C. 1283. The judgment was written by, Subba Rao, J., who did not see fit to advert 
to his statements in Chitralekha about the impermissibility of communal units. 
42. A.I.R. 1969 S.C. 1. 
43. Id., at 3. 
­198­ 
After defining class as a "homogeneous section" as in the Sagar case, quoted above, the Court 
cautions that 
for the purpose of Article 16(4) in determining whether a section forms a class, a test 
based solely on caste, community, race, religion, sex, descent, place of' birth or 
residence cannot be adopted, because it would directly offend the Constitution. . . .  44.   
Justice Shah's puzzling proviso appears here in a more elaborate from, listing all of the 
classifications forbidden in Article 16 (2). The Court has just acknowledged that a community or a 
caste may constitute a Backward Class. So its observation that caste (and the other forbidden 
criteria) cannot be the exclusive test of what is a class appears, then, to amount to no more than 
saying that the group taken as a class must share traits of backwardness as well as traits giving it an 
identity as a class.  45.   The proviso adds little to the discussion of what is a class, though it may have 
implications as to the permissble tests of backwardness. As we have seen, "caste" has different 
connotations in the context of the questions "what is a class" and "what is backwardness." In 
relation to the former it means a unit or community; in relation to the latter it means a status or rank. 
The Court may be implying that caste in the sense of status cannot be the exclusive test used to 
determine the backwardness of the class (i.e., the caste as a unit). 
These formulae from Sagar have become a judicial commonplace, recited at length in all 
discussions of backward classes. But there has been an interesting development in the way in which 
the language is interpreted and applied. In Periakaruppan v. State of Tamil Nadu, justice Hegde (as 
he then was) added a straightforward reading: "a caste has always been recognized as a class, "and 
cites P. Rajendran as standing for the proposition that caste units are permissible so long as they are 
shown to be socially and educationally backward.  46.   In State of AndhraPradesh v. Balaram
____________________ 
44. Id., at 3. 
45. Superficially, the Court appears to mean that, quite apart from the question of backwardness, 
caste (or the other forbidden classifications) cannot be employed to identify a class of potential 
recipients, unless some additional differentia is also applied. That it does not mean this is 
suggested by the Court's acknowledgement that it is discussing class "for the purpose of Article 
16(4)"­­that is, a Backward Class. The Court imports into the question of what is a class the 
standard of "no exclusive reliance on caste" developed in answering the question of what is 
backwardness because it does not distinguish the two questions. Furthermore, the additional 
differentia interpretation is textually suspect, since Art. 16(4) is an exception to Art. 16(2), which 
forbids classification on grounds of "caste . . . only." If a caste cannot be a class unless there is an 
additional differentia, Art. 16(4) would be redundant. 
46. A.I.R. 1971 S.C. 2303 at 2310. 
­199­ 
Pradesh v. Balaram, a two­judge bench reiterates that caste units are permissible: "one thing is clear 
that if an entire caste is, as a fact, found to be socially and educationally backward, their inclusion in 
the list of Backward Classes by their caste name is not violative of Article 15(4)."  47.   Furthermore, 
the problem of over­inclusion is not fatal to such a classification: a caste may be a backward class 
notwithstanding the presence in it of a "few individuals [who are] . . . both socially and 
educationally above the general average."  48.   
In Janki Prasad Parimoo v. State of Jammu and Kashmir, a Constitution Bench of the Supreme 
Court cites the Sagar "homogeneous section" language in the course of applying tests of 
backwardness to groups selected on the basis of residence, present occupation, grandfather's 
occupation, caste, and other characteristics.  49.   Like its predecessors, the Court has no difficulty 
with caste units: the only problem with these is the absence of adequate material on 4 of the 23 
caste groups named. But other groupings give rise to difficulty­­i.e., the state's listing of small 
cultivators and holders of small pensioners as backward classes is troubling to the Court. 
The Court finds it erroneous to place "economic consideration above considerations which go to 
show whether a particular class is socially and educationally backward."  50.   Neither small 
cultivators nor small pensioners is a class in the sense of "a homogeneous social section . . . with 
common traits and identifiable by . . . common attributes."  51.   Instead, these are "artificial groups" 
 52.   created by the state for the purpose of receiving these benefits. The nub of the Court's difficulty 
is revealed in its discomfort with the notion that one cultivator or pensioner might qualify as 
backward while his brother with a slightly large holding or a slightly higher pension would not.  53.   It 
strikes the Court as anomalous that those with similar early circumstances and formative 
experiences would end up on different sides of the backward line. Although the Court does not 
explain its unease, what seems implicit is the notion that what is supposed to be remedied by 
provision for 
____________________ 
47. A.I.R. 1972 S.C. 1375 at 1396. 
48. Id., at 1395; cf. Chait Ram v. Sikander, A.I.R. 1968 Pat. 337 at 339. 
49. A.I.R. 1973 S.C. 930. 
50. Id., at 941. 
51. Id., at 941. 
52. Id., at 942. 
53. Id., at 941. The notion that the units which are "socially backward" should not cut across kinship 
lines surfaces elsewhere. In Pradip Tandon v. State, A.I.R. 1975 All. 1,7, the Full Bench observes 
that where one branch of a family lives in the city and the other remains in the rural area, both 
branches remain the same "as far as social habits and norms . . . prevailing in the community to 
which the family belong . . . Then, to say that the branch which resides in urban centre is socially 
forward while the branch that resides in rural centre is socially backward would be 
contradictory." In Shetty v. State of Mysore, A.I.R. 1969 Mys. 48, 51, the Court observes that it is 
not possible to envisage a situation in which one member of a family is treated as backward and 
another is not. 
­200­ 
backward classes is not incidents of individual life history but the accumulated disadvantages of 
palpable social groups. Presumably, these groups may be castes, or territorial groups, or the 
followers of an occupation, but they must have some shared traits which can be measured in terms 
of social and educational backwardness. One wonders whether a group chosen by the income test 
commended by Balaji could pass muster with the Janki Prasad Court. 
This line of development is broken by some judicial observations in State of U.P. v. Pradip Tandon. 
 54.   In the course of addressing the measure of backwardness, the Court displays the full potential for 
confusion that is entailed by failure to distinguish the measurement question from the units question 
and to explicate the dual usage of caste. The Court asserts that "caste cannot be made one of the 
criteria for determining social and educational backwardness. . . . The socially and educationally 
backward classes of citizens are groups other than groups based on caste."  55.   In a particularly 
puzzling redaction of the Sagar "homogeneous section" language, the Court pronounces: 
The homogeneity of the class of citizens is social and educational backwardness. 
Neither caste nor religion nor place of birth will be the uniform element of common 
attributes to make them a class of citizens.  56.   
This phrasing seems the product of confusion rather than a desire to abandon existing learning on 
the matter. In any event, these observations are pure dicta as to caste units, since the scheme before 
the Court did not involve any communal units. 
The same standards that apply to caste units also apply to the use of religious communities as units. 
Preferences for religious groups have been struck down in a number of cases  57.   and upheld in others 
 58.   without 
____________________ 
54. State of U.P. v. Pradip Tandon, A.I.R. 1975 S.C. 563. 
55. Id., at 567 
56. Id., at 567. 
57. In Venkatramana v. State of Madras, A.I.R. 1951 S.C. 229, the Supreme Court rejected the 
inclusion of Muslims and Christians as Backward Classes but was silent as to the permissibility 
of using religious criteria. In State of Jammu and Kashmir v. Jagar Nath, A.I.R. 1958 J. & K. 14, 
aff'g A.I.R. 1958 J. & K. 1, a cabinet order authorizing direct appointment of Muslims to certain 
posts to remove the communal disparity was held void, since there was no designation of the 
preferred group as backward. Reserved seats for Muslims and Indian Christians on the Madras 
Corporation Council were held invalid in A.R.V. Achar v. State of Madras (High Court at Madras, 
Aug. 25, 1952, aff'd on other grounds, A.I.R. 1954 Mad. 563). Ramakrishna Singh v. State of 
Mysore, A.I.R. 1960 Mys. 338, invalidated a scheme of reservations in professional colleges 
which included reservations for Sikhs, Muslims, Jains, and Indian Christians. However, the Court 
implied that the Government was free to use religious as well as caste criteria in designating 
Backward Classes (id., 345). 
58. In Keseva Iyengar v. State of Mysore, A.I.R. 1956 Mys. 20. religious groups were allowed 
without discussion. (This case is now seriously discredited on other 
­201­ 
reaching the question of whether an entire religious community might be designated as backward. 
But from Balaji on, the Supreme Court implicitly treats the religious group on a parity with the 
caste group for this purpose. In Balaji the Supreme Court, while rejecting the inclusion of the 
Muslim community, indicates that preferences may be made for properly selected "communities." 
 59.   Indeed, since one argument for the rejection of the "caste" criterion is its inapplicability to non­
Hindu groups like Christians, Muslims, and Jains,  60.   it would appear that these religious 
communities could be the units or classes of backward citizens in the same fashion as caste groups. 
In Trikoll Nath Tiku v. State of Jammu and Kashmir [II] the Supreme Court struck down 
reservations for "Muslims of Kashmir" and "Hindus of Jammu" on, inter alia, the ground that the 
selection "could not be deemed to have been made on the backwardness of [these] classes." 61. 
Clearly, the Court reads "class" broadly enough to include religious communities. Members of an 
entire caste or community may be "treated as a backward class" because of their shared 
characteristics.  62.   
Although caste and community are permissible units, their use in determining the Backward Classes 
stands on a different footing than their use in designating Scheduled Castes. In the case of the 
Scheduled Castes, the caste unit is explicitly authorized by the Constitution, and its use is confined 
to the President and Parlia'ment.  63.   In the case of Backward Classes, it enjoys no such explicit 
constitutional sanction nor is its use similarly confined to the highest central authorities. Backward 
Classes for the purpose of a particular measure may be defined not only by central and state 
legislatures but by administrative departments and possibly even local authorities  64.   ­­agencies 
whose accountability is more parochial and who are more likely to share, or at least be responsive 
to, local prejudices and pressures. Scheduled Castes, for all the vagaries of the definition process, 
are composed of a limited number of definable groups, already in existence and forming a minority 
of the total population. But Backward Classes make up no such definite group, for it is a category 
indefinitely expandable, which might include a majority as well as a minority.  65.   The distribution 
and social position 
____________________ 
grounds. See chap. 8, §F, and chap 12, §B, below.) In State of Kerala v. Jacob Mathew, I.L.R. 
1964 (2) Ker. 53, Muslims and Latin Catholics were included among Backward Classes without 
discussion of this point (rev'g on other grounds A.I.R. 1964 Ker. 39, where these groups were 
disallowed). In Hariharan Pillai v. State of Kerala, A.I.R. 1968 Ker. 42, Muslims, Latin 
Catholics, and Backward Christians were included among the Backward Classes, while the Court 
observed that "classification should not be made on the basis of religion or caste" (id., 50) 
59. A.I.R. 1963 S.C. at 661. 
60. Id., at 659. 
61. A.I.R. 1969 S.C. 1 at 3. 
62. Id., at 3. 
63. Art. 341. 
64. See chap. 6, above. 
65. See, e.g., the situation in Mysore where Backward Classes included the 
­202­ 
of Scheduled Castes makes it unlikely that they could exercise a preponderant influence on the 
central government, but it is not improbable that Backward Classes could exercise such influence at 
the local or the state level.  66.   Different dangers are involved and different safeguards are required. 
Thus, judicial reluctance to review the exercise by the highest central authorities of their exclusive 
power to designate Scheduled Castes implies no similar deference to local agencies designating 
Backward Classes. 
In designating the Scheduled Castes not only may the President and Parliament use caste as a unit, 
but it was generally anticipated that they would use caste (in the sense of rank or standing) as the 
major criterion of inclusion. Scheduled Castes are composed of "untouchables," who traditionally 
suffered disabilities and restricted opportunities precisely on the grounds of membership in a 
particular caste. Since caste affiliation is crucially relevant to the determination of untouchability, 
the caste unit is accepted as an appropriate way to identify individuals who suffer that species of 
backwardness. But when dealing with Other Backward Classes, it is not clear that caste (in the sense 
of standing) in and of itself represents the same kind of barrier and source of disabilities; therefore, 
it is not clear that the caste unit has the same relevance in identifying the backward.  67.   
Similarly, the use of religious criteria in designating Scheduled Castes is not necessarily an 
appropriate precedent for the use of religious units in selecting Backward Classes.  68.   The power of 
the State to define Backward Classes is not necessarily as unfettered as the power exclusively 
granted to central authorities by the broad wording of Article 341. The relevance of religion in 
defining untouchability is apparent 
____________________ 
entire population except for Scheduled Castes, Scheduled Tribes, and Brahmins and constituted a 
majority of the total population (RCSCST 1958­59: 12). Scheduled Castes would seem to have 
little to gain from expansion of their numbers and the consequent dilution of benefits (the 
quantum of which is outside of their control). But a situation involving tangible incentive to 
existing Backward Classes to expand the category of Backward Classes is less difficult to 
imagine. 
66. A striking example of this is provided by the situation in Mysore where the state government 
restored the populous Lingayat community to the list of Backward Classes after the Nagan 
Gowda Committee, established by the state government to identify the Backward Classes, had 
excluded them from its list (RCSCST 1960­61: 1, 318­19). 
67. Cf. the observation of M. N. Srinivas ( 1956: 377) that "the barrier of untouchability" cuts off 
untouchables from the opportunities for group social mobility that do exist within the caste 
system. 
68. The President, it has been held, may use religious criteria in designating membership in 
Scheduled Castes. The exclusion from Scheduled Castes of persons professing a religion 
differing from Hinduism has been upheld as a valid exercise of his power to specify parts within 
castes as Scheduled Castes. For a full discussion, see chap. 9, § E. 
­203­ 
(though its application in marginal cases may be questionable); there is no self­evident reason to 
accord it such relevance in defining other species of backwardness. 
Although a sharper distinction between the question of units and the question of backwardness 
would permit greater clarity, it is by now well established that (1) caste and communal units may be 
used as classes whose backwardness is to be established; (2) caste or communal rank or status may 
be one of the tests or measures of backwardness by which these groups are selected; (3) caste or 
communal rank or status may not be the sole or exclusive measure of backwardness. 
B. THE PERMISSIBILITY OF NON­COMMUNAL CLASSES 
The most persevering adherents of the "modernist" view that communal units are impermissible 
might regard this erratic but basically consistent course of judicial development as merely a 
temporary deviation from the true constitutional position.  69.   But by the mid­1970s the expectation 
had receded that governments and courts might imminently adopt an austerely modernist view, 
eschewing the use of traditional ascriptive units in favor of classes based on income, occupaton, 
literacy, or other "neutral" tests applied to individuals rather than communities.  70.   The courts, in 
disappointing this appealing modernist urge to de­recognize communal divisions, have been 
responsive to the pull of a competing theme in the constitutional commitment to equality­­the 
commitment to pursue equality in the face of the institutionalized group inequality of Indian society. 
The Constitution authorizes, as Balaji puts it, preferences for "classes of citizens and not individual 
citizens as such."  71.   Does backward "classes" imply anything more than backward citizens­­i.e., 
citizens 
____________________ 
69. In his Telang Endowment Lectures, Professor Tripathi ( 1972 : 204) asserts that the framers did 
not intend to authorize determination of caste groups to be backward classes. In his view, ". . . 
Article 340 refers not to caste­or­religion based communities but to communities and groups 
sharing special conditions of hardship and difficulty due to their environment," like the 
inhabitants of the hill area of U.P. As history this assertion suffers from some difficulty: it flies 
in the face of the unmistakable understanding of the Constituent Assembly and of the same body 
which sat as the Provisional Parliament (see chap. 6, §B, above); and it runs counter to the clear 
understanding of knowledgeable and disinterested contemporary observers (see chap. 6, n. 69, 
above.) 
70. Expressions of this "modernist" view from diverse official, academic and journalistic source are 
found in chap. 6, §C, above. This theme, present though not dominant in Balaji, reaches its 
judicial apogee in Chitralekha and subsides with only occasional later outcroppings, as in the 
dicta of Pradip Tandon. 
71. A.I.R. 1963 S.C. at 659. A recent reiteration is in Jayasree v. State of Kerala, A.I.R. 1976 S.C. 
2381, 2386. 
­204­ 
selected because they possess the traits that define backwardness? It seems to imply that personal 
possession of the traits that define backwardness is neither a sufficient nor a necessary condition for 
membership in a class for purposes of Articles 15 (4) or 16 (4). For example, all five­year­olds are 
educationally backward, yet it is clear that five­year­old children are not an educationally backward 
class within the meaning of Article 15(A) ­­ if only because Article 15(4) is totally irrelevant to the 
State's power to make suitable provisions for overcoming this sort of backwardness. Similarly, 
among the applicants for a high government post, none are likely to be educationally backward by 
any standard of comparison with the general population. But the Constitution clearly envisages 
preferential treatment for some competitors for high posts, even though they will not be personally 
backward. And arguably, it envisages special provisions for some of the five­year­old ­­ those least 
likely to overcome their presently shared educational backwardness because of their membership in 
a group with other traits. Apparently, it is the backwardness of the class that is constitutionally 
relevant, not the backwardness of the individual beneficiary. Indeed, it has never been suggested that 
individuals be selected on grounds of personal backwardness, a procedure that would select those 
least likely to make use of the opportunity. Selecting individuals on the basis of the backwardness of 
the group involves the opposite problem: inevitably the benefits flow to the most able and 
advantage­­the least backward, personally­within the backward group. If the constitutional 
provisions are specifically designed to overcome the transmitted inequalities of past social 
distinctions, especially those distinctions whose recognition is otherwise forbidden, then the classes 
would be those groups that had suffered arbitrary hindrances to advancement because of past 
application of such distinctions. Clearly, caste and communal groups might fit as such classes. 
If castes and communities are such classes, the question remains whether other sorts of grouping are 
also possible. Are illiteracy, low income, or menial occupation sufficient likenesses to constitute a 
class for purposes of these provisions? May the State for purposes of its equalizing policies 
designate backward classes without reference to the historic social formations to which these 
provisions were addressed? This set of issues was nicely crystallized by two former Supreme Court 
Justices at a 1973 Jurists' Seminar on Backward Classes.  72.   As retired Chief Justice Subba Rao 
poses it, 
____________________ 
72. Speeches at the Jurists Seminar on Backward Classes, held in Bangalore in September 1973 are 
reprinted as Appendix 8 in vol. I, Part 2 of Government of Karnataka ( 1975). 
­205­ 
the real question is whether the Constitution provides only for the existing or 
subsequently formed castes or classes with definite characteristics and which are 
accepted by society as backward castes or classes or can the State form a new backward 
classes out of different castes and classes with specific qualities of backwardness.  73.   
Justice Hegde (who had recently resigned his seat on the Supreme Court after being passed over for 
the chief justiceship) puts forward what we have called the "historic" view: 
[T]here can be no doubt that the expression "class" in Article 15(4) as well as in Article 
340 denotes a well­knit and organized section of the society based on status or rank and 
not any classification made by the State on diverse considerations. . . . The "classes" 
mentioned in Article 15(4) refer to the existing organized sections of the society, and not 
any new groupings of individuals.  74.   
The "historic" view, as argued forcibly by several acute students of the subject, is that class can 
mean nothing other than caste or communal units.  75.   Thus, Mr. Havanur argues that Articles 15(4) 
and 16(4) do not permit "new classifications: they only permit specification of the already existing 
and recognized social classes."  76.   To this Chief Justice Subba Rao opposes another reading which, 
borrowing his term, we label the "elastic" view. 
[T]here is . . . another view, i.e., backward class is more comprehensive than the 
backward caste or community . . . The expression "class" is wider . . . in expression than 
"caste" or "community;" it takes in, in addition to "caste," other groups based on 
language, race, religion, occupation, location, poverty, sex, etc. Caste is also a class; . . . 
The expression "backward class" is an elastic and changing concept. It takes in not only 
the classes existing before the Constitution but also those formed after the Constitution. . 
. .  77.   
____________________ 
73. Id., at 65. 
74. Id., at 106. 
75. Havanur 1965: Iff, Radhakrishnan. 1965: 270. A more specific and less persuasive version of the 
"historic" view reads the provisions for "backward classes" as a commitment to a definite stratum 
of the population made up of certain castes. Thus, Mr. C. L. Sathi (n.d.), a lawyer, politician, and 
backward classes activist, decried the income test as violating Art. 340 of the Indian Constitution 
since "Art. 340 refers to all those castes which have been considered low like the barber, potter, 
blacksmith, carpenter, the household servants or shepherds, etc." But as our discussion of the 
background and drafting of the constitutional provisions indicates (chap. 6, §B), there was never 
any clear agreement about the identity of this stratum. Was it the 12% of Bombay or the 70% of 
Mysore? There was never any decision as to the cut­off point. Who were the Backward Classes 
was not pre­determined, but left subject to later determination. 
76. Havanur 1965: 107, 115. 
77. Government of Karnataka ( 1975) Vol. I, Part II, pp. 65, 67. 
­206­ 
There is some judicial support for the "elastic" view. In Chitralakha, the Supreme Court upheld a 
scheme that did not use castes or other historic social formations as the units whose backwardness 
was measured.  78.   And the broad definition of class propounded in Sagar and since reiterated by the 
courts would seem ample enough to accommodate departures from communal units. On the other 
hand, there is also support for a reading that is closer to the less permissive "historic" view.  79.   
The repetition of the requirement that a class display "certain likenesses or common traits . . . some 
common attributes" suggests that the recipients of preferences under Articles 15 (4) and 16 (4) have 
to share traits that make them a class other than the traits that constitute or define their 
backwardness. The Supreme Court in Janki Prasad draws close to the historic view when it rejects 
provisions for small cultivators and small pensioners on the ground that neither of these categories 
are classes in the sense of "a homogeneous social section . . . with common traits and identifiable by 
. . . common attributes."  80.   Instead, they are "artificial groups" created by the state for purposes of 
receiving benefits under a particular policy. The Court is looking for palpable social groups with 
perceptible traits like castes, territorial groups, or the followers of an occupation ­­ and there is a 
hint that such grouping should not cut across kinship lines.  81.   The judicial attachment to a 
requirement that such classes be shown to be socially as well as educationally backward also argues 
for groups whose social backwardness can be measured.  82.   
In part the problem is an artificial one, for even if "classes" in Articles 15(4) and 16(4) were 
confined exclusively to caste or community groups, there is nothing in Articles 15 (1) or 16 (2) or 29 
(2) to prevent the State from using other classifications (like income, literacy, etc.) instead of, or in 
conjunction with, caste.  83.   That is, Articles 15 (4) and 16 (4) empower the State to use these 
otherwise forbidden categories, but they do not 
____________________ 
78. Chitralekha v. State of Mysore, A. I. R. 1964 S.C. 1823. This result is anticipated in the strong 
commendation of the income criterion in Balaji v. State of Mysore, A.I.R. 1963 S.C. 649, 664. 
79. Cf. Viswanath v. Government of Mysore, A.I.R. 1964 Mys. 132 at 139; Garg v. State of Punjab, 
A.I.R. 1966 Putj. 476. 
80. A.I.R. 1973 S.C. 930, 941. 
81. Discussed at n. 53 above. 
82. See chap. 8, §C, below. 
83. E.g., Art. 15 does not reach a classification exempting the property of "agriculturists" from 
attachment and sale, since the classification does not depend upon any of the forbidden criteria ( 
Rura Ram v. Gurbachna, A.I.R. 1954 Punj. 254). Of course, if one placed sufficient emphasis on 
"only" in Arts. 15(1) and 16(2), one might conclude that so long as it was combined with other 
criteria, even caste would not require Art. 15(4) to legitimate it. 
­207­ 
compel the State to use them. Nor, when the State does use these categories, is it confined to using 
them without any admixture of other categories. 
The difference between the historic and the elastic views would be crucial in a doctrinal setting in 
which the provisions of Articles 15(4) and 16(4) were thought to exhaust governmental power to 
confer certain benefits (like reservations) or to give preferment to the disadvantaged. But at present 
the State is thought to enjoy broad powers to confer advantages, including reservations, on various 
categories of the disadvantaged quite apart from Articles 15(4) and 16(4).  84.   
Hence these rival readings have little constitutional import in the present doctrinal setting. But they 
are more than readings of the Constitution; they are, at least implicitly, prescriptions for policy. They 
speak both to the allowable means and to the goals to be pursued. The historic view is soundly 
anchored in the observation that the provisions for backward classes were intended to enable the 
government to address the problem of caste/communal disparities by adopting measures along caste/
communal lines. Reduction of caste/communal disparities was a central purpose of the inclusion of 
Articles 15 (4) and 16 (4): the contention that these provisions permit only measures which address 
other inequalities is unsupportable. But these provisions are not a detailed blueprint of how the 
government must pursue this objective; they are but an authorization to employ one method that 
would otherwise be barred to it. There is nothing to forbid the government from attempting to 
reduce caste/community differentials entirely by other means or to mix the specifically authorized 
means with other methods. Surely, the government may attempt to articulate its attempts to reduce 
communal disparities with measures directed at the general problems of poverty, backwardness, and 
inequality.  85.   Thus the Constitution would seem to provide wide leeway to the State. Whether to use 
caste units or not is a question of policy, not of constitutional prohibition or command. But if, in 
programs purporting to be empowered by these provisions, it abandons communal units in favor of 
classes defined in other ways, presumably it could be required to show that the classes employed are 
relevant to the elimination of the kind of backwardness that these provisions are designed to 
overcome. So far, no court has required such a showing. 
____________________ 
84. See chap. 12, §D, below, and the general doctrine of classification discussed in chap. 11, §D, 
below. 
85. To some of the participants the problem of communal disparities may have exhausted the 
purpose of these provisions, but it is clear that many others regarded them as part of an attack on 
a wider range of inequalities. See discussion at chap. 6, §B, above. 
­208­ 
C. THE UTILITY OF COMMUNAL UNITS 

The use of communal units has been the most intensely criticized aspect of the policy of 
compensatory discrimination ­­ both within the government and from intellectuals and scholars, 
who have overwhelmingly condemned it.  86.   If the use of caste (or communal) groups as the units 
for determining Backward Classes is a matter of policy rather than of constitutional prohibition or 
command, we should attempt to examine the arguments about their use. Our examination must 
remain, at this point, preliminary and abstract, for the arguments turn on factors we have not yet 
considered: the meaning of "social backwardness"; the feasibility of administering non­communal 
tests; and the wider consequences of using the caste units. Putting aside the argument of 
constitutional intent, which we have seen is inconclusive, let us concentrate on the arguments for the 
utility of communal units in selecting beneficiaries and distributing benefits.  87.   
The argument that caste units  88.   provide a convenient and accurate way of identifying the backward 
can rest on either of two assertions. The first of these, which we might call the correlational 
proposition, is that members of castes tend to share a common level of opportunities and resources. 
Because of this broad affinity of condition, membership in an appropriately selected caste provides 
a convenient way of identifying persons with a low level of opportunity and resources.  89.   The 
second assertion, which we might call the causal proposi­
____________________ 
86. See chap. 3, §D, and chap. 6, §C, above. 
87. The arguments elaborated in the text concern the contribution of communal units to the 
redistributive and nation building goals of the compensatory discrimination policy. But 
communal units also appeal to those who, despairing of creation of a noncommunal sphere of 
public life in the near future, advocate a regime of communal quotas. Thus Prof. A. M. 
Dharmalingam commends the extensive reservations of preIndependence South India. 
Reservations are "in the nature of tariffs" protecting developing groups. (The existing distribution 
of merit is the cumulative result of past "deliberate denial of opportunities to the mass of the 
people . . . not a thing to be proud of . . . ill gotten merit is equivalent to stolen property.") He 
proposes a regime of "proportional reservation and representation" in which "the rule must be the 
best from each section and not the best from the whole" (in Government of Karnataka 1975: I, 
Part II, 78). And, of course, there is a strain of virulent and bigoted anti­Brahminism that in the 
name of "castelessness" would transform these policies from means of inclusion into a device for 
the complete exclusion of Brahmins from educational facilities and public services. For example, 
see Ponnarangam 1958: 80 ff, 102. Other outcroppings of this substratum are noted in chap. 13, 
§A, below. 
88. The argument here is applicable to all communal units, but is stated in terms of caste to avoid 
additional cumbrousness. 
89. When the policy of preferences was first envisioned, it was taken for granted that it would be 
administered along communal lines. As Lelah Dushkin has pointed out, the notion of preferences 
along communal lines derives from the widespread (and unexcep­ 
­209­ 
tion, is that backwardness is the result of caste status, or more accurately, of a hierarchical caste 
system in which some groups were assigned low status and deprived of advantages.  90.   That is, that 
the caste group occupies a given status or standing in the social hierarchy; this status or standing is 
an important determinant of the general level of opportunities and resources of its members; 91. 
therefore membership in a group is a good predictor of backwardness. These propositions are 
independent: correlation may be maintained without causality and vice versa. The evidence to prove 
or disprove one or the other would differ. Typically, though, elements of both are merged in 
argument. To the extent that either of them is true, it can be argued that membership in a caste group 
is a good predictor of backwardness. 
In addition to their utility as a basis of selection, it may be argued that caste units are a useful way of 
distributing preferences. First, it might be argued (surprisingly, but with some factual basis) that 
communal membership is more readily ascertainable than, e.g., income level; so there is relatively 
little slippage in distribution ­­ the benefits get to the target group, if not necessarily to the most 
needful within it.  92.   Dissimulation about communal identity does take place ­­ both individually 
and on a collective scale ­­ but on discovery is usually treated as somewhat scandalous; 
misrepresentation of one's income, on the other hand, is easier to accomplish, harder to detect, and 
evokes little stigma. Second, it may be argued that distribution of benefits to members of an existing 
caste group may mobilize ties of kinship, loyalty, and mutual support to multiply the effect of the 
benefits more than does distribution to isolated individuals. Particularly, such communal groups 
have a greater potential capacity for political organization to facilitate the unimpeded flow of 
benefits through sluggish administrative channels. 
The correlational proposition obviously contains some truth. There 
____________________ 
tionable) assumption of a correlation between groups with the lowest social statuses and those 
which suffered the greatest lack of oportunities and resources. The correlation between the ritual 
status of a caste and the political, economic, and educational levels of the members of that caste 
was, she observed ( 1961: 1666), the "primary assumption on which government policy is based 
and the principal determinant of the official definition process." 
90. See, e.g., the assertion of Mr. O. K. Yadav that "preferences need to be given by the caste 
criterion for the simple reason that social backwardness is not only a serious handicap, rather it is 
the root cause of economic backwardness." ( Backward Classes Review 1( 4­5): 10) 
91. Thus, e.g., Havanur ( 1965: 64­65) argues that the relative advantages and disabilities which have 
accrued to members of different caste units are transmitted by caste loyalties, which lead to 
favoritism, even by the most well­intentioned. In this view open competition is tantamount to 
bias in favor of those castes whose members are responsible for selection; preferences for 
backward castes offset the formidable built­in preference for members of those castes already in 
favorable positions. 
92. See chap. 9, §G, below. 
­210­ 
is a sufficient affinity of conditions among members of caste groups so that caste units serve as 
better than random predictors of levels of education, literacy, income, or other indicia of 
backwardness. Just how strong the correlation will be for any given list of castes cannot be predicted 
in advance. It depends on the definition of backwardness employed, and on the way in which the list 
of castes is compiled. Because of local and sub­group variation it may be expected that the broader 
the groups used, the less the force of the correlational argument. Members of the same caste in 
different parts of the same state or even within the same district may differ greatly in social 
standing, resources, and opportunities ­­ as may component groups within a large cluster which goes 
by a common name.  93.   The correlation would be even weaker if such broad composites as varna 
categories were used.  94.   The religious unit might generally be somewhat more suspect than caste, 
since it commonly involves a larger group with more economic and social levels, while caste is often 
associated with a traditional occupation and a greater affinity of habits and social position among its 
members. The less differentiated the units, both socially and territorially, the less likely that a list of 
communities would isolate groups with distinctive social, economic, and educational characteristics. 
And the less perfect the correlation the greater the danger that the more prosperous and advanced 
within the favored groups would be the recipients of the lion's share of the benefits, thus subverting 
the purpose of helping the most disadvantaged. 
But the correlation is bound to be imperfect to some extent: some within the selected castes will 
have more resources; some outside them will have less. How one responds to failures of correlation 
depends on whether one also accepts the causal proposition. If it is believed that membership in a 
low caste is a major determinant of backwardness, then the absence of correlation is less troubling, 
for if the members of a low caste were in the past disabled to roughly the same degree as a result 
____________________ 
93. Cf. the situation in Mysore where the Nagan Gowda Committee used such large units as 
Lingayats, Vokkaligas, and Muslims without distinguishing among component groups of varying 
circumstances. See Havanur 1965: 93 ff. 
94. Although the units chosen as Backward Classes may be castes in the sense of "communities"­­
jatis, or clusters of such groups with a common name­­it is very doubtful that the State could 
employ caste in the sense of "varna" (i.e., the four divisions of society propounded by classical 
Hindu socio­legal theory). It may be assumed that "varna" is included within caste, as it appears 
in the prohibition of discrimination in Arts. 15, 16, and 29. But it might well be argued that 
"varna" does not fall within the meaning of "class" for purposes of Arts. 15(4) and 16(4), since 
the members of these vast categories hardly display the degree of homogeneity which Sagar and 
Triloki Nath require of a "class" In any event, the criterion is so unwieldy that no one has 
attempted to employ it: both the Backward Classes Commission and Mysore's Nagan Gowda 
Committee rejected the suggestion that Backward Classes be equated with Sudras. See 1 BCC 
44; Government of Mysore 1960: 2. 
­211­ 
of their shared low status, even the relatively well­off individuals within the caste may be thought of 
as worse off than they would otherwise have been. Ergo, preferential treatment to restore the 
deprivations suffered as a result of low caste is justified at whatever level of advantages a member of 
that caste now possesses. The differential treatment of those outsiders equally bereft of tangible 
resources becomes a distinction between those who are poor as a result of a heritage of disabilities 
and those who have failed to benefit from a favorable heritage.  95.   
Thus we see that these propositions tend to be associated with different views of the backwardness 
which requires and deserves remedy by preferential treatment. For those who argue the correlational 
proposition, backwardness is a low level of possession of present resources (such as education, 
income, cultural accomplishments, perhaps even confidence, poise, or status itself). Those who 
assert the causal proposition conceive of backwardness as the accumulated effects of past 
disablements due to low status. In practice, the two assertions merge: there are few who would 
assert the causal proposition without reliance to some degree on correlation. 
The causal proposition, too, contains at least some truth. Those characteristics which can be thought 
of as most directly associated with the low status of a caste ­­ lack of cultural resources, lack of a 
home environment favorable to educational achievement, unfavourable selfimage, lack of access to 
opportunities, subjection to disabilities and discrimination ­­ play a significant role in limiting the 
achievements of the members of that caste. But while the social position of a caste may be one of 
the factors affecting the extent of opportunities of its members, there is no reason to assume that it is 
the only one or that it is universally powerful. As in the case of the correlational proposition, the 
question is how much. 
____________________ 
95. The proponents of preferences on a caste basis are not as callous as the argument makes them 
appear. Typically, they ask only that preferences for Backward Classes be administered separately 
from assistance for the poor among the advanced communities. Thus, Mr. G. Latchanna, in his 
welcome address to the All­India Convention of Legislators and Leaders interested in the welfare 
of the "Other Backward Classes" ( September 1960), declared: "It is true that there are poorer 
sections even among the advanced classes of society, and we are not opposed to the continuance 
of special fianancial assistance to them. We want only that the poorer sections of the "Other  
Backward Classes" should not be deprived of social justice by mixing up their fate with that of 
the other poorer sections who are much more fortunate in their social and educational heritage." 

The intensity with which many Backward Class spokesmen feel the salience of the differential 
heritage is poignantly expressed by a Backward Classes activist, Shri Nagappa: "Even if [most of 
the 'advanced classes'] are backward economically, they are not poor mentally; the learning was 
their grandfathers' property" (speech at Conference of All­India Backward Classes Federation in 
New Delhi, 12 March 1966). Acute awareness of the crucial role of structural advantages 
combines with desperately low 

­212­ 
The arguments for caste units cannot be dismissed out of hand. The correlational and causal 
propositions both lead us to the need for data. How great is the degree of homogeneity or spread of 
conditions within caste groups? To what degree is caste standing (and the conditions intimately 
associated with it) a significant determinant of the kinds of backwardness to be remedied? Does a 
member of caste X with an income of Rs. 1,000 per annum encounter greater obstacles to mobility 
than a non­X with an equal or lesser income? The social scientists, Indian and foreign, who have 
condemned the use of caste units have not undertaken to disprove their utility in either the 
correlational or causal sense. They have instead based their objections on judgments of more far­
reaching consequences alleged to attend the use of caste units. 
Briefly, the use of caste units is said to produce detrimental effects on the recipients (alienation, 
aggravation of their sense of dependence), on non­recipients, (unfairness, lower morale, etc.), and 
on the society and polity generally. The distribution of preferences on caste lines is thought to 
perpetuate caste distinctions, accentuate caste consciousness, aggravate group tensions, and 
encourage political abuse of the preference device.  96.   Although a number of these assertions could 
be 
____________________ 
self­image. Thus Prof. A. M. Dharmalingam, an unrepentant advocate of a regime of communal 
quotas, addressed a Jurists' Conference on Backward Classes: "For instance a poor Brahmin or a 
poor Lingayat boy has enormous advantages derived from heredity and environment. A large 
number of his relatives and friends may be either rich or holding high position in government 
service or in the public life which combined to his higher intelligence will make him walk off 
with a seat or a post beating hollow the poor Kuruba or Yadava or Agasa. . . . The poor backward 
class individual is not only economically poor, but also poor in influence, heritage, environment 
and resourcefulness and in consequence goes to the wall every time he is pitched against a poor 
Brahmin or a poor Lingavat." (Government of Karnataka 1975: I, Part II, 86). 
96. Thus, e.g., it is often asserted that lists of backward communities will inevitably undergo a 
mushroom growth. Thus, Ghurye 1969:430 reports the ominous growth of the Maharashtra list of 
Other Backward Classes from 125 castes in 1953 to 160 in early 1967 to 178 in October 1967. In 
the absence of accompanying population figures, it is difficult to tell how substantial this 
increase really is. And in the case cited, one wonders how much of the "increase" was due to 
states reorganization. The Punjab Evaluation Committee details the growth of the Punjab 
Backward Classes lists from 14 communities in 1953 to 25 by 1958 to 63 in 1959. But there had 
been no further growth when the committee conducted its survey in 1965 ( Punjab, Welfare 
Department 1966: 142). Not surprisingly, non­communal lists of recipients also tend to grow over 
time. Thus, in the Punjab the 1956 list of 12 Backward Areas had grown to 25 by 1964 (id., at 
146­47) Possibly, expansion is more of a danger when any kind of "natural" groups are used to 
designate beneficiaries of preferences. It is then a question of whether their superior capacity to 
identify the deserving and their multiplier effect outweighs this additional danger. And the lists 
may shrink as well as grow. Thus in Andhra Pradesh there were 146 Communities on the 1956 
list of OBCs; 112 on the 1966 list, and 92 recommended by the state's Backward Classes 
Commission in 1970. But it would be misleading to assume any direct correlation of numbers of 
listed groups with population, since many entries are redundant. See n. 155 in chap. 8, §F. 
­213­ 
translated into testable propositions, there is a singular lack of data to support them (which, of 
course, does not prevent them from being widely believed). Even if it is conceded that all of these 
evils exist, surely not all of their presence can be attributed to the use of caste units in distributing 
preferences. For example, there is some evidence for the proposition that the use of caste units 
enhances caste consciousness, but it is clear that many other factors ­­ especially the suffrage ­­ play 
an even greater role. If the dependence of the recipients is aggravated by preferences, is it more 
aggravated by preferences on a caste basis than by preferences on, e.g., an income basis? We lack 
any reliable estimate of the net effect of the employment of caste units. Only with such an estimate 
can the alleged costs of this device be measured and weighed against the presumed benefits (and 
against the costs and benefits of the alternatives). 
We shall return to these broader questions. On the narrow question of the usefulness of the caste 
unit, the existing evidence does not permit us to condemn the use of these criteria per se. Each 
instance of their use will have to be considered on its own merits. So long as detailed data on the 
relative backwardness of different sections of the population do not exist, the use of relatively crude 
classifications such as caste groups might be considered a reasonable method of isolating backward 
groups. The larger and more undifferentiated the group the more suspect its use. However, such 
suspicion might be allayed to the extent that other "neutral" standards are combined with the 
forbidden classification (e.g., an economic ceiling for enjoyment of benefits), particularly if the 
designated group is large and varied in circumstances. 
Once it is accepted that communal units are to be used in conjunction with other indicia of 
backwardness, the question arises whether it would not be simpler to use the latter exclusively. If 
backwardness is conceived of as the lack of tangible resources, it is presumably possible to find tests 
which measure more directly the conditions to be remedied. But the answer would seem to depend 
upon the range and distribution of resources and opportunities within communal groups and on the 
extent to which caste status is determinative of them. If it turns out that there is actually 
considerable uniformity of circumstance or that shared status is an important determinant of 
mobility, caste units might appear a reasonable way of selecting Backward Classes. Paradoxically, 
the very data which would assure equitable administration of preferences along caste lines might 
permit dispensing with caste units altogether. But the data needed to have equitable administration 
on non­communal lines would permit more effective use of caste units. The choice would then turn 
on the relative benefits and disadvantages of alternative methods­­including the difficulties and 
costs of data col­ 
­214­ 
lection  97.   and administration as well as effectiveness in identifying the backward and the wider 
effects of the various tests. 
D. FORBIDDEN CLASSIFICATIONS AND THE BURDEN OF PROOF 
Before proceeding to the other half of the problem ­­ the measure of backwardness ­­ let us return 
briefly to the technical legal setting. Caste  98.   and religion,  99.   along with the other classifications 
mentioned in Articles 15 (1), 16 (2), and 29 (2) ­­ race, sex, place of birth, residence, descent, and 
language ­­ are forbidden criteria for State action.  100.   Articles 15 (4) and 16 (4), then, are exceptions 
to the prohibition on the use of these classifications; they may be used to define Backward Classes 
even though, when used for other purposes, their use may be struck down as 
____________________ 
100. The arrangement of forbidden criteria in the various provisions can conveniently be represented 
in tabular form: (see next page). 
97. Data­gathering along caste lines may, when conducted under official auspices, produce some of 
the ill effects attributed to preferences along caste lines. Cf. the experience of the Backward 
Classes Commission and the frequent observation of the intensification of caste feeling resulting 
from collection of caste data in the Census (e.g., Ghurye 1957: 192). 
98. Thus caste (and religion) cannot be employed for purposed of controlling crime ( Sanghar Umar  
v. State, A. I. R. 1952 Saur. 124), suppressing disorder ( State of Rajasthan v. Pratap Singh, 
A.I.R. 1960 S.C. 1208.), delimiting constituencies ( Bhopal Singh v. State. A.I.R. 1958 Raj.41), 
or composing electorates ( Nain Sukh Das v. State of U.P., A.I.R. 1953 S.C. 384). Indeed, quite 
apart from the ban on caste criteria, the use of other criteria for the purpose of promoting the 
identity and integrity of caste groups is itself invalid. The Supreme Court held unconstitutional 
laws providing for pre­emption on the basis of vicinage on the ground that the purpose of such 
laws was to promote and protect communal neighborhoods. "[Such] division of society . . . into 
groups and exclusion of strangers cannot [any longer] be considered reasonable . . . [or] in the 
interests of the general public" ( Bhau Ram v. Baij Nath, A.I.R. 1962 S.C. 1476 at 1481) For a 
review of the post­Independence posture of law toward caste, see Galanter 1968. 
99. Religion presents a somewhat more complicated instance, for while government is committed to 
eschew religious discrimination, it is simultaneously committed to protect the identity and 
integrity of separate religious groups. Thus although religious differentia may not be used for 
most purposes (see n. 98 above), it is permissible to use religion in determining the applicability 
of a law of religious trusts (e.g., Motidas v. S.P. Sahi A.I.R. 1959 S.C. 942) or in the application 
of family law. Though Art. 44 directs the eventual elimination of separate personal law for 
members of different religions, the continuing validity of disparate rules of personal law and the 
power of the State to create new rules of personal law applicable to members of particular 
religious communities has been upheld ( State of Bombay v. Narasu Appa, A.I.R. 1952 Born. 84; 
Srinivasa Aiyar v. Saraswati Ammal, A.I.R. 1952 Mad. 193). But of Sheokaran Singh v.  
Daulatram, A.I.R. 1955 Raj. 201, where a rule of Hindu law regarding recovery of interest was 
held invalid as discriminatory against non­Hindus. 
­215­ 
discriminatory.  101.   Other than caste, religion, and residence,  102.   these classifications are rarely used 
in designating the Backward Classes.  103.   While there is no constitutional ban on the use of these 
forbidden criteria for this purpose, their use is not prescribed nor is it immune from judicial review. 
When used, they are subject to review in accordance with the general standards of reasonable 
classification employed by the Court. 

Forbidden in Article: 
23(2) 29(2)
16(2)
Ground of 15(1) (Compulsory) (Admission to)
(Government)
Discrimination  (General)  Public State or Aided
Employment 
Service  Schools 
Religion  X  X  X  X 
Race  X  X  X  X 
Caste  X  X  X  X 
Sex  X  X 
Place of Birth  X  X 
Descent  X 
Residence  X 
Class  X 
Language  X 
____________________ 
101. Thus, in State of U.P. v. Pradip Tandon, A.I.R. 1975 S.C. 563, 569, the Supreme Court found 
reservation of admissions for rural students unconstitutional on the ground that it amounted to 
discrimination by "place of birth." This was in the context of a finding that inhabitants of rural 
areas did not constitute a socially and educationally Backward Class, so the question did not 
arise whether a legitimate Backward Class might be defined by, inter alia, place of birth. 
102. The uses of caste and religion have been discussed in the preceding section. Residence is a 
forbidden criterion for purposes of Art. 16 (see Kishori v. Board of Revenue, A.I.R. 1957 Raj. 
185). But it is allowable in educational matters ( Joshi v. Madhya Bharat, A.I.R. 1955 S.C. 334). 
It seems never to have been doubted that residence may be used in defining Backward Classes 
for all purposes (including reservations in government employment). Government lists of 
Backward Classes frequently employ territorial criteria­­e.g., the members of caste X in district 
Y. Indeed, location may itself be a constitutive part of backwardness that the State may validly 
take into account. See Chap. 8, §H, below. However, reservations and distributions on territorial 
lines which are not related to backwardness must meet the general tests of reasonable 
classification. 

District­wise distributions of medical college seats have been struck down as having no 
reasonable relation to the object to be achieved­­i.e., admission of best talent ( P. Rajendran v. 
State of Madras, A.I.R. 1968 S.C. 1012; State of Kerala v. Jacob Mathew, I.L.R. 1964[2] Ker. 53, 
62, aff'g A.I.R. Ker. 1964 39, 68; Abodha Kumar v. State of Orissa, A.I.R. 1969 Or, 80). 
However, distribution between regions on grounds of their different needs or histories is 
acceptable ( Sagar v. State of Andhra Pradesh, A.I.R. 1968 A.P. 165; Joseph Thomas v. State of 
Kerala, A.I.R. 1958 Ker. 33 [division of seats between Malabar and Travencore­Cochin based on 
historical reasons]). Cf Muralidhar v. State of Andhra Pradesh, A.I.R. 1959 A.P. 437 
(distribution between metroplitan area and hinterland acceptable). 

103. Descent: in effect the commonly used criteria of caste and tribe contain an element 
­216­ 
All governmental measures which distinguish between groups of citizens must meet the standards of 
"equality before the law" and "equal protection of the laws" laid down in Article 14.  104.   These tests, 
evolved in application of the general equal protection provision of Article 14, are well settled.  105.   It 
is also well settled that, owing to the intimate connection between Articles 14 and Articles 15 and 
16, similar standards of reasonableness are applicable to governmental measures challenged under 
the latter provisions.  106.   Thus where a measure is challenged as violative of Articles 15 or 16, 
____________________ 
of descent. See the discussion of converts to these groups in chap. 9, §F. Place of Birth: 
reservations for rural students have been invalidated as discrimination on grounds of place of 
birth ( State of U.P. v. Pradip Tandon, A.I.R. 1975 S.C. 563, 569, discussed in n. 101, above). 
Inhabitants of specified areas can be deemed Backward Classes if the place of birth criterion is 
avoided. Cf. Janki Prasad v. State of J. & K., A.I.R. 1973 S.C. 930, 943. Race: Tribal 
classifications may be thought of as "on racial ground [s]" ( Mahendra Nath v. State, A.I.R. 1970 
A. & N. 32, 34). Apart from this the use of racial classification is confined to Anglo­Indans 
(discussed in n. 50 of chap. 9, below). Cf. the recommendation of the Backward Classes 
Commission (I BCC 28) that the small Eurasian community of Travencore be deemed a 
Backward Class. Sex: It is clear that the State may take special provision for women 
independently of their membership in backward classes. Art. 15 is qualified by a proviso that the 
State may make "any special provisions for women" (Art. 15[3]). Unlike special provision for 
backward classes, that for women need not be favorable to them ( Yusuf Abdul Aziz v. State of 
Bombay, A.I.R. 1954 S.C. 321). Art. 15(3) qualifies only Art. 15 itself and does not engraft an 
exception on the prohibition of sex as a criterion in Art. 16. See Dattatraya v. State of Bombay, 
A.I.R. 1953 Bom. 311, and Venkataramana v. State of Madras, A.I.R. 1951 S.C. 229. Cf. State of 
Madras v. Champakam Dorairajan, A.I.R. 1951 S.C. 226 at 228, where the presence of a 
backward classes proviso in Art. 16 was thought to make the absence of one in Art. 15 
significant. Art. 16 concerns government employment specifically and overrides the more 
general Art. 15 in this area. See chap. 11, §B, below. Reserved posts for women would depend on 
their designation as Backward Classes. That such an eventuality is not too farfetched is indicated 
by the recommendation of the Backward Classes Commission that all women in India should be 
considered a Backward Class (1 BCC 31). While recommending special concessions in 
education, the Commission did not mention government service (id., at 32). Cf. the suggestion 
of the court in the Dattatraya case that special provisions for women might not be 
discriminatory on grounds of sex "only," since women as a group have been subject to other 
common influences and disabilities which make them a definable group. 
104. But cf. the holdings of the Madhya Pradesh High Court that non­statutory rules (e.g., university 
admission rules) are not "law" and therefore are not subject to the reach of Art. 14 (although they 
would be covered by Arts. 15 and 29). Vinod Sagar Sood v. State of Madhya Pradesh, A.I.R. 
1967 M.P. 182; Gokul Prasad v. Sohani, A.I.R. 1962 M.P. 126; Vishnu v. State of Madhya 
Pradesh, A.I.R. 1961 M.P. 247. 
105. Budhan Choudhry v. State of Bihar, A.I.R. 1955 S.C. 191; Bidi Supply Co. v. Union of India, 
A.I.R. 1956 S.C. 479. 
106. State of Mysore v. Narasinga Rao. A.I.R. 1968 S.C. 349. This has been specifically noted in 
litigation about preferences ( General Manager v. Rangachari, A.I.R. 1962 S.C. 36). In 
Devadasan v. Union of India, A.I.R. 1964 S.C. 179 at 185, the Court notes that the argument 
under Art. 14 is identical with that under Art. 16(1). For an early and clear 
­217­ 
the first test is that the classification on which [the government action] is founded must 
be based on an intelligible differentia which distinguishes persons or things grouped 
together from those left out of the group; and the second is that the differentia in 
question must have a reasonable relation to the object sought to be achieved by the rule 
or statutory provision in question. . . . In every case there must be some nexus between 
the bases of the classification and the object intended to be achieved by the statute.  107.   
Ordinarily, there is a presumption in favor of the constitutionality of any enactment. The burden lies 
on the challenger to show a clear transgression of constitutional limits.  108.   However, it is not clear 
that such a presumption arises where the State classifies citizens according to the criteria forbidden 
by Articles 15 and 16.  109.   When the State employs caste and religion to classify backward groups, 
who has the burden of showing the relevance and adequacy of the classification? 
The first High Courts to address this question took divergent courses. In Kerala a majority of a Full 
Bench clearly placed on the challenger the burden of proving that a scheme of reservations is not 
justified.  110.   The Andhra Pradesh Court, however, suggested that the onus of supporting the 
classification lies with the State.  111.   In upholding that court's decision in Sagar, the Supreme Court 
declared that "the conditions which justify departure [from the strictures of Art. 15 (1) and 29 (2)] 
must be 
____________________ 
application of the Art. 14 standards to the Art. 15 area, see Partha v. State of Mysore, A.I.R. 1961 
Mys. 220 at 229. The incorporation of Art. 14 standards has been criticized by Havanur ( 1965: 
115). 
107. Pandurangarau v. Andhra Pradesh Public Service Commission, A.I.R. 1963 S.C. 268 at 271. 
108. Moti Das v. Sani, A.I.R. 1959 S.C. 942; Mohd. Hanif Qureshi v. State of Bihar, A.I.R. 1958 S.C. 
731. 
109. Cf. the observations of Patanjali Shastri, J., in Kathi Raning Rawat v. State of Saurasthra, III 
(1952) S.C.R. 435 at 442, that while "equal protection claims under . . . Article [14] are 
examined with the presumption that the State action is reasonable and justified" where bias on 
any of the grounds mentioned in Arts. 15 or 16 is disclosed, "it may well be that the statute will, 
without more, incur condemnation as violating a specific constitutional prohibition unless it is 
saved by one . . .of the provisos to those articles." A Full Bench of the Punjab High Court has 
held that a classification on any of these forbidden grounds is "bad without anything more 
having to be proved," since a classification forbidden by the Constitution cannot possibly be said 
to be reasonable ( Pritam Kaur v. State of PEPSU, A.I.R. 1963 Punj., 9 at 17). (Provision of 
PEPSU Court of Wards Act allowing government to deprive a woman of management of an 
estate where similar mismanagement by a male would not be grounds for removal is invalid.) 
110. Hariharan Pillai v. State of Kerala, A.I.R. 1968 Ker. 42. Gopalan Nambiar, J., dissenting, would 
place the burden clearly on the State (id., at 53). 
111. Explicitly in Sukhdev v. Government of Andhra Pradesh, 1966 Andh. W.R. 294 at 303, 304, and 
implicitly, by adopting exacting standards for data to support the classification in Sagar v. State 
of Andhra Pradesh, A.I.R. 1968 A.P. 165. 
­218­ 
strictly shown to exist."  112.   The Court would not assume that constitutional criteria were employed; 
mere assertions by the state that relevant criteria had been taken into consideration were not 
sufficient. Subsequently, another Supreme Court bench declared that "the onus of proof is on the 
State to establish that the reservations are for socially and educationally backward classes of 
citizens."  113.   
Whether or not the presumption of constitutionality technically arises, it is clear that such 
classifications suffer under an onus of constitutional disrepute and will attract intense judicial 
scrutiny. The emergence of stringent judicial tests for sufficient data to support the classification 
indicates that wherever the burden of proof technically rests, it is the primary responsibility of the 
State to justify its classification  114.   ­­ at least where the challenger has an interest in the 
composition of the whole category of Backward Classes.  115.   
Where the matter at issue is not the eligibility of the beneficiary groups but other aspects of the 
scheme of preferential treatment, no such burden is carried by the State: the normal presumption of 
constitutionality would seem to obtain. Thus in a case involving the operation of reservations for 
Scheduled Castes, where designation of beneficiaries was not an issue, the Supreme Court observed 
that "the 
____________________ 
112. State of Andhra Pradesh v. Sagar, A.I.R. 1968, S.C. 1379 at 1384. 
113. State of U.P. v. Pradip Tandon, A.I.R. 1975 S.C. 563 at 569. But cf. Hridaya Narain Singh v. 
Mhd. Sharif, A.I.R. 1968 Pat. 296, discussed in n. 101 of chap. 8. 
114. Something more than mere assertions of unconstitutionality may be required to trigger the state's 
obligation. See Periakaruppen v. State of T.N., A.I.R. 1971 S.C. 2303 at 2310­11. In some 
instances the courts are willing to relieve the state of this burden by taking judicial notice of the 
notorious backwardness of a particular community (see Chait Ram v. Sikander A.I.R. 1968 Pat. 
337) or a composite group like Harijans ( Swain v. Secretary A.I.R. 1974 or. 115). 
115. In Chait Ram v. Sikander, A.I.R. 1968 Pat. 337, the Full Bench of the Patna Court dealt with a 
petitioner whose interest in challenging the Backward Classes designation was somewhat 
different from that of the frustrated competitor for a government job or medical college seat. The 
state had made immune from sale in execution of decrees the raiyati holdings of members of the 
Backward Classes. Petitioner was a creditor of a member of the Mallah community, a caste of 
fishermen and boatmen, and challenged their inclusion of the Backward Class list. The Court 
said that ordinarily the decision of the state government that a particular class or caste is socially 
and educationally backward would prevail, subject to the right of petitioner to satisfy the court 
that the test of backwardness adopted was irrational or irrelevant. Government's classification 
here was supported by gazetteers and common knowledge and petitioner offered no refutation. 
Even if the State carries an onus with respect to selection of the entire Backward Classes 
category, the creditor here, who is concerned specifically with discrediting the backwardness of 
the particular group to which his debtor belongs, need not be the beneficiary of a rule designed 
to make possible the vindication of rights by those with an interest in the composition of the 
entire Backward Classes category. 
­219­ 
burden of establishing that a particular reservation made by the State is offensive to Art. 16(1) is on 
the person" challenging it.  116.   
The stringency of the standards of reasonable classification in designating Backward Classes 
derives from the exceptional nature of the authorization for preferential treatment. "Being in the 
nature of an exception, the conditions which justify departure must strictly be shown to exist." 117. 
Articles 15(4) and 16(4) permit the use of otherwise forbidden criteria for the achievement of 
specific objectives. The courts have been anxious to confine their use to those objectives, to prevent 
the expansion of group preferences into a general principle of operation, and to reconcile these 
exceptions with their context of general fundamental rights. Thus the Supreme Court has 
continually emphasized the necessity of balancing the provisions for preferences against the general 
scheme of equality: 
It is implicit in . . . [Art. 16] that the doctrine of equality of opportunity shall be 
reconciled with that of reservation in such a way that the latter while serving the cause 
of the backward classes shall not unreasonably encroach upon the field of equality.  118.   
Presumably, if other (i.e., non­forbidden) criteria were used to select Backward Classes, the State's 
classification would still have to meet the general equal protection standards (intelligible differentia, 
reasonable relation to the object, etc.). But if it is the technically "exceptional" character of Articles 
15(4) and 16(4) that is the basis of the enhanced stringency we have observed, then it is an open 
question whether it would obtain where the scheme of "compensatory discrimination" did not 
technically exercise the powers conferred by these exceptions to deviate from the general provisions 
of Articles 15, 16, and 29. Would the ordinary presumption of constitutionality and the burden of 
proof on the challenger reassert themselves in the area of compensatory discrimination as reliance 
on forbidden communal criteria declined? 
The courts have, as one scholar remarked of Balaji, "virtually read the word 'reasonable' into" the 
provisions for compensatory discrimination.  119.   We shall in the sequel be examining various 
attempts of 
____________________ 
116. State of Punjab v. Hira Lal, A.I.R. 1971 S.C, 1777 at 1780. Cf. State of Kerala v. N. M. Thomas, 
A.I.R. 1976 S.C. 490 at 524, where Justice Beg clearly places on petitioner the burden of 
establishing "a constitutionally unwarranted discrimination." 
117. State of Andhra Pradesh v. Sagar, A.I.R. 1968 S.C. 1379 at 1383. 
118. Triloki Nath Tikku v. State of Jammu and Kashmir [I] A.I.R. 1967 S.C. 1283 at 1285; cf. State of 
Andhra Pradesh v. Sagar, A.I.R. 1968 S.C. 1379 at 1383; Balaji v. State of Mysore, A.I.R. 1963 
S.C. 649 at 662. 
119. Rama Rao 1967: 82. A clear example is provided by Partha v. State of Mysore, A.I.R. 1961 Mys. 
220, 229, where it is observed that the restraint on the fundamental rights of others "must be a 
reasonable restraint. . . . It is . . . well established that a 
­220­ 
the courts to decide what is a reasonable use of this power by balancing competing factors. At this 
point it should be recalled that judgments of reasonableness may differ, reflecting deeper 
differences about the context and purposes of compensatory discrimination.  120.   What is reasonable 
depends upon the range of factors which are considered relevant and the weights that are assigned 
them. 
____________________ 
restraint can be reasonable only when it is necessary to achieve a particular purpose and should 
not exceed what is necessary for the achievement of that purpose." 
120. "No abstract standard, or general pattern, of reasonableness can be laid down as applicable in all 
cases. The nature of the right alleged to have been infringed, the underlying purpose of the 
restrictions imposed, the extent and urgency of the evil sought to be remedied thereby, the 
disproportion of the imposition, the prevailing conditions at the time, should all enter into the 
judicial verdict. In evaluating such elusive factors and forming their own conception of what is 
reasonable, in all the circumstances of a given case, it is inevitable that the social philosophy and 
the scale of values of the judges participating in the decision should play an important part, and 
the limit to their interference with legislative judgment in such cases can only be dictated by 
their sense of responsibility and self­respect and the sobering reflection that the Constitution is 
meant not only for people of their way of thinking but for all, and that the majority of the elected 
representatives of the people have, in authorizing the imposition of the restrictions, considered 
them to be reasonable" ( State of Madras v. V.G. Row, A.I.R. 1952 S.C. 196, 200). 
­221­ 
8 The Backward Classes and the Judiciary: the Measure of Backwardness 

PREFERENCES FOR members of a particular group are not permissible if that group is not a 
Backward Class.    1. In Venkataramana v. State of Madras,    2. the Madras Government had filled a 
group of posts in accordance with its "Communal G.O.," which divided all available posts according 
to quotas among Harijans, Backward Hindus, Muslims, Christians, Non­Brahmin Hindus, and 
Brahmins. The Supreme Court held that reservation was permitted solely for backward classes and 
that it was "in the circumstances, impossible to say that classes of people other than Harijans and 
Backward Hindus may be called Backward Classes."     3. The denial to petitioner, a Brahmin, of the 
opportunity to compete for posts reserved for other non­backward groups was caste discrimination 
within the prohibition of article 16 (1) and (2). In State of Jammu and Kashmir v. Jagar Nath,     4. a 
cabinet order authorizing direct appointment of Muslims to certain posts, "to remove the communal 
disparity," was held invalid; since there was no designation of the preferred group as backward. The 
same point was made by the Supreme Court in regard to a subsequent Jammu and Kashmir scheme. 
 5.   There must be a finding that the group in question is backward, something more than "a general 
assertion, unsupported by acceptable data."    6. 

____________________ 
1. Cf Dr. Ambedkar's observation to the Constituent Assembly that "unless you use some such 
qualifying phrase as 'backward' the exception made in favour of reservation will ultimately eat up 
the rule altogether" ( VII CAD 702). Preferential treatment for non­backward groups is taken up 
in chap. 12, §D. 
2. A. I. R. 1951 S. C. 229. 
3. Id., at 230 Cf. A. R. V. Achar v. State of Madras, Writ Petition No. 568, High Court at Madras, 
Aug. 25, 1952, aff'd on other grounds A. I. R. 1954 Mad. 563. 
4. A. I. R. 1958 J. & K. 14, aff'g A. I. R. 1958 J. &. K. 1. An alternative ground for rejecting this 
scheme might have been that the device employed was not "reservation," but the Court did not 
deal with this question. 
5. Triloki Nath Tiku v. State of Jammu and Kashmir[I], A. I. R. 1967 S. C. 1283. 
6. Id., at 1284. 
­222­ 
Thus, some finding by the State that the group (for whom the reservations are made) is a backward 
class is required. But the term "backward" is undescriptive: what criteria may the State use in 
determining whether a group is backward? Furthermore, "backwardness" is obviously a relative 
term: what may the State use as a standard and where may it establish the cut­off point? What kind 
of evidence must it have to support its application of these criteria? What kind of control are courts 
empowered and competent to exercise over the standards employed by the State and over the way 
they are applied? 
A. EDUCATIONAL BACKWARDNESS 
The judicial response to these questions emerged in good part from a series of cases involving 
Mysore's schemes of reservation. Before the problem was encountered by the Supreme Court in the 
Balaji case, the Mysore High Court had already begun to elaborate the concept of educational 
backwardness. The old Princely State of Mysore had since 1921 provided extensive reservations for 
"backward communities," defined as all but Brahmins.     7. These arrangements were carried over into 
the post­Independence State of Mysore. When it was attempted to extend them into the enlarged 
Mysore (formed by the States Reorganization of 1956) the Order was quashed by the High Court in 
1958.  8.   This judicial setback led the government finally to list the Backward Classes, and it 
compiled a list of 164 communities­­all the Hindu communities in the state except Brahmins, 
Banias, and Kayasthas and all the non­Hindu groups except Parsis and Christians­­altogether over 
90% of the state's population. 
In Ramakrishna Singh v. State of Mysore, the High Court struck down a scheme in which 45% of 
9. In order to 
the seats in technical and professional colleges were reserved for the 164 communities.    
decide whether the reservations were constitutional, the High Court undertook to ascertain whether 
the beneficiaries of this scheme were "socially and educationally backward classes as envisaged in 
Article 15 (4). . . ."  10.   To qualify 
____________________ 
10. Id., at 346. 
7. For an enlightening account of these developments, see Dushkin 1974. The "all but Brahmins" 
system survived until the reorganization of the state in 1956 ( Mysore Backward Classes 
Committee 1961: 2). 
8. Writ petitions Nos. 369, 370, etc., of 1958, where the Court quashed the communal G. O. of 26 
July 1958 in an uncontested consent order. (This information is from A. I. R. 1960 Mys. 338.) 
Havanur ( 1965:88) explains the background of this unreported case: each of the five areas in the 
new state had retained its old list of Backward Classes. Members of communities which were 
backward in old Mysore (i.e., all but Brahmins) but not in other areas of the state, succeeded in 
pressuring the Government to extend the old Mysore order to the new areas in July 1958. 
9. A. I. R. 1960 Mys. 338. 
­223­ 
as such, they must have been selected by some intelligible principle" designed to further "the policy 
and object of the Constitution . . . to ameliorate the conditions of really backward classes . . . ." 11. 
Finding that communities with high percentages of literacy were included, the court struck down the 
state's classification as arbitrary on the ground that "literacy is the only possible test for determining 
educationally backward classes."  12.   
The Mysore Government responded by appointing a committee to suggest criteria of backwardness 
and the exact manner in which they should be applied, both for purposes of admissions to 
professional and technical colleges and for government posts. The Nagan Gowda Committee was 
formed on January 9, 1960, and submitted an Interim Report on February 19, focussing on a 
tentative classification for purposes of reservations in government service. It proposed and applied a 
dual test of literacy and representation in government service: communities above the state average 
on either of these measures were eliminated, leaving a list of 168 communities (an estimated 
35.34% of the state's population), for whom the Committee recommended a reservation of 36%. 13. 
Lingayats and the Bhunt section of Vokkaligas were excluded from the Committee's list by the 
literacy test; Muslims were excluded by the Government service test; the remaining Vokkaligas were 
on the list. On March 1, the Government issued an order adopting the Committee's list for purposes 
of reservations of government posts, but reducing the reservations from the recommended 36% to 
25%. As the season for admissions to medical and technical colleges neared and the Committee's 
final report remained unfinished, the Government on June 9 adopted the same list for reservations in 
admissions, ordering a reservation of 22%. When that year's admissions were challenged in the 
High Court, the interim list survived challenge. The Court found literacy an "intelligible" and, under 
the circumstances,  14.   a reasonable measure of educational backwardness and was tolerant of the 
Committee's extrapolation from old census data. "No one has suggested that the list includes any 
community which is not really socially and educationally backward."  15.   Indeed, there was an unsuc­ 
____________________ 
11. Id., at 347. 
12. Id., at 348. The Court indicated that it did not consider literacy in English a suitable test. 
13. The number of communities is somewhat misleading. The population figure of 35.34% is for 13 
major communities with 128 "sub­communities." In addition, there were 40 other groups for 
whom no individual population figures were available but whose total population was probably 
less than 3% of the state total. 
14. Partha v. State of Mysore, A. I. R. 1961 Mys. 220. 
15. Id., at 232. 
­224­ 
Article 15 (4) list but not on the government service list; Lingayats were on neither. 
The state government was unwilling to allow Lingayats to be omitted, so it raised the educational 
cut­off point from 6.9 per thousand to 7.1 for the purpose of including them (in the process they had 
to include the intervening Ganigas), bringing the total Backward Classes to 74% of the population 
(not counting the Scheduled Castes and Tribes).  20.   . The state government also refused to carve out 
the Bhunts from the Vokkaligas on the ground that Vokkaligas had to be treated as a whole. The 
Government retained the Committee's recommended figure of 50% for reservations in educational 
institutions. 
It was in this scheme that the Supreme Court encountered the question of educational backwardness 
in Balaji v. State of Mysore.  21.   The Court found it "doubtful if the test of the average of the student 
population in the last three High School classes is appropriate in determining the educational 
backwardness."  22.   But it did not press the merits of the student population as a measure,  23.   but 
directed its concern to the state's cut­off point. The Mysore scheme included as backward all 
communities whose enrolment in secondary schools throughout the state was less than the state 
average of 6.9 high school students per thousand of population. Assuming enrolment to be a rational 
test of backwardness, the Court asked "whether it was legitimate to treat castes and communities 
which are just below the state average as educationally backward classes?"  24.   It answered that it is 
not: only communities "well below" the state average may be deemed backward. (This standard is 
discussed below in §E.) 
In spite of the Balaji court's misgivings about the high school population test, composite tests along 
these lines have become the most com­ 
____________________ 
20. Havanur 1965: 96. 
21. A. I. R. 1963 S. C. 649. This was the first of these cases to go directly to the Supreme Court 
under the Article 32 writ jurisdiction. Counsel opposing the reservations, in spite of their earlier 
successes in the Mysore High Court, apparently felt that the composition of the bench hearing 
writ petitions was unfavourable to them and thought their chances were better in New Delhi. The 
earlier successes and the intense public debate surrounding the work of the Nagan Gowda 
Committee and the State Government's subsequent actions may have increased interest and made 
available the resources required for this more expensive undertaking. 
22. Id., at 660, The Court also opines that "the literacy test supplied by the Census Reports may not 
be adequate. . . ." (id.) 
23. It would seem that this test contains at least one serious probability of distortion due to difference 
in the age profile of different communities. A community with fewer youngsters of school age 
might appear more backward than one with many young people, even if a higher proportion of 
the former's young people were in high school than of the latter's. 
24. Id., at 660. 
­226­ 
mon way of determining educational backwardness. Thus, Kerala's 1965 Kumara Pillai Commission 
utilized a more sophisticated composite test of educational backwardness: communities were 
educationally backward if they had a low student population in the Xth standard or if the percentage 
of persons who had not completed primary education was low. Where Xth standard data was not 
available, communities were judged backward if they displayed an abnormally high fall­out (i.e. 
drop­out) rate. The cut­off points were pegged at slightly below the state averages.  25.   The 1970 
Kerala Commission, however, felt that a test focussing on the high school level was not appropriate, 
since many positions required advanced training. The Commission recommended that communities 
should be considered educationally backward if the percentage of persons in that group passing the 
Secondary School Leaving Certificate (S.S.L.C.) examinations, or the number of students in first 
degree programs in arts and science, law, medicine, ayurveda, agriculture, veterinary science, and 
engineering was lower than the group's percentage of the state population.  26.   
The Jammu and Kashmir Backward Classes Committee in its 1969 report determined educational 
backwardness by comparing the number of students in the 9th and 10th classes per thousand 
population in each group, with the statewide student population per thousand, but did not specify 
how the cut­off was arrived at.  27.   In 1970, the Andhra Commission, citing the Jammu and Kashmir 
Committee's use of standards IX and X, adopted a test utilizing student population in standards X 
and XI: communities whose student population in these standards was "well below" the state 
average were educationally backward.  28.   The Commission did not specify what it meant by "well 
below" the state average. In the cases of several communities which showed a "slightly higher level 
of education," the Commission used personal knowledge of the living conditions of the 
communities to conclude that they were in fact educationally backward.  29.   The inclusion of these 
groups was one of the grounds upon which the Andhra Pradesh High Court struck down the state's 
list. But the Supreme Court reversed on the ground that a few such instances was not sufficient to 
strike down the entire list.  30.   Balaji's doubts about student populations as a test, and notions about 
"well below the state average," had been "misap­ 
____________________ 
25. Kerala (Kumara Pillai Commission) 1966: 43­44. 
26. Kerala (Backward Classes Reservation Commission) 1971: I, 87. 
27. Government of Jammu and Kashmir 1969: 28. 
28. Andhra Pradesh Backward Classes Commission 1970: 58­59. 
29. Id., at 59. 
30. State of Andhra Pradesh v. Balaram, A. I. R. 1972 S. C. 1375, at 1396. The decision of the High 
Court in this case was unreported. 
­227­ 
plied;" for Balaji did not "lay down any hard and fast rule," but only indicated "broad principles to 
be kept in view."  31.   
In Karnataka (as Mysore was renamed in 1973), the Havanur Commission in 1975 used the S.S.L.C. 
level in its test of educational backwardness. It compared the number of students in a community 
passing the examination in April 1972 with the average number of students passing per thousand 
state population. Groups with less than the state average but more than 50% of the state average 
were classed as Backward Communities. Separate lists of Backward Castes and Backward Tribes 
were composed of communities with student averages below 50% of the state average.  32.   
Unfortunately, all of these measures of representation in the student population suffer from the 
speculative population figures on which they are calculated. But alternative tests of educational 
backwardness are still inchoate. 
One alternative measure was rejected in Dilip Kumar v. Government of U.P.  33.   The shortage of 
Higher Secondary schools in an area and low marks in the Pre­Medical Test obtained by residents of 
the area, were insufficient to classify all those residents as educationally backward, since the area 
undoubtedly included classes that were not educationally backward.  34.   The Supreme Court in a 
later encounter with this same U.P. scheme allowed that "lack of educational institutions and 
educational aids" might be one factor in the educational backwardness of an area, but the Court put 
more stress on attitudes toward education than on the absence of facilities: "traditional apathy for 
education on account of social and environmental conditions or occupational handicaps . . . 
[illustrates] educational backwardness."  35.   Thus people in the hill and Uttarkhand area of U.P. are 
educationally Backward Classes "because lack of educational facilities keep them stagnant and they 
have neither meaning and values nor awareness for education."  36.   Earlier this ingrained apathy 
principle had been elaborated by a Constitutional Bench in Janki Prasad, as a test which seems to 
measure both the social and the educational backwardness referred to in Article 15(4).  37.   The Court 
remarks the spread of facilities for education into 
____________________ 
31. Id., at 1397. 
32. Government of Karnataka 1975: I, 315­316. 
33. Dilip Kumar v. Government of U. P., A. I. R. 1973 All. 592. 
34. Id., at 595. In State of U. P. v. Pradip Tandon, A. I. R. 1975 S. C. 563, the Supreme Court concurs 
that the rural areas of U. P. can't be classified as educationally backward, on the curious ground 
that candidates from rural areas obtained 85 of 758 seats (12%) on merit. The Court thought this 
spoke well of the high standards of education in rural areas, although it estimated that 80% of the 
population of the state resided in those areas (id., at 568­69). 
35. A. I. R. 1975 S. C. 563, at 567. 
36. Id., at 567. 
37. Janki Prasad Parimoo v. state of J. & K., A. I. R. 1973 S. C. 930. at 938. 
­228­ 
rural areas and the growing sector of the village population which has embraced education as a 
means of social advancement. However, there remain "sectors of the population which show 
extreme apathy towards education due to age­old customs and habits of living, fostered by poverty, 
ignorance, superstition and prolonged social suppression."  38.   While other sectors in the rural areas 
deserve encouragement by the State, these apathetic sections "require to be goaded into the social 
stream by positive efforts of the State."  39.   Whether these attitudinal factors can be developed into a 
workable substitute for, or supplement to, representation in student bodies remains to be seen. The 
absence of census data suggests that surveys might be an appropriate way of gathering information 
relevant to educational backwardness. Once such survey apparatus were established, one would be 
free to devise more precise and differentiated measures of educational backwardness. 
B. SOCIAL BACKWARDNESS 
In addition to educational backwardness, the Balaji Court scrutinized the Mysore scheme for 
evidence of the "social backwardness" of the beneficiaries. While holding that a classification solely 
on the basis of caste would be unconstitutional, the Supreme Court suggested that "it may not be 
irrelevant to consider the caste of the . . . group of citizens."  40.   The use of castes as units or classes 
of citizens designated as Backward Classes has already been discussed in chapter 7 above. But caste 
may be used in another way: to determine the backwardness of groups and their eligibility for 
inclusion among the recipients of preferential treatment. These uses involve two different senses of 
the word "caste." The use of caste as the unit or community of citizens emphasizes the caste as an 
entity, an endogamous group or cluster of such groups with a common name, etc. The use of caste 
to determine backwardness emphasizes rank, status, prestige (or their opposites), and disabilities, 
the absence of resources and opportunities associated with them. Unfortunately, the Supreme Court 
does not make this dual usage explicit. 
It is quite clear that the Balaji Court finds it permissible to use castes as the units of citizens to be 
designated backward. It is in the second sense of caste that the Court limits its use­­i.e., in the sense 
of rank, status, or prestige of the group as a test of its social backwardness. The Court's first 
objection to the exclusive use of caste as a test of social 
____________________ 
38. Id., at 938. 
39. Id., at 938. 
40. A. I. R. 1963 S. C. at 659. 
­229­ 
backwardness is that the test would inevitably break down in relation to many sections of Indian 
society. 
How is one going to determine whether Muslims, Christians, or Jains, or even Lingayats 
are socially backward or not? . . . [Though] castes in relation to Hindus may be a 
relevant factor to consider in determining the social backwardness of groups or classes 
of citizens, it cannot be made the sole or dominant test in that behalf.  41.   
Since it is well known that these non­Hindu groups comprise (or are composed of) endogamous 
communities, the only reason this would not work would be the difficulty of applying to them the 
notions of caste rank and standing in traditional Hindu terms. Again, says the Court, poverty as well 
as caste is relevant: 
that social backwardness which results from poverty is likely to be aggravated by 
considerations of caste to which the poor citizens may belong but that only shows the 
relevance of both caste and poverty in determining the backwardness of citizens.  42.   
It is clear that the caste that is objected to here is not the caste as a unit, but caste in the sense of 
standing or status­­i.e., a level of opportunities and access to resources, respect, etc., which 
characterizes that group. The Court notes that not only poverty, but also such factors as occupation 
and place of habitation, go toward determining the backwardness "of a community of persons."  43.   It 
criticizes the Mysore Committee, which "was inclined to treat the caste as almost the sole basis in 
determining the question about the social backwardness of any community."  44.   "The predominant, 
if not the sole test that weighed in their minds was the test of caste."  45.   And the "classification of 
the socially backward classes of citizens made by the State proceeds on the consideration only of 
their castes without regard to the other factors which are undoubtedly relevant."  46.   Thus, "the social 
backwardness of the communities . . .has been determined in a manner which is not permissible 
under Article 15 (4)."  47.   
Inspection of the Mysore scheme suggests that the role of the "caste test" in determining social 
backwardness was neither as dominant nor as clear as the Supreme Court indicates. To be sure, the 
Nagan Gowda Committee announced that "the status accorded to [the] community in society" 
would be its test of social backwardness.  48.   But it 
____________________ 
41. Id., at 659. 
42. Id., at 659. 
43. Id., at 659. 
44. Id., at 660. 
45. Id., at 660 
46. Id., at 660. 
47. Id., at 660. 
48. Mysore Backward Classes Committee 1961: 19. 
­230­ 
never specified any findings about status and it never applied any such test. Its list of backward 
communities is the list produced by application of its education (high school enrolment) test. No 
group was dropped from the list by the caste status test, nor was any group added. In the case of the 
three large borderline groups (Lingayats, Vokkaligas, and Muslims) the Committee merely 
announced its opinion of their social backwardness. In each case that opinion corresponded exactly 
with the results of the education test: Lingayats and Bhunts forward, Vokkaligas and Muslims 
backward. The "caste test" made no visible difference in the listing of these groups. The only real 
effect of the caste status test was in the case of the many minor communities for whom population 
figures were not available and to whom, consequently, the school enrolment test could not be 
applied. Presumably, it was reputation of their low status that was used to identify 40 of these 
groups as backward. These groups were probably no more than 5% of the total population of 
Backward Classes. The ghostliness of the caste status test is clearly demonstrated by the way in 
which the Committee divided its list into Backward and More Backward ­­ it deemed as More 
Backward those "socially backward communities whose standard of education is less than 50 per 
cent of the state average."  49.   The Supreme Court notes at one point how hollow the purported status 
test is, when it observes that in relation to Muslims the finding of social backwardness was "stated 
merely as a conclusion and no data or reasons are cited in support of it."  50.   But this is literally the 
case for all communities ­­ there were no findings about their social backwardness.  51.   
It is difficult to escape the conclusion that the state government's subsequent additions to the 
Committee's list were more result­oriented than guided or impelled by any test of caste status. The 
state altered the educational test so that it accommodated Lingayats  52.   and announced that they 
were also socially backward. Even its stated rationale ­­ that 
____________________ 
49. Id., at 22­23. 
50. A. I. R. 1963 S. C. at 661. The Committee decided Muslims were backward by majority vote. See 
Havanur 1965: 93­94. 
51. The most elaborate entry in the Committee's chapter on social backwardness reads in its entirety: 
"(ii) Vokkaligas ­­ All sections of Vokkaligas excluding Bhunts are socially backward. Bhunts are 
socially forward. The standard of education among Bhunts is above the State average, being nine 
per thousand according to the figures available to the Committee" ( Mysore Backward Classes 
Committee 1961: 20). It is particularly revealing that the only piece of evidence the Committee 
cites is a repetition of its educational test. 
52. The state government professed concerti that it was not possible to obtain absolute mathematical 
precision in the matter, so it decided the figures should be rounded to the nearest integer. It thus 
raised the state average from 6.9 to 7.0 and the Lingayats' average was lowered from 7.1 to 7.0 (A. 
I. R. 1963 S. C. at 660­61). (A slightly different account is given in Havanur 1965: 96.) 
­231­ 
"it observed that a large percentage of Lingayats live in rural areas arid most of them are engaged in 
agriculture and menial labour and suffer from all the consequences of illiteracy and poverty"  53.   ­­ 
does not reveal exclusive reliance on a test of caste status. 
But if the "caste test" (in the sense of measurement of group status) played a restricted role in the 
making of the Mysore list, caste politics was of the essence. The use of caste units, the strong flavor 
of caste politics, and the announcement that the caste status test was being applied ­­ all made it 
easy to overestimate the extent to which it was in fact applied. Perhaps the Court's alarm at the 
obtrusion of caste politics and its discomfort with the use of caste units (to which it put forward no 
objection) found a tempting target in the "caste test" of social backwardness ­­ a target which 
loomed larger than life because the Court had not arrived at a clear distinction between castes as 
units and caste status as a factor in backwardness. The state's use of castes as units is equated with 
their selection by caste status. Having labelled the scheme as one of exclusive application of a caste 
test, the Court passed without mention several problematic aspects of Mysore's use of this test. May 
caste status be determined as casually as the Committee and the state government appear to have 
done it, without any supporting evidence, merely on opinion of the committee members? May this 
status test (and other tests) be applied to such large and diverse composite clusters as Lingayats, 
Vokkaligas, and Muslims in which homogeneity of status, much less any close correlation with 
social conditions, is unlikely?  54.   
If caste standing cannot be the sole or dominant criterion of social backwardness, Balaji provides a 
catalog of other factors which the State might employ: poverty, occupation, habitation.  55.   It is not 
clear 
____________________ 
53. A. I. R. 1963 S. C. 649 at 654. Cf. the argument of Shri Patil, the leading proponent of the 
Lingayat cause on the Nagan Gowda Committee that Lingayats were not only educationally 
backward (see n. 19 above) but also socially backward. His argument emphasizes traditional and 
current occupation and rural residence and concludes that these are the sources of the group's 
social inferiority ( Mysore Backward Classes Committee 1961: 33­38). 
54. Havanur 1965: 93­94. 
55. A. I. R. 1963 S. C. 649 at 659. An unusual opportunity to observe a judge putting his precepts 
into practice was provided by the circumstance that in 1968, former Chief Justice Gajendragadkar 
(the opinion­writer in Balaji) constituted a one­man commission to make recommendations 
about the revision of the list of Backward Classes in Jammu and Kashmir. In terms that contain 
recognizable echoes of Balaji, the Commission recommended the following criteria for selection 
of the Backward Classes: (1) economic backwardness of a class; (2) occupation(s) of a class; (3) 
habitation; (4) average student population per thousand in that class; (5) caste, in relation to 
Hindus. The Commission does not define what it means by a class, but presumably communal 
units are included. ( Reported in Backward Classes Review 1 ( 3): 23.) It is notable that 
­232­ 
whether any of those other measures may be used solely or dominantly, although Balaji's 
commendation of an income test implied that they might.  56.   The extent to which these have been 
sufficient to support a designation as backward is discussed in §§ G and H below. 
Neither Balaji nor any subsequent case has addressed the problem of the incommensurateness of the 
various kinds of backwardness to which the state may legitimately attend in designating 
beneficiaries. There is no simple way in which low income, low status, undesirable habitation, etc., 
can be, added together. But if the State may take all of them into account, are there any limits on the 
formulae it may use in cumulating them? Perhaps part of the appeal of economic tests ­­ quite apart 
from the desire to avoid caste ­­ is their susceptibility to quantification. 
It is established that it is not necessary for all of these factors to be used simultaneously. When 
Mysore later used only income and occupation as tests of backwardness, the High Court advised it 
that it was necessary to add tests of caste and residence.  57.   In Chitralekha, the Supreme Court 
repudiated this view, allowing that caste might be omitted entirely if backwardness could be 
ascertained by other relevant criteria.  58.   Subsequent Supreme Court decisions have concurred that 
caste (in the sense of standing or rank) may play some part in the determination of social 
backwardness, though it may not be the sole criterion.  59.   But they have had little to add about the 
way in which caste status is to be employed. 
In Balaji, the most complete judicial discussion of this point, the Court finds that "caste" is "not 
irrelevant" to social backwardness, because "considerations of purity," "feelings of superiority and 
inferiority," "narrow caste loyalties," and the rigidity and inflexibility of the caste system lead to 
caste's playing an unfortunately large part in determining the status of a citizen. The "social 
backwardness" which results from poverty is likely to be aggravated by "considerations of caste." 60. 
The relation between low status and social backwardness remains somewhat equivocal. Is low 
status itself to be taken as constituting the social backwardness to be remedied by preferential 
treatment?  61.   Or is 
____________________ 
in spite of Balaji's embrace of the income test, the former Chief Justice does not recommend its 
exclusive use here. These recommendations provided the foundation for the work of the 
Backward Classes Committee that reported in November 1969 (Government of Jammu & 
Kashmir 1969). 
56. A. I. R. 1963 S. C. 649 at 664. 
57. Viswanath v. Government of Mysore, A.I.R. 1964 Mys. 132. 
58. A. I. R. 1964 S. C. 1823. 
59. State of Andhra Pradesh v. Sagar, A. I. R. 1968 S. C. 1379; P. Rajendran v. State of Madras, A. I. 
R. 1968 S. C. 1012. 
60. A. I. R. 1963 S. C. 649 at 659. 
61. The Balaji Court's argument for the relevance of poverty and occupation seems to be that they 
too confer low status (A. I. R. 1963 S. C. 649 at 659). 
­233­ 
low status merely an indicator of a low level of tangible resources and opportunities which are the 
object of government policy? 
Both views point to factual questions on which the courts have been presented with little systematic 
data. If caste status is an indicator of the absence of resources and opportunities, the question arises 
how strong is the association? To what extent are those characteristics which can be thought of as 
intimately associated with caste status­­low self­image, disabilities, etc. ­­ a significant determinant 
of the low level of resources or opportunities or of an impaired capacity to utilize them? If low 
status is in itself tantamount to backwardness, how large a part does caste play in the determination 
of the status of individuals, families, and groups? How does it interact with occupation, wealth, 
education, residence, and cultural traits as a determinant of status? 
The answers will turn, in part, on what is meant by status, a question that is not without conceptual 
and practical difficulties of its own. Status can be measured in innumerable ways, and the relative 
status of a particular caste may differ with the measure employed.  62.   Is status to be measured by the 
self­image of a group? or by the regard in which it is held by those in the locality? the region? Or is 
status to be measured by observation of patterns of interaction with other groups? Or is it to be 
measured by the possession of attributes which give standing in some hierarchic scale ­­ local, 
regional, or pan­Indian? Presumably, a measure might be any of these types if there were some 
plausible connection with the backwardness to be remedied and if the status were ascertainable with 
some accuracy and without excessive cost. 
Presumably, though, there are some measures of status that are not sufficiently reliable or relevant. 
It has occasionally bee suggested that the socially backward should be equated with Sudras  63.   ­­ the 
fourth or lowest varna (class, estate) in the traditional fourfold classification of Hindu socio­legal 
theory. Varna has long been recognized as both 
____________________ 
62. On the multiplicity and ambiguity of status rankings, see Silverberg 1967. 
63. Many would accept this proposition as a general description. Thus the Tamil Nadu Backward 
Classes Commission ( 1971: I, 37) noted: "The castes included in the Backward Classes are 
regarded as 'Savarnas,' i.e., inside the Chatur­Varna, but largely falling within the Sudra fold." 
But taken prescriptively, it is more troublesome. This criterion of backwardness was advocated 
by Shri S. D. Singh Chaurasia, a member of the Backward Classes Commission ( III BCC 22 ff), 
but it was rejected by the full commission on the ground that there were a small number of Sudra 
communites which "have made some advance in recent decades" ( I BCC 44). The equation of 
backwardness with Sudra status seems to carry more plausibility to northern Indians than to 
southerners. The late Dr. Ram Manohar Lohia was a proponent of the Sudra theory ( Lohia 1964 
: passim). A similar proposal was rejected by the Nagan Gowda Committee, which found it 
difficult to justify, "especially when the Committee is aware that some sections of the so­called 
Sudras are socially well­advanced today" ( Mysore Backward Classes Commission 1960: 2). 
­234­ 
inexact in its correlation with social conditions and exceedingly difficult to determine in many 
cases.  64.   
Determinations of caste status are rendered difficult by the tendency to describe them in the 
traditional idiom of purity and pollution. Thus, for example, the Mysore Government characterized 
farming as a low occupation. But such traditional measures may play a less important role than it 
appears to those who view rural India from the state capital.  65.   For example, studies of evaluations 
of occupations in rural India suggest that purity/impurity values may have relatively little influence 
in the evaluation of occupations: compared with other societies, farming ranks unusually high in 
rural opinion, the priesthood unusually low.  66.   Although the ritual or ceremonial standing of a caste 
may be one of the factors affecting its status in situ and thereby the extent of opportunity open to its 
members, there is no reason to assume any exact correspondence.  67.   
There are few examples of courts and governments attempting to assess the social backwardness of 
a community ­­ particularly that associated with caste status. In Jacob Mathew v. State of Kerala, 
the single judge found fault with the state's assessment of the social backwardness of the populous 
Ezhuva caste.  68.   The Government claimed that it classified the Ezhuvas as backward because they 
had been denied access to the temples of high­caste Hindus, were regarded as "untouchables" and 
polluting in use of tanks, their traditional occupation was the degraded one of toddy­tapping, there 
was no intermarriage between them and higher­caste Hindus, and they had been latecomers in the 
field of education. The Court, upon consulting the state's District Gazetteers published a year 
before, concluded that except for the lack of intermarriage, all of these disabilities were more 
historical than current. Lack of intermarriage, the court asserted, was an expression of social 
distance not sufficient to establish social backwardness.  69.   Reversing, the Division Bench was 
skeptical that the legal abolition of untouchabil­ 
____________________ 
64. On the indeterminacy of varna classifications, see chap. 9, § D. 
65. See Marriott 1959: 104. 
66. Bopegamage and Veeraraghavan 1967 ; Harjinder Singh 1967. Cf. Rama Rao ( 1967: 83), who 
questions Balaji's assertion that traditonal notions of caste status are relevant to determining 
backwardness: "Obviously, in the context of the present Indian society, ritual purity of a caste has 
nothing to do with social status of a community which depends upon several factors like the 
extent of wealth possessed by its members, the political pull it can exert, etc. A poor, orthodox, 
tufted Brahmin priest is more often an object of ridicule than of veneration, and at any rate, 
sociologists have observed, and as everyday observation of rural India will bear out, that the non­
landowning Brahmins in the villages are subservient to the members of the dominant castes." 
67. Determination of status by governmental authorities may involve a problem of reactivity, as 
status may be affected by governmental patronage in the form of preferences. (See chap. 10, 
below.) 
68. A. I. R. 1964 Kerala 39. 
69. Id., at 60 
­235­ 
ity and opening of temples had altered the character of the Ezhuva community.  70.   
[T]ime alone transmutes the ideals of law into the realities of everyday life. No one can 
say that the introduction of progressive measures is the end, not the beginning, of a 
process of amelioration. Habits of thought die hard and slow, and occupations like 
toddy­tapping carry their special stigma from one generation to another and through the 
decades of conduct and behavior.  71.   
The government and the courts have to make do with the data that are available. But what data 
becomes available will depend in part on the kind of standards of precision and pertinence that the 
courts establish. We shall take this up in §F below. 
After Balaji, discussion of social backwardness gained in sophistication and rigor, moving away 
from simple estimates of caste standing to analysis of conditions that impede group advancement. In 
a thoughtful discussion of the relation of status and backwardness, the 1965 Kumara Pillai 
Commission in Kerala equated social backwardness with subjection to disabilities ­­ i.e., those 
invidous social distinctions which hinder members of a group by denying them access to facilities 
for improvement. Social backwardness, then, can be measured by popular conceptions of the status 
of a community because those conceptions are translated into disabilities which impinge on the 
tangible opportunities of the group.  72.   
In 1967, Kerala set up another commission, in response to judicial prodding,  73.   this one to 
determine who were the Backward Classes for purposes of reservations in state services. While the 
Kerala Commission was sitting, Andhra set up a Backward Classes Commission to redesignate its 
Backward Classes after its Cabinet Sub­Committee's list had, like its predecessor, been declared 
unconstitutional. Upon the advent of the Dravida Munnetra Kazagham (DMK) government in Tamil 
Nadu in 1969, a mandate to consider the question of who should be the Backward Classes was 
pointedly withheld. That state set up a Backward Classes Commission limited to reviewing the 
progress of the then listed Backward Classes. All of these commissions reported in 1970 
____________________ 
70. State of Kerala v. Jacob Mathew I. L. R. 1964 (2) Ker. 53. 
71. Id., at 58­59. Subsequently, the Kumara Pillai Commission carefully weighed recent advances 
against the heritage of disability and concluded that the Ezhuvas' disabilities had not completely 
disappeared, that caste prejudice still persisted and impinged on the lower income group among 
the Ezhuvas (Kerala [ Kumara Pillai Commission] 1966: 46 ff.). 
72. Kerala ( Kumara Pillai Commission) 1966: 29­34. 
73. See §F, below. 
­236­ 
and all had something to say about the measurement of social backwardness.  74.   
The Tamil Nadu Commission moved the shortest distance away from the simple equation of social 
backwardness with caste standing, which it measured in terms of public opinion: "Social 
backwardness is a matter of local or regional opinion based on taboos and other concepts of a 
hierarchical society."  75.   The Commission was emphatic that poverty is not the sole cause of social 
backwardness; it is "only one factor affecting social position."  76.   
The Kerala Commission thought that Backward Classes under Article 16 (4) should be delineated 
by four tests: education, economic position (because prosperity produced a "home atmosphere 
congenial to education"), participation in governmental service (and consequent participation in 
governmental power and the benefits that it affords)"  77.   and "social backwardness due to historical 
reasons."  78.   The latter includes such factors as stigmatized occupations, vestiges of caste 
discrimination, a heritage of purdah, and aversion to education in English. The Commission 
proposed that from communities that met these tests, members of families with incomes below Rs. 
8,000 should be considered backward for purposes of Article 16 (4). 
The Andhra Pradesh Commission was more historically minded and more focussed on occupation. 
Social backwardness was defined in terms of the heritage of restricted opportunities resulting from a 
group's traditional association with an occupation which was "inferior or unclean or undignified or 
unremunerative or . . . [without] influence or power."  79.   These occupational characteristics were 
applied along with "general poverty of the community" and "caste in relation to Hindus." By the 
latter, the Commission referred to restrictions on intercourse with the advanced classes that had 
"hindered social progress in the past by denial of access to educational institutions, public offices 
and offices 
____________________ 
74. The Kerala Commission was appointed on Oct. 17, 1967, and reported on Nov. 30, 1970. The 
Andhra Pradesh Commission was appointed on April 12, 1968, and reported on June 20, 1970. 
The Tamil Nadu Commission was appointed on Nov. 13, 1969, and reported on Nov. 26, 1970. 
The three commissions exchanged views. The Andhra Pradesh report mentions a visit to Madras 
and Kerala ( 1970: 51), and there was a meeting between the Tamil Nadu and Kerala 
Commissions ( Tamil Nadu Backward Classes Commission 1971: I, 231). 
75. Tamil Nadu Backward Classes Commission 1971: I, 36. 
76. Tamil Nadu Backward Classes Commission 1971: I, 37. 
77. The Commission carefully distinguished its use of representation in services in measuring 
backwardness and in assessing under­representation in the services (Kerala [ Backward Classes 
Reservation Commission] 1971: I, 76). 
78. Kerala ( Backward Classes Reservation Commission) 1971: I, 73 ff. 
79. Andhra Pradesh Backward Classes Commission 1970: 50 ff. 
­237­ 
of public resort . . . and have left vestiges which . . . [constitute] a serious impediment to social life." 
 80.   This segregation was imposed upon the followers of lowly occupations who in turn were 
members of "certain castes and communities."  81.   
The Andhra Pradesh classification was struck down by the High Court of that state on the ground 
that the Commission had classified Backward Classes mainly on the basis of caste and had failed to 
collect the data necessary to ascertain the social and educational backwardness of the group.  82.   The 
Supreme Court, reversing, found acceptable the use of caste units and thought the Commission had 
collected "sufficient material to enable [it] to be satisfied that the persons included in the list are 
really socially and educationally backward."  83.   
Earlier, a Constitution Bench had expressed some doubt about the criterion of traditional occupation 
when detached from evidence of present circumstance. In Janki Prasad v. State of Jammu and 
Kashmir the state deemed as Backward Classes 62 "traditional occupational classes" including 
various kinds of artisans, peddlers, and laborers. Traditional occupation was defined as "the main 
occupation of his . . . grandfather's."  84.   Finding that it would be wrong to give benefits where the 
father has been mobile, the Court thought the validity of such a classification required a showing 
that the family livelihood is still earned from this occupation.  85.   
But if present economic circumstances cannot be ignored in determining social backwardness, 
social backwardness is not to be equated with poverty. Balaji had confidently asserted that social 
backwardness resulted from poverty. But later courts have not seen the connection as so direct. Even 
if poverty and backwardness are so related that the well­off members of a community might not 
suffer from the social backwardness that characterized the community in general, "poverty is not the 
decisive and determining factor of social backwardness."  86.   In Janki Prasad the Court indicates that 
it will not take low income as in itself constitutive of social backwardness.  87.   Although the measure 
of social backwardness is never articulated, it seems to lie in the direction of 
____________________ 
80. Id., at 51. 
81. Id., at 51. 
82. See State of Andhra Pradesh v. Balaram, A. I. R. 1972 S. C. 1375 at 1394. 
83. Id., at 1396. 
84. A. I. R. 1973 S. C. 930 at 940. 
85. Id., at 940. 
86. Jayasree v. State of Kerala, A. I. R. 1976 S. C. 2381 at 2384. 
87. Janki Prasad v. State of J. & K., A. I. R. 1973 S. C. 930 at 937: "In this country, except for a 
small percentage of the population, the people are generally poor­­some being more poor, others 
less poor. Therefore, when a social investigator tries to identify socially and educationally 
backward classes, he may do it with confidence that they are bound to be poor. His chief concern 
is, therefore, to determine whether the class or group is socially and educationally backward." 
­238­ 
cultural traits­­shared attitudes and dispositions that render it unlikely that members of a group will, 
without special help, take advantage of opportunities to improve their position.  88.   Thus the Court 
notes that some sectors of the village population have taken to education, and their embrace of 
education is "almost the measure of social advance they have made recently." The socially 
backward, it seems, are those "sectors of the population which show extreme apathy toward 
education due to age­old customs and habits of living, fostered by poverty, ignorance, superstition 
and prolonged social suppression."  89.   Thus the trend is to think of social backwardness in terms of 
a heritage of deprivation. This is usually visualized in terms of the accumulated effects of low 
position in a social hierarchy. But one Supreme Court Bench, upholding Uttar Pradesh's designation 
of the hill and Uttarkhand areas as backward, included in its catalog of "the traits of backwardness" 
not only the lack of economic development but the absence of social hierarchy.  90.   Generally, 
however, social backwardness is visualized in terms of the accumulated effects of low position in a 
social hierarchy.The most ambitious attempt to characterize social backwardness was the work of 
the Karnataka Backward Classes Commission, under the chairmanship of L.G. Havanur, a 
committed and scholarly student of the subject. In its 1975 report, the Commission set forth a 
catalog of factors which contribute to social backwardness: 
i.  [Residence in] rural, isolated and segregated areas. 
ii.  Their economic condition is so poor that the majority . . . are incapable of owning land, house 
or other property. 
iii.  . . . they are engaged in very unremunerative . . . occupations [that] are considered unclean or 
inferior, and therefore low. 
iv.  People of the advanced castes and communities entertain prejudice against them. 
v.  Low status or inferiority associated with their castes makes it difficult 
____________________ 
88. Id., at 941. Social backwardness in this view is clearly a trait of a group­­not an "artificial" group 
like pensioners or small cultivators, but some palpable historic social formation. See the 
discussion of Janki Prasad in chap. 7, §B, above. 
89. Id., at 938. 
90. State of U.P. v. Pradip Tandon, A. I. R. 1975 S. C. 563, at 567, contains a somewhat puzzling 
catalog of the traits of social backwardness. "There is no social hierarchy. There are no means of 
controlling the environment through technology. There is no organization of the society to create 
inducements for uplift of the people and improvement of economy. Building of towns and 
industries [and] growth of cash economy . . . are absent . . . . [t]hey do not make effective use of 
resources. When large areas . . . maintain a sparse, disorderly and illiterate population whose 
property is small and negligible . . . social backwardness is observed." 
­239­ 

. . . to have access to places of cultural training, or to have religious and secular education. 
vi.  Many . . . have been segregated from the advanced communities due to social taboos against 
inter­dining and inter­marriage, and against similar opportunities of association. 
vii.  [They are] prevented from enjoying status in society due to the age­old social customs. 
viii.  [consequent] . . . apathy for education  91.   
Although there are formidable conceptual and measurement problems in operationalizing such a 
catalog, it is clear that the notion of identifying the Backward Classes has moved far beyond 
summary assertions about their status. The Commission was emphatic that "we have ignored the 
principle of caste ­­ the principle of status ­­ the principle of high and low ­­ in deciding social 
backwardness."  92.   Instead, the Commission devised a complex test based on poverty and isolation, 
social enforcement and reinforcement of that poverty and isolation, and the development of cultural 
traits that are adaptive to poverty and exclusion. 
The basic principle for ascertaining social backwardness was "economic backwardness based on 
occupational and residential backwardness."  93.   Applying its test of backwardness, the Commission 
included among the Backward Classes several castes of temple functionaries as well as a group 
regarded as Kshatriyas, and it excluded some groups with lower ritual status on ground of their 
"economic and educational advancement."  94.   
The Havanur report, the most comprehensive and thoughtful of the Backward Classes reports, is 
marked by a conceptual mastery of the subject that has so far evaded the courts. The Commission 
moved beyond the confusions imparted by the inability to separate caste as a unit and caste as a 
status. It set a challenging agenda for conceptualization and measurement of backwardness that may 
help to address the real and perplexing questions of whether, how, and at what cost disparities 
among communities can be dispelled. 
C. CUMULATIVE BACKWARDNESS 
The relation between educational and social backwardness is entwined with the relationship 
between Articles 15 (4) and 16 (4). Article 16 (4) refers to "any backward class," but Article 15 (4) 
mentions classes 
____________________ 
91. Government of Karnataka 1975: I, 313­14. 
92. Id., at 307. Cf. id., at 287. 
93. Id., at 267. 
94. Id., at 307. 
­240­ 
that are "socially and educationally backward." If the difference in wording  95.   can support any 
inference, it is that Article 15 (4) covers a narrower spectrum of groups while Article 16 (4) includes 
also those who are backward in other respects than "socially and educationally."  96.   Governmental 
bodies ­­ including the Backward Classes Commission and the Home Ministry ­­ have never seen fit 
to make any distinction whatsoever.  97.   
No emphasis was put on these qualifying words until the high court in Ramakrishna Singh, in 
striking down the scheme of reservations there, objected that the scheme provided no standard for 
the determination of social backwardness.  98.   "It would not be enough" said the court, "to say that 
these communities are educationally backward. It will also have to be seen whether they are socially 
backward."  99.   The court thus read the phrase "socially and educationally backward classes" 
conjunctively: to be entitled to preferences, a group must be both socially backward and 
educationally backward. It would seem equally plausible to read the phrase disjuntively, 100. 
allowing government to give preference to groups which are either socially backward or 
educationally backward.  101.   But the Balaji Court, anxious to impose a more restrictive notion of the 
legitimate beneficiaries in opposition to the expansive Mysore view, found this conjunctive reading 
convenient. "The backwardness under Article 15 (4) must be social and educational. It is not either 
social or educational, but it is both social and educational."  102.   
____________________ 
100. The parliamentary debate over Art. 15 (4) provides no indication about the preference of the 
drafters. Near the end of the debate, Prime Minister Nehru, rejecting an amendment that would 
have added "economically" explained that these words were chosen "because they occur in 
Article 340 and we wanted to bring them bodily from there. Otherwise, I would have had not the 
slightest objection to add 'economically'. But if I added 'economically' I would at the same time 
not make it a kind of cumulative thing but would say that a person who is lacking in any of these 
things should be helped" (Parliamentary Debates, Vol. XII­XIII [Part II], col. 9830 [1 June 
1951]). 
101. This result may be approximated by keeping the burden of disproving backwardness on the party 
who complains about a scheme. Cf. Hridaya Narain Singh v. Mhd. Sharif, A. I. R. 1968 Pat. 296 
(where educational backwardness of a community is beyond question, lack of a showing of 
social backwardness is not fatal, where there was no data to show that they were not socially 
backward). 
102. A. I. R. 1663 S. C. 649 at 658. The Court provides no argument for this reading, but an 
interesting argument in its favor is put forward by the Kumara Pillai Commis­ 
95. See the explanation of Prime Minister Nehru (n. 55 to chap. 6, above) that the intention was to 
align the wording of Art. 15 (4) with that of Art. 340. 
96. Art. 16 (4) could, of course, be narrower instead of broader and include only those who are 
backward in all respects, not only socially and educationally. 
97. Cf. the Nagan Gowda Committee's conclusion that the backward classes referred to in Art. 16 (4) 
are the socially and educationally backward classes of Art. 15 (4) ( Mysore Backward Classes 
Commission 1961: 17). 
98. A. I. R. 1960 Mys. 338. 
99. Id., at 348. 
­241­ 
The conjuctive reading propels the courts into serious difficulties, both textual and substantive. 
First, there is the textual paradox. The great weight which the conjunctive reading places on the 
qualifying adjectives in Articles 15 (4) suggests that their absence in Article 16 (4) is significant and 
that the latter provision covers a much wider range of groups. Yet generally the courts have refused 
to attribute any importance to the distinction in phrasing. Thus they have held the Balaji tests of 
backwardness applicable to the government employment area, governed by Article 16. In Dasa 
Rayudu v. Andhra Pradesh Public Service Commission the High Court says flatly that the "absence 
of some words in Article 16 (4) hardly makes any difference."  103.   In Triloki Nath Tiku v. State of 
Jammu and Kashmir [I] the Supreme Court clearly indicates that Article 15 (4) and Article 16 (4) 
backwardness are commensurate. The principles of Balaji, which turned on Article 15 (4), are 
"equally applicable" to the area of government employment, regulated by Article 16 (4).  104.   It 
seems somewhat paradoxical for the Court put so much emphasis on "socially and educationally" in 
Article 15 (4) and then to turn around and say that their absence in Article 16 (4) makes no 
difference.  105.   
Second, the conjunctive reading would severely restrict the flexibility of State policy. It seems to 
imply that it would be unconstitutional for the government to provide special educational aid to a 
community which was educationally backward if it were not also socially backward; nor would it be 
able to give aid, even of a non­educational type (e.g., loans or housing), to a group that was not 
educationally backward but might suffer unusual economic or other disadvantages. Finally, the 
conjunctive reading would seem to constrict ­­ or perhaps invalidate altogether ­­ the use of purely 
economic tests of backwardness. If, for example, a test of family income (of the sort that Balaji 
itself commends) 
____________________ 
sion. Contrasting the State's duty to provide educational facilities for all with the special 
commitment to provide educational aid to those classes whose educational backwardness is the 
result of the denial of opportunities entailed by their social backwardness, the Commission 
concludes that special provisions under Art. 15 (4) are intended only for those who suffer this 
conjunction of disadvantages (Government of Kerala 1965: 15­16). 
103. A. I. R. 1967 A. P. 353 at 361. 
104. A. I. R.1967 S. C. 1283 at 1285. 
105. But cf. the dissenting judgment of Gopalan Nambiar, J., in Hariharan Pillai v. State of Kerala, 
A. I. R. 1968 Ker. 42 at 53, which draws a sharp distinction between the coverage of Arts. 15 (4) 
and 16 (4). He asserts that while backwardness for purposes of Art. 15 (4) must be "cumulative" 
(i.e., both social and educational), for purposes of Art. 16 (4) any sort of backwardness will 
suffice. Even the majority in that case suggests that Art. 16 (4) "has perhaps a wider import than 
Article 15 (4)," but is sure that "the 'backward class' mentioned therein refers to the weaker 
sections of the citizens " (id., at 48). 
­242­ 
were used to identify the backward, there would be no way of knowing that the benficiaries were 
drawn from groups or families that were socially and educationally backward. 
Notwithstanding these conceptual and practical difficulties, the conjunctive reading has become 
firmly established.  106.   It is not clear that it adds much of a hurdle where a court is inclined to credit 
a group's backwardness, for as the Supreme Court suggests in Janki Prasad, failure to take 
advantage of educational opportunities may be the best indicator of social backwardness. 107. 
Conversely, the Court noted that "if a class as a whole is educationally advanced, it is generally also 
socially advanced because of the reformative effect of education of that class."  108.   Perhaps the 
enduring appeal of this formula lies in the scope that it gives courts for assessing schemes: 
"educationally" enables a court to demand evidence of backwardness that can't be reduced to the 
low status of a group; "socially" enables the court to ask for something more than merely some 
educational data about a group. One cost of this firmer judicial control is to tie compensatory 
discrimination schemes firmly to such groupings as caste, locality, or occupation whose social and 
educational backwardness can be assessed separately­­unlike income strata. 
The formula also supplies a rationale for court approval of income cut­offs which eliminate from 
benefit coverage the affluent sections of a backward group. In Jayasree v. State of Kerala, the Court 
found it unobjectionable to eliminate the rich because they were not socially backward.  109.   This 
raises the possibility that a court itself might impose such a cut­off or require the state to impose 
one in order to ensure that only the socially backward were included. 
D. UNDER­REPRESENTATION 
The courts have read the qualifying words "socially and educationally" in Article 15(4) as an 
implicit explanation to Article 16(4). However, the designation of recipients of reservations in 
government employment under Article 16(4) is distinctive. For Article 16(4) permits reservations in 
favor of any "backward class of citizens which in the opinion of the State is not adequately 
represented in the services under the State." 
____________________ 
106. Thus in Dilip Kumar v. Gov't. of U.P., A. I . R. 1973 All. 592, 595, the Court finds the 
government designation insufficient as to the educational backwardness of the areas and totally 
lacking "application of mind" as to the social backwardness. 
107. Janki Prasad v. State of Jammu and Kashmir, A.I.R. 1973 S.C. 930 at 938. 
108. Id., at 937­38. 
109. A. I. R. 1976 S. C. 2381 at 2384. 
­243­ 
A determination of inadequacy of representation is a necessary but not a sufficient condition for 
reservation. Before reservation is permissible, the inadequately represented class first must be 
determined to be backward. In Triloki Nath Tiku v. State of Jammu and Kashmir [I] the Supreme 
Court noted that the power to make reservations under Article 16 (4) may be exercised only in favor 
of backward classes:  110.   "[W]hether a particular class of citizens are backward is an objective factor 
to be determined by the State."  111.   The State's determination of backwardness "is not the final word 
on the question: it is a justiciable issue."  112.   As to under­representation, however, the opinion of the 
State "may ordinarily be accepted as final, except when it is established that there is an abuse of 
power."  113.   Thus the determination of the inadequacy of representation is left to the discretion of 
the State­presumably subject to some minimum standard of plausibility. Just where the minimum 
lies is not yet clear. A percentage of representation in government posts lower than a group's 
population in the state is the commonly proffered measure of inadequate representation.  114.   A more 
exacting standard has sometimes been applied: thus the Andhra Pradesh Commission imported 
(from Balaji's discussion of educational backwardness) the notion that representation in the services 
must be "well below the state average" to determine inadequacy of representation.  115.   But the 
general assumption is that disparity with population figures is sufficient. The disparity need not be 
massive to support a judgment of under­representation. Thus in the Thomas case although it did not 
turn technically on a finding of under­representation under Article 16(4), the Court concludes that 
the fact that Scheduled Castes and Scheduled Tribes made up only 8.7% of employees in the 
Registration Department (while they were approximately 10% of the population of the state) "clearly 
shows [they] . . . were under­represented."  116.   Provisions to overcome such under­representation are 
not automatically invalid if they produce "over­representation" of the beneficiaries in a particular 
grade.  117.   
Is the backwardness required to support reservation under Article 16(4) the same as that required for 
measures under Article 15(4)? It 
____________________ 
110. A. I. R. 1967 S. C. 1283. 
111. Id., at 1285. 
112. Id., at 1285. 
113. Id., at 1285. 
114. Typically, such representation is measured by the total of appointments secured, but the Kerala 
Backward Classes Reservations Commisson ( 1971:I, 105) ascertained adequacy of 
representation by counting only posts gained in open competition, excluding those gained by 
reservation to prevent any "relapse into inadequacy" by cutting off reservations. 
115. Andhra Pradesh Backward Classes Commission 1969: 107. 
116. A. I. R. 1976 S. C. 490 at 554. 
117. Harpartap Singh v. Union of India, 1970 S. L. R. 40 (Punj. & Haryana). 
­244­ 
might be argued that under­representation is itself a species of backwardness and that Article 16(4) 
which, unlike 15(4), is not confined only to the "socially and educationally backward" authorizes 
reservations for groups which are backward in this specific respect.  118.   Inadequate representation 
has been employed by the Backward Classes Commission and by other bodies as one criterion in 
designating Backward Classes.  119.   However, it is clearly used as an indicator and component of 
educational and social backwardness, rather than as in itself a kind of backwardness sufficient to 
justify reservations.  120.   But representation in government service is not invariably regarded as a 
suitable measure of backwardness. The Havanur Commission in Karnataka deliberately omitted it 
on the ground that "inadequacy of representation is the result of social and educational 
backwardness and not the cause."  121.   
Even if the absence of "socially and educationally" in Article 16(4) might conceivably indicate that 
a backward class for purposes of that Article need not meet the same tests of social and educational 
backwardness required by Article 15 (4), it does not justify dispensing with some requirements of 
backwardness beyond mere under­representation.  122.   To permit under­representation as a sufficient 
justification for reservations would in effect read the word "backward" out of Article 16(4) and 
would open the door for a general regime of communal quotas.  123.   It would also, as the Supreme 
Court points out in Triloki Nath, confer a windfall on groups which had taken to occupations other 
than government service. The Court there concludes that there are two distinct conditions for the 
application of Article 16(4): a class of citizens 
____________________ 
118. In the now discredited case of Keseva Iyengar v. State of Mysore, A. I. R. 1956 Mys. 20, the state 
had designated as backward "all communities other than Brahmins who are not adequately 
represented in the services." This was upheld as within Art. 16 (4), the court declining to 
scrutinize the adequacy of this classification. 
119. I BCC 45­47. 
120. Thus the Kerala Backward Classes Reservation Commission regards representation in 
government service as "a sign of ability to stand on their own, a sign of non­backwardness of 
that class of citizens" ( 1971: 75). The Commission carefully distinguished its use of 
representation in services as an indicator of backwardness from its use in assessing "adequate 
representation" in the services ( 1971: 76). 
121. Government of Karnataka 1975:I, 315. 
122. Thus in State of Jammu and Kashmir v. Jagar Nath, A. I. R. 1959 J. & K. 14, the Government 
Order, presumably based on a finding that Muslims were "underrepresented," was not sufficient 
to satisfy the court that they thereby comprised a backward class. 
123. If the State may apportion reservations to specific recipients (see chap. 13, §C), these might be 
quotas much like those condemned in the Venkataramana case, with the difference that there 
each group was assured a certain number of seats, but now, since reservations are minimums but 
not ceilings (see chap. 13, §B) it would be possible for members of the favored groups to 
compete for and obtain additional seats. 
­245­ 
who are backward by the Balaji tests and who lack adequate representation in the services.  124.   
The justification for accepting under­representation as itself a conclusive criterion of backwardness 
is the notion that members of any communal group are entitled by right to a portion of posts 
commensurate with their ratio to the total population. The Constitution repudiates this notion by 
making it a Fundamental Right of every citizen not to have another preferred on the ground of his 
caste or religion.  125.   The only exception to this Fundamental Right is in favor of members of 
backward classes. However broad the discretion enjoyed by the State to designate backward classes, 
it does not include the power to deem as backward any group which may be "under­represented," for 
this would turn the exception into a general principle. 
E. ABSOLUTE VS. RELATIVE BACKWARDNESS 
The main contribution of the Balaji opinion to the problem of designating the Backward Classes 
lies, not in the provision of clear and applicable standards for testing backwardness, but in its severe 
limitations on the range of groups that it will allow to go by that name. In Ramakrishna Singh v. 
State of Mysore, the state's scheme was invalidated on the ground that some of the communities 
designated as "backward" were not "really backward" but indeed were "relatively forward"  126.   In 
the Balaji case the Supreme Court, visibly perturbed at the large portion of the population included 
among the beneficiaries, clearly indicates that those who may be deemed backward under Article 
15(4) are a restricted group. 
[T]he concept of backwardness in Article 15(4) is not intended to be relative in the sense 
that any classes who are backward in relation to the most advanced classes of the society 
should be included. . . .  127.   
Indeed, the Backward Classes intended to be covered by that Article are those who "are in the 
matter of their backwardness comparable to the Scheduled Castes and Scheduled Tribes."  128.   And 
the Constitution contemplates that such backward groups will be "treated as being similar to the 
Scheduled Castes and Scheduled Tribes."  129.   
In its handling of educational backwardness, the Supreme Court 
____________________ 
124. A. I. R. 1967 S. C. at 1285. 
125. Earlier proposals for reserved seats and posts for religious minorities were eliminated from the 
final version of the Constitution (IX CAD 701). 
126. A. I. R. 1960 Mys. at 349 ff. 
127. A. I. R. 1963 S. C. at 658. 
128. Id., at 658. 
129. Id., at 658. 
­246­ 
gives an indication of which groups might qualify as these "really backward" classes. The Mysore 
scheme included as educationally backward all communities whose enrolment in secondary schools 
throughout the state was less than the state average of 6.9 high school students per thousand of 
population. All communities whose enrolment was below 6.9 per thousand were deemed backward; 
those whose enrolment was less than 50% of 6.9 per thousand were deemed "more backward." 
Assuming this to be a rational test of educational backwardness, the Court asks 
Whether it was legitimate to treat castes or communities which are just below the state 
average as educationlly backward classes? If the state average is 6.9 per thousand, a 
community which . . . is just below the . . . [average] cannot be regarded as backward. It 
is only communities which are well below the state average that can properly be 
regarded as educationally backward classes of citizens.  130.   
What is "well below?" The Court goes on to concede that classes of citizens whose average of 
student population is below 50% of the state averages (i.e., less than 3.45 per thousand) are 
obviously Backward Classes of citizens. Muslims, whose student population is 5.0 per thousand, 
"are not so below the state average that the community could be treated as educationally backward 
in the State. . . ."  131.   If the test of student enrolment is used, "the legitimate view . . . would be that 
the classes of citizens whose average is well or substantially below the State average can be treated 
as educationally backward."  132.   While it avoids prescribing any hard and fast rule and says that 
drawing the line is up to the State, the Supreme Court clearly indicates that it will accept as 
"backward" only those whose resources and accomplishments are considerably below the local 
average. 
The Balaji Court does not indicate how its two tests of "absolute" backwardness­­"well below" state 
averages and "comparability to Scheduled Castes and Scheduled Tribe"­­are to be combined. Nor 
has there been any further development of either standard by subsequent courts. While the "well­
below" test has not recurred, the test of backwardness comparable to that of the Scheduled Castes 
and Tribes has exercised an attraction on some courts.  133.   
____________________ 
130. Id., at 660 (emphasis added) 
131. Id., at 661. 
132. Id., at 661. 
133. This theme of resolution of the Backward Classes into the Scheduled Castes surfaces in 
Chitralekha v. State of Mysore, A. I. R. 1964 S. C. 1823, where Subba Rao, J., after 
animadversions against the use of caste units says, "If . . . [an] entire sub­caste, by and large, is 
backward, it may be included in the Scheduled Castes by following the appropriate procedures 
laid down by the Constitution" (id., at 1834). 
­247­ 
The heaviest use of this standard has been in Andhra.  134.   In Sukhdev v. Government of Andhra 
Pradesh the court in striking down the Backward Classes list noted that the Backward Classes for 
whom special provisions are contemplated are those comparable in their backwardness to the 
Scheduled Castes and Scheduled Tribes.  135.   In Sagar v. State of Andhra Pradesh, where the 
principle ground for striking down the revised list was the State's exclusive reliance on caste in 
constructing the list, the High Court explicitly applied this criterion.136 The court noted that the 
criterion of backwardness set forth in Balaji, that the Other Backward Classes be similar to the 
Scheduled Castes and Scheduled Tribes, was not applied in the preparation of the list in question. 
"There is no material from which we could determine that the classes shown in the list are similar to 
the Scheduled Castes and Scheduled Tribes."  137.   The Supreme Court, upholding the judgment, 
repeated and presumably endorsed this admonition.  138.   Nowhere do these courts indicate exactly 
what kind of comparability they have in mind­­whether of status, of disabilities suffered, of 
economic or educational condition, or of representation in government services. Particularly since 
the assertion of this standard is coupled with strictures against selection based solely on caste, it is 
difficult to see how such comparability could be shown. 
In response, the Andhra Pradesh Backward Classes Commission formulated a rather ingenious 
reading of the "comparable to Scheduled Castes and Tribes" standard. They found the communities 
comprising the OBCs to be those which are hindered at present by the accumulated results of a 
heritage of past disablement which, like that of the Scheduled Castes, "has hindered social progress 
in the past by denial of access to educational institutions [etc.] . . . because of traditional 
occupation . . . and . . . left vestiges which . . . constitute . . . a serious impediment to social life." 139. 
In the next round of Andhra litigation, the Supreme Court relaxed its comparability requirement 
without giving it further specificity. Backward Classes must be comparable to Scheduled Castes and 
Tribes, but if they are "really . . . backward" they need not be "exactly similar in all respects."  140.   In 
Janki Prasada Constitution Bench asserted that this "comparability" was a "fundamental 
requirement." The Court sug­ 
____________________ 
134. Cf. the dissenting judgment of Gopalan J. Nambiar, in Hariharan Pillai v. State of Kerala, A. I. 
R. 1968 Ker. 42, which found troubling the fact that Balaji's suggestions had not been adverted 
to in making the reservations (id., at 55). 
135. 1966(1) Andh. W.R. 294. The Court cites the provisions of Art. 340(1) as evidence for the 
"comparability" test. 
136. A. I. R. 1968 A.P. 165. 

137. Id., at 187. 
138. A. I. R. 1968 S. C. 1379 at 1383. 
139. Andhra Pradesh Backward Classes Commission 1970: 51. 
140. State of A.P. v. Balaram, A. I. R. 1972 S. C. 1375 at 1395. 
­248­ 
gests that the groups may be equated in terms of failure to develop "enthusiasm for education." That 
is, it is those "sectors of the population which show extreme apathy toward education due to age­old 
customs and habits of living, fostered by poverty, ignorance, superstition and prolonged social 
suppression" that need to be "goaded into the social stream by positive efforts of the State." The 
Scheduled Castes and Tribes exemplify such social and educational backwardness, and their 
backwardness in these respects provides the model for the selecting of Other Backward Classes.  141.   
Subsequently, the Supreme Court reiterated the comparability formula in the course of upholding 
Kerala's income cut­off provisions, but there was no further development of specific bases of 
comparison.  142.   So far it is difficult to see that this doctrine has made any difference in outcome, 
but it clearly provides the Court with doctrinal potential for resisting expansive listing of Backward 
Classes. It also provides a basis for eventual merger of the Backward Classes and Scheduled Caste 
categories and undermines attempts to achieve a radical separation between them.  143.   
F. THE SUFFICIENCY OF DATA 
In Kesava Iyengar v. State of Mysore,  144.   the court declined to examine the adequacy of the 
classification, finding that "[d]oubtless the State is the sole authority to classify the communities as 
'backward classes.'"  145.   The state "list" was actually a Government Order passed in 1921­­34 years 
before­­and the fact that the Government had not seen it to revise it was accepted as sufficient 
evidence of its current accuracy. With the Ramakrishna Singh case, however, the Mysore Court 
became much stricter in scrutinizing the data used to make the determination of backwardness. 146. 
It found the state's population and 
____________________ 
141. Janki Prasad v. State of J. & K., A. I. R. 1973 S. C. 930 at 938. The Court is not suggesting that 
there is a definite line between the backward groups and others. Its discussion of comparability 
follows immediately after this observation: "The words 'advanced' and 'backward' are only 
relative terms­­there being several layers or strata of classes hovering between 'advanced' and 
'backward,' and the difficult task is which class can be recognized out of these several layers as 
being socially and educationally backward." 
142. Jayasree v. State of Kerala, A. I. R. 1976 S. C. 2381 at 2383. 
143. E.g., justice Krishna Iyer attempt in State of Kerala v. N. M. Thomas to separate Scheduled 
Castes and Tribes as a "superclassification" for which treatment quite distinct from Other 
Backward Classes is warranted. See discussion in chap. 11, §D, below. 
144. A. I. R. 1956 Mys. 20. 
145. Id., at 28. 
146. A. I. R. 1960 Mys. 338. 
­249­ 
literacy figures, taken from the 1941 census, defective because "considerable changes have taken 
place . . . [so that] the census report of 1941 cannot be any basis for the determination of Backward 
Classes in 1958"  147.   However, in the Partha case, the Mysore Court relented to the, extent of 
finding reasonable the projections of figures on caste population and literacy from older censuses­­
from the 1941 census where figures were available, but in some cases the 1931, 1911, and even 1901 
figures. The court found this not to be too unreasonable.  148.   
Balaji envisioned much higher standards in regard to the data upon which determinations of 
backwardness were to be made. The Court observes that 
evolving proper criteria for determining which classes are socially backward is 
obviously a very difficult task; it will need an elaborate investigation and collection of 
data and examining the said data in a, rational and scientific way. That is the function of 
the State which purports to act under Article 15(4).  149.   
Several high courts embraced and applied this high standard. In Sukhdev v. Government of Andhra 
Pradesh,  150.   a single judge struck down a scheme of reservations in medical college admissions on 
grounds, inter alia, of the absence of evidence that the government had acted on concrete and 
reliable data or had held any inquiry to ascertain the social and educational backwardness of the 
listed groups. Dismayed by the fact that the list in question was 25 years old and that there were not 
even state averages against which to measure the listed groups, the court found this procedure a 
violation of Article 15.  151.   Subsequently, in Sagar v. State of Andhra Pradesh, the same judge (now 
Chief Justice), sitting on a Division Bench, had an opportunity to refine and extend these standards. 
 152.   After the earlier list was struck down, the state switched to an income test while a Cabinet Sub­
Committee was established to determine state policy regarding preferences for Backward Classes. 
The Cabinet Sub­Committee, after soliciting information on criteria from various states, directed 
that the Director of Social Welfare should check the lists of Backward Classes to select 
those castes or communities which, apart from the point of view of caste are considered 
backward on account of the low standard of living, education, poverty, place of 
habitation, inferiority of occupation followed, etc.  153.   
____________________ 
147. Id., at 347. 
148. A. I. R. 1961 Mys. 220 at 232. 
149. A. I. R. 1963 S. C. 649 at 659. 
150. 1966(1) Andh. W. R. 294. 
151. Id., at 304, 305. 
152. A. I. R. 1968 A.P. 165. 
153. Id., at 185. 
­250­ 
The Director accordingly drew up a list of 113 communities, which was submitted to the Cabinet 
Sub­Committee, which in turn recommended them to the cabinet, which deleted certain items from 
the list, overruled the priorities suggested by the Director of Social Welfare, and promulgated the 
list. 
The High Court expressed dissatisfaction with this procedure. It found no indication that materials 
had been placed before the cabinet from which it could be concluded that the Balaji criteria had 
been used in preparing the list. 
[W]e are not able to ascertain whether any material, and if so what material, was placed 
before the Cabinet Sub­Committee, upon which the list of Backward Classes was drawn. 
. . . [I]t was stated below that the Law Secretary and the Director of Social Welfare sat 
down together and drew up a list, the former specifying the legal requirements and the 
latter as an expert advising on the social and educational backwardness of class or 
classes.  154.   
The Director of Social Welfare, who disclaimed personal knowledge that the 113 communities were 
socially and educationally backward, deposed that he arrived at his assessment by his direct 
knowledge and discussion with officials, reinforced by consultation of standard works.  155.   
To accept the contentions . . . that we should rely on the expert knowledge of the 
Director of Social Welfare would be tantamount to saying that we have to be satisfied 
with the personal knowledge of an officer of the government who himself is frank 
enough to admit that the classification is a difficult one and that he cannot say that he 
has personal knowledge of the life conditions and needs of these classes and that his 
knowledge was acquired as a result of his considerable touring of many districts . . . and 
information derived from books, like Thurston and Siraj­ul­Hasan, which information . . 
. is available to everyone.  156.   
The Court pointed out that these authors wrote nearly 60 years before, and that they had not 
ascertained the educational backwardness of these castes, nor would their books disclose what were 
the criteria adopted in determining social and educational backwardness. Indeed, 
____________________ 
154. Id., at 186. 
155. A measure of the accuracy of these methods is provided by the work of the Commission, which 
recommended a new OBC list for Andhra Pradesh in 1970. It found that of the 32 groups listed 
by the Director as "Aboriginal Tribes," 15 were synonyms, titles, mistaken renderings of, or 
subgroups within, groups already on the Scheduled Tribes or Scheduled Castes lists; 3 were 
unlocatable; 2 were groups that were well­off. This left 12 groups out of 32 to be included in the 
new list of OBCs ( Andhra Pradesh Backward Classes Commission 1970: 60 ff.). 
156. A. I. R. 1968 A.P. 165 at 187. 
­251­ 
according to some of these books, the listed castes are relatively forward.  157.   
Neither the Director of Social Welfare nor the Cabinet Sub­Committee nor the Cabinet 
had before them the population of each of these classes, their economic conditions, 
percentage of literacy or education, inferiority of their occupations, their habitation, or 
their social and economic status vis­a­vis the Scheduled Castes and Scheduled Tribes. . . 
. The fact that the information cannot be easily obtained from the Census reports can 
afford no valid ground of justification for sustaining the list of backward classes under 
Article 15(4). This information could be obtained, if the state required it, and it is not for 
us to suggest the manner in which it can be done, a duty which it should perform 
however inconvenient or difficult it maybe. . . .  158.   
Before the Sagar case reached it, the Supreme Court had encountered the data problem in two 
Article 32 petitions (in which it sat as a court of first instance). In Triloki Nath Tiku v. State of 
Jammu and Kashmir [I],  159.   it found among other defects of the scheme that there was no material 
on the relevant facts: the population and state of backwardness, social, economic, or cultural, of the 
groups listed.  160.   "What is placed before us is a general assertion, unsupported by any acceptable 
data, that all Muslims . . . and the majority of the Hindus . . . are backward."  161.   Census figures 
from 1941 had been placed before the Court. But the Court cautioned that "the 1941 figures may not 
afford any workable guide, as a quarter of a century has passed by since then and there must have 
been revolutionary changes during the period."  162.   Rather than invalidate the scheme in default of 
this information, the Court adopted the unusual course of directing the High Court to gather the 
necessary material and report back to the Supreme Court. 
In P. Rajendran v. State of Madras, the Court was quite receptive to Madras's claim that the main 
criterion upon which its list of castes was chosen was "the social and educational backwardness of 
the caste based on occupations."  163.   There is nothing in the judgment from which it might be 
gathered that the Court was exacting in its demands for data to substantiate this; indeed, it seems not 
to have probed to see whether it was actual or traditional occupation that the state was employing. 
When it encountered Sagar, the Supreme Court put considerable stress on the absence of evidence 
of the state's use of criteria other than 
____________________ 
157. Id., at 186. 
158. Id., at 187. 
159. A. I. R. 1967 S. C. 1283. 
160. Earlier, the Jammu and Kashmir High Court struck down medical college reservations on the 
ground that the State had provided no data on either educational or social backwardness ( Lalita  
Shuri Tiku v. State of J. & K., A. I. R. 1966 J. &. K. 101). 
161. Id., at 1284. 
162. Id., at 1284. 
163. A. I. R. 1968 S. C. 1012. 
­252­ 
caste and appeared to endorse the High Court's standards on data.  164.   "Being in the nature of an 
exception [i.e., reservations under Article 15 (4)] the conditions which justify departures must be 
strictly shown to exist."  165.   The mere assertion by the state that its officers had taken into 
consideration relevant criteria was not sufficient. Whether such criteria had been employed was up 
to the courts, who would not assume that constitutional criteria had been employed. The High 
Court's decision striking down the state list was affirmed; the state had not made a suitable 
investigation nor collected the material necessary to ascertain the social and educational 
backwardness of the listed groups. 
Two weeks after the Supreme Court decision in Sagar, the Andhra Pradesh Government appointed a 
nine­member Backward Classes Commission, chaired by the retired Chief justice of the High Court, 
including six legislators and an I.A.S. officer as secretary. Unlike its discredited predecessor, this 
Commission worked openly, sending out questionnaires, touring the districts, recording evidence. 
Its attempt to apply measures of school enrolment and other quantitative standards was severely 
"handicapped for want of up­to­date statistical information in regard to population of the several 
communities in this State."  166.   The Commission was forced to rely on projections from old census 
data: "[t]he population figures for each caste for 1968 have been arrived at by inflating the 1921/ 
1931 census figures by the percentage of the increase of the total population."  167.   Data on student 
enrolments were based on projections from returns by about half of the schools. Occupation and 
income data were derived from the testimony of informants, either at hearings or on the 
Committee's questionnaires. Although, unlike the Kerala Commissions, this one does not set out its 
statistical findings, but merely reports its conclusions, it represents an advance in precision and 
system over the earlier Andhra efforts. 
After two years of work, the Commission recommended a list of 92 communities, estimated to 
constitute over 38% of the State's population, to be the Backward Classes. It recommended an array 
of programs to benefit them, including reservations of 30% of places in government service and in 
admissions to professional education, these reservations to be reviewed after ten years. The state 
government accepted the Commission's recommended list of Backward Classes and reserved 52% 
of places for them. When the validity of the new rules was challenged, the High Court once again 
struck down the Government Order on the ground that no proper investigation and collection of data 
____________________ 
164. A. I. R. 1968 S. C. 1379. 
165. Id., at 1384. 
166. Andhra Pradesh Backward Classes Commission 1970: 12. 
167. Id., at 13. 
­253­ 
had taken place. On appeal, the Supreme Court thought that this time the High Court had gone 
astray in reading Balaji to forbid the use of castes as units in designating Backward Classes.  168.   
The petitioners who challenged the state's Order had not specifically attacked the inclusion of any of 
the groups on the list, and the Supreme Court was indisposed to accept an attack on the overall 
process of the Commission.  169.   The Court was sanguine about the inevitable crudeness of any such 
investigation. Thus the use of 1921 and 1931 census data where it was the best available was not 
fatal.  170.   Nor was the use of "the personal impression gathered by members of the commission . . . 
to augment the various other materials gathered as a result of detailed investigation" in any way 
improper. Indeed, the Commission was commended for its efforts to find out about actual living 
conditions.  171.   The Commission's efforts were not to be judged by inappropriate standards: 
In spite of best efforts that any Commission may make in collecting materials and data, 
its conclusions cannot be always scientifically accurate in such matters. . . . [T]he proper 
approach . . . should be to see whether the relevant data and materials referred to in the 
report of the Commission justify its conclusions. . . . [T]here was sufficient material to 
enable the Commission to be satisfied that the persons included in the list are really 
socially and educationally backward.  172.   
Thus in the setting of the conscientious performance of the Commission, the Court relented on the 
sterner pronouncements about data inspired by its earlier encounters with more political 
commissions. 
The course of developments in Kerala illustrate a different judicial approach to the data problem. In 
Kerala, immediately after Balaji, a single judge struck down the scheme of reservations in 
professional colleges.  173.   Among the Court's reasons for rejecting findings of backwardness was 
that no investigation had been made or data collected. "[N]o conclusion can be arrived at in 1963 
regarding the educational backwardness . . . of any group on the basis of figures collected in 1941." 
 174.   The data must refer to circumstances existing "within a reasonable time of the passing of the 
order declaring a particular group as a backward class."  175.   The Division Bench, reversing the 
striking down of reservations for several groups, found the paucity of up­to­date data a source of 
considerable worry.  176.   For the short run they were 
____________________ 
168. State of Andhra Pradesh v. Balaram, A. I. R. 1972 S. C. 1375. The High Court decision was not 
reported. 
169. Id., at 1396. 
170. Id., at 1399. 
171. Id., at 1399. 
172. Id., at 1396. 
173. Jacob Mathew v. State of Kerala, A. I. R. 1964 Ker. 39. 
174. Id., at 60. 
175. Id., at 61. 
176. State of Kerala v. Jacob Mathew, I. L. R. 1964 (2) Ker. 53. 
­254­ 
willing to rely on a kind of personalized judicial notice.  177.   "It is impossible to say that our 
conclusion had not been influenced to some extent at any rate by our experience of life and work in 
this state."  178.   But no enduring conclusion could be based upon outdated data or "on judicial 
experience which such data may disprove or modify."  179.   It was essential that the state embark 
immediately upon a fact­finding inquiry. When the case was appealed to the Supreme Court, the 
state agreed to do just this, and the appeal was withdrawn after argument.  180.   
The resulting Commission,  181.   chaired by a retired High Court judge, assembled and analyzed in a 
thorough and workmanlike way a variety of data: mainly on educational accomplishments and the 
existence of disabilities, but with some reference to occupation, income, and landholding. Although 
the Commission relied upon projections from old census data for its population figures, it not only 
collected data from government and educational institutions, but commissioned the Bureau of 
Economic and Statistics to conduct a sample survey of certain aspects of socio­economic 
conditions. In addition, it collected an enormous number of submissions by individuals and groups. 
Recommending the continuation of reservations in educational institutions on a diminished scale for 
ten years, it proposed retention of communal units with the addition of an income ceiling. 
On the basis of these recommendations, Kerala revised its lists for educational reservations, but 
continued to use the old ones for reservations in government service. When the service reservations 
came before a full bench in Hariharan Pillai v. State of Kerala,  182.   the dissenting judge, responding 
to numerous defects of the scheme, found the data meagre and out of date. He found it "impossible 
to say that the assessment of backwardness had not been either solely or predominantly on 
considerations of caste or community, or that [it] has proceeded [along the lines indicated in Balaji." 
 183.   However, the majority, while equally troubled by the paucity of current data, opted for the 
strategy of stimulating an inquiry that had worked so well in the admissions area. The Court was 
particularly concerned lest there were non­backward sections within the backward communities and 
that investigation be made to insure against this and to insure that the groups denominated as 
backward were indeed "by and large backward, socially and econo­ 
____________________ 
177. Cf. judicial notice of a group's notorious backwardness in Chait Ram v. Sikander, A. I. R. 1968 
Pat. 337. 
178. I. L. R. 1964(2) Ker. 53 at 61. 
179. Id., at 61. 
180. Interview with Shri P. Govinda Menon, M.P., New Delhi, Dec. 15, 1965. 
181. Government of Kerala 1965. 
182. A. I. R. 1968 Ker. 42. 
183. Id., at 56. 
­255­ 
mically."  184.   The Court was unwilling, on the basis of the materials before it, to hold that this was 
not the case, but found it "necessary to strike a serious note of warning."  185.   The data (a 1935 
Committee, the 1941 Census, a 1957 Committee) were out of date. "It is essential that relevant data 
be collected periodically."  186.   Action under Articles 15(4) and 16(4) must be "modulated from time 
to time." "This can only be done if surveys are made at regular intervals and detailed information 
collected."  187.   It was desirable that the state undertake a detailed survey as early as possible. "There 
will be no justification in continuing to apply the principles embodied in Rules 14 and 17 . . . after 
31­3­1968 without a fresh appraisal of the questions involved."  188.   
Less than nine months later, the Kerala Government established the Backward Classes Reservation 
Commission. Its six members included legislators, two advocates, and (a revealing innovation) the 
Director of the Bureau of Economics and Statistics as its member­secretary. Unlike all the other 
commissions to this date, the Nettur Commission abandoned reliance on extrapolations from old 
census figures and ascertained group population by a sample survey.  189.   It also collected some data 
on household income as well as educational and employment data from government and schools. By 
an extremely circuitous route, the Commission recommends that 38% of posts be apportioned 
among nine Backward Classes. (These are nine of twelve groupings into which all the communities 
in the state are sorted; the Backward Classes are those members of each grouping with annual 
family income of less than Rs. 8,000. These nine, having obtained fewer posts on merit than their 
proportion in the population, are deemed inadequately represented.)  190.   The Commission's 
recommendations, "accepted" by the government in 1973 but later rejected by a different ministry, 
had not been put into effect as of 1979.  191.   
The Karnataka Backward Classes Commission was established in 1972 in response to persistent 
criticism of the "income­cumoccupation" test that prevailed there from 1963, under the 
chairmanship of its most redoubtable critic, L.G. Havanur.  192.   Setting out forthrightly to work on 
the basis of community units, the Commission compiled an immense fund of data on the various 
communities of the 
____________________ 
184. Id., at 51. 
185. Id., at 51. 
186. Id., at 51. 
187. Id., at 51. 
188. Id., at 51. The Hariharan judgment was delivered on Jan. 31, 1967, 14 months before the date 
mentioned. The Commission was established on Oct. 17, 1967. 
189. Kerala( Backward Classes Reservation Commission) 1971: I, 87; II, 462. 
190. Id., at I, 60 ff. 
191. See "The Politics of Reservation" in The Hindu, 29 Jul. 1977. Namboodiripad 1976: 336. 
192. See the chronology in Government of Karnataka 1975: I, 7 ff. 
­256­ 
state. Unfortunately, it reverted to the use of projections to calculate the populations of these 
communities.  193.   But in spite of lack of cooperation,  194.   the Commission carried out a massive 
survey. Two hundred villages, at least one from each taluk, and 204 blocks or wards from urban 
areas were selected. Of these, 193 villages and 185 urban blocks were actually surveyed. 
Every individual in the selected villages and urban blocks/wards was covered under the 
said survey. . . . About 63,650 families with a population of about 3,55,000 have been 
covered. 378 investigators collected the information for more than six months under the 
supervision of 20 District Social Welfare Officers and 19 District Statistical Officers. In 
addition, the Deputy Secretary of our Commission, who is also a Joint Director of 
Bureau [sic] of Economics and Statistics, supervised the survey work. Training was 
given to the Investigators in the respective headquarters and two meetings of all the 
District Social Welfare Officers were held in the office of the Commission for giving 
them necessary instructions. Out of the 404 villages and town/city blocks the 
Commission received the survey material from 193 villages and 185 urban blocks. In 
addition to the office staff, the Commission appointed 100 Tabulators. The complexity 
and magnitude of the work was such that the entire staff had to work beyond office 
hours for more than one­and­half years.  195.   
Those surveyed were 1.18% of the entire population of the state.  196.   Household interviews elicited 
information on 
personal characteristics, education, employment, household particulars, land holdings, 
ownership of houses, living conditions, health conditions, livestock, agricultural 
implements, vehicles, radios and news­papers, membership in co­operatives and 
political organisations, cottage industries, income and expenditure and traditional 
occupation.  197.   
Information was also collected at the village, ward, and taluk levels about panchayats, cooperatives, 
money lending, educational institutions, forced labor, and a host of other matters.  198.   
Notwithstanding the thoroughness and rigor of its work, the Commission's recommendations were 
the focus of heated controversy. As in Kerala, it was the exclusions as well as the inclusions that 
delayed 
____________________ 
193. Government of Karnataka 1975: II, 1­2. Cf. the note on estimated population by the Andhra 
Pradesh Backward Classes Commission 1970: 156. 
194. The Commission notes that it could not get the Directorate of Social Welfare, the Bureau of 
Economics and Statistics, or the Census Department to conduct the survey ( I, 308). But the 
Commission did secure additional funding to enlarge its data collection activities. 
195. Id., at I, 308. 
196. Id., at I, 310. 
197. Id., at I, 309. 
198. Id., at I, 309­10. 
­257­ 
acceptance. When the Government did implement the Commission's recommendations, it added to 
the 185 communities recommended by the Commission some 20 groups that the committee had 
omitted, excluding the relatively prosperous and politically potent Lingayat community.  199.   The 
measure and moderation of the Commission was also abandoned in the portion of places reserved: 
where the Commission recommended 32%, the state government reserved 50% (thus returning to 
the pre­Balaji level). 
A Backward Classes commission in Gujarat, sitting contemporaneously with the Havanur 
Commission in Karnataka, marked a new high point in the utilization of social science expertise. 
 200.   The five­member commission included two eminent social scientists, and other highly qualified 
social scientists were utilized in staff and advisory positions. In addition to industriously assembling 
data by hearings and inspections, the Commission set out to conduct a comprehensive two­stage 
sample survey of households throughout the state. The first phase ­­ a statewide socio­economic 
survey of 6,449 household ­­ as conducted, but the second phase ­­ a purposive sample to augment 
the frequency of observations of groups not adequately included in the first phase ­­ was 
unfortunately abandoned owing to time pressure. 
The Karnataka and Gujarat reports mark an advance in the scope of data collected and the rigor 
with which it is analyzed. They suggest possibilities for development of systematic surveys as a tool 
to identify Backward Classes and to monitor programs for the amelioration of their conditions. 
Government need not employ such Commissions when it designates Backward Classes. Of course, 
the evidentiary standards applied to commission data are equally applicable when the Government 
makes the determinations without benefit of such an inquiry. "[T]he State should produce material 
before the Court to show that there was a proper inquiry, relevant criteria were adopted and the 
decision on [who are backward] is reasonable. Mere expression of opinion is not sufficient." 201. 
Thus a State scheme was struck down where "there was no application of mind" in determining the 
social backwardness of a group and the criteria for educational backwardness were faulty.  202.   
____________________ 
199. Government of Karnataka: G.O. No. DPAR­1 SBC, 77, dated March 4, 1977. 
200. Government of Gujarat 1976. 
201. Dilip Kumar v. Government of U.P., A. I. R. 1973 All. 592 at 595. These requirements may be 
applied with leniency. Thus a state may act on incomplete data ­­ it need not wait until all the 
data is in before acting ( Viswanath v. Government of Mysore, A.I.R. 1964 Mys. 132 at 136). Nor 
need all the supporting data be included in the order providing for reservation. Justification to 
satisfy the court can be provided later and separately ( Jacob Mathew v. State of Kerala, A. I. R. 
1964 Ker. 39 at 51). 
202. A. I. R. 1973 All. 592 at 595. 
­258­ 
All of the recent commissions recommend that reservations be made for a fixed period and 
then reviewed. In Periakarrupan the Supreme Court made it clear that periodic re­assessment 
was obligatory: 
. . . Government should not proceed on the basis that once a class is considered a 
backward class it should continue to be backward class for all time. . . . The Government 
should always keep under review the question of reservation . . . only the classes which 
are really socially and educationally backward should be allowed to have the benefit of 
reservation.  203.   
Candidates on the existing list of Backward Classes had secured about half the seats in the general 
pool. This showed that "the time has come for a de novo comprehensive examination of the 
question."  204.   Thus the courts have moved from a position of unquestioning acceptance of 
government action to one of close, if intermittent and uneven, probing to insure that the government 
has acted on data sufficient to support its claim to have used acceptable criteria. Some courts have 
applied such standards unrelentingly to strike down an order, others have invoked them to prod the 
state into upgrading its performance. In so doing, the courts have helped institutionalize periodic 
review and pushed the states beyond primitive data collection and simple assertion into systematic 
generation and analysis of data. The generation and analysis of data suggests new possibilities for 
designing programs for the Backward Classes and monitoring their implementation. And it 
foreshadows a new generation of problems for judicial control. The data in surveys varies in quality 
and in the inferences that may be drawn with confidence from it. Problems of sampling error, 
coding reliability, intervening variables, and such threaten the inferences that may be drawn. As the 
designation of Backward Classes involves the collection and interpretation of survey data, courts 
will have to devise some minimum standards for the handling of empirical data. 
The first target of such standards might well be the practice of extrapolating population figures from 
old census data. The preIndependence enumerations of population by caste were open to some 
suspicion about their accuracy ­­ perhaps most in the case of groups of low status.  205.   And there is 
reason to think that different communities have experienced different rates of fertility, mortality, and 
migration in the intervening years.  206.   As time goes by, the speculativeness of such 
____________________ 
203. A. I. R. 1971 S. C. 2303 at 2311. 
204. Id., at 2311. 
205. The Andhra Pradesh Backward Classes Commission ( 1970: 14) observes that in pre­preferential 
treatment days there may have been systematic understatement of the population of communities 
"about whom there was some stigma." Thus projections would understate their population and 
overstate, e.g., their current educational accomplishments. 
206. On differential fertility, see e.g., Mandelbaum 1974: 42­59; Driver 1963: 89; cf. K. Davis 1951: 
179. 
­259­ 
estimates is compounded and the interpretation of good current data confounded. Continued 
judicial indulgence of this practice seems out of line with the judicial thrust toward more 
adequate data in designating the backward. But a substitute is not easy to come by. Are 
governments to be required to carry out surveys? 
If there are to be Backward Classes defined in part by caste units, the absence of population figures 
for these units renders difficult the task of assessing their condition. Central and state governments 
have been reluctant to generate data broken down along caste lines.  207.   The Census did not collect 
caste data (except regarding Scheduled Castes) in the post­Independence censuses. The caste 
enumeration and the census recording of social precedence was perceived as a device of colonial 
domination, designed to undermine as well as to disprove Indian nationhood. There was revulsion at 
the unseemly scramble to use census listings to upgrade caste status and to inflate numbers for 
political advantage. And abandonment of caste enumeration seemed a step toward the 
disestablishment of caste. In 1955 the Backward Classes Commission remarked on the shift in 
emphasis in Census operations from ethnographic to economic. Finding itself adrift without 
accurate data, the Commission recommended collection of caste data on the 1961 census.  208.   But 
Chairman Kalelkar in his covering letter denounced the recording of caste as a violation of 
voluntarism and beset by immense practical difficulties.  209.   That a caste enumeration would 
aggravate caste consciousness and undermine modernization has remained the received view. 
The practical difficulties raised by the caste­blind policies of the Census have bedevilled Backward 
Classes commissions. The Andhra Pradesh and Karnataka Commissions have urged resumption of 
caste enumeration.  210.   For some, governmental unwillingness to collect caste data is a deliberate 
move to "preserve the status quo of disabilities, distinctions and injustices."  211.   Thus Havanur urges 
that census data be required to implement the commitment to uplift the weaker sections. He 
dismisses the disestablishment argument with the observation that government has not found it 
necessary to cease recording data on the basis of any of the other suspect criteria ­­ religion, 
language, sex, etc. The argument that omission of caste enumeration lessens caste consciousness he 
finds "amusing and absurd as when a merchant omits labelling grain bags to prevent bandicoots 
from eating the grains."  212.   
____________________ 
207. Cf. Government of Karnataka 1975: I, 309. 
208. I BCC II. 
209. I BCC: xvii ff. 
210. Government of Karnataka 1975 : I, 332 ff; Andhra Pradesh Backward Classes Commission 
1970: 15. 
211. Government of Karnataka 1975: I, 336. 
212. Id., at 336. 
­260­ 
The controversy about collection of caste data echoes the larger conflict over the use of 
communal or of neutral units in designating Backward Classes (discussed in chapter 7). 
Avoidance of communal units obviates the need for caste data; to measure the backwardness 
of communities requires data on caste lines. Since the courts have accepted the use of caste 
units, their insistence on data of higher quality implies some responsibility to assemble 
appropriate data to justify their designations of backwardness and the desirability of data to 
assess the results of programs. Such data might be produced by regular decennial census, by 
special census operations, or by surveys conducted by other agencies. Not only have courts 
insisted on fresh and detailed investigation as a prerequisite to preferential treatment,  213.   but 
also the collection of "necessary materials" under judicial auspices has been directed by the 
Supreme Court.  214.   The possibilities of judicial action to require assembly of an appropriate 
data base have not been pursued, although apparently such possibilities are present. 
G. ECONOMIC TESTS 
Economic tests of backwardness have been highly praised by critics of community­based 
preferential treatment. Income and occupation tests of various kinds are used in many schemes for 
Other Backward Castes and to a lesser extent in schemes for Scheduled Castes and Tribes. It is 
important to distinguish three different ways in which such economic tests can be employed in 
designating the Backward. 
First, income and occupation may be used as criteria for selecting backward communities. Second, 
income and occupation tests may be employed in conjunction with community membership to 
determine individual eligibility for preferential treatment. Third, income and occupation may be 
used independently as criteria for selecting individuals who compose a Backward Class. 
1. Income and Occupation as Criteria for Selecting Backward Communities 
The thrust of the Balaji case was that backward communities should be selected by criteria other 
than merely caste standing ­­ especially po­ 
____________________ 
213. Hariharan Pillai v. State of Kerala, A. I. R. 1968 Ker. 42 at 51. 
214. Triloki Nath Tiku v. State of Jammu and Kashmir [I], A. I. R. 1967 S. C. 1283 at 1286. The Court 
directed the High Court, either directly or through the district court, to gather material on 
population, community strengths in the services, the extent of backwardness of the various 
communities, etc. 
­261­ 
verty, occupation, and residence.  215.   In P. Rajendran v. State of Madras the Supreme Court upheld 
the list where the state claimed that occupation was the principal ground upon which the 
communities had been selected.  216.   The Court apparently did not ascertain whether it was actual or 
traditional occupation that the state had employed.  217.   Apparently, it was the latter  218.   ­­ a standard 
which seems not very different in its result than the "caste test" that the Supreme Court found 
objectionable in Balaji.  219.   The courts have never formulated any standards for the application of 
income or occupational tests to communities. It is not cleat whether such limitations as "well below 
the state average" and "comparable to Scheduled Castes and Tribes," which were formulated in 
respect to educational backwardness, would apply to economic factors as well. The absence of 
castewise census data would make it exceedingly difficult to apply such tests. Nevertheless, 
assessments of the income and occupational distribution of com­ 
____________________ 
215. Shri Havanur, a staunch proponent of the use of communal units, objected to the Nagan Gowda 
Committee's failure to apply these tests. He suggested that a caste should be considered 
backward if its per capita income is less than half the state average, or if more than 50% of its 
members are engaged in occupations regarded as inferior ( 1965: 110). 
216. A. I. R. 1968 S. C. 1012. 
217. If it was actual occupation, this involves an intriguing problem of data. How did the State 
ascertain this? Castewise occupational figures had not been collected in censuses since 1941 or 
earlier. 
218. Hardgrave ( 1969: 144) recounts the vacillation of the Madras Government in regard to the 
listing of Nadars as a Backward Class: "The government wanted to make it clear that the 
concession . . . contemplated for the Shanan community is admissible only to persons whose 
traditional occupation is toddy­tapping.' This only compounded the difficulties, as the 
traditional occupation of all Nadars was toddy­tapping. In 1962 the government clarified its 
position, saying that the concessions 'should be given only to the children of those persons who 
had been actually toddy­tappers until the introduction of total prohibition in their native 
districts.' In a memorandum to the chief minister, the president of the D. M. N. S. [Dakshina 
Mara Nadar Sangam] opposed 'actual' occupation as the criteria [sic] for eligibility to 'backward 
classes' benefits. 'This is patent injustice,' he wrote. 'In the case of no other community is actual 
occupation of the individual parent made the criterion!'" In 1963, Nadars were put on a par with 
other Backward Classes in this regard. 
219. It appears that traditional occupation was the single factor most heavily relied on by the 
Backward Classes Commission in the determining the backwardness of various communities in 
a way that was labelled as a caste test. Traditional occupation is not always readily ascertainable, 
particularly if the groups are large. Cf. the rejection of this standard by the Nagan Gowda 
Committee on the ground that it was too difficult to define (Government of Mysore 1960: 2­3). 
If a test of disabilities or low status is employed, traditional occupation may be redundant. Thus 
the Kumara Pillai Commission rejected occupation as a separate test on the ground that the 
incidence of the stigma attaching to certain occupations was on the low income groups of those 
communities which had traditionally followed those occupations, groups that were already 
included by the "caste" test ( Kerala[ Kumara Pillai Commission] 1966: 37 ff.). 
­262­ 
munities play a prominent role in the assessment of backwardness by some of the state Backward 
Classes commissions. Traditional occupation and current employment in unclean, undignified, and 
unremunerative occupations bulk large, and commissions have attempted to take income into 
account in spite of formidable difficulties of ascertaining it. Of course, income and occupation here 
are used to characterize the group. The individual beneficiaries of programs may not share these 
characteristics, and indeed in the case of high­echelon benefits they often (but not always) will not. 
2. Income and Occupation Tests in Combination with Communal Criteria 
Economic tests may be combined with communal criteria in another way: the economic test may act 
as a ceiling, excluding from preferences the most prosperous section of the backward community. 
Thus, e.g, Kerala adopted a "means­cum­community" test after the Kumara Pillai Commission 
concluded that the heritage of disabilities did not seriously impinge on the well­off families within 
the listed communities. Medical­college seats were reserved for members of listed communities 
whose annual family income was less than Rs. 6,000.  220   This scheme was challenged from two 
directions: by an applicant who met the income test but was not a member of one of the listed 
communities, and by members of the listed communities with higher incomes. The High Court 
rejected the claim of the low­income non­member on the ground that it was open to the state to give 
caste some relevance along with income ­­ in effect, to apply the income test only to members of 
certain castes.  221.   
Surprisingly, the second challenge was initially successful. A single judge of the Kerala High Court 
struck down the income limitation on the ground that the absence of poverty cannot be the sole 
determinant of the absence of social and educational backwardness.  222.   (If it is a fallacious 
inference that since poverty is not a sufficient ground for inclusion its absence cannot be a sufficient 
ground for exclusion, then it is one that stems from a perception that this is income discrimination 
among the backward.) The Court also finds the Rs. 6,000 limit arbitrary: it is improperly low, for it 
would exclude a family with two wage earners in Class IV government employment, and it is less 
than would enable a family to support a child in medical college.  223.   On appeal, a 
____________________ 
220 The Kumara Pillai Commission had recommended an income ceiling of Rs. 4,200, but the state 
government raised it to Rs. 6,000 when putting the recommendations into effect. 
221. Laila Chacko v. State, A. I. R. 1967 Ker. 124. 
222. Shameem v. Medical College, A. I. R. 1975 Ker. 131 at 138. 
223. Id., at 138. 
­263­ 
Division Bench had no difficulty in finding that the more prosperous members of backward 
communities might be excluded on the ground that they were lacking the requisite social 
backwardness to entitle them to preferential treatment.  224.   Since poverty is relevant to the 
determination of social and educational backwardness, its absence may be taken into account. To the 
arbitrariness point, the court responds that the inevitability of borderline cases does not preclude 
dividing the group to exclude those who are not backward, shifting benefits from the more advanced 
to less advanced sections of the listed community. The question is whether the Commission has 
material from which it could conclude that the better­off among the listed castes were not socially 
backward. Having satisfied itself that the Commission kept in mind the guiding legal principles, the 
court declined to second­guess the Commission's weighing of the evidence.  225.   
The following year the matter was taken directly to the Supreme Court in Jayasree v. State of 
Kerala.  226.   In the meantime the state government had raised the cut­off point from Rs. 6,000 to Rs. 
10,000 annual income. Both caste and poverty are relevant in determining the social backwardness 
of a group. Here the Commission's finding ­­ that the wealthy sections of the listed groups were not 
stigmatized or disabaled by their group membership and were thus not socially backward ­­ was a 
permissible application of relevant tests. In the course of its judgment, the Court makes some 
tantalizingly obscure observations about the interaction of community and poverty. It suggests that 
not only is the classification of Backward Classes solely on the basis of caste impermissible, but 
that "classification . . . based solely on poverty . . . will not be logical."  227.   The argument, never 
made explicit, seems to be that the "social backwardness which results from poverty is likely to be 
magnified by caste considerations,"  228.   and thus "special provision is contemplated for classes of 
citizens and not for individual citizens."  229.   Thus the Court seems to move away from the notion 
propounded by Balaji that income might be used alone ­­ "neither caste alone nor poverty alone will 
be the determining tests [of backwardness]"  230.   ­­ at the same time that it commends Chitralekha's 
upholding of Backward Classes based on income combined with occupation.  231.   
If income and community are both relevant, they may be combined in a different way to produce 
what might be called a staggered income test. In accordance with a policy of a higher ceiling for 
members of 
____________________ 
224. State of Kerala v. Krishna Kumari, A. I. R. 1976 Ker. 54 at 60. 
225. Id., at 61. 
226. A. I. R. 1976 S.C. 2381. 
227. Id., at 2385. 
228. Id., at 2385. 
229. Id., at 2386. 
230. Id., at 2386. 
231. Id., at 2386. 
­264­ 
communities "socially looked down upon,"  232.   the Punjab Government reserved seats in 
medical colleges for members of listed communities with annual family incomes of less than 
Rs. 1,800 or any other person with annual family income of less than Rs. 1,000. This staggered 
income test has been upheld.  233.   
Among the advantages ascribed to income cut­offs is that, by excluding the well­off within the 
backward communities, they insure that preferences go to the really disadvantaged and deserving 
within these communities. Just how many are actually excluded depends, of course, on where the 
line is drawn and on how tightly the test is administered. Usually the limits have been fixed so high 
that only a small porton of these communities would be excluded. An extreme example would be 
Kerala's limit of Rs. 6,000 (later 10,000) annual income, which probably eliminates no more than 
1% in the populous Ezhuva community and even less in other communities.  234.   Of course, this 1% 
may account for a far higher proportion of those who receive benefits like medical school 
admissions ­­ so there may be a real redistributive effect.  235.   Where the community is poor and the 
means test high, the number of those excluded may hardly justify the cost of administration of the 
income test.  236.   But in so far as such tests assuage the fears of other communities that well­off 
sections within the backward are the undeserving beneficiaries of preferences, such tests perform an 
important symbolic function ­­ possibly their most important one.  237.   
Income cut­offs, usually at far lower levels, are a common feature of many schemes for fee 
concessions and scholarships and other benefits of a less exalted sort than medical admissions. For 
example, the largest scheme of all, scholarships for Scheduled Castes, is limited to those SC 
____________________ 
232. RCSCST 1958­1959: 1, 12. 
233. Gurinder Pal Singh v. State of Punjab, A. I. R. 1974 P. & H. 125. See Ramesh Chander Garg v. 
State of Punjab, A. I. R. 1966 Punj. 476. 
234. The Kumara Pillai Committee estimated that its original ceiling of Rs.4,200 would exclude 
about 4% of the Ezhuvas, Muslims, and Latin Catholics (App. VIII). Presumably, the Rs. 6,000 
standard would restore the great majority of these. 
235. As noted by the Division Bench in State of Kerala v. Krishna Kumari, A. I. R. 1976 Ker. 54 at 
60, 61. 
236. Thus the Study Group on the Weaker Sections of the Village Community (1% 1: 1, 12) 
dismissed as an idle exercise the application of a Rs. 1,000 test to Scheduled Castes. and Tribes. 
237. Preferences in higher echelon positions (medical school admissions, legislative seats, etc.) are in 
the nature of things a symbolic payment to groups for whom the government cannot do much in 
more tangible currency. It is those within these groups who are already better off who are most 
likely to serve as symbols of group participation in the achievements and control of the society. 
A policy of eliminating them by income ceilings might, if the line were drawn low enough, 
exclude most of the likely candidates for this symbolic function. 
­265­ 
students whose parents' or guardians' income is less than Rs. 750 per month.  238.   
The income cut­off in these cases has never been challenged, although it rests on somewhat 
different premises than the cut­off upheld in the Kerala cases just discussed. There the Commission 
concluded, and the courts concurred, that income serves as a measure that social backwardness 
associated with membership in the listed group had been dispelled. It is not self­evident that this 
conclusion would be sustainable at the much lower levels typically used for fee concessions, or that 
it would apply in the case of Scheduled Castes. 
It may seem curious that competitors for high­echelon benefits are aided at much higher levels of 
income than those who aspire to avail themselves of elementary school. In part, the explanation is 
that income cut­offs are really used for two distinct purposes. At low levels, they are used to define 
priorities: they provide a crude measure of ability to pay­­or at least they isolate those with the most 
pressing and unrefusable need. But where they are used for higher­echelon benefits their purpose is 
quite different: they are used as an indicator of the presence of family advantages that offset the 
disablement associated with membership in a particular community.  239.   
3. Income and Occupation as Independent Criteria of Backwardness. 
Proponents of economic tests have sought not only more accurate identification of backward 
communities and the exclusion of the well­off from benefits, but the total abandonment of the 
communal element in selection of Backward Classes. There was considerable sentiment for this in 
the late 1950s. The Ministry of Home Affairs recommended to all the states an occupational test 
with an income ceiling.  240.   Erstwhile 
____________________ 
238. RCSCST 1973­74: 84. 
239. Thus the Tamil Nadu Backward Classes Commission ( 1971: I, 104) recommends a higher 
income level for eligibility to apply for admissions and reserved posts than for scholarships on 
the the ground that "parents with less than the income . . . may not be able to give the 
environmental advantages. . . ." 
240. The Ministry's proposal was that a list of backward occupations be drawn up. This list was to 
comprise: "(a) any non­agricultural occupation in any state of India in which 50% or more of the 
persons belong to the Scheduled Castes or the Scheduled Tribes, (b) any non­agricultural 
occupation in which general literacy percentage of the persons depending thereon is less than 
50% of the general literacy in the state" (RCSCST 1958­1959: 1, 12). This was to be used with a 
family income ceiling of Rs. 1,000 per year (a figure based on the notion that the average family 
consisted of five members and per capita income in the countryside did not exceed Rs. 200). 
Presumably, the occupational test was to apply to actual rather than traditional occupation. This 
proposal was subsequently declared unworkable. See Letter of the Minister of Home Affairs to 
the Chief Secretaries of all State governments/Union Territories, 14 August 1961, reprinted at 
RCSCST 1960­1961: II, 366. 
­266­ 
Bombay did adopt an income test for some scholarship aid.  241.   The central government continued to 
urge economic tests on the states.  242.   This sentiment was given a boost by Balaji in an aside on the 
family income test used in Maharashtra. The Court observed that if any state adopts such a test and 
attempts to enlarge educational opportunities by increasing the number of institutions, it can claim 
to proceed on a rational, broad and scientific approach which is consistent with, and true to, the 
noble ideal of a secular welfare democratic State set up by the Constitution of this country.  243.   
The Court notes that such an approach might, if necessary, be supplemented by "special provision 
by way of reservation to aid the Backward Classes and Scheduled Castes and Tribes."  244.   Although 
some who welcomed the Balaji endorsement seemed to believe that Maharashtra had abandoned 
communal tests entirely, the income test had in fact been adopted only for the giving of financial aid 
in education. It was not employed for distributing seats in medical colleges or posts in government 
services, which continued to be distributed to the Backward Classes as communally defined. 
Although they enjoy Supreme Court commendation, such income tests are by no means mandatory. 
 245.   Nevertheless, such "pure" economic tests have proved attractive to a number of states and 
agencies.  246.   In 1963 the principal central scheme for Other Backward Classes, post­matriculation 
scholarships, shifted from using the state lists of backward communities to a test of family income. 
Scholarships were to be given first to students with family incomes of less than Rs. 1,500 per year; 
second, if any funds remained, to those with family incomes of Rs. 1,500­2,000 per year (in the case 
of technical courses, Rs. 1,500­2,4­00).  247.   
____________________ 
241. Ironically, Bombay's adoption of the income test may have been inspired less by an animus 
against communal tests than by a desire to accommodate the new converts to Buddhism 
(interview with Professor R. D. Bhandare, Bombay, 1966). 
242. See the letter cited in n. 240. 
243. A. I. R. 1963 S. C. 649 at 664. 
244. Id., at 664. 
245. As indicated by all the subsequent litigation in which the absence of economic tests has not 
troubled the court­­e.g., P. Rajendran v. State of Madras, A. I. R. 1968 S. C. 1012. This non­
mandatory character of economic tests is even more clear in the case of Scheduled Castes and 
Scheduled Tribes. In Devadasan v. Union of India, A. I. R. 1964 S. C. 179 at 188, the Court 
rejected the argument that the Government was obliged to implement Art. 16(4) by selecting 
classes by income rather than along communal lines. 
246. Particularly in disbursement of benefits in primary and secondary education. See, e.g., RCSCST 
1960­61: 1, 148­9 (free schooling to those with annual income of less than Rs. 1,200 in 
Maharashtra and Gujarat and, up to Form III, in Madras). 
247. No. F 5­3/63 Sch. 4, Government of India, Ministry of Education. These scholarships were to be 
distributed without a "merit test" and in "the order of 
­267­ 
In Andhra, after the High Court struck down the list of backward communities, benefits were given 
to Backward Classes defined as those with incomes below Rs. 1,5 00.  248.   West Bengal provided 
educational concessions for those whose parents' annual income was below Rs. 3,000.  249.   Mysore, 
in reaction to judicial invalidation of its earlier schemes, switched over to an income­cum­
occupation test; the Backward Classes were those families who had annual incomes of less than Rs. 
1,200 and pursued the following occupations: 
1.  Actual cultivator; 
2.  Artisan; 
3.  Petty businessman; 
4.  Inferior service (i.e. Class IV in Government service and corresponding class or service in 
private employment) including casual labour, and 
5.  Any other occupation involving manual labour.  250.   
In the course of determining the boundaries of these notions, the Mysore Court has indicated that it 
is "the habitual occupations of families" rather than "casual or temporary occupations" that are 
intended.  251.   But no court there or elsewhere has addressed the problem of which occupations the 
State might find to be indicators of backwardness. It was suggested by the Mysore Court that the 
object of this 
____________________ 
poverty ­­ i.e., students from families with lower incomes will receive preference over those from 
families with larger incomes." 
248. Andhra Pradesh Backward Classes Commission 1970: 7. 
249. RCSCST 1964­65: 313. 
250. Govt. of Mysore 1965: 32 ff. The 1963 education order is reprinted in its entirety in Shetty v. 
State of Mysore, A. I. R. 1969 Mys. 48. Havanur (1965: 98­99) eloquently portrays some of the 
consequences of the then new Mysore test: "The notification states that a student is backward if 
his parent's or guardian's income is less than Rs. 1,200 per annum and that he is either an 
agriculturist, petty business­man artisan or a menial servant. If the student owns lands and 
buildings fetching income, that does not matter. If the parents have a dozen of children and the 
income is Rs. 101 per month, all his children are forward. Whereas the father of the only child 
whose income is Rs. 100 per month the child is backward. A primary school teacher whose pay 
is Rs. 70 per month with half a dozen children is forward. A peon whose pay is Rs. 90 with a 
child is backward. A landlord who owns hundreds of acres of land is backward if the Tahsildar 
certifies that the income was Rs. 1,200 per year. A widow who gets Rs. 15 by way of rent from 
her small house is forward. The backwardness is of an individual, and on the date of the 
certificate. If the business­man gets profits to any extent after the date of certificate, the child is 
still backward. In the matter of reservation of posts, the candidate may be a gazetted officer 
whose pay may be Rs. 500 per month, but he is backward when he applies for a higher post or 
other posts, provided the parent or guardian of the gazetted officer secures a certificate of the 
type described above from a Tahsildar. The gazetted officer is backward, but the children of the 
backward gazetted officer are forward because their father is a gazetted officer." 
251. Id., at 50­51. 
­268­ 
classification is to "help those who are at the bottom rung of the ladder of social and economic 
development."  252.   But it is not suggested that a parent's being a cultivator or an artisan has in itself 
stultifying effects on children. Rather, it seems to be assumed that those who engage in these 
callings occupy positions of low status and command few resources, so that opportunities for 
personal development are less likely to be present in their homes. Therefore, it may be assumed that 
their children will enjoy a less favorable environment for educational accomplishments. The 
assumption that the low status of these occupations can be taken as an indicator of the unfavorable 
developmental environment is the same as that which supplies the rationale for the communal units. 
Whether occupation or community correlates more highly with affinity of conditions remains to be 
tested. 
The Supreme Court's only discussion of occupational backwardness provides little guidance. In 
Janki Prasad v. State of Jammu and Kashmir,  253.   the Court says it is "quite open to the State to 
declare that persons belonging to low income families following a traditional occupation should be 
regarded as persons belonging to a backward class."  254.   But it is not enough that the occupation be 
traditionally associated with the family: "his family must be still following the traditional 
occupation."  255.   The Court then proceeds to confine "traditional occupations" to those which 
"require special skills developed by tradition,"  256.   thus excluding such menial callings as 
agricultural laborers, bearers, load carriers, and cart drivers. But since it is open to the State to 
define Backward Classes by combining income with actual as well as traditional occupation, and 
since these callings would surely be associated with backwardness, it is difficult to see the purpose 
of interposing this skill requirement. 
In addition to excluding the well­off among the backward communities, a further and more 
important virtue is claimed for a purely economic test: that it allows benefits to reach the deserving 
in all communities, including those members of "advanced" communities who live in depressed 
circumstances. Thus in Hariharan Pillai v. State of Kerala the Court cautioned the Government that 
not only was it necessary to eliminate the non­backward sections of the backward communities, it 
was also necessary to include the backward sections within other communities. The Backward 
Classes, said the Court, must be drawn from "all the weaker sections of citizens. "  257.   
The application of economic tests involves a number of problems 
____________________ 
252. Sridhara v. Director of Technical Education, 1966 (2) Mys. L. J. 744 at 777. 
253. A. I. R. 1973 S. C. 930. 
254. Id., at 940. 
255. Id., at 940. 
256. Id., at 940. 
257. A. I. R. 1968 Ker. 42 at 51. 
­269­ 
which raise some doubts about how efficiently they accomplish their avowed objectives. Such terms 
as "manual labor" and "family income" give rise to a host of conceptual problems. Joint families, 
income in kind and services rather than cash, and generally the difficulties of determining 
agriculturists' net income ­­ all of these add formidable problems of accurate assessment.  258.   And 
all these difficulties are compounded by a widespread inclination to understate income.  259.   If a 
means test applied with precision might be thought to have unwelcome disincentive effects, the 
difficulties have not been in this direction. In the absence of income and tax records, government 
has relied on income certificates, issued by government officers, who have neither the resources nor 
the inclination to make accurate assessments.  260.   In many areas income certificates have been 
generally available to almost anyone except those who, like schoolteachers or inferior governments 
servants, had incomes that were matters of general knowledge. Of course, the opportunities for 
obtaining such certificates are not spread evenly: they are greater for those who have contacts with 
government officials (and their clerks). Although the prosperous have on the whole succeeded in 
obtaining such certificates, in some instances the truly destitute have experienced considerable 
difficulty. 
Economic tests, like communal tests, entail the "lion's share" problem. Those who get the benefits 
are those among the poor who are relatively better off and have the resources to take advantage of 
the opportunities offered them.  261.   Whatever criteria are used in selecting Backward Classes, there 
is the problem of getting benefits to those at the bottom. When caste units are employed, it is the 
elite within the caste who receive a disproportionate share of the benefits. Income tests in fact 
broaden the field against which the most backward and deprived members of these communities 
have to compete: it substitutes for their more prosperous caste fellows a much larger array of poor 
(or allegedly poor) families among the more prosperous and powerful groups. It was apparently with 
such considerations in mind that the Mysore High Court insisted that the state should have 
supplemented its income and occupational tests with a caste test, "There is hardly any doubt that the 
scheme . . . has not helped the really Backward Classes amongst the Hindus."  262.   The court 
substantiates this with some details from the 
____________________ 
258. One report mentioned as instances of "rural ingenuity" that "income was invariably interpreted 
as net savings after all the normal family living needs were met," while employed persons 
omitted dearness allowances, etc. in computing their income. 
259. Cf. the Nagan Gowda Committee's rejection of an income test on the ground that "it is difficult 
to enforce in practice" ( Mysore Backward Classes Committee 1960: 3). 
260. Problems of issuance of such certificates are discussed in chap. 9, § G below. 
261. These resources might include contacts, high self­esteem, a tradition favorable to education, or 
other by­products of membership in a high or prosperous caste. 
262. A. I. R. 1964 Mys. 132 at 139. 
­270­ 
1963 selection for Mysore engineering colleges, which tend to suggest that the more advanced 
a group was educationally, the more it benefitted from the application of the income­cum­
occupation test ( table 21 ). While the test seemed to help the less well­off in the advanced 
communitites and the middle strata, the communities at the bottom were almost entirely 
excluded. Of course, the differences which emerge here may have been produced, not by the 
operation of the income­cumoccupation test, but by selective processes at earlier stages which 
limited the numbers of Korubas and Bedars who entered this competition.  263.   But evidently, 
distribution of preferences along income/ 
TABLE 21 RELATIVE SUCCESS OF SELECTED CASTES ON MERRF AND ON 
SELECTION BY INCOME­CUM­OCCUPATION TEST IN MYSORE ENGINEERING 
COLLEGES, 1963 
Ratio 
of % 
of 
Merit
Number  Seats  No. of 
of Seats Won  Seats  % of  Ratio of %  Classification 
% of  Gained  % of  to %  Won in  Reserved  Reserved by
Population on  Merit  of  Reserved  Seats  Seats Won  Nagan Gowda 
Group  of State  Merit  Seats  Pop.  Pool  Won  to % of Pop  Committeea 
BRAHMINS  4.28  285  35.9  8.40:1  22  15.5  3.6:1  Forward 
LINGAYATS  15.57  111  13.9  0.89:1  35  24.6  1.6:1  (Borderline)b 
VOKKALIGAS  12.98  32  4.0  0.3:1  35  24.6  1.9:1  Backward 
More 
KORUBAS  6.88  6 .75  0.11:1  2  1.4  0.2:1 
Backward 
More 
BEDARS  4.74  3  0.38  0.08:1  1  0.7  0.15:1 
Backward 
TOTALS  794  142 
Source: Calculated from data provided in Viswanath v. Government of Mysore, A. I. R. 1964 Mys. 132
at 137, 139. 
a Mysore Backward Classes Committee 1961, Statement 9. 
bClassified as forward by the Committee, backward by state government. 

____________________ 
263. Table 21 includes all the figures that the court provided, so we cannot exclude the possibility that 
the pattern for these five castes was atypical. Nor do we really know how many Koruba and 
Bedar applicants failed because they were (1) excluded from the reserved seats by the income­
cum­occupation test, or (2) beaten out for reserved seats by members of more advanced groups 
who met the income­cum­occupation test. 
­271­ 
occupational lines did little to offset the effects of earlier inequalities ­at least at the bottom of 
the scale. 
The Mysore scheme here, like most economic tests, combined income/occupation with merit. 
Income/occupation determined eligibility to enter the reserved­seat competition; among the 
eligibles, the seats were awarded on the basis of merit. Presumably, the tendency of economic tests 
to exclude the lowliest could be mitigated by the use of a "pure" income test ­­ i.e., using income to 
determine both eligibility and the order of distribution. This is done, e.g., in the central 
postmatriculation scholarship scheme, where the student with the lowest income gets his scholarship 
first, and the remainder are distributed in "order of poverty." This kind of arrangement seems 
workable where the economic test does not involve incommensurables (like income and occupation) 
and where it is assumed that all of the group are equally eligible and deserving. Typically, however, 
that assumption is not present, and it is felt that merit inter se should play a role in the selection. 
In considering the suitability of economic tests, we should distinguish what we might call higher­
echelon preferences (which require for their utilization a degree of prior success and some matching 
resources) from what we might call early­stage preferences, which can be utilized without prior 
success or matching resources. Reservations in medical college admissions and higher government 
posts are examples of higherechelon preferences; fee concessions in primary school and free lunches 
for schoolchildren are early­stage preferences. Obviously, the distinction is not absolute, but a matter 
of degree. While economic tests may be suitable for distributing fee concessions or free lunches to 
schoolchildren, it is not clear that they are as well adapted for distributing higherecholon 
preferences. Since some degree of prior success is required even to enter the competition for 
medical colleges or high posts, it is not, with rare exception, the most destitute who will utilize these 
preferences. Thus an economic ceiling for such higher­echelon preferences would have to be set 
fairly high in order to encompass a significant number of potential competitors. For example, the 
family income limit (for medical college admissions) of Rs. 10,000 in Kerala or Rs. 3,000 in Bihar 
excludes only a small fraction of the population. It has been estimated that over 80% of village 
families have incomes of less than Rs. 1,000 per year.  264.   
____________________ 
264. The Study Group on the Welfare of the Weaker Sections of the Village Community ( 1961:5­6) 
cites two 1961 studies, one in Orissa and one in Soilth India, both of which concurred that about 
80% of rural households had annual incomes of less than Rs. 1,000. In the Punjab, where the 
ceiling was Rs. 1,000, a single judge questioned its appropriateness on the ground that 90% of 
the Indian population would­fall within the Backward Classes thus defined (High Court of 
Punjab, Civil Writ No. 1917 of 1965, referred to Division Bench on 13 August 1965; order and 
opinion by R. S. Narula, J.). 
­272­ 
The Backward Classes Reservations Commission in Kerala exposed the underlying problem of 
the income ceiling. Having recommended that Backward Classes be families with aggregate 
annual incomes of less than Rs. 8,000, it then turns its attention to higher government posts: 
[W]hen promotion to a selection grade becomes due to an employee in a feeder category 
for the Gazetted post, he would almost have reached the maximum of his scale of pay, 
which when combined with the other ingredients of his aggregate annual family income, 
would be above the Rs. 8,000 per annum. So by the time he is ripe for promotion, he 
goes out of the Backward Class, because the economic limit for a Backward Class 
family is Rs. 8,000 per annum. If this is the case of officers in the feeder category for 
each of the lowest Gazetted categories of posts, no mention need be made about cases of 
those in the feeder category of higher gazetted posts. To illustrate: In the case of posts of 
Section Officers which are in the feeder category for posts of Under Secretaries in the 
Government Secretariat, Public Service Commission, etc., the economic limit may have 
to be raised at least to Rs. 9,000 per annum per family. In the case of posts of Under 
Secretaries, which are in the feeder category to selection posts of Deputy Secretaries, 
the economic limit may have to be raised at least to Rs. 13,000 per annum per family. In 
the case of posts of Deputy Secretaries, which are in the feeder category for promotion 
to selection posts of Joint Secretaries, the economic limit may have to be raised at least 
to Rs. 15,000 per annum per family. There are certain other feeder categories, the posts 
which carry scales of pay, the maximum of which is Rs. 1,300 per mensem with the 
usual D.A. in certain Departments. In the case of this feeder category, the economic 
limit may have to be raised to Rs. 17,000 per annum per family. Therefore, when the 
State finds that the Backward Classes are inadequately represented in the Gazetted 
category in the services under the State and wants to make provision of reservation of 
posts in their favour, it can effectively do so, only if the economic limit of Rs. 8,000 is 
suitably raised to different limits on the lines indicated above.  265.   
The Committee envisions that the Balaji pronouncements against layers of Backward Classes may 
cause difficulties.  266.   But of course the layering here is an ironic reversal of the attempt to preserve 
some part of the benefits of the least well­off. Here, what is proposed is a layering to enable benefits 
to be extended to those who are doing well by any all­India standard. In such a scheme a high school 
graduate applying for a lower post would not be backward if he had a family income of Rs. 8,500, 
whereas a highly placed government officer earning twice 
____________________ 
265. Kerala ( Backward Classes Reservation Commission) 1971: 128­129. 
266. Id.,129­134. Cf. the discussion of layers in chap. 13, §C, below. 
­273­ 
that would be backward when he sought promotion to a yet higher position. We reach what is 
surely the reductio ad absurdum of the income test­­in which income is not used to identify 
deserving beneficiaries, but the deservingness of beneficiaries is used to fashion the income 
test. 
In somewhat milder form, all income tests entail this paradox. If the cut­off is set high, so many are 
included that few benefits will go to the members of the most disadvantaged groups. If they are set 
low, it is difficult to employ preferential treatment to include historically disadvantaged groups in 
the upper reaches of national life.  267.   Thus different income tests are needed for different kinds of 
programs. Were an income test designed to enable the more mobile members of backward 
communities to obtain medical college admissions used to distribute early stage benefits, it would 
amount to a provision for virtually the entire population. A state committed to such a definition of 
Backward Classes could not in effect provide them with any benefits it was not in a position to 
supply to the population at large. Or, it could only give some late­stage higher­echelon benefits in 
which most potential recipients would have been eliminated by their own lack of resources and 
earlier failures. 
But in the transition from early stage to higher echelon, the function of income tests is transformed. 
In the case of early­stage benefits, the income test is a measure of need or at least priority in 
responding to need. At later stages, income becomes an indicator of cultural disadvantage. But the 
later the stage at which a particular preference devolves on the individual ­­ and most higher­echelon 
preferences devolve fairly far on in life and depend on prior success ­­ the more questionable the 
relevance of such tests as family income. If the father's income has some relevance to the 
deservingness of an applicant to medical college, why should his father's present income affect the 
chances of a lawyer applying for a munsif's post or the chances of a gazetted officer applying for a 
promotion?  268.   If the father's low income is thought to reflect a constric­ 
____________________ 
267. The Supreme Court, noting the insufficiency of poverty as a test of backwardness, observed that 
"if poverty is the exclusive test, a very large proportion of the population of India would have to 
be regarded as socially and educationally backward. . . . An untenable situation might arise. In 
this country except for a small percentage of the population the people are generally poor­­some 
being more poor, others less poor. Therefore when a social investigator tries to identify socially 
and educationally Backward Classes, he may do it with confidence that they are bound to be 
poor. His chief concern is, therefore, to determine whether the class or group is socially and 
educationally backward" ( Janki Prasad v. State of Jammu & Kashmir, A. I. R. 1973 S. C. 930 at 
937). 
268. If it were the applicant's own income that was measured, either the scale would have to be set 
high or one would confine the competition to lawyers who were either failures or dissimulators. 
­274­ 
tion of opportunities earlier in life, it is income at that earlier time that would be relevant. 
All preferences that directly redistribute access to higher­echelon positions (legislative seats, 
medical college admissions, higher Government posts) are going to go to individuals who already 
have more resources and opportunities than the average of the population. Presumably, the 
justification for conferring preferential treatment of this kind on an individual is that he comes from 
a class that enjoys fewer resources and opportunities than the average, and preference for him is 
expected to benefit that class. It might do so in various ways, depending on the way in which the 
classes are chosen. When backward communities are employed, it is the more prosperous and 
mobile within them who receive these higher­echelon benefits. Presumably, they are linked to their 
less­fortunate caste fellows by ties of kinship, loyalty, and mutual support, so there is some reason to 
hope for a "trickle­down" that will spread and multiply the benefits. But where income tests are 
used, there is little reason to think that those at the bottom will be indirectly benefitted. Besides the 
spread of tangible benefit, the recipient might be expected to play a representative function.He 
might be a representative of the interests of the disadvantaged group ­­ as in legislative seats and to 
a certain extent even in higher government posts ­­ providing them with a spokesman or access or 
leadership. Or he might be a representative in a symbolic sense, giving the group a sense of 
participation in the achievements of the society and a model of possible mobility (as in higher 
professional training). There is no self­evident reason to think that the recipients under an economic 
test will serve this kind of representative function, tangible or symbolic, for those in the lowest 
groups.  269.   
Pure economic tests would in several respects be incompatible with some of the doctrines developed 
in Balaji and accepted by subsequent courts ­­ which need not, of course, be taken as a defect of 
economic tests. First, as we have seen, many of the most prominent benefits are at levels of 
accomplishment that would require any economic ceiling to be placed fairly high.  270.   Such a test 
could not possibly conform to Balaji's notion of "absolute backwardness" ­­ that the Backward 
Classes be 
____________________ 
269. Unfortunately, in spite of numerous assertions about the broad effects of preferences, Indian 
social scientists have not attempted to measure their impact on the recipients, nor to compare the 
impact of alternative policies. So the statements above must remain speculative. 
270. For example, in Jayaswal v. Principal, A. I. R. 1968 Pat. 504, the court concluded that it was not 
unreasonable to consider an annual family income of Rs. 3,000 (far above the national average) 
as "poor" for purposes of attending a medical college. Cf. the response of the judge to the 
arbitrarily low cut­off of Rs. 6000 per annum in Shameem v. Medical College, A. I. R. 1975 Ker. 
131 at 138. 
­275­ 
confined to those "well­below the state average" or comparable in backwardness to the 
Scheduled Castes and Tribes.  271.   These standards could only be used with economic tests of 
the first type ­­ i.e., economic tests applied to determine the backwardness of communities. 
Second, if it were recognized that the economic ceiling for purposes of, e.g., medical admissions, 
would have to be higher than that for free lunches, economic tests could be used only by sacrificing 
the notion of a single uniform standard for all of the Backward Classes.  272.   
Third, economic tests make it difficult if not impossible to apply Balaji's notion that the Backward 
Classes are both socially and educationally backward. How is one to know that the low­income and/
or low­occupation families are both socially and educationally backward? In Viswanath v. 
Government of Mysore, the High Court found it difficult to conclude that the high­scoring 
successful students chosen in the "Backward Classes Pool" were educationally backward. 273. 
Presumably, the high scores of individuals were not to the point ­­ the individuals need not be 
educationally backward, but supposedly they belong to an educationally backward class. But there is 
no assurance that they are members of such a class unless it is assumed that the whole low­income 
and low­occupation class is socially and educationally backward. But if that is the case, then the 
"lion's share" problem will be reproduced on a vastly larger scale, for it will be these low­income 
families which are least backward (in the sense of being least deprived of other advantages) who 
will get the benefits. At least in part, this will include those families with higher status, better 
contacts, and the other advantages that are associated with higher caste. So economic tests, when 
used in conjunction with merit, may be expected to minimize the convergence with other forms of 
social backwardness.  274.   
We have seen that the courts have erected increasingly stiff requirements of accurate data to support 
Backward Classes' classifications. So far the courts have only invoked these standards where a state 
has employed the normally forbidden communal criteria. But in so far as economic (or other non­
communal) tests are used to identify the recipients of preferences which are purported to be 
authorized by Articles 15(4) and 16(4), it would appear that courts could require from states a 
____________________ 
271. See §E above. 
272. See chap. 13, §C, below. 
273. A. I. R. 1964 Mys. 132. 
274. Furthermore, the economic test may well run afoul of the notion that reservations should be 
guaranteed minimums rather than provisions for places "over and above" the places gained on 
merit. (See chap. 13, §B.) Presumably, some of the backward (economically defined) will win 
places on merit. If they have to be subtracted before distributing of the reserved places, it will be 
necessary for the state to obtain income certificates from all candidates, not only from those 
candidates who claim to be backward. 
­276­ 
convincing showing of the relevance of such tests to remedy the sorts of backwardness that 
these provisions were designed to overcome. 
H. GEOGRAPHICAL CRITERIA 
As we noted earlier, Balaji said that the State might take into account habitation in identifying the 
Backward Classes.  275.   But a residence test, like economic tests, raises many problems with existing 
doctrine. 
Reservations in admissions to medical colleges in favor of residents of backward areas started to 
come before High Courts in northern India in the late 1960s. The judicial response was mixed. The 
Jammu and Kashmir Court allowed reservation of places for residents of Ladakh,  276.   and one bench 
of the Allahabad Court allowed reservations for candidates from rural areas, hill areas, and 
Uttarkhand Division because the residents of these areas undeniably formed a "socially and 
educationally backward class" in relation to medical education.  277.   
However, another Allahabad bench found these reservations troubling on the ground that the 
government had failed altogether to indicate the basis for regarding these classes as socially 
backward, and its criteria for finding them educationally backward were insufficient. The Court 
appeared willing to take judicial notice that Uttarkhand was so backward that the designation of its 
residents as a socially and educationally backward class was justified, but it struck down the rural 
and hill areas reservations.  278.   Similarly, in Gurinder Pal Singh v. State of Punjab,  279.   reservations 
for candidates from "Backward Areas" were struck down on the ground that such preference on 
residential grounds did not distinguish between "[a] millionaire and a pauper living in such areas," 
 280.   and thus did not direct benefits to "the really deserving residents of such areas." 
In the meantime the Supreme Court had encountered backward areas in the modified Jammu and 
Kashmir scheme which designated as Backward Classes two "areal" groups: residents of areas 
within five miles of the cease­fire lines, and residents of 696 villages in ten "bad pockets" (about 
8% of the total population of the state). The border areas were characterized  281.   by inhibitions on 
mobility and investment 
____________________ 
275. A. I. R. 1963 S. C. 649 at 659. 
276. Sardool Singh v. Principal, Medical College, A.I.R. 1970 J. & K. 45. 
277. Subhash Chandra v. State, A. I. R. 1973 All. 295, 297. 
278. Dilip Kumar v. Government of U.P., A. I. R. 1973 All. 592. 
279. Gurinder Pal Singh v. State of Punjab, A. I. R. 1974 P. & H. 125. 
280. Id., at 127. 
281. Janki Prasad Parimoo v. State of J. & K., A. I. R. 1973 S. C 930. 
­277­ 
as well as severe restrictions on free movement; the bad pockets were distinguished by difficult 
terrain, absence of vehicular communication, and general inaccessibility, as well as scant 
irrigation, lack of electric power, and so forth. The Supreme Court held it allowable to regard 
the residents of the bad pockets and cease­fire areas as socially and educationally backward, 
but required that the government cure defects in the rules to insure that outsiders would not 
secure advantages that should be confined to "genuine residents."  282.   
When the U.P. scheme was litigated again in Pradip Tandon,  283.   a Full Bench of the Allahabad 
High Court struck down the reservations on the ground that the government had indulged in 
unwarranted generalization, not justified by any investigation. Even if the paucity of candidates from 
these areas reflected their educational backwardness, there was no data on their social 
backwardness, and to regard 80% of the population as socially backward would be "an appalling 
opinion, staggering to one's imagination."  284.   Basically, the reservation amounted to an 
impermissible classification by place of birth.  285.   
In November 1974, the Supreme Court decided appeals from the confliciting Allahabad decisions. 
 286.   The Court was satisfied that both the populations of the hill areas (2.5 millions) and of the 
Uttarkhand Division (0.75 millions) were indeed socially and economically Backward Classes. 
Because these areas displayed traits of economic undevelopment, isolation, and absence of 
education institutions, their populations might be regarded as socially and educationally backward 
classes. But the attempt to deem the 75 million inhabitants of Uttar Pradesh's rural areas (80% of the 
state's population) a Backward Class was undone by the lack of heterogeneity of the class: "They are 
not of the same kind. Their occupation is different. Their standards are different."  287.   
We are left with the notion that the population of a locality or a region may form a Backward Class. 
But as the area gets larger, it will become increasingly difficult to show that its inhabitants display 
sufficient homogeneity to qualify as a class. As in the use of communal units, some over­inclusion 
is unavoidable ­­ who are not personally backward may be included. But geographical classes may 
be even more suscepti­ 
____________________ 
282. Id., at 942­43. 
283. Pradip Tandon v. State of Uttar Pradesh, A. I. R. 1975 All. 1. 
284. Id., at 7. 
285. Id., at 8. 
286. State of Uttar Pradesh v. Pradip Tandon, A. I. R. 1975 S. C. 563. (This was an appeal from 
Pradip Tandon, n. 283 above, and from Subhash Chandra, n. 277, above.) 
287. Id., at 569. In addition, the Court stresses that the rural population is a majority. Cf. chap. 12, 
§E, below. 
­278­ 
ble to over­inclusion than communities. While the inhabitants of geographical areas do share 
similar conditions of bad roads, isolation, lack of schools, etc., they may differ sharply in their 
resources and in their capacity for coping with these shared problems. Extending benefits that 
require prior training and advantages (like medical admissions) to geographical classes would, 
one might expect, lead to the enjoyment of these benefits by the most advantaged within the 
beneficiary group, those who are personally least socially and educationally backward. 
Geography can be used in conjunction with other factors as well as independently; and it can be 
used to disqualify as well as to qualify for benefits. Lists of eligible groups are typically based on 
area and contain area restrictions.  288.   Location may be used not only to select beneficiaries but to 
screen otherwise eligible beneficiaries ­­ analogous to the income cut­off provisions discussed 
above. Government may impose area restrictions on the enjoyment of certain benefits. The Supreme 
Court held that it is permissible for the State to give favorable tax treatment to income of Scheduled 
Tribes earned within specified tribal areas and to withhold that beneficial treatment of income 
earned outside those areas.  289.   
I. NON­INCLUSION CLAIMS 
Of the 46 reported litigations involving preferences for Backward Classes, 29 were concerned with 
the designation of those classes by the State.  290.   Almost all involved challenges to the preferential 
policies on the ground that the list of Backward Classes was too inclusive. As we have seen, a 
significant portion of these claims have been successful. Of the 24 cases in which such a challenge 
was raised, the government's action was struck down (or modified) in 17, and in several others the 
government was warned to take corrective action to avoid future invalidation of its list.  291.   
On the other hand, claims that the lists of Backward Classes were insufficiently inclusive have not 
found a receptive audience in the courts; they have been reluctant to consider claims by members of 
an allegedly backward class who assert that they should have been included. In Partha v. State of 
Mysore, a scheme for reservations was attacked 
____________________ 
288. See discussion of area restrictions at chap. 5, §B, above. 
289. I. T. Officer v. Rymbai, A. I. R. 1976 S. C. 670 [NIC]. 
290. These figures are based on a universe of 110 reported cases concerning compensatory 
discrimination from 1951 to 1977. The derivation of that universe is described below in chap. 14, 
n. 45. 
291. To appreciate the significance of this rate of success, see chap. 14, §C, below. 
­279­ 
on the ground that the Lingayat community should have been included in the list of Backward 
Classes, which included others who were no more backward than they.  292.   The court rejoined 
that so long as the communities which are included are all backward ones, "the fact that the 
list may not be exhaustive of all the Backward Classes in the State" does not require that it be 
struck down, since the claim for inclusion is, in effect, a claim under a Directive Principle (i.e., 
Article 46) and therefore non­justiciable.  293.   The contention in Partha was, from the court's 
point of view, that the line had been drawn at the wrong place; but it was not a case of 
demonstrated comparability with the groups under the line. 
Similarly, in the cases challenging the validity of the income cut­off, the issue was the validity of 
drawing a line separating members of listed communities whose annual incomes were more than 
Rs. 10,000 and those with smaller incomes. Could the state include the latter but omit the former? 
In Jayasree v. State of Kerala, the Supreme Court emphasized that it was a function of the State to 
"evolv[e] . . . proper criteria" for "determining who are socially and educationally backward 
classes,"  294.   while it is the jurisdiction of the courts "to decide whether the tests applied are valid." 
 295.   The state's determination that social and economic backwardness did not extend to the more 
affluent sections of the listed communities was found to be an application of valid tests. 
With the development of requirements of objective standards supported by sufficient data, it is 
doubtful that the courts would permit the exclusion of a group which met the standards which the 
state purported to use in choosing Backward Classes. The claim that group X meets all the criteria 
by which groups Y and Z were designated as backward would seem to be litigable on equal 
protection grounds. The more "objective" the criteria the courts demand of the state, the less reason 
for any reluctance to admit such claims for inclusion. 
Claims that the standards should be modified to include group X as well as groups Y and Z might 
still be dismissed as non­justiciable claims under Article 46, as Partha suggests. But that case was 
decided at the very beginning of judicial supervision of state selection of Backward Classes. That 
jurisprudence, requiring that the state employ valid tests of backwardness, provides many ways to 
challenge the validity of the state's tests. And the emergence of the theme that compensatory prefer­ 
____________________ 
292. A. I. R. 1961 Mys. 220. 
293. Id., at 238. 
294. The rationale of the State for the classification was that Lingayats had been excluded because 
their percentage of literacy was above the state average. See Govenment of Mysore 1960. 
295. A.I.R. 1976 S.C. 2381 at 2385. 
­280­ 
ence is itself part of the Fundamental Right to equality guaranteed by Articles 14, 15, and 16 
suggests fertile doctrinal ground for erecting claims for inclusion.  296.   So far, however, such 
claims on behalf of excluded groups have not been forthcoming. 
There is, however, another judicial path to questions of inclusion, one that has been more heavily 
trafficked. Claims for inclusion may be put forward not only on behalf of groups who are not 
designated, but also by individuals whose inclusion within the groups that have been designated has 
been challenged. Such claims may take several forms: group X of which I am a member is really 
part of (or equivalent to) group Y, which is being provided benefit A; I am really a Y and I have not 
ceased to be a Y because I have become an X. Although I was formerly an X, I have now become a 
Y. It is typically more manageable for litigants to assert or contest personal membership in a listed 
group than to challenge the structure of the state's list. When raised in the form of disputes about 
individual membership, courts cannot avoid the task of determining the boundaries of the 
designated groups. Thus some problems of inclusion are litigated in the guise of questions of 
individual membership. It is to these questions that we now turn. 
____________________ 
296. See chap. 11, §D, below. 
­281­ 
9 Membership in Groups Entitled to Preferential Treatment 
IN THE previous chapters we have been concerned with the problems of choosing the groups that 
may receive preferences. In this chapter we shall examine the corollary problems of determining 
whether an individual is a member of one of these preferred groups. We have seen that Scheduled 
Castes, Scheduled Tribes, and Backward Classes have for the most part been designated in terms of 
communal units: caste, tribe, and religion have been the primary differentiae. Some diversification 
of criteria has occurred: in some instances communal units have been supplanted by, or combined 
with, economic or occupational tests. The prospect is that in spite of such diversification, communal 
criteria will continue to be at least one element in the designation of the recipients of preferences ­­ 
especially of the Scheduled Castes and Tribes. 
With membership in specified communal groups a qualification for preferment of various kinds, it 
is not surprising that disputes have arisen about such membership. In numerous cases the courts 
have had to pass on the question of whether a person was in fact a member of such a preferred 
group. This is not to say that such determinations are normally either disputable or disputed. 
Typically, for purposes of qualifying for most schemes of preferences, membership in the preferred 
group is evidenced either by simple declaration or by obtaining a certificate to that effect from some 
1. These declara­ 
government official.    
____________________ 
1. Where there is sufficient material to show that a person belongs to a preferred group, failure to 
submit the required certificate is not grounds for withholding the benefit. But to regard a person 
as Scheduled Caste, Scheduled Tribe, or Backward Class "without any material whatsoever" is 
arbitrary and capricious and will be struck down ( Mahendra Nath Pathak v. State of Assam, A. I. 
R. 1970 A. & N. 32, 34). (Petitioner had made a bare assertion: "I am belonging to the Scheduled 
Tribe community. The certificate will be submitted later on." It could not even be ascertained 
which tribal community he claimed to belong to.) However, eligibility does not depend entirely 
on claiming membership in an application, but depends on whether a candidate supplies 
information which, if verified, would show membership in the listed group ( P. Susila v.State of 
Madras 
­282­ 
tions and certificates remain unchallenged in the great majority of cases. Virtually all of the 
cases in which membership has been disputed involve election contests in which a successful 
candidate for a reserved seat is challenged on the ground that he is not a member of the group 
for which the seat is reserved. The election situation offers unique incentive to disprove 
membership of a particular individual: there is a single affected party for whom the stakes are 
high enough to justify the expense of litigation and who is likely to have the resources (and 
2. 
backing) to undertake it.    
The cases raise many puzzling questions: Is membership in a caste or tribe to be determined solely 
by birth? or by allegiance? or by the opinion of its members or of the general neighborhood? Does 
one lose (or gain) caste membership by joining a sect or another religion? by excommunication? by 
assimilation? Does one lose tribal membership by claiming or achieving caste status? Who is a 
communicant of a particular religion? those born into it? those who have been converted to it? those 
who adhere to its precepts? What is the effect of unorthodoxy? of excommunication? Are the tests 
used for the application of personal law appropriate in the area of preferences? 
Before taking up these questions, which involve instances in which membership in a listed group is 
rendered problematic by the existence of an additional affiliation or by some factor which appears to 
vitiate membership in the group, there is the preliminary question of the meaning of the lists 
themselves. 
A. READING THE SCHEDULES 
Just which groups are on the lists of Scheduled Castes and Scheduled Tribes is not as simple a 
matter as it might appear. The nomenclature of communal groups in India is equivocal and 
ambiguous, fluid and shifting. A high­level official committee noted that 
____________________ 
State of Madras, A. I. R. 1970 Mad. 399, 400). Inconsistent assertions on other occasions will not 
necessarily defeat a claim of membership ( Bhaiya Ram v. Anirudh, A. I. R. 1971 S. C. 2533). 
2. Most challenges of claimed group membership that reach the courts are initiated by disappointed 
rivals. It is, of course, possible for such challenges to originate with governmental authorities 
monitoring the distribution of preferences (for an apparent example, see Kajari Saha v. State, A. 
I. R. 1976 Cal. 359) or with organizations acting on behalf of genuine recipients. The Evening 
News [ New Delhi] of Jan. 19, 1974, reports that police registered a charge of cheating against 
two doctors who were alleged to have obtained admission to medical college on false certificates 
of being members of Scheduled Castes. The case was registered following a complaint lodged by 
the Society for the Welfare of Children and Women. 
­283­ 
the names by which certain groups and sub­groups are known vary from district to district 
and sometimes from area to area within a district. Moreover, there is a tendency for castes and 
sub­groups to give themselves new and fanciful names from time to time, and to resent being 
3. 
called by any other name.    
The lists of castes and tribes that make­up the Scheduled Castes and Tribes Orders make a valiant 
attempt to grapple with this by including some synonyms and specifying some sub­groups, but other 
synonyms and sub­groups are omitted ­­ by design or otherwise. To make matters worse, many of 
the names used have a shifting denotation, being used at different levels of generality ­­ e.g., the 
same term may be used to refer to an endogamous caste group and to a cluster of related castes. 4. 
Sometimes more than one community bears the same name.     5. The "name" of one group may be 
6.
used as titles or surnames in another.     A group's name may connote a profession followed by others 
7. The lists also include many entries that seem to refer to no group at all ­­ some are merely 
as well.    
general appellations (e.g., Chandal, Panchama), others are groups that cannot be found by the 
8. The confusion is compounded by problems of inconsistent 
census­takers in the area specified.    
translation, transliteration, and spelling. The problems of nomenclature and enumeration are 
detailed by the Superintendant of Census Operations in Madras: 
Synonyms used by certain castes are the names of certain other main castes. For 
instance, the Chakkiliyans in certain areas are known under the synonym Dombars. 
Dombara is also a different Scheduled Caste with the synonym of Dombars. Again in 
Thanjavar district, some Pallars are known by the synonym of Panikkars. But Panikkars 
in other areas always do not belong to Pallars. Similarly, Pandaram is a synonym of 
Valluvan, but all Pandarams are not Valluvans. Again, Andi Pandarams do not belong to 
the Scheduled Castes. Jogi is a synonym of Kuravan. But it is also a synonym of Dom or 
Dombara in Kanyakumari area. Uralis and Pulayans are found both in the Scheduled 
Castes and Scheduled Tribes list. In spite of clear instructions issued by us, generic 
names like Harijans, Scheduled Castes and Scheduled Tribes have been used in the 
9. 
Census return.    
____________________ 
3. Department of Social Security 1965: 20. 
4. Cf. Fox 1969. 
5. Depatment of Social Security 1965: 13. 
6. RCSCST 1961­62: I, 5. 
7. The Punjab Evaluation Committee Welfare Department ( 1966: 141) reports that in such cases 
higher caste "persons not acutally belonging to Backward Classes try to avail of the benefits by 
displaying their professions and making a caste out of it." 
8. Department of Social Security 1965: 15. 
9. Census of India, 1961, Vol. IX, Madras, Part IV­A (1), Scheduled Castes and Tribes (Report and 
Tables), p. 13. 
­284­ 
And, of course, the very existence of the lists acts to induce movement across the line into 
favored categories by manipulation of equivocal nomenclature.  10.   
In view of this confusion, it is sometimes a problem merely to know whether a group to which 
somebody belongs is a group that is listed on one of the Schedules. A review of the Supreme 
Court's experience gives some idea of the intractability of these problems to judicial solution. 
In Basavalingappa v. Munichinnappa,  11.   election of a candidate to a seat reserved for Scheduled 
Castes was challenged on the ground that he was not a member of a Scheduled Caste. The candidate 
claimed to belong to the Bhovi caste, a group listed in the Scheduled Caste Order, but his 
challengers claimed he was a member of the Voddar caste, for which there was no listing in the 
Order. The Election Tribunal found that Bhovi, as listed in the order, was a sub­caste among 
Voddars, and since the candidate did not belong to this sub­caste he was not eligible for election. 
The High Court reversed, finding that the Bhovi caste mentioned in the Order was no other than the 
Voddar caste. The Supreme Court, cautioning that it was not ordinarily open to a person to offer 
evidence to prove that his caste A is part of a caste B listed in the Order, decided to evaluate the 
evidence in view of the peculiar circumstance ­­ i.e., that caste B is otherwise nonexistent. There 
had been no Bhovis anywhere in Mysore before the States' Reorganization of 1956. But the 1950 
Order listed Bhovis in Mysore, and it must have been intended to include some group. Thus the 
Court felt justified in examining the evidence to determine which caste was meant by this term. The 
evidence showed that a Voddar conference had in 1944 resolved to change their name to Boyi and 
had communicated this to the Mysore Government, which had in 1946 passed an order that the 
community be called Boyi in all government communications and records. The caste was also 
sometimes spelled Bovi or Bhovi.  12.   The Voddars, it was found, were indeed the Bhovis referred to 
in the Order.  13.   
A few months later, in Bhaiya Lal v. Harikishan Singh,  14.   the Supreme Court made good its 
warning that it would not ordinarily permit members of group A to show that that group was a part 
of listed caste B. 
____________________ 
10. A dramatic instance of such movement appears in the case of the Sunris and Sahas discussed 
below at nn. 18 ff. 
11. A. I. R. 1965 S.C. 1269. 
12. Cf. Konkani v. Shankar Rao, 3 E. L. R. 409 (1953), where the Election Tribunal found that the 
Konkni community, the Konkna caste, and the Konkna tribe were in fact the same. 
13. More than a decade later, when the lists were revised, a Karnataka M.P. vainly attempted to have 
"Waddars" included in the Scheduled Castes list as a synonym for Bhovi (Shri G. Y. Krishnan at 
LSD [5th Series], Vol. 64, No. 16, cols. 103­4 [2 Sep. 1976]). 
14. A. I. R. 1965 S.C. 1557. 
­285­ 
A candidate elected to a Scheduled Caste seat had declared in his nomination paper that he was a 
Chamar. (The Scheduled Caste Order contained an entry "Chamar, Jatav or Mochi".) The Election 
Tribunal and the High Court declared his election invalid on the ground that he was not a Chamar 
but a Dohar ­­ a group not on the Scheduled Caste list in that district. The candidate claimed that he 
was a Dohar Chamar (and that this group was also called Mochi.) The Court refused to allow him to 
offer evidence that Dohars were a sub­caste of Chamars on the ground that it could only determine 
whether or not he was a member of the listed caste ­­ i.e., whether he was a Chamar.  15.   Evidence 
showed that Dohars did not interdine or intermarry with Chamars, that they lived in different 
mohullas, that they had their own Dohar samaj. The Court concluded that he was not a Chamar. 
These cases perturbed the authorities in charge of administering preferences, who had always acted 
on the assumption that every listing included any sub­groups and that sub­groups need not be 
specified separately.  16.   The Lokur Committee, pointing to the need for accurate listing of all 
synonyms, phonetic variations, and sub­groups, predicted that it would be "impossible to include all 
sub­groups."  17.   
The elusiveness of the facts in these cases and the Supreme Court's despair of ascertaining them is 
demonstrated in the next case to reach the Supreme Court, Abhoy Pada Saha v. Sudhir Kumar 
Mondal.  18.   The Scheduled Caste Order list for Bengal included an item "Sunri excluding Saha." 
One Mr. Saha, a successful candidate whose nomination paper described him as a member of the 
Sunri caste, was challenged on the ground that he was not a Scheduled Caste. The Election Tribunal 
thought that Sahas, though originally a group within the Sunris, had developed into a distinct caste 
that had severed all connections with the Sunris.  19.   According to the Tribunal, Mr. Saha belonged to 
the Sunris 
____________________ 
15. But cf. Balchand v. Laxminarain Mateh, 8 E. L. R. 465 (1953), where Jatavas were held to be a 
part or group within the Scheduled Caste of Chamars, where the item included "Jatia or Jatav 
Chamar." 
16. Department of Social Security 1965: 19. The Ministry of Law had opined that all parts of a listed 
group were entitled to benefits, even sub­divisions which were not listed (RCSCST 1958­59: I, 
10). The establishment of the Lokur Committee is attributed to the pressure on government of the 
Supreme Court's requirement of comprehensive listing (RCSCST 1965­66: 65). 
17. Id., at 20. 
18. Abhoy Pada Saha v. Sudhir Kumar Mondal, A. I. R. 1967 S. C. 115. 
19. Census figures suggest that the Tribunal may have been on the right track. The reported Sunri 
population declined steadily from 114 thousand on the 1901 census to 45,000 on the 1951 census, 
presumably because members of this group were reporting themselves as Saha. But the Sunri 
population rose to 106,000 in 1961 and 166,000 in 1971. See West Bengal Tribal Welfare 
Department Bulletin of the Cultural Research Institute II ( 1): 43, for figures to 1951; Census of 
India, 1961, Vol. XVI, Part V­A (i), p. xliv, for 1961; Census of India, 1971, Series 1, Paper 1 of 
1975, Scheduled Castes and Scheduled Tribes ( 1975): 92, for 1971. 
­286­ 
and therefore was not a member of the independent Saha caste, which had, because of their 
historical association, been excluded by the Order as a precaution. The High Court, on the other 
hand, thought that the effect of the listing was to exclude all Sunris with the surname of Saha, a 
procedure it found within the President's power under article 341 to exclude parts of castes. 
The Supreme Court rejected both views and expressed extreme dissatisfaction with the Tribunal's 
reading of the Order. The Tribunal was wrong in thinking Saha an independent caste. Since Sunris 
are a caste, the Court deduces from the phrasing of the listing, Saha must be a sub­caste. The error 
of the Tribunal 
lay in interpreting the Order in the light of the evidence before it. There was no 
justification for doing that. After all, the evidence in a case may be imperfect. . . . A 
method of interpreting a statutory provision which might lead to such uncertainty cannot 
be correct. . . . Evidence cannot alter the natural interpretation of the words in the Order. 
 20.   
Sahas then were a smaller caste within the Sunri caste. Since there was evidence that the candidate 
was within the larger, but none that he was in the smaller, he is a Sunri and the election is declared 
valid.  21.   
In its understandable reluctance to become embroiled in the evidentiary quicksand, the Court 
attempts to escape by a rigid formalism which assumes that the "text" (i.e., the Order) contains 
answers which can be elicited by proper interpretation. The "text" here is admittedly a very crude 
approximation to the social actualities. Yet the Supreme Court rejects the empirical finding of the 
Tribunal about the actuality, because it does not conform to the logical structure of the Order, and 
substitutes a finding in accordance with that logic. 
In Laxman Siddappa Naik v. Kattimani Chandappa Jampanna,  22.   the Supreme Court attempted to 
formulate some standards for dealing with these troubling cases. A successful candidate for a seat 
reserved for the Scheduled Tribes was challenged on the ground that he was not a member of the 
Nayaka tribe, as declared in his nomination paper. The applicable list of Scheduled Tribes, which 
was adjusted after the transfer of the district from Bombay to Mysore in the States' Reorganization 
in 1956, had an entry for "Naikda or Nayaka including Cholivala Nayaka, Kapadia Nayaka, Mota 
Nayaka and Nana Nayaka." The 
____________________ 
20. A. I. R. 1967 S. C. 115 at 118. 
21. A further episode in the Sunri­Saha story is found in Kajari Saha v. State, A. I. R. 1976 Cal. 359, 
where a medical college applicant who claimed to be Sunri with the surname Saha was found by 
welfare department investigators to be a member of the Vaishya Saha caste and not a Sunri at all. 
The disappointed applicant's father was a government servant, hired against a reserved place, of 
more than 20 years service. 
22. A. I. R. 1968 S. C. 929. 
­287­ 
challenger claimed that the candidate was not a Nayaka, but a member of the Bedar caste. The 
candidate claimed that "he was not a Naikda but a Nayaka and that Nayakas were also known as 
Bedars."  23.   The High Court, after examining census reports and other documentary evidence, 
concluded that there were no Nayakas in this district and that the candidate was a Bedar. 
The Supreme Court read the earlier cases as making the determination of membership a question of 
fact to be proved by evidence ­­ a reading not immediately apparent. Where membership was 
contested, the onus lay on the challenger to prove that it was not as claimed. The challenger might 
disprove the claim to belong to a group by evidence of different customs or by showing social 
acceptance by the other community. Here, in spite of suspicious circumstances surrounding the 
candidate's certification as a Nayaka, the challenger had failed to discharge this burden. "There was 
no evidence one way or the other. In these circumstances, the election petitioner could not succeed 
because of the weakness of the appellant's case."  24.   
Here the Court moves from statutory interpretation to the rules of evidence. But if an individual's 
membership in a group is a question of fact, are not the relations that obtain between groups ­­ 
whether they are co­extensive, distinct, overlapping, one within the other ­­ also questions of fact 
which cannot be answered by deductive reasoning? 
Apparently not, for a few months later the Supreme Court encountered yet another case of this kind. 
In Pars Ram v. Shivchand, a candidate's nomination paper was rejected on the ground that he was 
not, as claimed, a Chamar, but was a Mochi. (Chamar was on the Punjab's Scheduled Caste List, 
Mochi was on the state's list of Backward Classes.) The High Court found that Mochis and Chamars 
were distinct groups and that apparently "Hindu Mochis have developed into a higher caste or caste­
group than that of the Chamars, particularly in the matter of social status."  25.   Following Bhaiya Lal, 
the High Court held that since Mochi was not mentioned as a synonym of Chamar, it is not the same 
caste; since it is not mentioned in the Order, it is not a Scheduled Caste and the candidate was not 
eligible to stand for a Scheduled Caste seat. On appeal the Supreme Court affirmed, holding that the 
claim that both appellations referred to the same caste "is not open 
____________________ 
23. Id., at 932. This dispute appears to be an artifact of a freak in the process of delimiting Scheduled 
Tribe constituencies. It was discovered too late that the four districts in the constituency 
contained only 2% Scheduled Tribes population. There are no members of the Nayaka tribe in 
the area, but the Bedars, who predominate in the area, are also called Nayakas. (Personal 
communication to the author from a knowledgeable Mysore lawyer.) 
24. Id. 
25. A. I. R., 1968 P. & H. 331, 335. 
­288­ 
to agitation by evidence."  26.   Since the determination of the lists lies within the exclusive power of 
the President, "it is not for us to examine it and come to the conclusion that if a person was in fact a 
Mochi, he could still claim to belong to the Scheduled Caste of Chamar. . . ."  27.   Thus the Supreme 
Court reasserts its unwillingness to review the factual determination of the relation between groups. 
 28.   It assumes an ordering of groups that corresponds to the ordering of names in the presidential 
list. 
Yet, two years later, in Bhaiya Ram Munda v. Anirudh Patar,  29.   the Supreme Court did not hesitate 
to review evidence and conclude that Patars, although not mentioned among the sub­tribes of 
Mundas included in the Scheduled Tribes Order, "are it appears regarded as the lowest in the social 
order amongst the Mundas but they are still Mundas."  30.   The fact that some sub­tribes are 
enumerated in the Order does not give rise to an inference that those not enumerated are not 
Mundas. The Court refused to read earlier cases as foreclosing a reading of the list to match what 
the evidence disclosed to be the factual situation. 
These cases do not involve any large general questions; the facts are elusive and obscure and 
difficult for a court to elicit. (Indeed, the benefits of appeal on the facts are not readily apparent.) 
Exposed to such an imponderable situation, there is a temptation to impose order 
____________________ 
26. A. I. R. 1969 S. C. 597, 600. 
27. Id., at 600. Mochi was listed as a synonym of Chamar in the Scheduled Castes lists for 11 states 
­­ all but Assam, U.P., and Punjab. When Punjab was divided in 1966, the item remained 
unaltered for Punjab and Haryana, but Mochis were listed along with Chamars in both Delhi and 
Himachal Pradesh. The Supreme Court cites this difference in treatment as evidence "that the 
question of inclusion of Mochis in the Scheduled Castes was considered" by the President as 
recently as 1966 (id., at 599). 
28. In striking contrast to this timid retreat from the tangled facts is the sensitivity to the fluid and 
ambiguous character of these identities displayed in another case involving the same cluster of 
groups in the Punjab. In Didar Singh v. Sohan Singh, A. I. R. 1966 Punj. 282,288, a candidate 
who claimed to be a Ramdasia by caste was declared unqualified by the Election Tribunal on the 
ground that he was a Rehtia. (The Scheduled Caste Order for the Punjab included an item 
specifying Chamar, Jatia Chamar, Rehgar, Raigar, Ramdasi or Ravidasi ­­ but Rehtia was not 
listed.) The court declined to take Rehtia as mutually exclusive of Ramdasia. One could not 
exclude the possibility of "a Ramdasi being also called a Rehtia. In this part of the Country [the 
position of groups like Chamars] . . . seems to have been traditionally in a fluid state, in so far as 
their description goes. The expressions Ramdasi, Ramdasia or Ravidasia . . . do not seem to 
possess or convey any clearcut or crystallized distinctive characteristics. . . . [Nor do] these 
expressions respectively carry precise, exact and distinctive meanings excluding the possibility of 
some individuals . . . using any one of these epithets interchangeably. . . . [T]he position has all 
along been somewhat confused and certainly imprecise." 
29. A. I. R. 1971 S. C. 2533. 
30. Id., at 2538. 
­289­ 
on the factual muddle by adopting formal deductive solutions. The choice between grappling with a 
very elusive and muddled social actuality or invoking a formal logical order to encompass it is a 
theme we shall meet again as we explore the judicial treatment of various membership problems. 
But they point to a general phenomenon that deserves reflection. These cases involve a contest 
about some coveted prize ­­ usually an electoral contest, sometimes a medical admission ­­ and the 
candidate seeks to establish his eligibility or to disqualify a competitor. These efforts highlight the 
opaque and malleable qualities of the lists, qualities that are present even when less­exalted benefits 
are being distributed and there is less incentive for anyone to mount a challenge (or defense) of a 
particular assertion of group identity. Manipulative practices that we see in bold relief in cases 
involving important prizes may exist in less visible but widespread form, availed of routinely and 
normally without challenge.  31.   The adequacy of the judicial response to questions of group 
membership cannot be assessed solely by the incisiveness and subtlety with which it addresses 
litigated disputes. Such an assessment must also consider the framework that judicial treatment 
provides for routine administration of preferences. 
B. CASTE AND SECT 
The tests for determining group membership were first given extensive consideration by the 
Supreme Court in Chatturbhuj Vithaldas Jasani v. Moreshwar Parashram.  32.   An Election Tribunal 
had rejected the nomination papers for a reserved seat submitted by a Mahar who had joined the 
Mahanubhava Panth, a Hindu sect which repudiated the multiplicity of gods and the caste system. 
Reversing the Tribunal, the Supreme Court held that the candidate remained a Mahar and was thus 
entitled to stand for the seat reserved for Scheduled Castes. To determine whether adherence to this 
sect made the candidate cease to be a Mahar, the Court specified three factors to be considered: "(1) 
the reactions of the old body, (2) the intentions of the individual himself and (3) the rules of the new 
order."  33.   
____________________ 
31. E.g., the Tamil Nadu Backward Classes Commission ( 1971: I, 160) notes that "some forward or 
social higher castes have sometimes a sub­section whose caste name is more or less similar or 
identical with a name found in the list of Backward Classes. Without compunction such 
loopholes are freely made use of." Comparable routine misrepresentation of income is discussed 
in §G, below. 
32. (1954) S. C. R. 817. Hereafter referred to as Jasani. 
33. Id., at 838. 
­290­ 
The candidate was admitted to all Mahar caste functions and had been allowed to marry within the 
community. He twice married Mahar girls, neither of whom were Panth members at the time of the 
marriage. He always identified himself as a Mahar. The Panth, in spite of its doctrinal repudiation of 
caste, had not penalized him for his adherence to the caste. 
The Supreme Court concluded that "conversion to this sect imports little beyond an intellectual 
acceptance of certain ideological tenets and does not alter the convert's caste status."  34.   It is clear 
that the primary consideration was the second test­­i.e., the intentions of the convert himself­­
intention not in the sense of mere declaration but as evidenced by a consistent course of conduct and 
dealings.  35.   The Court applied the "broad underlying principle" of Abraham v. Abraham, 36. 
decided a century before by the Privy Council in determining the law of inheritance applicable to a 
Hindu convert to Christianity: "[h]e may renounce the old law by which he was bound, as he has 
renounced his old religion, or, if he thinks fit, he may abide by the old law, notwithstanding he has 
renounced his religion."  37.   Applying this principle, the Supreme Court found that "if the 
individual . . . desires and intends to retain his old social and political ties," and if the old order is 
tolerant of the new faith and does not expel the convert, the conversion has no effect. "On the other 
hand, if the convert has shown by his conduct and dealings that his break from the old order is so 
complete and final that he no longer regards himself as a member of the old body and there is no 
reconversion and readmittance to the old fold . . . [he cannot] claim temporal privileges and political 
advantages which are special to the old order."  38.   
Although the test is primarily one of the convert's intention or behavior, this intention must be 
confirmed through acceptance by the old group. The inclusion of this additional test is important 
because, as 
____________________ 
34. Id., at 840. The Panth is a devotional sect, founded in the 11th century, which not only eschewed 
caste and polydeism, but also challenged the validity of the Vedas, image worship, and the 
system of asramas (stages of life). On the role of this sect in Mahar tradition, see Miller 1966; 
Zelliott 1969: 26 ff. 
35. Thus unsupported contentions that "I am belonging to the Scheduled Tribe community" from 
which "it cannot be ascertained to what tribal community he claims to have belonged" are 
insufficient to establish membership ( Mahendra Nath Pathak v. State of Assam, A. I. R. 1970 A. 
& N. 32, 34). 
36. Abraham v. Abraham, 9 M. I. A. 199, (1863). The rule was subsequently overturned by the 
Indian Succession Act, 1865 (now Indian Succession Act, 1925, §58). But the courts are still 
divided over whether the Hindu rule of survivorship is applicable to Christian families who 
continue to be joint after conversion. 
37. Reporter's note at 9 M. I. A. 196. Cf. id., at 242­44. 
38. (1954) S. C. R. at 838. 
­291­ 
the Court says, "the only modification here is that it is not only his choice which must be taken into 
account but also the views of the body whose religious tenets he has renounced, because here the 
right we are considering is the right of the old body, the right conferred on it as a special privilege 
to send a member of its own fold to Parliament."  39.   
The third test, "the rules of the new order," is of minor significance. Since it is the legal and political 
rights of the old body that are being considered, "the views of the new faith hardly matter."  40.   "The 
new body is free to ostracise and outcaste the convert from its fold if he does not adhere to its tenets 
but it can hardly claim the right to interfere in matters which concern the political rights of the old 
body, when neither the old body nor the convert is seeking legal or political favours from the new 
body as opposed to purely spiritual advantages."  41.   If this test has to be taken into account at all, it 
is only as indirect evidence of the intentions and conduct of the convert. For example, continued 
acceptance by a new group which was notoriously intolerant of the retention of the old ties might 
well evince an intention to break with the old group. But here the Court found it "evident that 
present day Mahanubhavas admit to their fold persons who elect to retain their old caste customs." 
 42.   
In Shyanzsundar v. Shankar Deo Udgir,  43.   the principles of the Jasani case were applied to decide 
whether a candidate for a reserved seat had lost his membership in the Samgar caste by joining the 
Arya Samaj, a Hindu sect which rejects idolatry and ascription of caste by birth.  44.   He had been 
accepted for membership in a local Arya Samaj organization, had paid membership dues, had 
married a girl of Sonar caste in accordance with Arya Samaj rites, and had reported himself as an 
"Arya" in the 1951 census. The Mysore High Court, citing the Jasani case, announced that there 
would be no deprivation of caste unless there was either (1) expulsion by the old caste or (2) 
intentional abandonment or renunciation by the convert. There being no evidence of expulsion or 
ostracism by the old caste, the question was whether there had been a break from the old order "so 
complete and final that . . . he no longer regarded himself as a member of the Samgar caste."  45.   The 
____________________ 
39. Id., at 839. 
40. Id., at 838. 
41. Id. 
42. Id., at 840. For a situation in which the rules of the new order played a more prominent role, see 
Rhagava Dass v. Sarju Bayamma, A. I. R. 1942 Mad. 413, where by joining the Byragi sect a 
person ceased to belong to his original caste. 
43. A. I. R. 1960 Mys. 27. 
44. On the Arya Samaj, see Lajpat Rai 1915. The caste referred to as "Samgar" is evidently the same 
as that listed as "Samagara" in Census of India 1961 (Paper No. 2: 84). The Sonar caste, 
mentioned below, are traditionally goldsmiths and of a higher social standing, usually associated 
with Vaishya status. See below, n. 54. 
45. A. I. R. 1960 Mys. at 32. 
­292­ 
court found that his activities evinced that he regarded himself as a Samgar, as did his testimony that 
he believed in idolatry and in texts repudiated by the Samajists. The court found no evidence that he 
could not have married the Sonar girl "in the ordinary way" and thus the marriage was not 
inconsistent with his membership in the caste, nor was the census report, since the court refused to 
accept Arya as equivalent to Arya Samajist. Almost as an afterthought, the court notes that the 
Samaj did not expel him for departure from their tenets. Such expulsion would only have reinforced 
the court's conclusion, where the absence of it (if there ever were expulsions) might indicate an 
acceptance inconsistent with his remaining in the caste. The test that emerges, somewhat inchoately, 
is that so long as the person identifies himself with the old caste and is accepted by the caste­­no 
matter if he is accepted by the new group or not­­he remains a member of the caste for purposes of 
eligibility for nomination to a reserved seat. 
C. TRIBE 
In Kartik Oraon v. David Munzni,  46.   the challenged candidate was a member of an Oraon family 
that had been Christian since his grandfather's time. His eligibility to stand for a seat reserved for 
Scheduled Tribes was challenged on the ground that as a Christian he was no longer an Oraon, since 
he had abandoned the animistic faith, did not follow the manners and customs of the tribe, and had 
no affinity of interests or aspirations with the tribal people. The High Court found that the 
candidate's active participation in the civic life of the tribe belied the charge of lack of common 
interests and aspirations. As to manners and customs, there was evidence that although Christian 
tribals did omit certain observances of tribal religion, they retained such practices as exogamy based 
on totemistic lineage, certain harvest rites, ceremonial eating of first fruits, birth and marriage 
observances, and style of writing surnames. Even if he omitted to observe certain festivals or 
observed some in a manner different than other tribals, the court concluded, the "most importnt 
thing . . . is that the non­Christian tribals treat the converted Oraons as tribals, calling them 
"Christian Oraons.  47.   Christian and non­Christian Oraons intermarry and their descendents are 
treated as full members of the tribe. Christian Oraons are invited to feasts and participate in them. 
Applying the Jasani tests, which it found "fully applicable," the court concluded that conversion did 
not extinguish membership in the tribe.  48.   
____________________ 
46. A. I. R. 1964 Pat. 201. On the Oraons, see Rov 1928. 
47. Id., at 203. 
48. A similar approach was applied by an Election Tribunal which had to decide 
­293­ 
In Wilson Reade v. C. S. Booth,  49.   an Election Commission had rejected nomination papers for 
a Scheduled Tribe seat from a candidate whose father was English and whose mother was a 
Khasi. In pre­Independence days he had accepted for himself and his children (his wife was a 
Khasi) privileges restricted to Anglo­Indians. But he was accepted as a Khasi by the 
tribespeople, the group being matrilineal and anyone born of a Khasi mother being regarded 
as a member of the tribe; he had followed "the customs and the way of life of the tribe," was 
treated by them as one of themselves, and had been active in Khasi politics. The Assam High 
Court found that even though he was an Anglo­Indian within the constitutional definition, 50. 
this did not prevent him from being a member of this tribe or some other community. 
Whether he was in fact a Khasi depended not on purity of blood but on his conduct and on 
acceptance by the community. 
More recently the Supreme Court, in Horo v. Jahan Ara,  51.   extended this line of analysis to permit 
assimilation into a tribe by one without any tie of descent. A prominent Scheduled Tribe politician 
married a Tamil Christian woman in 1954. Upon his death in 1970 she sought to fill his seat in 
Parliament, from a constituency reserved for Scheduled Tribes. Her nomination papers were rejected 
by the Returning Officer on the ground that she was not, as she claimed, a member of the Munda 
tribe, since the status of a Munda could be acquired only by birth and not by marriage. In a careful 
assessment of the evidence, the Supreme Court concluded that although Mundas are normally 
endogamous and marriage with non­Mundas is often accompanied by excommunication of the 
Munda partner, such marriages may be sanctioned and those excommunicated and their spouses 
admitted to the tribe upon the performance of certain ceremonies. Here there was abundant 
evidence 
____________________ 
whether a Konda Dora who had converted to Christianity at the age of ten "for the purpose of his 
education" was a member of the Tribe and thus qualified to stand for a reserved seat. There was 
no evidence that the convert had been excommunicated by the Tribe. Finding that converts 
observe the same customs and habits, intermarry, and are treated as members of the tribe, the 
Tribunal held that "mere acceptance of Christianity is not sufficient to make him cease to be a 
member" ( Gadipalli Parayya v. Boyina Rajayya, 12 E. L. R. 83 [1956]). 
49. A. I. R. 1958 Ass. 128. 
50. Art. 366(2) defines an Anglo­Indian as "a person whose father or any of whose other male 
progenitors in the male line is or was of European descent but who is domiciled within the 
territory of India and is or was born within such territory of parents habitually resident therein." 
Some preferences for Anglo­Indians on a declining scale for a ten­year period were provided by 
the Constitution (Arts. 331, 333, 336, 337). On the Anglo­Indian community, see Grimshaw 
1959: 227­40. On this tribe, consult Gurdon 1914. 
51. A. I. R. 1972 S. C. 1840. 
­294­ 
that the marriage was accepted as valid and was approved by the elders of the community 
assembled in the Parha panchayat. Upon such acceptance she became a member of the tribe 
and was entitled to stand for the reserved seat.  52.   
The situation in the Jahan Ara and Wilson Reade cases is the reverse of that in the Jasani, 
Shyamsundar, and Kartik Oraon cases. In the former, it is the "new" group that is the politically 
privileged group rather than the "old" one. The question is not whether the new identificaiton 
precludes the old, but whether the old one precludes the new. In Wilson Reade and Jahan Ara, as in 
the other cases, the courts looked to the views of the privileged group to confirm the individual's 
claimed membership. It was the Khasis who were entitled to special representation, and it was their 
acceptance of Reade that was determinative. Neither birth nor the possibility that the Anglo­Indian 
community might also have accepted him was considered relevant. Similarly, in Jahan Ara it was 
acceptance by the Mundas that was determinative. All of these cases permit overlapping and 
multiple group affiliations. The possibility that an individual might be accepted by a second group 
is not taken to automatically remove him from the first. 
D. TRIBE AND CASTE 
The empirical approach of the Jasani case has not been applied in the problem of tribals attaining 
caste status. The case of V. V. Giri v. D. Suri Dora  53.   arose out of an election to a seat in Parliament 
reserved for a member of a Scheduled Tribe. The candidate was born a Moka Dora, and his family 
had described itself as such in all documents from 1885 to 1928. Since that time they had described 
themselves as Kshatriyas.  54.   There was evidence that the family had adopted Kshatriya customs, 
celebrated marriages in Kshatriva style, was connected by marriage to Kshatriya families, employed 
Brahmin priests, and wore the sacred 
____________________ 
52. There is some uncertainty about the application of the Jahan Ara analysis in the Scheduled Caste 
setting. A provision to accord Scheduled Caste status to women who marry Scheduled Caste men 
was recommended by the Joint Parliamentary Committee on Revision of the Scheduled Castes 
and Tribes Orders, but was not included in the revision of the lists that was enacted in 1976. On 
the Scheduled Caste situation, see text at n. 255, below. 
53. A. I. R. 1959 S. C. 1318. Hereafter referred to as the Dora case. 
54. Classical Hindu legal and social theory divides society into four great varnas (literally "colors") 
or orders: Brahmins (priests and scholars); Kshatriyas (rulers and warriors); Vaishyas (merchants 
and agriculturists); Sudras (menials). Varna distinctions are influential but bear no direct 
relationship to existing social divisions ( Srinivas 1962; chap. 3; Fox 1969). 
­295­ 
thread in the manner of Kshatriyas. His election was challenged on the ground that he was no 
longer a Moka Dora and was therefore ineligible to stand for the seat. 
Applying the tests set forth in the Jasani case, the Election Tribunal concluded that the candidate 
was no longer a Moka Dora, finding that "he has expressed unequivocal intention of drifting away 
from the clan, has totally given up feeling himself to be a member of the Moka Dora tribe and 
considers himself a Kshatriya."  55.   Apparently, the candidate's family was one of a number of 
families of Mokasadars or large landowners who, according to the Tribunal, "would not like to be 
called Moka Doras but considered themselves Kshatriyas."  56.   The Tribunal found support for its 
finding in the observation that "persons of [this] type . . . who have drifted away from their old clan 
and renounced the tribal customs and manners and chosen to adopt the prevailing practices of the 
higher caste [sic] of the Hindu community could not be entrusted with the task of representing the 
genuine grievances" of the tribal communities; to do so would amount to a denial of the benefit of 
special representation conferred on the tribals by the Constitution.  57.   Since the Tribunal had found 
against the candidate on the question of his intention to remain a Moka Dora, it was not necessary to 
go into the question of acceptance by either the tribe or the Kshatriyas. 
So far the case, although reaching the opposite result, proceeded along the lines laid down in the 
Jasani case. However, the case took a radically different turn when it reached the High Court of 
Andhra Pradesh,  58.   which addressed itself, not to the question of whether he had remained a Moka 
Dora, but to the quite distinct question of whether he had become a Kshatriya. Starting with the 
principle that caste is a matter of birth rather than choice and that higher caste cannot be gained, 59. 
the court conceded that 
it is possible that a member of a Scheduled Tribe may in course of time adopt certain 
customs and practices in vogue among the Hindus, but in order to bring them within the 
fold of Hinduism it would take generations. Even if they came within the fold of 
Hinduism, [a] question would arise 
____________________ 
55. 15 E. L. R. 1 at 38 (1957). 
56. Id. 
57. Id. 
58. A. I. R. 1958 A.P. 724. 
59. The cases cited by the High Court in support of this point (id. at 735) are readily distinguishable. 
Sahdeonarain v. Kusumkumari, A. I. R. 1923 P. C. 21, and Chunku Manjhi v. Bhabani Majhan, 
A.I.R. 1946 Pat. 218, are concerned with whether tribals are governed by Hindu personal law. 
Maharajah of Kolhapur v. Sundaram Iyer, A. I. R. 1925 Mad. 497 is concerned with varna status 
for purposes of finding the applicable rule of inheritance. 
­296­ 
whether they have formed a separate sect among themselves, or [whether] they would belong 
to the [Sudras] or to the twice­born class.  60.   
Having thus indicated the severe limitations of possible mobility, the court proceeded to lay down, 
as requirements for proving that such movement had taken place, its version of the tests employed in 
the Jasani case. These tests­­intention, reaction of the old group, rules of the new group­­are used in 
the Jasani case to test whether the individual has remained in his old group. But here the court uses 
them as tests of assimilation to the new group. They become in effect a set of binding requirements, 
which must all be satisfied in order to prove a case of successful mobility. 
The High Court found no evidence of the reaction of the old tribe (which, by the Jasani approach, 
would have been irrelevant had he failed on the test of intention) and 
no evidence as regards the reaction of the new fold, except that some of the Kshatriyas 
recognize appellant as a Kshatriya. We can understand this if this had been the result of 
generations, but the acceptance of the appellant as a Kshatriya by one or two families 
would not . . . be sufficient.  61.   
Since he thus failed to attain Kshatriya status, the High Court assumed that he therefore remained a 
Moka Dora and found him eligible for nomination to the reserved seat. 
The Supreme Court, rather than reasserting the Jasani tests and disengaging them from the High 
Court's theories about caste mobility, took a third tack. Where the Election Tribunal had addressed 
itself primarily to the "intention" test laid down by Jasani, and the High Court had insisted that all 
three factors mentioned in Jasani were required to prove mobility, the Supreme Court fixed its 
attention only on the third­­and originally least important­­of the Jasani tests: the reactions of the 
new group. 
The Supreme Court found the evidence insufficient to demonstrate that the candidate was a 
Kshatriya, since "the caste status of a person in this context would necessarily have to be determined 
in the light of the recognition received by him from the members of the caste into which he seeks an 
entry."  62.   Finding no evidence of such recognition, the Court said "unilateral acts cannot be easily 
taken to prove that the claim for the higher status which the said acts purport to make is 
established."  63.   The Court concluded that the candidate had not be­ 
____________________ 
60. A. I. R. 1958 A.P. at 735. The "twice­born" are the three higher varnas­Brahmins, Kshatriyas, 
and Vaishyas­­who have supposedly undergone a second or intellectual birth upon investiture 
with the sacred thread. The Sudra varna is only once­born. 
61. Id., at 736. 
62. A. I. R. 1959 S. C. 1318 at 1327. 
63. Id. 
­297­ 
come a Kshatriya and had therefore remained a Moka Dora, eligible for the reserved seat.  64.   
In spite of the similarity in outcome, the course of reasoning taken by the Supreme Court here is in 
sharp contrast to that in the cases discussed earlier. The Dora case agrees with them that neither 
birth nor mere intention is determinative of group membership; the conduct of the individual and 
the attitudes of the groups must be considered. But which groups? and their attitudes about what? In 
the cases discussed above, when the question was whether X had, by joining a new group B, ceased 
to be a member of privileged group A, it was the group A whose reactions were consulted­­the 
Mahars in the Jasani case, the Samgars in the Shyamsundar case, the Oraons in the Kartik Oraon 
case. But when the question was whether a person had become a member of a privileged group B, 
then the views not of the old group A but only of group B were pertinent­­the Khasis in the Wilson 
Reade case. It was irrelevant what the Anglo­Indians might have thought of Reade, just as the views 
of the Mahanubhava Panth, the Arya Samaj, and the Christians received only a passing glance. Had 
the courts in these cases seriously considered acceptance by the non­privileged group as 
incompatible with membership in the privileged group, the cases would most probably have had 
different outcomes. Following on these lines, one would have expected the Supreme Court in the 
Dora case to address itself to the views of the Moka Doras. But, like the High Court below, they 
consider only the views of Kshatriyas. 
True, the Jasani case did mention "the rules of the new order" as one of the factors to be considered. 
But it is clear that this was not only the least important factor, but was intended to mean the rules of 
the new order respecting the retention by X of his membership in the old group. It was not the views 
of the Bs as to X's membership in the Bs that counted, but their views as to his membership in the 
As. In the Dora case, when the Supreme Court consults the attitudes of the new group, it is on the 
question of X's membership in the Bs. The practical effect of these divergent approaches can be 
easily seen in tabular form ( table 22 ). Let us imagine that X, a member of privileged Group A, has 
somehow aspired to membership in Group B. Table 22 merely restates the requirement of Jasani 
that to be an A (for the purpose of filling a 
____________________ 
64. Another election tribunal dealing with a similar Mokasadar family reached the same conclusion 
by anticipating the approach followed by the Supreme Court. The candidate remained a tribal 
since "there was no evidence that the Kshatriya community as a whole recognized him as 
belonging to their class" ( Gadipalli Parayya v. Boyina Rajayya, 12 E.L.R. 93 [1956]). One may 
wonder who are "the Kshatriya community as a whole"­­whether this is defined in terms of the 
locality, the district, the state, or all­India? 
­298­ 
reserved seat) X must fulfill both tests I and II. This occurs only in cases 1 and 2. It is clear 
that according to the Dora court the yes on test III in case 1 would make the answer "no." And 
in cases 4 and 6 and 8 the Dora method would make him an A by virtue of his not being a B 
(i.e., failing test 111), whereas the Jasani method would find him not an A because of failing 
either test I (case 6), test II (case 4), or both tests (case 8).  65.   
Since the Supreme Court never discusses the question of the candidate's membership in the Moka 
Doras, one can only gather that there is implicit in the Court's view a logical incompatibility 
between membership in the two groups. Had his Kshatriya status been upheld, he would ipso facto 
have not been a member of the Moka Doras for the purpose of standing for the reserved seat. Faced 
with the question of whether X remained an A, the Court addressed itself to the question of whether 
he had become a B. But this course of reasoning is only plausible if it is assumed, first, that the two 
memberships exhaust the possibilities, and second, that they are mutually exclusive. 
In assuming that they exhausted the possibilities the Court seems to deny the possibility that the 
candidate's family had, although failing in some sense to become Kshatriyas, so separated 
themselves from the tribe as to lose acceptance as members. Such an intermediate possibility was 
considered by the Election Tribunal and to some extent by the 
TABLE 22 CONTRASTING RESULTS OF DIFFERENT TESTS OF MEMBERSHIP 
Case
No.  Test I Test II Test III Is X an A?  Jasani
Did X intend Did the As Did the Bs tests
to remain an accept X accept X
A and so conduct as an A? as a B? 
himself?

Dora
test

1.  yes  yes  yes  yes  no 


2.  yes  yes  no  yes  yes 
3.  yes  no  yes  no  no 
4.  yes  no  no  no  yes 
5.  no  yes  yes  no  no 
6.  no  yes  no  no  yes 
7.  no  no  yes  no  no 
8.  no  no  no  no  yes 
____________________ 
65. It should be noted that using the approach of the High Court in the Dora case the answer would 
be yes in every instance except for case 7 and then only upon the additional condition that the As 
accepted X as a B. 
­299­ 
High Court.  66.   Such splitting off is one of the classic and welldocumented methods by which 
new castes are formed.  67.   
The Jasani line of cases had allowed overlapping and multiple affiliations. It was possible to be 
simultaneously a Mahar and a member of the Mahanubhava Panth, a Samgar and an Arya Samajist, 
a Christian and an Oraon, an Anglo­Indian and a Khasi. Why is it not possible to be both a Moka 
Dora and a Kshatriya? The Court does not indicate the source of its notion that these affiliations are 
mutually exclusive. But it seems that this incompatibility is felt because Indian society is visualized 
as consisting of groups with unique corporate ranks in some definite rank ordering. Thus 
membership in one such group entails occupying such a rank and is inconsistent with membership 
in another group, which would mean simultaneously holding a lower rank in the same system of 
ranks. The Court refers to the claim here as one for "higher status," which presumably cannot be 
achieved without giving up membership in the group with lower status. 
In part, this notion of Indian society as consisting of mutually exclusive groups ranked in a definite 
and unique order is a carry­over from the area of personal law. The courts have long applied to 
members of different religious communions their respective laws in matters of marriage, divorce, 
inheritance, succession, and religious endowments. Since in Hindu law there were some differences 
in the rules applicable to the three higher varnas on the one hand and Sudras on the other,  68.   the 
courts had from time to time to determine which rules were applicable to particular persons or 
groups. For this purpose Hindu society was visualized as if it consisted of four ranked 
compartments­­the lowest being residual­­and any of the actually existing caste groups would be 
assigned, if need be, to one of these theoretical compartments.  69.   Since 
____________________ 
66. 16 E. L. R. at 38; A. I. R. 1958 A.P. at 735. 
67. See e.g., Muthusami Mudaliar v. Masilamani, I. L. R. 33 Mad. 342 (1909); Hutton 1961: 50 ff. 
68. These differences are concisely summarized by Derrett 1958: 383­85. 
69. In order to make such assignments, the courts evolved various tests: lists of diagnostic customs 
(see e.g., Gopal v. Hanmant, I. L. R. 3 Bom. 278 [1879] where the tests of Sudra status are 
widow remarriage and admission of illegitimate sons to dine and marry within the caste and to 
inheritance), or alternatively, tests of reputation (see e.g., Subrao v. Radha, I. L. R. 52 Bom. 497 
[1928], where it is held that varna depends on the consciousness of the caste as to its status and 
the acceptance of this estimate by other castes. According to the theory of varna, Hinduism 
comprised the four varnas and every caste group could be assigned to one of these; caste and 
varna were co­extensive with Hinduism. But departures from the symmetry of this scheme are 
found in many instances where courts modified it to account for the actualities of the situations 
before them. Thus it is possible to have varna standing without belonging to a caste group ( 
Sunder Devi v. Jheboo Lal A. I. R. 1957 all. 215 [convert to Hinduism]; Upoma Kuchain v 
Bholaram I. L. R. 15 Cal. 708 [1888] [daughter of outcaste]; cf Ratansi v. AdministratorGeneral 
­300­ 
the Constitution, the courts have continued to make such varna assignments when necessary. 70. 
Now that the various Hindu Code acts of 1955­56 have eliminated almost all of the instances in 
which varna might make a difference in applicable law,  71.   the courts can look forward to the day 
when they will no longer be faced with the task of making these imponderable and often fictitious 
varna identifications. 
In personal law cases the question was not whether an individual was or was not a member of some 
existing social group, but whether he should be assigned the status of one or the other varna. 72. 
Ordinarily, the individual was indisputably a member of some actual caste or group, and the 
proceedings took the form of determining the varna of this group. It was assumed for this purpose 
that all caste groups could be assigned to one or the other of the varnas. Since the purpose of 
determining varna was to ascertain the appropriate rule of law, and since varnas clearly stood in a 
ranked order, the whole object of the proceeding was to arrive at a unique determination of status. 
The question before the courts in these election cases is quite different in kind. It is whether in fact a 
person is for a particular purpose to be considered a member of some existing group. There is no 
necessary relationship between membership in such a group and the holding of 
____________________ 
General, A. I. R. 1928 Mad. 1279). Caste and varna may apply to persons who are not strictly 
Hindus ( Inder Singh v. Sadhan Singh, I. L. R. [1944] 1 Cal. 233 [Sikh Brahmins]). Caste groups 
have been recognized which have no varna and are not Hindu in any sense ( Abdul Kadir v. 
Dharma I. L. R. 20 Bom. 190 [1895]). Again, members of the same caste may hold different 
varna statuses ( Subrao v. Radha, above). 
70. After the advent of the Constitution the administration of separate personal law to the respective 
religious communities was challenged as discriminatory and Ultra vires Art. 15. Although Art. 
44 directs the eventual elimination of separate personal laws, the continuing validity of disparate 
rules of personal law and the power of the state to create new rules applicable to particular 
religious communities has been upheld (e.g., State of Bombay v. Narasu Appa, A. I. R. 1952 
Bom. 84). The assignment of a community to a varna has been held not to constitute a 
deprivation of rights to equality before the law, nor is it religious discrimination ( 
Sangannagonda v. Kallangonda, A. I. R. 1960 Mys. 147). 
71. I.e., the Hindu Marriage Act of 1955, the Hindu Succession Act of 1956, the Hindu Minority and  
Guardianship Act of 1956, and the Hindu Adoptions and Maintainance Act of 1956. Derrett 
( 1958: 383­85) suggests that the only instances in which varna might continue to have effect are 
succession to sanyasis and determination of the maximum age for adoption. 
72. The judicial treatment of the relation between varna and caste was plagued by confusion, 
engendered in part by the use of "caste" to refer at once (1) to the four great classes or varnas 
into which Hindu society is theoretically divided by the Sanskrit lawbooks, (2) to the multitude 
of existing endogamous groups or jatis, and (3) to intermediate clusters of jatis known by a 
common name (cf. Fox 1969 ). The listings in the Scheduled Castes Order and the various state 
lists of Backward Classes are a mixture of the second and third types. 
­301­ 
varna status.  73.   Apparently, the candidate in the Dora case­­and his family and possibly the whole 
group of Mokasadars­­claimed to be Kshatriyas. It is unclear whether they were merely asserting 
Kshatriya varna status or whether they were claiming membership in some particular endogamous 
group of Kshatriya families. It would seem possible to achieve varna status without necessarily 
becoming effective members of a caste with that status.  74.   In any event, it is unclear whether 
attainment of the varna status claimed in the Dora case would have been felt by the tribe to be 
incompatible with continued membership. Whatever theoretical incompatibility there may be in 
belonging to two varnas, it is not impossible to be accepted as a member of two actual social 
groups. 
But varna theory is not the sole source of the Supreme Court's notion of mutually exclusive group 
membership.  75.   This notion is supported by its picture of Indian society. The Court indicates that it 
bears in mind "the recognized features of the hierarchical social structure prevailing amongst the 
Hindus" and the "inflexible and exclusive nature of the caste system."  76.   The Court is, of course, 
only giving its view of the conditions that obtain; it expresses the hope "that this position will 
change, and in course of time the cherished ideal of a casteless society . . . will be attained." 77. 
Nevertheless, in its anxiety not to be "unrealistic and utopian," the Court seems unnecessarily to 
give currency to the view that all groups in Indian society are ranked in some unique and definite 
order.  78.   Thus the Court is impelled to formulate a general 
____________________ 
73. Cf Mulai v. Lal Dan Bahadur Singh, 9 E. L. R. 9 (1952), where an Election Tribunal found that 
the proclamation of a former prince had not transformed the Gonds of Rewa State into 
Kshatriyas. "But even if the Gonds . . . could be deemed to be Kshatriyas, they would not cease 
to be members of Scheduled Tribes." Even "their aristocratic sub­division known as Raj Gonds, 
still continue to be Gonds and . . . belong to the Scheduled Tribes." 
74. See n. 69, above. 
75. Ironically, the Supreme Court judge who delivered the opinion in Dora was a man of outspoken 
reformist views who opposed caste distinctions and had in his writings eloquently denounced 
varna determinations as a blot on the administration of Hindu personal law ( Gajendragadkar 
1951:516 ff.) The continuing pull of varna notions on judges of the most austerely anti­hierarchic 
views is displayed in Justice Krishna Iyer's suggestion that untouchables are those "outside the 
four­fold Hindu division" ( State of Kerala v. N. M. Thomas, A. I. R. 1976 S. C. 490 at 535). 
76. A. I. R. 1959 S. C. at 1327. 
77. Id. 
78. This tendency to picture Indian society as a series of graded corporate ranks seems congenial to 
the courts, perhaps because of the felt necessity of having some conceptual means for reducing 
the immense variety of Indian society to terms which could be applied without extensive 
investigation in each individual instance. Cf. Marriott's suggestion that urban and educated 
Indians (and foreigners) tend to conceive of caste in terms of criteria which constitute or imply a 
scale of Hindu ritual values, rather than according to the structure of interaction among various 
groups ( Marriott 1959:104). 
­302­ 
standard for assigning standing in such a rank ordering. Noting that whatever may have been the 
case in ancient times, "status came to be based on birth alone," the Court says "it is well known that 
a person who belongs by birth to a depressed class or tribe would find it very difficult, if not 
impossible, to attain the status of a higher caste amongst the Hindus by virtue of his volition, 
education, culture and status."  79.   Thus Hindu society is not only hierarchic but inflexible as well. If 
the Court meant literally that caste status was determined by birth alone, it would of course, be 
redundant to consider the views of any group in order to determine it. But the Court apparently 
means that birth is determinative in the first instance and that this can be varied­­not by an 
individual's conduct but only by the unanimous recognition of his claims by members of the higher 
group. J. L. Kapur, J., dissenting, vigorously rejected the primacy of birth and put forward another 
general theory of assignment of rank. Holding that caste varies as a consequence of the gunas, 
karma, and subhavana and is dependent on actions, he found that the candidate had "by his action 
raised himself to the position of a Kshatriya and he was no longer a member of the Scheduled . . . 
Tribe. . . ."  80.   
Either of these general theories may prove embarrassing. Acceptance of the dissenting judge's 
theory that caste (or varna) status may be gained wholly by individual action could expose the 
courts to a torrent of litigation in which they would be faced with the necessity of setting up legal 
tests for caste standing and assigning it to individuals and communities.  81.   The majority's 
acceptance of birth as the primary determinant of group membership avoids this difficulty and is no 
doubt accurate in the overwhelming majority of instances. But by making membership in the old 
group dependent on failure to achieve the purported membership in the new group, the majority's 
theory may disincline the courts from giving legal recognition to existing patterns of mobility, 
which ordinarily involve a period of conflicting claims and overlapping 
____________________ 
79. A. I. R. 1959 S. C. at 1327. 
80. Id., at 1331. Gunas, etc., means roughly material nature, deeds, and temperament. R. Dhavan 
( 1977: xxix) points out that this is an Arya Samaj theory of caste. 
81. Of course, such litigation occurs without judicial encouragement. See e.g., Sankaram Namboodri 
v. Madhavan, A. I. R. 1955 Mad. 579, where one section of a family of Embrandiri Brahmins 
changed its name and claimed that they were Nambudiri Brahmins and therefore governed by the 
special personal law of the latter. The court was willing to concede that had they voluntarily 
effected a partition in accordance with Nambudiri law, this would help establish their status as 
Nambudiris for purposes of personal law in the future, but held that the change of name was a 
mere unilateral declaration of no effect on their status in the absence of evidence of renunciation 
by the whole family of their present status or of their acceptance by the Nambudiri community as 
a whole. 
­303­ 
identifications. Successful separation from an old group may be overlooked, with the result of 
imposing on a privileged group a candidate who is not an accepted member of it. More generally, 
existing channels of mobility may be discredited. For if acceptance by a new group removes one 
from the old, the Hinduization of tribals and the formation of new sects would be accompanied by 
the danger of disqualification for receipt of preferences. Since the system of preferences is designed 
to increase flexibility and mobility within Indian society, there seems little reason to make 
abandonment of older and slower methods of mobility a condition for the utilization of the new 
ones. 
The dilemma posed by these opposing theories can be solved by eschewing any general theory of 
assignment and deciding questions of group membership by the pragmatic approach of the Jasani 
case. Since this permits multiple identifications, the courts would not pose for themselves the kind 
of either/or puzzles that have no satisfactory answer. They would be concerned only with whether 
for the purpose of the particular measure the individual concerned ought to be counted a member of 
the privileged group.  82.   As far as it is compatible with the particular legislative policy, membership 
would depend entirely on the voluntary affiliation of the individual  83.   ­­as confirmed, where 
necessary, by voluntary acceptance by the group's members. Consonant with the constitutional 
principles of freedom of association and the autonomy of social groups, State imposition of 
standards of membership would be minimized,  84.   and judicial determination of the relative status of 
groups would be eliminated. 
____________________ 
82. In Jankilal v. Jabarshingh, A. I. R. 1957 Nag. 87, the court found that "a person can be a Hindu 
and also be deemed a member of an aboriginal tribe" for the purposes of the provisions of the 
Central Provinces Land Alienation Act, which outlawed conditional sale provisions in mortgages 
made by listed tribes. 
83. A right to benefits is not dispelled by casual claims to "forward" status. Where some applicants 
"have chosen to mention [in application forms] that they belong to the forward community, while 
making it clear that they were Nadar Christians from Kanyakumari Dt., belonging to the Church 
of South India . . . the classification of the petitioners as backward did not depend upon what 
they stated in the application against the relevant column, but whether they had mentioned in the 
application that they belonged to the Church of South India. Once that appeared on the record, it 
would be obvious that they belong to the backward class . . . ." ( Susila v. State of Madras, A. I. 
R. 1970 Mad. 399). 
84. It has been argued with some persuasiveness that the Constitution withdraws all governmental 
power to determine whether an individual is a member of a particular caste. "Can a secular 
government force a citizen to belong or not to belong to a particular caste?" queries the chairman 
of the Backward Classes Commission in considering whether census clerks may put down the 
caste of an individual according to their conception of it or whether the individual's conception is 
determinative. The chairman contends that if caste is to be a voluntary affiliation, government 
must refrain from assigning it ( Report of the Backward Classes Commission 1956: I, xviii). But 
once 
­304­ 
E. RELIGION 
The Constitution is openly and determinedly secular.  85.   Freedom of religion is guaranteed. 86. 
Religious discrimination on the part of the State is forbidden.  87.   The courts have been vigilant in 
invalidating governmental measures framed along religious lines.  88.   Nevertheless, in some 
instances religion has been made a qualification for preferential treatment. The President's 1950 
Order specifying Scheduled Castes provided that "no person professing a religion different from 
Hinduism shall be deemed a member of a Scheduled Caste."  89.   In 1956 this was broadened to 
include Sikhs.  90.   
Who meets this religious qualification? The legal definition of Hinduism, developed for the purpose 
of applying appropriate personal law, was neither a measure of religious belief nor a description of 
social behavior as much as a civil status describing everyone subjected to the application of "Hindu 
law" in the areas reserved for personal law.  91.   
____________________ 
caste and tribe have been accepted as appropriate units for the distribution of preferences, some 
governmental determinations, in order to prevent abuses, are unavoidable. It would seem possible 
to have more or less objective standards, even for determining a voluntary affiliation and to 
refrain from any assignments other than the determination of whether or not an individual is 
within the group to which he claims to belong. 
85. On Indian secularism generally, see Smith 1963 ; Luthera 1964 ; Seminar 1965; Smith 1965 ; 
Sharma 1966 ; V. K. Sinha 1968 ; Derrett 1968; Galanter 1972. For a discussion of the problems 
treated in this section, see Smith 1963:322 ff. 
86. Art. 25, 26, 28, 30(1). 
87. Art. 15, 16, 29(2), 30(2), 325. 
88. State of Rajasthan v. Pratap Singh, A. I. R. 1960 S. C. 1208; Nain Sukh Das v. State of U.P., A. I. 
R. 1953 S. C. 384; State of Jammu and Kashmir v. Jagar Nath, A. I. R. 1958 J. & K. 14. But cf. 
n. 70, above, and n. 113, below. 
89. Constitution (Scheduled Castes) Order, 1950, Para. 3. An exception was included for Sikh 
members of 4 of the 34 Scheduled Castes listed for the Punjab. Cf. the Government of India 
(Scheduled Caste) Order, 1936, para. 3, which provided that "[n]o Indian Christian shall be 
deemed a member of a Scheduled Caste." The Constitution (Scheduled Tribes) Order, 1950, 
contains no religious provision. 
90. The Scheduled Castes and Scheduled Tribes Orders (Amendment) Act, 1956 (63 of 1956), para. 3. 
The main provision now reads "No person who professes a religion different from the Hindu or 
Sikh religion shall be deemed a member of a Scheduled Caste." The inclusion of Sikhs has an 
interesting history going back to dissatisfaction with their exclusion from the Scheduled Castes 
under the 1936 Order. In the Constituent Assembly, Sikhs gave up their demands for political 
safeguards for minorities in return for assurances that their backward classes would be 
recognized and listed as Scheduled Castes (there was some opposition to this inclusion from 
among the Scheduled Castes). See the speech of Sardar Hukum Singh at IX CAD 235; remarks 
of Sardar V. J. Patel, IX CAD 247. Sardar Hukum Singh moved an amendment to this purpose in 
the Constituent Assembly, but it was withdrawn with the understanding that the Sikhs would be 
considered in the making of the lists of Scheduled Castes ( VII CAD 552 ff.). 
91. Or, more accurately, all who would be subject to Hindu law in the absence of proved special 
custom or of a contingency such as marriage under the Special Marriage 
­305­ 
Heterodox practice, lack of belief, active support of non­Hindu religious groups,  92.   expulsion by a 
group within Hinduism  93.   ­­none of these removed one from the Hindu category, which included all 
who did not openly renounce it or explicitly accept a hostile religion. The individual could venture 
as far as he wished over any doctrinal or behavioral borders; the gates would not shut behind him if 
he did not explicitly adhere to another communion.  94.   The same negative definition remains today 
for purposes of application of personal law. In Chandrasekhara Mudaliar v. Kulandaivelu Mudaliar, 
 95.   the Supreme Court had to decide on the validity of a consent to adoption by one who disavowed 
belief in the religious efficacy of adoption, in Hindu rituals and scriptures, in "atma" [soul] and 
salvation. The Court found that "the fact that he does not believe in such things does not make him 
any less a Hindu. . . . He was born a Hindu and continues to be one until he takes to another religion 
. . . whatever may be his personal predilections or views on Hindu religion and its rituals."  96.   
In the post­constitutional cases involving preferences the same latitudinarian conception of 
Hinduism has been carried over from the area of personal law. To "profess" Hinduism merely means 
to be a Hindu by birth or conversion.  97.   Unorthodox practice or lack of personal belief in its tenets 
does not mean lack of profession for this 
____________________ 
Act (III of 1872, now 43 of 1954). But persons who are not Hindus may attract the application of 
Hindu law. These include some Christians (see e.g., Chinnaswammy v. Anthonyswamy, A. I. R. 
1961 Ker. 161. Cf. n. 36, above). Tribals may be subject to Hindu law although they are not 
Hindus. See Mira Devi v. Aman Kumari, A. I. R. 1962 M.P. 212. Hindu personal law was applied 
to some Muslim groups until the passage of the Muslim Personal Law (Shariat) Application Act 
(XXXVI of 1937). 
92. Bhagwan Koer v. Bose, 30 I.A. 249 (1903). 
93. Ratansi D. Morarji v. Admr. General of Madras, A. I. R. 1928 Mad. 1279. 1283. 
94. The concept of "Muslim" is treated somewhat differently. There are expressions of the same 
negative (no conversion) test. See Bhagwan Bakhsh Singh v. Drigbijai Singh, I. L. R. 6 Luck. 487 
(1931). Generally it is held that adherence to some minimum of beliefs (the unity of God, the 
mission of Mohammed as his prophet, and the authority of the Koran) is necessary and sufficient 
to make one a Muslim ( Narantakath v. Parakkal, I. L. R. 45 Mad. 986 [1922]; Jiwan Khan v. 
Habid, I.L.R. 14 Lah. 518 [1933]). Repudiation of these beliefs, even without conversion to 
another religion, makes one not a Muslim ( Resham Bibi v. Khuda Bakhsh, I. L. R. 19 Lah. 277 
[1938]). Compare the test for Sikhs, n. 100 below. 
95. A. I. R. 1963 S. C. 185. 
96. Id., at 200. See also the definition of "Hindu" set out in the Hindu Marriage Act, 1955, Section 2, 
and discussion in Derrett 1963, §§17­20. A similar latitudinarianism may be observed in the tests 
for whether a tribe is sufficiently Hinduized to attract the application of Hindu law. Orthodoxy is 
unnecessary; it is sufficient that the tribe acknowledge themselves as Hindus and adopt some 
Hindu social usages, notwithstanding retention of non­Hindu usages ( Chunku Manjhi v. 
Bhabani Majhan, A. I. R. 1946 Pat. 218). 
97. Michael v. Venkataswaran, A. I. R. 1952 Mad. 474. 
­306­ 
purpose­­one may eat beef and deny the authority of the Vedas. In effect the test seems to amount to 
a willingness to refrain from calling oneself something else. Thus where the election to a reserved 
seat of an active supporter of Dr. Ambedkar's Buddhist movement  98.   was challenged on the ground 
that he was not a Hindu, the court found that "it has to be established that the person concerned has 
publicly entered a religion different from the Hindu . . . religion."  99.   Mere declarations falling short 
of this would not be sufficient. The candidate had supported the movement for mass conversion by 
serving on the reception committee, editing a newespaper supporting the movement, and attending a 
rally where an oath, "I abandon the Hindu religion and accept the Buddha religion," was 
administered by Dr. Ambedkar. When those who wished to convert were asked to stand, the 
candidate stood. But there was no evidence that he did in fact take the oath; the court held that in 
the absence of such a declaration, he remained a Hindu.  100.   
The same test of public declaration was subsequently upheld by the Supreme Court in Punjabrao v. 
Meshram,  101.   where there were eyewitnesses that the candidate did convert to Buddhism at a mass 
ceremony. In spite of their political enmity to the candidate, the Court finds the testimony of these 
witnesses credible in view of the fact that the candidate was a prominent follower of Dr. Ambedkar 
and "it would be highly improbable that [he] would have remained aloof' from the Buddhist 
movement.  102.   This direct evidence was corroborated by a later written declaration and by certain 
acts and omissions by the candidate. The "strongest circumstance" corroborating the evidence of the 
conversion is that as chairman of a local committee, the candidate presided over the conversion of a 
small Shiva temple into a Buddhist temple. "[H]owever great the admiration or regard that a Hindu 
may have for Lord Buddha, he would shudder at the idea of desecrating a Shiva Linga in this 
manner or even of converting what was once a Shiva Temple into a 
____________________ 
100. Id., at 299. The vagaries of the declaration test are illustrated in Rattan Singh v. Devinder Singh, 
7 E. L. R. 234 (1953), 11 E. L. R. 67 (1955), where the candidate had at various times described 
himself as a Mazhabi Sikh, a Harijan Hindu, a Balmiki, and a Balmiki Hindu. The Tribunal, 
holding that the minimum qualification for being a Sikh is willingness to declare "I solemnly 
affirm that I believe in the ten gurus and that I have no other religion," found him to be a 
Balmiki Hindu in 1953 and a Mazhabi Sikh in 1955. Any objective evidence was rigorously 
excluded since "the question of . . . religion . . . is a matter of personal faith and cannot be the 
subject of any evidence of a third party." 
101. A. I. R. 1965 S. C. 1179. 
102. Id., at 1181. The witnesses, who were also Buddhist converts, were members of a rival faction 
within Dr. Ambedkar's Republican Party (id., at 1180). 
98. On this Buddhist movement, see text at n. 163, below. 
99. Karwadi v. Shambharkar, A. I. R. 1958 Bom. 296 at 297. 
­307­ 
Buddhist Temple."  103.   Second, the candidate willingly married his daughters to Buddhists, upon 
which the Court reflects that "if he were a Hindu . . . it is unlikely that he would have reconciled 
himself with the idea of giving his daughters in marrige to non­Hindus, more particularly when the 
bridegrooms' side insisted on following the Buddhist ritual."  104.   Third, the wedding announcements 
substituted a picture of the Buddha and an invocation of his blessing for the usual picture and 
invocation of the Kuladaivata (household deity). Rejecting the candidate's contention that he was 
treating Buddha as "the 11th incarnation," the Court observed that he was not of that "sophisticated" 
class that have discarded altogether the picture and the blessing, and "had [he] considered himself to 
be a Hindu, he would have followed the usual practice."  105.   
All of the above indicate to the Court that a conversion did in fact occur. The test of conversion is 
public declaration. Entry into another religion must be so public "that it would be known to those 
whom it may interest."  106.   Therefore, if one publicly declares that he has ceased to belong to his old 
religion and has embraced a new one, he will be accepted as professing the new one. There is no 
religious test for membership in the new group beyond the declaration. "In the face of such an open 
declaration it would be idle to enquire further whether the conversion . . . was efficacious."  107.   
A decade later, in its next encounter with the Buddhist issue, the Supreme Court coupled a more 
latitudinarian view of Hindu practice with a condescending if sympathetic view of the Buddhist 
movement. In Ganpat v. Presiding Officer,  108.   the eligibility of several candidates for a Scheduled 
Caste seat in the Maharashtra Assembly was challenged on grounds of their conversion to 
Buddhism. The challenged candidates claimed to be Hindus, and the question was whether this 
claim was belied by other evidence. For example, the alleged departures of one candidate from 
Hindu practice consisted of his being listed on a distinctively Buddhist invitation to his brother's 
daughter's wedding and 
____________________ 
103. Id., at 1183. 
104. Id., at 1182. 
105. Id., at 1182. The incarnations (avataras) of Visnu are commonly accounted to number ten, 
Buddha being the last to have appeared. However, in general Buddha receives little attention 
from Vaishnavite Hindus. Apparently, the candidate here either had a different list or was 
unaware that Buddha was already included. 
106. Id., at 1184. 
107. Id., at 1184. The formalities of conversion are to be tested by the new religion itself, not by 
external standards. In the case of these Buddhists it appeared that officiation by a Bhikku (priest) 
was not required and recitation of three vows and five precepts was considered sufficient (see id. 
at 1184). On the problem of a declaration test, seen. 100, above. 
108. A. I. R. 1975 S. C. 420. 
­308­ 
participating in the wedding, during which pictures of Dr. Ambedkar and Buddha were garlanded, a 
Buddhist Bhiku officiated, and mantras were recited in Pali. The Court was unpersuaded that these 
instances established that the candidate, a born Hindu, had ceased to profess Hinduism. 
The Court's skepticism is based in part on its view that religious identification is "a highly personal 
matter" and that "open assertion," especially by an educated person, should be given greater weight 
than the interested testimony of others based on "stray instances."  109.   But it also finds support in 
the Court's views of Hinduism and of the Buddhist movement. The Court remains unpersuaded that 
there is any significant departure from Hinduism here. 
Hinduism is so tolerant and Hindu religious practices so varied and eclectic that one 
would find it difficult to say whether one is practicing or professing Hindu religion or 
not. Especially when one is born a Hindu the fact that he goes to a Buddhist temple or a 
church or a durgah cannot be said to show that they are no more Hindus unless it is 
clearly proved that they have changed their religion from Hinduism to some other 
religion.  110.   
Hinduism's tolerance of syncretism combines with the persistence of custom to make it difficult to 
abandon. "[I]f a certain community in a spirit of protest says they would like to give up Hinduism 
and adopt Buddhism it is not likely to make much change either in their beliefs or in their practice." 
If in circumstances of extreme oppression, "some members of the Scheduled Castes in their protest 
resort to desperate measures to erase the indignity of untouchability one cannot blame them. But 
whether it produces any result is a different question however well meaning such efforts may be." 
 111.   
Hindu tolerance of Buddhist practices raises the question of whether Buddhism is "a religion 
different from the Hindu . . . religion" within the meaning of the Scheduled Castes Order. Converts 
to Christianity and Islam are, of course, non­Hindus.  112.   But Buddhists, Jains, and Sikhs are treated 
as Hindus for many purposes.  113.   Hindu is an 
____________________ 
109. Id., at 425. 
110. Id., at 424. 
111. Id., at 424. 
112. Michael v. Venkataswaran, above n. 97. But one may remain a Hindu even while belonging to a 
religious sect that includes non­Hindus among its members ( State of Bombay v. Yagna Sastri 
Purushadasji, 61 Bom. L. Reporter 700 [1960]). 
113. These groups are Hindus for purposes of personal law. See, e.g., Hindu Marriage Act, 1955, §2. 
They are specifically included among Hindus for purposes of authorizing the state to open 
Hindu temples to untouchables. See the explanation II to Art. 25 (2) of the Constitution. But 
their separateness is recognized in some contexts; e.g., Jains were not of "the same religion" as 
Hindus for the purposes of temple­entry legislation ( State v. Puranchand, A. I. R. 1958 M.P. 
352; Devarajiah v. Padmanna, A. I. R. 1958 Mys. 84). 
­309­ 
equivocal term, sometimes used with reference to adherents of more or less "orthodox" vedic and 
brahmanical communions and at other times used to embrace the full array of "heterodox" sects, 
including Sikhs, Jains, and Buddhists. Yet Jains and Buddhists are considered nonHindus for 
purposes of preferences, and Sikhs are mentioned separately.  114.   The candidate in Meshram argued 
that "Hindu" in the Scheduled Caste Order should be given a broad meaning that would include 
Buddhist, in line with much other official usage. The Supreme Court finds this clearly negatived by 
the language of the Order: the separate mention of Sikhs would otherwise be redundant. This 
reading of the Order clearly accords with legislative intention to exclude the major "heterodox" 
communions. But the Court's observation that "the word 'Hindu' is used in the narrower sense of the 
orthodox Hindu religion, which recognizes castes and contains injunctions based on caste 
distinctions,"  115.   is troubling. It comes close to suggesting a positive definition of Hinduism which 
goes beyond merely excluding Sikhs, Jains, and Buddhists to the point of throwing doubt on the 
inclusion of all sects which in one fashion or another repudiate caste distinctions in doctrine or 
practice. 
This is particularly discomfiting because the Court, while inquiring what it means to "profess" a 
religion "different from the Hindu," reverses the question to examine what it means to profess 
Hinduism. It reads the Order as requiring that to be treated as a member of a Scheduled Caste "a 
person . . . must be one who professes either Hindu or Sikh religion."  116.   And it notes, "the word 
'profess' in the Presidential Order appears to have been used in the sense of open declaration of 
practice by a person of the Hindu (or the Sikh) religion."  117.   
The Court is obviously not attempting to establish a positive religious test of eligibility for 
preferences. But the unfortunate juxtaposition of these dicta on Hinduism and its profession with 
the evidentiary exami­ 
____________________ 
Differences in regard to the organizatin of religious trusts may be recognized by having a Hindu 
religious trusts law which covers Jains and Buddhists but not Sikhs ( Moti Das v. S. P. Sani, A. I. 
R. 1959 S. C. 942). 
114. They were excluded until 1956, see note 90 above. ( Gurmukh Singh v. Union of India, A. I. R. 
1952 Pun. 143; Rattan Singh v. Devinder Singh, n. 100, above). 
115. A. I. R. 1965 S. C. at 1184. The Supreme Court has more recently enunciated views of Hinduism 
which seem to conflict with its description here. In Sastri Yagnapurushdasji v. Muldas  
Bhundardas Vaishya, A. I. R. 1966 S. C. 1119, it was held that the Swaminaravan sect was, in 
spite of its protestations, part of Hinduism and subject to Bombay's temple­entry act. The 
Supreme Court observed that the sectarians' belief in caste pollution "is founded on superstition, 
ignorance and complete misunderstanding of the true teachings of the Hindu religion. . . ." 
116. A. I. R. 1965 S. C. at 1184. 
117. Id., at 1184. 
­310­ 
nation of conventional Hindu practice and attitude with regard to intermarriage, invitations, and 
respect for idols comes uncomfortably close to suggesting that these Hindu proprieties are a 
requirement for preferences. Not only converts, but anti­caste sectarians, atheists, and militant 
iconoclasts would feel the pinch of such a test. It would not only be extremely difficult to apply, but 
it would no doubt violate the understandings of many Hindus, impede innovation and change within 
Hinduism, and present serious problems of religious discrimination and freedom of religion.  118.   
In spite of the Court's remarks, it is clear that the Scheduled Caste Order itself does not establish or 
sanction such a positive religious test. It does not require anyone to profess Hinduism, much less 
practice it. It merely requires that he does not profess a different religion. Professing Hinduism and 
professing a non­Hindu religion may be mutually exclusive categories, but they are not exhaustive. 
 119.   The question is whether one professes a non­Hindu religion; the profession and practice of 
Hinduism need be considered, as they are in Meshram, only for their evidentiary value in this 
inquiry. 
It is clear, than, that Buddhists are excluded by the presidential Order. The Meshram Court merely 
notes that when a person "has ceased to be a Hindu he cannot derive any benefit from that Order." 
 120.   The Court never reaches the question of whether the "Hinduism" test for recipients of 
preferences infringes the constitutional ban on religious discrimination by the State.  121.   
____________________ 
118. While the Order's religious distinction itself is open to these objections, it might conceivably 
correspond to some difference in conditions. Any relevant differences between those who 
observe Hindu proprieties and those who do not is even more implausible. 
119. One wonders whether someone who repudiated Hinduism, or an avowed athiest, would be held 
to "profess a religion other than Hinduism." Cf. Chandrasekhara Mudaliar v. Kulandaivelu 
Mudaliar. A. I. R. 1963 S. C. 185. 
120. A. I. R. 1965 S. C. at 1184. 
121. The failure to reach the constitutional Issue was due to the preoccupations of Meshram's counsel 
in the High Court. The Election Tribunal had decided the factual question of a conversion 
against the candidate. Much as the Supreme Court did later. On appeal to the High Court. 
Meshram's counsel never raised the question of the constitutionality of the religious 
classification. Instead, counsel were eager to argue that Buddhism was in fact not a religion 
different from Hinduism. The High Court held that this was a factual matter which should have 
been pleaded and proved by evidence and refused to allow the argument to go forward. 
Assessing the evidence, the High Court found no evidence of authority to effectuate a 
conversion to Buddhism and a predominant political rather than a religious motive. The court 
concluded that the ceremony was a protest against caste and demonstrated petitioner's sympathy 
with Buddhism but was not tantamount to the requisite profession of non­Hindu religion. On 
appeal, the Supreme Court re­assessed the evidence once more and re­instated view the of the 
Election Tribunal that the conversion was effectuated. 
­311­ 
The constitutional challenge had been raised in several earlier cases. In Gurmukh Singh v. Union of 
India,  122.   a Bawaria Sikh protested his exclusion from the Scheduled Castes, in which the 
presidential Order had included Hindu Bawarias. A Full Bench of the Punjab High Court found that 
Article 341 empowered the President to select those "parts of castes" which he felt should be 
included and that he could select those parts on the basis of religion without violating Article 15(1), 
the ban on religious discrimination, since Article 15(4), which authorized preferential treatment for 
the backward, operated as an exception to the prohibitions of 15(1). The Court conceded that 
Scheduled Castes were to be designated on the basis of their backwardness. But, finding that the 
Constitution vested in the President the entire power to make such determinations, the Court refused 
to review his order by considering whether the Sikh Bawarias were in fact sufficiently backward to 
be included. 
In Michael v. Venkataswaran,  123.   the religious requirement was upheld against a Paraiyan convert to 
Christianity who wished to stand for a reserved seat. Even if there are cases in which both the 
convert and his caste fellows consider him as still being a member of the caste, the court found, "the 
general rule, is [that] conversion operates as an expulsion from the caste . . . a convert ceases to have 
any caste."  124.   The presidential Order, according to the court, proceeds on this assumption and 
takes note of a few exceptions.  125.   The court declined to sit in judgment on the President's 
determination that similar exceptional conditions do not prevail in other instances. Thus the 
presidential Order was upheld not because of an absence of judicial power to review it but because 
of its accuracy in the general run of cases. 
In In re Thomas,  126.   another bench of the Madras Court considered a convert case which did not 
involve the presidential Order. The Madras Government had extended school­fee concessions to 
converts from Scheduled Castes "provided . . . that the conversion was of the . . . student or of his 
parent." A Christian student whose grandfather had converted could not, it was held, complain of 
discrimination on grounds of religion. The basic criterion was not religion but caste, and converts 
did not belong to the Scheduled Castes. By conversion they had "ceased to belong to any caste 
because the Christian religion does not recognize 
____________________ 
122. A. I. R. 1952 Punj. 143. 
123. A. I. R. 1952 Mad. 474. The continuing efforts of Christian groups to secure the extension of 
benefits to recent converts and to established Christians are usefully compiled in National 
Council of Christian Leaders ( 1978). 
124. Id.,at 478. 
125. I.e., the four Sikh groups listed in the 1950 Order. See n. 90, above. 
126. A. I. R. 1953 Mad. 21. 
­312­ 
a system of castes."  127.   The concessions to recent converts were merely an indulgence, and the State 
could determine the extent of this indulgence.These cases indicate the three distinct grounds on 
which the religious disqualification has been upheld: (1) the inappropriateness of reviewing the 
President's Order; (2) the constitutional authorization for using religious criteria in the designation 
of beneficiaries; (3) the theory that no religious discrimination is involved, since non­Hindus have 
no caste. 
1.  Is the Scheduled Castes Order unreviewable? Executive action, even in pursuance of expressly 
granted and exclusive constitutional powers, is not immune from judicial review for conformity 
with constitutional guarantees of fundamental rights. The special status of the Order as an act 
of the President is no bar (if it ever was),  128.   for since 1956 the religious requirement has been 
promulgated by Parliament. In any event, the posture of the Gurmukh Singh and Michael cases 
was one of judicial restraint rather than judicial powerlessness.  129.   Is such restraint 
appropriate? Since those cases were decided, judicial power to review the government's 
designation of beneficiaries of preferences has been firmly established. Since 1959 the courts 
have subjected the standards used by state governments to designate backward classes to close 
and detailed scrutiny.  130.   There is no indication that the delineation of Scheduled Castes by the 
President and Parliament enjoys an immunity from similar judicial scrutiny where challenged 
as contravening fundamental rights.  131.   
____________________ 
127. Id., at 22. The exclusion of Buddhists from the preferences for Scheduled Castes has been 
similarly justified. See discussion below in text at nn. 163 ff 
128. Art. 12, 13. in Karkare v. Shevde A. I. R. 1952 Nag., 330, the court found that the immunity 
conferred on President and state governors by Art. 361 "does not place the actions of the 
Governor purporting to be done in pursuance of the Constitution beyond the scrutiny of the 
Courts. . . . Unless there is a provision excluding a particular matter from the purview of the 
Courts [as in Arts. 122, 212, 263, 329 (a)] it is for the Courts to examine how far any act done in 
pursuance of the Constitution is in conformity with it." But cf. Biman Chandra v. Governor, A. 
I. R. 1952 Cal. 799, holding that Art. 361 removes the acts of a governor from judicial review 
unless there is evidence of dishonesty or bad faith. 
129. Art. 12 provides that for purposes of the Fundamental Rights provisions of the Constitution, 'the 
State' includes the Government and Parliament of India and the Government and the Legislature 
of each of the States and all local or other authorities. . . ." In Gurmukh Singh the court concedes 
that the President is included and that therefore his action is governed by the requirements of the 
chapter on fundamental rights. 
130. See chaps. 7 and 8, above. 
131. See Art. 12, n. 129 above. The Hinduism requirement is an expression of the power conferred by 
Art. 341 to select "castes, races or tribes or parts of or groups within castes, races or tribes" 
(emphasis added). Is the power of the President and of Parliament (which has exclusive power to 
modify the list) subject to review on the 
­313­ 

2.  Article 15(4) permits special provisions in favor of Scheduled Castes notwithstanding the 
provisions of Article 15(1), which ban discrimination by government on grounds of caste or 
religion. Does Article 15(4) authorize the use of religious criteria in selecting the Scheduled 
Castes? Caste and religion are normally forbidden bases of classification.  132.   Constitutional 
authorization for special provisions to disadvantaged groups has been held to authorize the use 
of such otherwise forbidden classifications.  133.   However, their use is still subject to the 
standards applicable to any governmental classification of citizens. The classification "must be 
founded upon an intelligible differentia which distinguishes persons or things that have been 
grouped together from those left out . . . and . . . the differentia must have a rational 
relationship to the object sought to be achieved."  134.   In selecting Backward Classes, caste units 
may be employed, but caste standing may be used as a criterion only to the extent that it is 
useful in identifying social backwardness and cannot be the sole or dominant test for that 
purpose.  135.   Religious groups, too, may presumably be the units used in designating Backward 
Classes, but it is clear that religion cannot be the sole criterion of their backwardness.  136.   Non­
Hindus cannot be made the recipients of preferences on the basis of purely religious 
classification in the absence of other evidence of backwardness. One wonders why they can be 
excluded from preferences solely on the basis of religion in spite of the presence of such 
evidence. Perhaps the State's power to define Backward Classes is not as broad as the 
President's (and Parliament's) to define Scheduled Castes. And perhaps the relevance of 
religion to untouchability, notwithstanding its inappropriateness in marginal cases, is more 
apparent than to defining backwardness. But it cannot be 
____________________ 
question of whether the delineation of "parts" or "groups" is reasonably related to the object of 
the classification? There is no indication that this power is exempt from such review. See Art. 
13(2) and (3). 
132. On the religious classification, see cases noted at 88 above. On caste classifications, see 
Sanghar Umar v. State, A. I. R. 1952 Saur. 124; Bhopalsingh v. State, A. I. R. 1958 Raj. 41; State 
of Madras v. Champakam Dorairajan, A. I. R. 1951 S. C. 226. 
133. The power to designate "classes" in Art. 15(4) and Art. 16(4) operates as an exception to the 
prohibition on the use of all grounds of classification in Arts. 15(1) and 16(2) respectively. E.g., 
Ramakrishna Singh v. State of classification in Arts A. I. R. 1960 Mys. 338, 349. See chap. 7, § 
D, above. 
134. A. I. R. 1960 Mys. at 346­48, A. I. R. 1961 Mys. at 229. These cases represent the application to 
the field of preferences of the general standards for the constitutionality of classifications, 
firmly, establsihed by numerous rulings of the Supreme Court. See, e.g., Budhan Chaudhry v. 
State of Bihar, A. I. R. 1955 S. C. 191; Bidi Supply Co. v. Union of India, A. I. R. 1956 S. C. 479. 
135. Balaji v. State of Mysore, A. I. R. 1963 S. C. 649 at 658; Chitralekha v. State of Mysore, A. I. R. 
1964 S. C. 1823 at 1833. For full discussion, see chap. 7, § A. above. 
136. See chap. 7, above. 
­314­ 
inferred that this exempts the power from review in those cases where the use of such standards 
is challenged as inappropriate. 

The religious requirement is an expression of the power which Article 341 confers on the 
President and Parliament to determine which "caste, race or tribe or part of or group within any 
caste, race or tribe" shall be included in the list of Scheduled Castes.  137.   If the standards of 
reasonable classification are to be applied, the crucial consideration would be whether the 
division into Hindu and non­Hindu corresponds to some difference in conditions, resources, or 
the incidence of disabilities so that the division is rationally related to the object of the 
preferences. Existing precedents would not seem to foreclose such an approach, since the cases 
explicitly upholding religious tests were all decided before Judicial review was firmly 
established in the preference area and before the Jasani case and others had developed an 
empirical approach to questions of group membership.  138.   At least one court seems to have 
moved along these lines. In an unreported 1967 judgment, the Madras High Court warned the 
state government that the claim of a person to belong to the Backward Classes could not be 
rejected merely on the ground of his recent conversion to a religion like Christianity.  139.   

3.  The argument that religious criteria are appropriate rests squarely upon the third ground­­that 
acceptance of a non­Hindu religion operates as loss of caste. It is castelessness which 
constitutes the crucial difference of non­Hindus from Hindus. In Rajagopal v. Armugam a two­
judge bench of the Supreme Court subscribed to this notion of the de­castifying effects of 
conversion.  140.   Upon his youthful conversion to Christianity, the candidate 

lost his membership of' the Adi Dravida Hindu caste. The Christian religion does 
not recognize any caste classifications. All Christians are treated as equals and 
there is no distinction between one Christian and another of the type that is 
recognized between members of different castes belonging to Hindu religion . . . . 
The tenets of Christianity militate against persons professing Christian faith being 
divided or discriminated against on the basis of any such classification as the caste 
system. It must therefore be 

____________________ 
137. Italics added. 
138. Punjabrao v. Meshram, A. I. R. 1965 S. C. 1179, clearly does not foreclose such an approach, 
since it decides only the disputed fact of whether there was a conversion and never reaches any 
of the constitutional arguments against the religious classification. 
139. Meera Bai v. Director of Medical Education (High Court of Judicature at Madras, 20 September 
1967. Anantanaravanan, C. J., and Veeraswami, J.). 
140. A. I. R. 1969 S. C. 101. In the litigation between these parties their names are variously spelled 
Rajagopal and Rajagopal, Armugam and Arumugam. In referring to the reported cases, I have 
used the spellings found in the reports; in the narrative account, I have used the longer spellings. 
­315­ 

held that, when the appellant got converted to Christianity in 1949, he ceased to 
belong to the Adi Dravida caste.  141.   

Does acceptance of Christianity, Islam or Buddhism invariably evidence a loss of membership in 
the caste group to which the convert belonged at the time of conversion?  142.   This can be read as 
either a question of fact or a question of law­­a question about his observable interactions with 
others or about his legal status. The finding of castelessness in Rajagopal I is based, not on any 
evidence about the candidate's social interactions with others after his conversion, but solely on 
Christian doctrine­­and oblivious of common knowledge that Christian practice often falls short of 
these aspirations.  143.   The Court acknowledges that in Jasani the sect's anti­caste views were "only 
ideological and involved no change of state."  144.   But it omits to subject the candidate's experience 
as a Christian to the same test of operative effect, repeating only that Christianity "is a religion 
which militates against recognition of caste."  145.   Christian refusal to "recognize" (in the sense of 
sanction or legitimate) caste identities among Christians is taken to foreclose the possibility that the 
court might "recognize" (in the sense of perceive or identify) caste identities that may exist among 
Christians. 
Why in such cases do the courts insist that the act of conversion has as a matter of law deprived the 
convert of membership? In other instances of conversion by members of privileged groups, the 
courts have addressed themselves to the factual side of the question. In the Jasani case the Supreme 
Court set out to determine "the social and political consequences of such conversion . . . in a 
common­sense practical way rather than on theoretical grounds."  146.   Thus in the Jasani and 
Shyamsundar cases the courts addressed themselves to whether the adherence to sects within 
Hinduism had actually removed the person from his caste. 
A similar empirical approach has been applied to conversions among Scheduled Tribes.  147.   The 
Scheduled Tribes Order has never contained any religious disqualification, but some tribal 
politicians have 
____________________ 
141. Id., at 107. 
142. One of the paradoxes of the religious test as now applied is that an individual convert from 
Hinduism to Sikhism or vice versa automatically remains within the Scheduled Castes without 
any consideration of his intention to remain in the caste, continued acceptance by his caste 
fellows, or change in his condition. He is accounted a member of the caste even though, 
according to the loss of caste by conversion theory, he has renounced it. 
143. In a sequel to this litigation, another bench of the Supreme Court displays a thorough 
acquaintance with the frequency of caste identity among South Indian Christians, but comes to a 
similar conclusion that converts do not suffer from disabilities ( Arumugam v. Rajgopal [II], A. I. 
R. 1976 S. C. 939, 949). See discussion in § F, below. 
144. A. I. R. 1969 S. C. at 108. 
145. Id., at 108. 
146. (1954) S. C. J. at 326. 
147. See § C, above. 
­316­ 
pushed for a bar on converts. Thus, Mr. Kartik Oraon (who back in 1963 failed to disqualify his 
Christian opponent on the ground that he was no longer an Oraon)  148.   cited as a favorable precedent 
the protection of Scheduled Castes from "a more aggressive, vocal and advanced force, i.e., those 
converted to other religions."  149.   The 10% of converts, he claimed, took 90% of the benefits 
intended for the tribals. In 1969 the joint Committee on the Scheduled Castes and Scheduled Tribe 
Orders Amendment Bill proposed a provision that 
no person who has given up tribal faith or faiths and has embraced either Christianity or 
Islam, shall be deemed a member of any Scheduled Tribe.  150.   
This proposal provoked heated controversy during the 1970 debate on the Amendment Bill, since it 
proposed a glaring distinction between Hindu tribals and those who converted to Christianity and 
Islam. (It was pointed out that only 4% of the Scheduled Tribes had declared themselves in the 1961 
census to be following a tribal religion.) Ironically, the government opposed the proposal on the 
ground that it constituted unconstitutional discrimination "between converts to Christianity and 
Islam on the one hand and converts to other religions on the other."  151.   Mr. Kartik Oraon expressed 
puzzlement that the government found a religious test unconstitutional in the case of Scheduled 
Tribes but permissible in the case of Scheduled Castes.  152.   Similarly, government was alert to the 
fact that conversion did not automatically change the conditions of tribals, but insistent that 
conversion by Scheduled Castes did effectuate a fundamental change.  153.   In any event, the version 
of the bill introduced and passed in 1976 did not contain any such provision. 
Like the government, the courts have dealt differently with questions of religious change among the 
Scheduled Castes than among the Scheduled Tribes.  154.   Why, in dealing with religious 
classification among the Scheduled Castes, have the courts forsaken the empirical approach they 
found congenial in other sittings? Why, rather than looking to the facts of the individual case, have 
they chosen to apply this theoretical rule about loss of caste? As the Supreme Court observed in the 
Jasani
____________________ 
148. See Kartik Oraon v. David Munzni, A. I. R. 1964 Pat 201, discussed in § C, above. 
149. LSD (4th series), Vol. 16, No. 50, col. 3147 ( 25 Apr. 1968). 
150. RCSCST 1969­70: 93. 
151. Shri K. Hanumanthaiya, Minister of Law and Social Welfare, LSD (4th series), Vol. 45. No. 6, 
cols. 318­19 ( 17 Nov. 1970). 
152. Shri Kartik Oraon at place cited in n. 149; also at LSD (5th series). Vol. 44, No. 16, col. 53 ( 2 
Sep. 1976). 
153. See Shri B. K. Daschowdhury, LSD (4th Series), Vol. 45, No. 10, col. 265 ( 23 Nov. 1970). 
154. See § C, above. 
­317­ 
case, "conversion . . . imports a complex composite composed of many ingredients. Religious 
beliefs, spiritual experience and emotion and intellectual conviction mingle with more material 
considerations such as severance of family and social ties and the casting off or retention of old 
customs and observances. The exact proportions of the mixture vary from person to person."  155.   It 
is surprising that the courts both in old cases like Michael and Thomas and in recent ones like 
Rajagopal have accepted a picture of conversion which corresponds more closely to missionary 
aspirations than to observable consequences. 
This acceptance reflects the continued force of a view of caste groups which sees them as units in 
an overarching sacral order of Hinduism. To abandon the whole is to abandon the part. According to 
this view, which prevailed prior to the Constitution, caste groups were visualized as occupying a 
unique place in an integrated but differentiated religious order in which the different parts enjoyed 
rights and duties, privileges and disabilities as determined by their position in this order.  156.   From 
this view of caste derived the long­standing reluctance of the courts to give legal effect to caste 
standing among non­Hindu communities.  157.   But the courts have always recognized castes among 
non­Hindus when the claim was not an assertion of standing in this sacral order, but was a claim 
regarding the corporate autonomy or internal regulation of the group.  158.   For the latter purposes 
"caste" included "any well­defined native community governed for certain internal purposes by its 
own rules and regulations" and was thus not confined to Hindus.  159.   
In the preference area the question of non­Hindus having caste arises in two sorts of factual 
situations: first, those involving individual converts; second, those involving a caste group or a 
section of a caste made up of members who are non­Hindus. In the case of individual converts, the 
question facing the court would seem to be whether the individual's acceptance of Christianity, 
Islam, or Buddhism evidences a loss of membership in the caste group to which he belonged at the 
time of conversion. There is evidence that in at least some cases of conversion the convert continues 
to regard himself as a member of the old caste and 
____________________ 
155. (1954) S. C. J. at 326. 
156. For an analysis of this and competing conceptualizations of caste, see Galanter 1968. 
157. See, e.g., Michael Pillai v. Barthe, A. I. R. 1917 Mad. 431, where a claim of high­caste 
Christians for restoration of a wall separating them from low­caste Christians in church was 
rejected on the ground that such a claim for precedence could not be enforced among Christians, 
since it was based on Hindu notions of pollution. 
158. Abdul Kadir v. Dharma, I. L. R. 20 Bom. 190 (1895); Yysuf Beg v. Maliq, A. I. R. 1927 Mad. 
397; Inder Singh v. Sadhan Singh, I. L. R. 1944 (l) Cal. 233. 
159. Abdul Kadir v. Dharma, I. L. R. 20 Bom. 190 at 192 (1895). 
­318­ 
to be so regarded by others.  160.   The unwillingness of the courts to credit such membership derives 
not from an examination of the facts but from an identification of castes as components in the sacral 
order of Hinduism. Caste and Hinduism are regarded as coterminous; when Hinduism is abandoned, 
so is caste membership. 
The second type is somewhat more troublesome. There is little dispute that such persons as, e.g., the 
Sikh Bawarias in the Gurmukh Singh case are, in fact, members of a caste group of the kind whose 
existence among non­Hindus has been well known and long recognized by the judiciary.  161.   There 
is evidence that some non­Hindu castes or parts of castes are in circumstances equivalent to those of 
the Hindu untouchables.  162.   To refuse to recognize caste membership in such cases implies that the 
"caste" to which the court is addressing itself is not caste in the sense of a body of persons bound by 
social ties, but caste in the sense of a body of persons which occupies a given place in the ritual 
order of Hinduism. 
A more troublesome case is presented by the adherents of the Buddhist movement initiated in 1956 
by Dr. Ambedkar.  163.   More than three million persons, mostly Mahars from Maharashtra but 
including members of other Scheduled Castes in many states, had become Buddhists by the time of 
the 1961 census. In addition, a large but unknown number have refrained from, or concealed, 
conversion in order to remain eligible for the preferences they enjoyed as Scheduled Castes.  164.   
Exclusion of Buddhists is of course mandated by the "Hinduism" requirement of the Scheduled 
Castes Order: 
____________________ 
160. The. reports are replete with cases in which converts have lived so indistinguishably with their 
caste fellows that the Courts retrospectively infer a tacit reconversion without either formal 
abjuration of the new religion or formal expiation and re­admittance to Hinduism ( Durgaprasda 
Rao v. Sundarsanaswami. A. I. R. 1940 Mad. 513; Gurusami Nadar v. Irulappa Konar, A. I. R. 
1942 Mad. 193). And cf. Muthusami Mudaliar v Masilamani, I. L. R. 33 Mad. 342 (1909), where 
Christian wives were accepted as members of a Hindu caste. The "indulgence" extended by the 
State in the Thomas case, n. 126 above, seems to reflect an awareness that recent converts, if not 
effective members of their old castes, are at least subject to similar disabilities. Cf. the Kerala fee 
concessions in education for Christian converts from the "Backward Classes" (RCSCST 1960­61: 
320). 
161. See I BCC 28 ff; cases cited n. 158, above. 
162. See e.g., the observations of the Commissioner regarding Christian and Muslim sweepers, who 
do not receive special treatment (RCSCST 1960­61: 316). 
163. On this Buddhist (or neo­Buddhist, as it is sometimes called) movement, see Zelliott 1966 and 
1969 ; Keer 1962; Isaacs 1965, Miller 1966. 
164. "The majority [of Buddhists] consider the special privileges their right on grounds of past 
suffering and present economic restrictions, and do not list their children as Buddhists until they 
have made their way through the educational system and into government jobs" ( Zelliott 
1969:292). 
­319­ 
As Buddhism is different from the Hindu religion, any person belonging to the Scheduled Castes 
ceases to be so if he changes his religion. He is not, therefore, entitled to the facilities provided 
under the Constitution specifically for the Scheduled Castes.  165.   
This is reinforced by the notion that converts have no caste since "Buddhism [does] not recognize 
castes."  166.   Shortly after the first conversions the Home Secretary announced: 
The legal position is fairly clear; namely, that if the fact of conversion is proved, then 
obviously the persons concerned cease to belong to any caste much less to the 
Scheduled Castes. Where a right, privilege or preference has been specially recognized 
under the Constitution specifically for members of the Scheduled Castes, such treatment 
cannot obviously continue if the persons have ceased to belong to the Scheduled Castes 
by reason of the conversion.  167.   
Recognizing that conversion was of itself unlikely to improve the condition of the converts, the 
central government recommended that the state governments accord the Buddhists the concessions 
available to the Backward Classes. Such preferences, less in scope and quantity than those for 
Scheduled Castes, have been granted in some states, but others have withdrawn all preferences.  168.   
Persistent efforts of Buddhists to regain the preferential treatment that they formerly enjoyed have 
achieved only partial success. The State of Maharashtra, where most of the Buddhists are 
concentrated, restored to them those of the benefits enjoyed by Scheduled Castes which are 
conferred by the state. But the central government (and other 
____________________ 
165. RCSCST 1957­58: I, 25. This ruling is based squarely on the "Hinduism" requirement of the 
President's Order. See the statement of Pandit Pant, Times of India, Aug. 21, 1957, p. 12, col. 3. 
Compare the observation of the Supreme Court that "Hindu" for purposes of the Scheduled 
Caste Order includes that religion which "recognizes caste and injunctions based on caste 
distinctions" ( Punjabrao v. Meshram, A.I.R. 1965 S.C. 1179 at 1184). This finding, along with 
the non­inclusion of Buddhists among Hindus, was gratifying to most Buddhists, who felt that 
the decision confirmed both the efficacy of their conversion ceremony and the correctness of 
their conviction that they are no longer Hindus ( Zelliott 1966: 207). 
166. Statement of Mr. B. N. Datar in the Rajya Sabha, reported in Times of India, Aug. 27, 1957, p. 
10, col. 1. 
167. Letter of Home Secretary to Mr. B. K. Gaikwad, M.P., of June 13, 1957, reported in U. P. I. 
dispatch from Poona, June 22, 1957. The Home Secretary went on to say that, as to "the actual 
circumstantial need of these converts to Buddhism for special treatment toward their educational 
and economic advancement . . . the issue is one of pure fact and would be whether they are 
sufficiently backward to be eligible for special assistance as some of the weaker sections of the 
people." That is, they would be treated as backward classes but not as Scheduled Castes. 
168. RCSCST 1957­58: I, 25; II, 60. 
­320­ 
states) have remained adamant in their unwillingness to include Buddhists on a parity with 
Scheduled Castes. The major benefits outside the purview of the state governments are reservations 
in legislative bodies, post­matriculation scholarships, and reservations in central government 
employment. Since most Buddhists oppose the continuation of legislative reservations, the dispute 
has centered on the latter benefits.  169.   After several years of vacillation the central government in 
1969 acceded to Buddhist demands in regard to post­matriculation scholarships, leaving the other 
practices unchanged.  170.   When in 1976 the Scheduled Castes Order was amended by Parliament, 
Buddhists were not included.  171.   
It is clear that although Buddhists are not members of a "caste" in the sense of a component of the 
Hindu sacral order, they constitute a "caste" or community in the minimal sense of an identifiable 
group who share religious tenets and social practices and who marry and socialize chiefly among 
themselves. If this community can be shown to comprise persons who were subject historically to 
the deprivations suffered by the Scheduled Castes and who suffer at present from disabilities and 
invidious treatment, what distinguishes them from the Scheduled Castes for purposes of preferential 
treatment designed to overcome these disadvantages? (There is some evidence that they do suffer 
comparable 
____________________ 
169. Most Buddhist spokesmen contend only for economic, educational, and employment benefits. 
See Republican Party 1964: 20. However, Professor Bhandare ( 1965) argues that Buddhists are 
entitled to all benefits, including legislative reservations. Other Buddhist spokemen have 
expressed opposition to any reservations­­especially in government service because it removes 
educated members of the community from participation in political activity (interview with B. P. 
Maurya in Chicago, 28 June 1963). 
170. A bill that would have restored benefits to Buddhists was defeated in the Lok Sabha ( New York 
Times, Aug. 30, 1961, p.2, col. 6). Restoration of benefits was one of the demands in the great 
Buddhist agitation of 1964­65, which resulted in over 300,000 arrests ( Republican Party 1964; 
Khobrogade 1966). Subsequently, the central government appeared to move close to this 
position. In 1966, Mr. Y. B. Chavan, then defense minister, told a rally that the central 
government had accepted in principle the extension of concessions given to Scheduled Castes to 
neo­Buddhists also ( Indian Express, April 15, 1966). However, the Congress agreement in 
principle to drop the Hinduism test for Scheduled Castes was later explained as a proposed shift 
from caste to economic criteria to be implemented by 1981 ( Rudolph and Rudolph 1967: 149). 
Educational concessions at the college level were finally extended to Buddhists in November, 
1969 ( Phadnis 1969: 1897). 
171. The proponents of inclusion of the Buddhists took the tack of trying to incorporate the language 
of Explanation II (to Art. 25 of the Constitution) into the Scheduled Caste Order­­i.e., that "that 
reference to Hindu shall be construed as including a reference to persons professing Sikh, Jain or 
Buddhist religion" LSD (5th Series), Vol. 64. No. 17, cols. 147 ff. ( 2 Sep. 1976). Cf. Professor N. 
M. Kamble, RSD, Vol. 97, No. 18, cols. 17 ff. ( 3 Sep. 1976). 
­321­ 
disabilities and that the condition of most converts has not changed appreciably.)  172.   In effect, their 
disqualification lies in their failure to acknowledge their low ritual status by the acceptance of caste 
identity. Such acceptance and acknowledgment is made a condition of receiving benefits designed to 
dispel their depressed social and economic conditions. Exclusion of Buddhists, then, involves a 
curious transposition of ends and means. Historically, low ritual status was employed to identify the 
recipients of preferences directed toward secular betterment because of a presumed correlation 
between ritual impurity and depressed socio­economic conditions.  173.   Here, evidence of the 
existence of those conditions which preferences are designed to dispel is never considered. 
Acknowledgment of low ritual status is transformed from an instrument for identifying the 
deserving into an independent prescriptive requirement. As a result, the preference policy is 
transformed into an instrument for extracting such acknowledgment. 
The argument has been advanced that existing constitutional provisions foreclose extension of 
preferences to the Buddhists and that an amendment would be required to empower the government 
to do this. But Articles 15(4) and 16(4) clearly contemplate a plurality of backward classes. 174. 
There would seem to be nothing to prevent the government from denominating the Buddhists as 
Backward Classes and conferring upon them preferential treatment in education, government 
employment, distribution of land, etc., at the same level that it is conferred upon the Scheduled 
Castes­­as Maharashtra has done. 
Since legislative reservations alone are constitutionally specified exclusively for the Scheduled 
Castes, the inclusion of Buddhists would depend upon whether they could be encompassed within 
the Scheduled Caste category. The history of the Scheduled Caste classification militates against the 
notion that the present definition of Scheduled Castes represents the farthest constitutional limit of 
that category. The first Scheduled Castes Order of 1936 did not contain a Hinduism test (although 
Christians were excluded).  175.   The Hinduism test was not specified in the Constitution.  176.   Its 
subsequent formulation can 
____________________ 
172. On the continuing disabilities experienced by Mahars who have become Buddhists, see Karve 
1968: 97. Nair ( 1962: 187) reports of Sewagram ( Gandhi's ashram and model village), where 
the Mahars have become Buddhists: "There has been no improvement in the status of the 
Harijans since they embraced a new religion. They continue to be treated as untouchables." 
173. Dushkin 1961: 1667. 
174. See chap. 13, §§C, D. 
175. See n. 89, above. 
176. At one point in the drafting of the Constitution, Mr. K. M. Munshi moved an amendment which 
described the Scheduled Castes as "a section of the Hindu community" and which was accepted 
but subsequently omitted in accord with the agreement to include some Sikhs among the 
Scheduled Castes ( Shiva Rao 1968: 760). 
­322­ 
hardly be considered an unalterable expression of the constitutional mandate.  177.   It was certainly 
not so regarded in 1956 when the government expanded the requirement to include Sikhs.  178.   The 
power of Parliament is presumably wide enough to designate as Scheduled Castes any "castes" or 
parts of castes. It is not clear that this is confined to "caste" in the sense of a group that is a ranked 
component of the Hindu sacral order or whether "caste" in the sense of community would suffice. 
The argument for the more restrictive interpretation is that the classification is based upon 
disabilities consequent upon the occupation of a low position in that sacral order and applies only to 
those within it. But if the group is identifiable as one whose present disabilities stem from former 
membership in that order, the welfare objectives of the preference policy would seem to recommend 
that "caste" be interpreted widely enough to include all who had been victimized by that order. 
____________________ 
177. The Order was issued in August 1950 (P. T. I. Dispatch from New Delhi dated August 16, 1950). 
178. See n. 90 above on the process of including Sikhs. The formidable difficulties which their 
inclusion poses to the argument against inclusion of the Buddhists is nicely displayed in the 
following exchange from the Rajya Sabha debate over the 1976 revision of the Scheduled Castes 
and Scheduled Tribes Order: 

Shri Shyam Lal Yadav ( Uttar Pradesh): Sikhs are also a different community, a 
different religion. Why the Sikhs are included? 

Shri K. Brahmananda Reddi [ Home Minister]: They were included long ago. 

Shri Shyam Lal Yadav: . . . Sikhism is also a different religion than Hinduism as 
Buddhism is a different religion. 

Shri K. Brahmananda Reddi: Sikh religion is treated as coming under Hinduism. 

Shri Shvam Lal Yadav: How? Sikhism is quite different. 

The Vice­Chairman ( Shri Lokanath Misra): You cannot expect the Hon. Minister to 
go into the philosophy of Hinduism at this stage. 

Shri Kalp Nath Rai ( Uttar Pradesh): I want to know whether Hinduism includes 
Jainism, Buddhism and Sikhism. . . . 

(Interruptions) 

Shri K. Brahmananda Reddi: I have just now read what it means. 

The Vice­Chairman ( Shri Lokanath Misra): It is because of the Presidential Order 
that the effect has been given and, therefore, that is the finality in the matter. 

Shri Shyam Lal Yadav: They can change it. 

Shri K. Brahmananda Reddi: Whether Parliament can change it or not, is a different 
issue altogether. . . . 

Prof. N. M. Kamble: . . . Now, the quesion is whether the Buddhists are governed by 
the Hindu Law. Hon. Minister knows it. 

Shri K. Brahmananda Reddi: I am not going into that question. I have just now 
answered on the basis of information that we have that on consideration these 
Buddhists do not form a part of them so as to be entitled to be considered as 
Scheduled Castes. 

Shri K. K. Madhavan: What about Sikhism? 

(Interruptions) 

The Vice­Chairman ( Shri Lokanath Misra): Order, order, please sit down. RSD Vol. 
97, No. 18, cols. 49­51 ( 3 Sep. 1976). 

­323­ 
The exclusion of non­Hindus from preferences appears to give expression to a view of caste that is 
at variance with the constitutional and statutory "dis­establishment" of the sacral concept of caste 
since Indian independence. Varna distinctions have been abolished in the personal law; 179. 
untouchability has been abolished; imposition of caste disabilities have been prohibited;  180.   all 
support for claims of precedence based on caste standing has been withdrawn.  181.   Caste groups are, 
of course, free to entertain any views they wish concerning their station within (or without) Hindu 
society. But they may no longer call on the government or the courts to confirm and enforce these 
views. The law recognizes caste groups as corporate entities, bound by manifold ties of association. 
The nature of these ties, religious and social, varies from group to group in accordance with its own 
internal order. But the Constitution bars governmental recognition of hierarchical ordering, the 
apportioning of sacramental honor among caste groups, or the imposition of religious tests for 
recognition of a group as a caste.  182.   
Since preferences for Scheduled Castes are meant to alleviate the disadvantages historically 
associated with low standing in the Hindu sacral order, an attenuated recognition of this hierarchical 
order is utilized for the purpose of deciding who is to receive these preferences. Scheduled Castes 
are indentified in part by considerations of varna, pollution, and ritual distance.  183.   But it is clear 
that these otherwise discredited legal categories are appropriate only to the extent that they serve to 
identify the needy and deserving. They are appropriate only as convenient descriptions of practical 
conditions, not as prescriptive requirements. 
Conversion, both to sects within Hinduism and to religions outside Hinduism, has traditionally been 
and continues to be one of the common expedients of those at the bottom of the caste hierarchy 
seeking to improve their position.  184.   Historically, the disadvantages associated 
____________________ 
179. Hindu Marriages Validity Act, 1949 ( XII of 1949) and the Hindu Code Acts, n. 71, above. 
180. Art. 17, Art. 15(2), Untouchability (Offenses) Act, 1955 ( XXII of 1955) and kindred state 
legislation. On the operation of these prohibitions, see Galanter 1972. 
181. See Galanter 1968. 
182. It is open to a caste group itself to impose religious requirements for membership. See Devchand 
Totaram v. Ghaneshyam, A. I. R. 1935 Bom. 361. 
183. In fact, attempts to identify the Scheduled Castes in terms of criteria of pollution, impurity, and 
disabilities proved inadequate to isolate the groups which local administrators felt deserving of 
inclusion. Additional criteria of poverty and illiteracy had to be added. Thus the Scheduled 
Castes were determined by an uneven mixture of ritual and socio­economic criteria. See chap. 5, 
above. 
184. Titus 1959; Pickett 1933; Hazari 1951; Keer 1962; Isaacs 1965. 
­324­ 
with their low position have often followed converts across religious line; they continue to do so in 
at least some cases in the present  185.   ­­ a fact not unknown to the government.  186.   How effective or 
ineffective a means of escape conversion has proved to be in a particular case should be evaluated in 
empirical terms. To deem conversion to non­Hindu communions an automatic disqualification for 
aid violates the constitutional command of equal treatment for different religions. It also restricts 
freedom of religion,  187.   which might be thought to require that government refrain from 
administering its welfare schemes so as to put a heavy price­tag on its exercise.  188.   And, apart from 
its dubious constitutionality, it is at variance with the stated welfare objectives of the policy of 
preferences for the disadvantaged.  189.   
The Hindu requirement seems to reflect a hostility toward conversions which is anachronistic. 190. 
Conversions do not carry the same import of political and national division that they did in pre­ 
____________________ 
185. For some recent and pungent observations on the failure of conversion per se to alter the 
condition of low caste converts, see Andhra Pradesh Backward Classes Commission 1970: 
202­3. 
186. Before the Buddhist movement was on the scene, Home Ministry spokesmen were fully 
cognizant of the continuing incidence of disabilities on members of nonHindu religions. Thus in 
the Rajya Sabha debate on the scope of the Untouchability (Offences) Act, Mr. B. N. Datar 
observed that untouchability is carried across religious lines and pursues converts ( Times of 
India, Sep. 19, 1954). Cf. the assurances to Parliament of the then Home Minister, Pandit Pant, 
that the Untouchability (Offences) Act would apply among Christians and Muslims as well as to 
Hindus (quoted in Banerjee 1960: 100). 
187. Apparently, the Hinduism requirement has curtailed conversion; Lelah Dushkin ( 1957: 97) 
reports that "the fact that it was only the Hindu lowest castes that were included (in the 
Scheduled Castes lists) virtually stopped the movement toward conversion to other religions, in 
most areas." 
188. Cf. the American case of Sherbert v. Verner, 374 U.S. 398 ( 1963). 
189. In Ganpat v. Presiding Officer, A.I.R. 1975 S.C. 420, a two­judge bench of the Supreme Court 
advances a curious argument that the religious test impedes the perpetuation of preferences. 
"The attempt of persons that have changed their religion from Hinduism to Buddhism" to retain 
benefits as Scheduled Castes "only shows that otherwise these persons might get a vested 
interest in continuing to be members of the Scheduled Castes" (id., at 425). The religious test of 
the Scheduled Castes Order is said to be salutory "from the point of view of discouraging that 
tendency" (id.). Of course anything that cuts off benefits can be said to prevent those excluded 
from developing a vested interest in them. But if denial of benefits to non­Hindus may 
discourage conversion, it is difficult to see how it decreases the urge to retain benefits (except in 
circles where they might be a stigma of rejected Hindu identity). Indeed, it is instances of the 
non­occurrence of such an effect that the court cites as evidence that it obtains. 
190. Cf. the disabilities imposed upon converts from Hinduism in the various Hindu code acts. See, 
e.g., Sundarambal v. Suppiah, A. I. R. 1963 Mad. 260. Derrett ( 1958: 393) remarks that it is 
"strangely inconsistent with the claim to be a secular state" that a Hindu who changes his 
religion is liable to be divorced by his wife, may forfeit an 
­325­ 
Independence India,  191.   and anxiety inspired by the threat of separate electorates is similarly out of 
date. The test is certainly not designed to bolster Hindu orthodoxy, since one can remain a Hindu 
while embracing the most heterodox beliefs and practices. Not only is the test inconsistent with the 
ban on religious discrimination but, ironically, it seems to impose on Hinduism the notion of a hard 
and fast line between creeds and communions. This is not a characteristically Hindu notion, but one 
more consonant with the exclusivist creeds of the West, which require the convert to abjure his 
previous faith. It is rather surprising that Indian jurisprudence should give currency to this notion 
rather than to views more reflective of the Indian tradition of religious tolerance.  192.   
Preferences are designed to remedy certain conditions ­­ particularly to offset or dispel the 
disadvantages created by the imposition of social disabilities and the lack of economic and 
educational resources. The usefulness of the theory that converts are casteless in describing these 
conditions is questionable. Where the preference is one that devolves on the members of the group 
as individuals, surely the appropriate question is whether the disabilities and disadvantages have, in 
the particular case, been effectively dispelled by the conversion. Where the preference is a political 
one that devolves on the group corporately, the appropriate question would be whether the group 
still accepts the convert as a member and is willing to have him as a representative. In neither 
instance would it be necessary to decide whether he is a member of the group for all purposes or 
whether he is a Hindu. 
F. GROUP INTEGRITY: ADMISSION, EXPULSION, AND RE­ADMITTANCE 
In the cases discussed earlier, where the tests of affiliation (i.e., individual intention and conduct) 
and group acceptance were em­ 
____________________ 
existing claim to maintenance, may lose the right to give his child in adoption, may lose the right 
to be the guardian of his own issue, and may have his issue deprived of the right of inheritance 
from unconverted relations. See also D. Smith 1963, chap. 6. 
191. It should be recalled that the "Hinduism" requirement was formulated in the aftermath of the 
controversy produced by Dr. Ambedkar's 1936 threats that untouchables would convert from 
Hinduism to an unspecified religion. Hindu leaders were particularly eager to get assurances that 
it would not be Islam. There was even talk of assurances that if it were Sikhism the converts 
might retain the benefits accorded under the Government of India Act, 1935. See Zelliott 1969: 
214 ff.; Keer 1962: 277ff; Anonymous 1935­37. 
192. For an example of judicial expression of Hindu cosmopolitanism, see Kolandei v. Gnanavarum, 
A. I. R. 1944 Mad. 156, where a gift by the manager of a Hindu joint family to a Roman Catholic 
church was upheld as binding on his heirs, since "the very idea 
­326­ 
ployed to decide questions of group membership, their conjunction was sufficient to confirm a 
previously existing membership.  193.   But can affiliation and acceptance operate to create a new 
membership? or to restore membership that has been lost? 
The latter problem has arisen as an extension of the loss­of­caste­byconversion theory employed in 
Rajagopal v. Armugam.  194.   After his youthful conversion to Christianity, candidate Rajagopal had 
effected a reconversion to Hinduism by a series of acts "clearly amount[ing] to a public declaration 
of his professing the Hindu faith."  195.   In line with the requirements of the Meshram case, he had 
made "a complete public declaration" in such a way that the change would be known to those whom 
it might interest.  196.   Therefore, at the time of the contested election, he indisputably met the 
"Hinduism" requirement of the Scheduled Castes Order. 
However, since conversion to Christianity had divested him of caste membership,  197.   the question 
arises whether upon reconversion he had again become a member of the Adi Dravida caste and thus 
regained his eligibility to stand for a reserved seat.  198.   To answer this, the Court 
____________________ 
that pious and charitable purposes should . . . mean objects that can be said to promote a 
particular faith, creed, or dogma is foreign" to the ideals of Hinduism (id., at 157). Cf. State of 
Bombay v. Yagna Sastri Purushadasji, 61 Bom. L. Reporter 700 (1960), where admission of 
non­Hindus to membership was held not incompatible with a sect being "Hindu." While such 
cosmopolitan tolerance should not be imposed on groups which find it uncongenial, certainly the 
courts should not impose a narrower view on those groups which are willing to accept members 
of different religions. 
193. §§ B, C, above. 
194. A. I. R. 1969 S. C. 101. 
195. Id., at 107. The acts of the candidate included marriage to a Hindu woman, bringing up his 
children as Hindus, entering them as Hindus on school records, applying for alteration of his 
service record to show he was a Hindu. 
196. See text at n. 106, above. 
197. Without the interposition of the loss­of­caste­through­conversion theory, the question can be 
addressed in straightforward factual terms. In a similar case where a candidate had converted to 
Christianity for purposes of marriage, and later reconverted, a single judge of the Andhra 
Pradesh High Court, applying the Jasani tests, concluded: "He continued to retain his caste, 
though he lost his religion. . . . After his reconversion to Hinduism . . . he regained his religion 
and became a full­fledged 'bindla' . . . qualified to stand for election [from a reserved 
constituency]" ( Narasimha Reddy v. Bhupathi, 31 E. L. R. 211, 223 [1967]). 
198. This two­level inquiry into (1) reconversion and (2) caste membership was unknown to earlier 
administrative policy. For purposes of eligibility for preferences, reconverts to Hinduism who 
were born in Scheduled Castes have been deemed members of the Scheduled Castes ( RCSCST 
1953: 132). See also the statement of the Union Home Minister, reported in the Times of India, 
Sept. 12, 1959. However, a draft directive suggests that the Rajagopal approach has been 
accepted for administrative purposes: "a convert or reconvert to Hinduism or Sikhism may be 
accepted as a member of the Scheduled Castes if he has been received back and accepted as a 
member by the concerned Scheduled Caste" ( CWSCSC [4th] 2: 59 [ 1969]). 
­327­ 
turns to a doctrine developed in the application of the personal law that reconversion might be 
effectuated by acceptance by a Hindu community upon undergoing whatever formalities it saw fit to 
impose­­even none at all.  199.   In these cases, acceptance by the caste was employed as evidence of 
successful reconversion. In Rajagopal, however, the reconversion is conceded: the question is 
whether he regained his caste membership. The old cases' test for reconversion is transformed into a 
test for caste membership. Assuming, without deciding, that re­entry into a caste is possible, the 
Court finds no evidence that it had occurred. If re­entry were possible, it would be by virtue of 
acceptance by the caste. 
In matters affecting the well­being or composition of the caste, the caste itself is the 
supreme judge. It was on this principle that a reconvert to Hinduism could become a 
member of the caste, if the caste itself as the supreme judge accepted him as a full 
member of it.  200.   
Here, the Court found no evidence to satisfy this requirement. The various acts of the candidate that 
amounted to his public declaration that he was a Hindu do not show that the Adi Dravidas re­
admitted him or started to recognize him as one of them. Since his marriage (to an Adi Dravida 
woman) "was not performed according to the rites observed by members of that caste,"  201.   it was no 
proof of admission into the caste by "members of the caste in general."  202.   There was no evidence 
that at any stage "any step was taken by members of the caste indicating that . . . [the candidate] was 
being accepted as a member of this caste."  203.   Acceptance by his immediate family and a few 
relatives "cannot be held to be equivalent to a recognition by the members of the caste in general." 
 204.   
This opinion raises many puzzling questions. By "members of the caste in general" does the Court 
refer to the endogamous jati (kula) to which the convert previously belonged? or to all who call 
themselves Adi Dravidas ­­ a numerous group spread over much of South 
____________________ 
199. No proof of formal abandonment of his new religion is necessary for the convert to effect a 
successful reconversion to Hinduism. While a mere declaration is not sufficient to restore him to 
Hindusim ( Ramayya v. Josephine Elizabeth, A. I. R. 1937 Mad. 172), acceptance by a Hindu 
community with whatever formalities it deems proper­even none at all­­is sufficient. 
Durgaprasada Rao v. Sundarsanaswami, A. I. R. 1940 Mad. 513; Gurusami Nadar v. Irulappa 
Konar, A.I.R. 1934 Mad.630. However, cf. Marthamma v. Munuswami, A.I.R. 1951 Mad. 888 at 
890, where the primary test is the "intention" of the reconvert; the court says "the religious 
persuasion of a man now­a­days depends on his 'subjective preference' for any religion." Cf. the 
declaration test discussed in n. 100 above. 
200. A. I. R. 1969 S. C. at 109. 
201. Id., at 110. 
202. Id., at 110. 
203. Id., at 110. 
204. Id. 
­328­ 
India?  205.   or to the Adi Dravidas of the constituency in question? or to some other group larger than 
his endogamous jati but smaller than all Adi Dravidas? The Court does not specify what it would 
consider adequate evidence of acceptance: would a pronouncement by caste dignitaries or a 
resolution of a caste association suffice? Nor is there any indication that the caste (whatever unit is 
involved) possesses any organs through which it might express itself on the question.  206.   Unlike the 
small communities involved in the personal law reconversion cases, Adi Dravidas are a widespread 
composite group. Whatever adequate evidence of acceptance might be, it is clear that the Court 
places on the "re­convert" the burden of proving it. Since the Supreme Court had earlier ruled that 
the burden of proof rests on the challenger to disprove a candidate's claimed membership in a group, 
 207.   it appears that conversion, even when followed by reconversion, shifts the burden of proof to the 
candidate. 
Rajagopal did not tarry long in discharging this burden. Three months after his appeal was 
dismissed by the Supreme Court, a conference was held under the auspices of the Kolar Gold Fields 
Adi Dravida Scheduled Castes Conference. This conference, which Noranha, J., later referred to as 
Rajagopal's "magnum opus,"  208.   attracted at least 
____________________ 
205. According to the 1961 Census Adi Dravida is one of several names used by parts of the vast 
Parayan cluster of Madras (others are Paraiyan, Samban, Panchama, Vettiyan), whose total 
population in 1961 was 3.4 millions, over 56% of the entire Scheduled Caste population of 
Madras ( Census of India 1961, Part IX Madras, Part V A, pp. 13, 19). There were also 37,439 
Adi Dravidas listed in Andhra. The term "Adi Dravida" (lit., original Dravidian) was adopted by 
many Tamil untouchable castes in the 1900s. The census return under this heading increased 
from nil in 1911, to 49,978 in 1921 to 1,546,520 in 1931, when it was an accepted census 
classification. In 1922 the Madras Government undertook to adopt the terms Adi Dravida and 
Adi Andhra for future record­keeping purposes. ( G. O. No. 817 (Law General), dated 35th 
March 1922, reprinted in People's Herald 2( 3):5 ( July 15, 1969). As the name suggests, its use 
implied that this group were the original inhabitants of South India who were subsequently 
reduced to their present low estate by Aryan invaders. For an exposition, see Rajah 1925, chap. 
4. Thus, Adi Dravidas are a composite group, the components of which retain their 
distinctiveness to some extent. Gough ( 1960: 17, 244) describes Adi Dravidas as not "a single 
organized group" but "rather a named caste category of similar culture and occupation 
comprising a number of endogamous groups each in turn divided into local communities." In 
the village she studied these groups "retained separate administrative organization for each sub­
caste community." Similarly Bèteille ( 1965b: 92) observes that although higher castes and the 
government often treat Adi Dravidas as a single unit, "there are divisions and subdivisions 
among them." 
206. Indeed, there is some question whether a composite group like Adi Dravidas may exercise the 
traditional prerogatives of a caste. Cf. Ellappa v. Ellappa, A. I. R. 1950 Mad. 409. 
207. Laxman Siddappa Naik v. Kattimani Chandappa Jampanna, A. I. R. 1968 S. C. 929 (discussed 
in § A, above). 
208. Rajagopal v. Armugam [II] 1973(2) Mys. L. J. 240 at 250. 
­329­ 
a thousand persons, mostly Adi Dravida Hindus, and addressed various questions of interest to 
Scheduled Castes. At the conference 
a religious ceremony was performed . . .it was a "Shuddh" or cleaning ceremony to 
remove the stigma of [ Rajagopal's] being considered as a Christian pursuant to the 
judgment of [the Mysore High] Court and the Supreme Court and their publication in 
some newspapers and it was not a ceremony to convert him, already a Hindu, into 
Hinduism.  209.   
Accordingly, the conference resolved to 
set right the doubt arising in [ Rajagopal's] election petition, recognize and also admit 
him to our Caste, Sri S. Rajagopal, who was born in our Caste, brought up amidst us, 
and who represented our Caste in Mysore Legislature for the past 6­­7 years, serving 
efficiently and honestly.  210.   
As the 1972 general elections approached, Rajagopal and Arumugham squared off for another 
electoral contest and another round of litigation (which we shall call Rajagopal II). Arumugham 
objected to Rajagopal's nomination papers on the ground that upon his reconversion, Rajagopal had 
not re­obtained his membership in the caste. The Returning Officer rejected Rajagopal's papers. 
Arumugham was duly elected and Rajagopal challenged the election. The Returning Officer's 
rejection, according to the High Court, was "hasty, unwarranted and erroneous."  211   The Adi 
Dravida conference was "Vox Populi. . . abundant proof that he belongs to the mainstream of that 
community."  212.   He showed that he was a member of the caste "without a semblance of reasonable 
doubt."  213.   
On appeal the Supreme Court seemed about to reopen the fundamental question of the relation of 
religion to caste. Aware that the correspondence of caste and Hinduism is imperfect, it observes that 
caste identity among Christians is "not an infrequent phenomenon in South India."  214.   Whether 
converts to Christianity remain members of a caste "ultimately . . . must depend on the structure of 
the caste and 
____________________ 
209. Id., at 251. 
210. Id., at 253. 
211.Id., at 255. 
212. Id., at 253. Arumugam argued that the conference was an improper body for such a 
manifestation of caste assent and that the only legitimate arbiters of acceptance into the Adi 
Dravida fold in the Kolar Gold Fields were the Nattamdars [headmen] of each block in the 
mining area and the Adi Dravida Jana Sabha. After inspection of the evidence the court 
concludes that "the institution of Nattamdars and the existence of the Adi Dravida Jana Sabha in 
KGF can be dismissed as belated inventions tot the purposes of defense" (id., at 254). (The court 
observes that panchayats sponsored by the mine management and "unrecognized panchayats 
formerly functioned in the blocks but are now reported to be defunct [id., at 254­55.]) 
213. Id., at 255. 
214. Arumugam v. Rajgopal [II] A. I. R. 1976 S. C. 939 at 945. 
­330­ 
its rules and regulations."  215.   Does the caste expel converts or does it tolerate them within its fold? 
The Court suggests that such toleration seems to prevail among Adi Dravidas. There are both 
Hindus and Christians, and intermarriages between them, so it seems prima facie that on conversion 
to Christianity Rajagopal "did not cease to belong to the Adi Dravida caste."  216.   This line of 
argument, which appears to undermine the "loss of caste" notion, would bring the Court squarely to 
the question of the constitutionality of the Hinduism test of the Scheduled Caste Order. But the 
Court veers away, saying it need give no final opinion on whether he remained in the caste, since it 
finds he was not only reconverted but re­admitted. 
Taking the other path, the Court characterizes a caste as a "social combination of persons governed 
by its rules and regulations," which need not be written or formalized.  217.   The caste may, "if its 
rules and regulations so provide, admit new members just as it may expel an existing member." It 
may specify particular rites or ceremonies for re­admission, but if it does not, then "the only thing 
necessary for readmission would be the acceptance of the person . . . by the other members of the 
caste."  218.   The question then is whether the Adi Dravidas accepted Rajagopal after his reconversion. 
Of the battery of evidence he had assembled, the Court found his participation in Hindu weddings 
and funerals unpersuasive. Nor was it impressed by the fact that he was looked upon as a leader and 
peacemaker by the Adi Dravidas of the locality. It was most impressed by the 1968 conference, 
which had been held with the object, inter alia, of 
readmitting [Rajagopal] . . . into the fold of Adi Dravida caste and not only was a 
purificatory ceremony performed . . . with a view to clearing the doubt which had been 
cast upon his membership . . . by the decision of this court [in Rajagopal I] but an 
address . . . was also presented to [Rajagopal] . . . felicitating him on this occasion.  219.   
The Court also found persuasive his laying the foundation stone for constructions of a new wall of 
the Hindu temple, and his honorific participation in temple ceremonies was "strongly indicative of 
the fact that he was accepted and treated as a member of the Adi Dravida community."  220.   
Yet the Court's estimation of this array of evidence is puzzling. Recall that it was earlier established 
that Rajagopal had reconverted to Hinduism and indisputably met the Hinduism requirement of the 
____________________ 
215. Id., at 945. 
216. Id., at 947. 
217. Id., at 948. 
218. Id., at 948. 
219. Id., at 951. 
220. Id., at 950. 
­331­ 
Scheduled Caste Order; the only question was of his caste membership. But since the Court 
indicated its awareness that Christians were accepted as members of the Adi Dravida caste, it is not 
clear what remained to be proved with this evidence. Its emphasis on Hindu ceremony suggests that 
the Court was interested in caste acceptance of him as a Hindu ­­ for if he were accepted as a 
member of the caste but not viewed as a Hindu, the problem of the Hinduism requirement would be 
unavoidable. (Of course, he might be deemed a Hindu for purposes of the Scheduled Caste Order 
even if he was not recognized as one by his former caste people. Judicial deference to caste 
autonomy on the question of caste membership has not been extended to the question of whether 
someone is a Hindu.) The tendency to fuse the caste membership question with that of Hinduism 
need not derive from such deliberate avoidance of the constitutional question; more likely it flows 
from the same view of caste that underlines the "loss of caste" theory. 
The Court not only avoids the question of the constitutionality of the Hinduism requirement, it 
supplies a rationale for that test which elaborates on the "loss of caste" theory. It is because 
orthodox Hindu society . . . attaches social and economic disabilities to a person belong 
to a Scheduled Caste . . . that . . . favorable treatment is given to him by the Constitution. 
Once such a person ceases to be a Hindu and becomes a Christian, the social and 
economic disabilities arising because of Hindu religion cease and hence it is no longer 
necessary to give him protection and for this reason he is deemed not to belong to a 
Scheduled Caste. But when he is reconverted to Hinduism the social and economic 
disabilities once again revive and become attached to him because there are disabilities 
inflicted by Hinduism. A Mahar or a Koli or a Mala would not be recognized as 
anything but a Mahar or a Koli or a Mala after reconversion to Hinduism and he would 
suffer from the same social and economic disabilities from which he suffered before he 
was converted to another religion.  221.   
In this version, it is not that caste is lost by conversion but that the disabilities attendant upon it are 
dissipated. This may be as much a fiction as is the "loss of caste" theory. But merely to set it out 
raises some puzzling questions about the Court's combination of caste autonomy doctrines with 
preference policy. For if disabilities impinge because of the response of "Hinduism," then why 
should the revival of protections be dependent on acceptance by the caste group? 
The caste autonomy doctrine serves nicely where individual intention and conduct and group 
acceptance are reinforcing. But suppose 
____________________ 
221. Id., at 949. 
­332­ 
they diverge? Should group acceptance outweigh intention and conduct? In the Jasani case the test 
for retention of membership in the privileged group was primarily one of the convert's intention. 
However, it was necessary that this intention be confirmed by acceptance by the old group. The 
inclusion of this additional test is important because, as the Court says, 
the only modification here is that it is not only his choice which must be taken into 
account, but also the views of the body whose religious tenets he has renounced, 
because here the right we are considering is the right of the old body, the right to send 
one of its own fold to Parliament.  222.   
Since the general electorate choose from among the candidates standing for the reserved seat, the 
only way that representation of the privileged group can be assured is by seeing that only those 
accepted by it are permitted to stand for these seats. But since seats are reserved not for particular 
Scheduled Castes but for the Scheduled Castes as a whole, one might question the relevance of 
looking specifically to his own community, to see whether he was accepted. For example, if the 
Scheduled Castes comprised As, Bs, and Cs, and X's membership in the As is in doubt, whether X 
is a member of the Scheduled Castes might depend equally not only on whether the As accept him 
as an A but on whether the Bs and Cs do so as well or instead. 
Unlike legislative representation, such benefits as scholarships, housing, and agricultural loans do 
not devolve on the group in any corporate or collective capacity; membership in the group is merely 
a convenient device for identification of deserving beneficiaries. Thus, if the preference in question 
devolves not on the Scheduled Castes collectively but on their members individually, the rationale 
for giving greater weight to group acceptance would be absent. The primary test of the intention and 
conduct of the individual would apparently be conclusive, and it would not be necessary to confirm 
this by acceptability to the group. 
Much the same may be said of the situation in which affiliation and group acceptance are most 
directly in conflict ­­ that of expulsion or excommunication. What effect should expulsion or 
excommunication from the group have on eligibility for preferential treatment? The Constitution 
leaves unchanged the power of a caste group to excommunicate those who offend it.  223.   In religious 
denominations ­­ and 
____________________ 
222

(1959) S. C. R. at 839. 
223Section 9 of the Civil Procedure Code forbids the courts to take cognizance of "caste questions." 
. It is still a good defense to a criminal action for defamation to assert the privilege of 
communicating news of an excommunicaton to one's caste fellows ( Panduram v. Biswambar, A. 
I. R., 1958 Or. 259; Varadiah v. Parthasarathy, I. L. R. 1964 [2] Mad. 417. Cf. Hadibandhu v. 
Banamali, A. I. R. 1960 Or. 33). The power to excommuni­ 
­333­ 
these include at least some caste groups  224.   ­­this right of excommunication is constitutionally 
protected in so far as it is used for the enforcement of essential religious discipline.  225.   In spite of 
the language in Rajagopal that the caste is the "supreme judge" of its membership, it seems doubtful 
whether ostracism or excommunication would effectively deprive one of membership in the group 
for purpose of qualifying for benefits that were not directed to the group collectively. In such a case 
the relevant question would be whether the expulsion was associated with any conditions which 
effectively dispelled the disabilities or other backwardness which had caused the group's members 
to be singled out for preferential treatment. Presumably, if the individual still suffers these 
disadvantages he would remain a member of the privileged group in the sense of being a legitimate 
recipient of these preferences. So while the group's opinion of the membership of the individual 
should be decisive in the case of preferences that devolve on the group collectively, where the 
preferences are distributive (i.e., devolve on the members of the group individually), the State may 
legitimately intervene to protect the individual from group power.  226.   The appropriate question 
would be whether his condition has changed from that which caused members of the group to be 
given preferences. 
Is membership merely a convenient device for identifying deserving beneficiaries? It might be 
argued that preferences are directed to members of particular groups not only to benefit the 
individual members directly, but to utilize the ties of kinship, loyalty, and cooperation among them 
to share and multiply these benefits. But while group cohesion is an instrument of welfare policy, it 
is not necessarily an objective of that policy or a necessary condition for its effectiveness. 
Preferences which include among their objectives the promotion of the cohesion and integrity of the 
group must be distinguished from those 
____________________ 
cate is subject to two important restrictions. It is an offense to undertake any disciplinary action 
directed toward the enforcement of "untouchability" (Untouchability Offenses Act, §7). It is a 
corrupt election practice to threaten a candidate or elector with excommunication 
(Representation of the People Act, 1951, § 123 [2]). 
224. See e.g., Sri Venkataramana Devaru v. State of Mysore, A. I. R. 1958 S. C. 255. 
225. Saifuddin Saheb v. State of Bombay, A. I. R. 1962 S. C. 853. Some caste groups among the 
Scheduled Castes or Backward Classes might regard themselves as religious denominations and 
invoke their constitutionally protected right to excommunicate for the purpose of upholding 
religious discipline. However, since preferences to these groups are for the purpose of 
ameliorating their condition rather than promoting their religious unity, such expulsion should 
not outweigh the right of the expelled individual to share in these preferences ­­ at least so far as 
his enjoyment of them does not detract from the religious freedom of the group. 
226. This would seem analogous to the notion that excommunication is protected only so far as it 
withdraws religious rights and does not extend to deprivation of civil rights (except where 
consequent upon the former). See Saifuddin Saheb v. State of Bombay. A. I. R. 1962 S. C. 853. 
­334­ 
where group cohesion is used only as an instrument. It is clear that preferences for the Scheduled 
Castes and Backward Classes are for the purpose of ameliorating the deprivations of the lowly; they 
are not designed to promote the cohesion and integrity of the groups that make up these categories. 
On the other hand, the preferences for Scheduled Tribes include among their purposes the 
promotion of the cultural intergrity and autonomy of the tribal groups.  227.   While expulsion should 
not outweigh the right of the individual to enjoy preferences among the Scheduled Castes and 
Backward Classes, in the case of the Scheduled Tribes the group's view of the individual's 
membership might be given greater weight. 
If, as we have seen, individual affiliation and group acceptance may determine who remains a 
member or is restored to membership, the question arises whether affiliation and acceptance may 
determine who becomes a member. May someone who was not a putative member become a 
member of a Scheduled Tribe, a Scheduled Caste, or a caste or religious group listed as a Backward 
Class and receive the preferential treatment extended to its older members? 
In Jahan Ara we saw art instance of assimilation of an outsider into a Scheduled Tribe.  228.   In 
Rajagopal II the Supreme Court suggested that a caste might admit to its fold even those not born 
within it.  229.   Subsequently, a Constitution Bench of the Supreme Court faced the question in 
Guntur Medical College v. Mohan Rao.  230.   The aspiring medical Student was the son of Madiga 
parents who had converted to Christianity before he was born. Converts to Christianity from 
Scheduled Castes were treated by the State of Andhra Pradesh as members of the Backward Classes. 
When the applicant did not succeed in securing admission to medical college as a member of the 
Backward Classes, he underwent conversion to Hinduism and, claiming to be a Madiga, applied for 
admission to another medical college as a Scheduled Caste. His provisional selection was cancelled 
on the ground that he did not meet this state requirement: 
No candidate other than a Hindu including a Sikh can claim to belong to Scheduled 
Castes. No candidate can claim to belong to the Scheduled Caste except by birth.  231.   
____________________ 
227. See chap. 5, §C, above. 
228. See discussion in §C. 
229. See A. I. R. 1976 S. C. 939 at 948. 
230. A. I. R. 1976 S. C. 1904. 
231. The Principal of the Medical College had misinterpreted this to require that the applicant be a 
Hindu by birth, an interpretation which the Supreme Court deemed "plainly erroneous" before 
proceeding to hold the requirement void as properly interpreted. The grounds for this voiding 
seem to be that the power to define Scheduled Castes is exclusive with the President and 
Parliament, and it is not open to other authorities to apply more restrictive definitions. 
­335­ 
The Supreme Court found the "birth" requirement void as beyond the provisions of the Scheduled 
Castes Order. Hence, it was open to the applicant to show that he had become a Madiga upon his 
conversion to Hinduism. He claimed that his parent remained Madigas even after their conversion, 
so he was in fact a Madiga by birth ­­ a possiblity the Court is willing to countenance, since it says 
that loss of caste upon conversion is not an absolute rule, but is dependent upon "the structure of the 
caste and its rules and regulations."  232.   But the Court is equally willing to contemplate the 
possibility that even if his parents lost membership in the Madiga caste upon conversion, and the 
applicant was not a Madiga by birth, he might have become a Madiga. 
It is for members of the caste to decide whether or not to admit a person within the 
caste. . . . It may if its rules and regulations so provide, admit a new member just as it 
may expel an existing member. The only requirement for admission of a person as a 
member of the caste is the acceptance of the person by the other members of the caste. 
 233.   
Although there is no finding that that is the case here, and although it is said in reference to a child 
of former members of the caste, the Court does not limit this rule to those with a history of descent 
from the caste in question. Presumably, it is applicable to someone who is by family history a total 
stranger to the caste.  234.   
The rejection of the "birth" requirement raises the question of how preference policies will be 
protected against exploitative assertions of group identity. It is not surprising that there have been 
attempts to effect changes in personal status for the purpose of establishing eligibility for 
preferential treatment.  235.   Earlier we noted some instances of deceptive self­characterization by 
individuals and groups, often in marginal and ambiguous cases.  236.   Here, however, we are dealing 
with 
____________________ 
232. Id., at 1907. 
233. Id., at 1908. The Court gives no indication of whether the caste's "rules or regulations" allowing 
admission need be long established or whether freshly minted ones would suffice. 
234. Cf. the further developments noted below in text at notes 251 ff. 
235. Although documented cases of such deception are not plentiful, it is perceived in many quarters 
that such deception is widespread. Thus, a venerable member of the Rajya Sabha complained 
that "there are many IAS and IPS officers. They change their names or add some Scheduled 
Caste or Scheduled Tribe names and come into this list and get jobs thereby, cheating the 
Government." Specifically, he complains, in Andhra Christians with superior access to 
education, "by calling themselves Mala or Madiga or Chamar are getting all the facilities and 
they are coming into power. His father is a Christian; his grandfather is a Christian; his religion 
is Christianity; he goes to Church everyday; he marries a Christian girl. But for the purposes of 
these benefits, he calls himself a Scheduled Caste man. Many such cases are there. Everyday it 
is happening" (RSD Vol. 97, No. 18, col. 44 [3 Sep. 1976]). 
236. See §A, above. 
­336­ 
deliberate and explicit change in status. This rather drastic stratagem has been adopted in regard to 
those very scarce and critical opportunities, medical and professional school admissions. It seems to 
take two forms: (1) conversion to a religion whose members are entitled to reserved places (thus we 
have, in addition to reconversion to Hinduism, instances reported of conversions to Buddhism in 
Maharashtra and to Islam in Jammu and Kashmir);  237.   (2) high caste Hindus securing adoption in 
Scheduled Caste families.  238.   
In Ashok v. Dean, three higher caste students had embraced Buddhism and described themselves as 
Nava­Buddhas for purposes of securing admission to a medical college.  239.   (The state's list of Back 
ward Classes included Nava Buddhas, which it defined as "Scheduled Castes converted to 
Buddhism.") The court found them ineligible and ordered their admissions withdrawn on the 
ground that Nava­Buddhas meant Scheduled Castes converted to Buddhism "and not just any one 
belonging to any caste whatsoever who has embraced Buddhism."  240.   Here the court managed to 
resolve the matter handily, with an assist from the Government Order, but it might well have reached 
the same result unaided. 
The adoption stratagem has proved more difficult to resolve. In 1968 there were reports that students 
in Delhi and elsewhere in nothern India were securing admissions by having themselves adopted by 
____________________ 
237. In September 1966 it was reported ( organizer, Sept. 11, 1966) that eleven young Kashmiri 
Hindus had converted to Islam on the ground that their religion was an obstacle to obtaining 
technical education and suitable jobs. According to the Organizer ( Oct. 30, 1966), "Thousands 
of letters poured in on the converts from far and near. Arya Samaj, RSS, Dogra Sabha, Kashmiri 
Pandit Sabha leaders met the eleven misguided youths and impressed on them the impropriety of 
changing their religion for worldly considerations. Union Home Minister Nanda is . . . believed 
to have written to the Kashmir Government that the students should not be given admission just 
because they had changed their religion. After all this persuasion the eleven youngsters came 
back to the Hindu fold on Oct. 23. The shuddhi ceremony was performed in the Arva Samaj 
Mandir Srinigar." 
238. These are courses that go beyond the mere procurement and use of a false "caste certificate." 
Such certificates, attesting to membership in a Scheduled Caste or Scheduled Tribe, are issued 
by various officials­­upon production of a document purporting to be from a caste organization 
(attesting to the membership of the applicant) and a certificate from a special executive 
magistrate. The three metropolitan magistrates courts in Bombay, for example, are reported to 
issue "nearly 10,000 caste certificates every year [but] do not have the machinery to verify their 
authenticity" ( Times of India, 24 June 1977). There are numerous reports of the fraudulent use 
of such certificates to secure admissions and jobs, promotions, loans, and other benefits ­­ e.g., 
CWSCST (5th) 21: 33­35 ( 1973). The practice is parallel to the false procurement of income 
certificates discussed in §G, below. 
239. Ashok v. Dean, 69 Bom. L. Reporter 603 ( 1967). 
240. Id., at 612. 
­337­ 
Scheduled Castes.  241.   The Elayaperumal Committee reported that in Rajasthan, non­Scheduled 
Caste students had themselves adopted by Scheduled Caste persons, then obtained certificates from 
magistrates, on the basis of which they were admitted to engineering and medical colleges against 
seats reserved for Scheduled Castes. The Committee reported that in 1968­69, of the 28 Scheduled 
Castes' seats in medical colleges, 16 were occupied by non­Scheduled Caste candidates on adoption 
certificates.  242.   One response to these abuses is a "birth" test, an option now closed by the Supreme 
Court.  243.   What alternative means are there for checking exploitative acquisition of membership in 
beneficiary groups? In response to inquiries, the Department of Social Welfare provided "legal 
advice" notable for its formalism and lack of resourcefulness: 
[T]he legal status of a high caste child adopted by a Scheduled Caste family would make 
him eligible to treatment as a member of the adopters' caste if the adoption satisfied the 
conditions laid down in the Hindu Adoptions and Maintenance Act, 1956. In the case of 
Scheduled Tribes, however, the said enactment does not apply and the guiding factor in 
such cases will be whether the customs of the concerned tribe permit the adoption of a 
nontribal into the tribe and whether the particular adoption in question is covered by the 
customs.  244.   
This advice seems at variance with the Rajagopal notion that the caste is the "supreme judge" of its 
membership, which would seem to imply that group acceptance rather than the formalities of 
adoption law would be decisive. Castes could admit or confirm new members according to their 
lights. But while Rajagopal requires some group confirmation, the 
____________________ 
241. RCSCST 1967­68: 33; Dept. of Social Welfare 1969: 207. 
242. Department of Social Welfare 1969: 207. Most reports of fake adoptions concern professional 
admissions, but such practices are not unknown in regard to posts. See the report of a Brahmin 
who had himself adopted by a Scheduled Caste and appeared in the I.A.S. examination 
(RCSCST 1970­71: 127). 
243. E.g., the Ministry of Labor instructed employment exchanges to register as Scheduled Castes 
reconverts to Hinduism but not converts who were born in a different religion "whatever may be 
their family connections" (CWSCST [4th] 5: 27 [ 1970]). See, generally, the Circular sent by the 
Govenment of India to all state governments that they should admit to reserved seats in medical 
colleges only those "born in a Scheduled Tribe/Scheduled Caste family" (CWSCST [4th] 14: 19 [ 
1670]). 

These "birth" requirements have been around long before the recent concern with exploitative 
acquisition of membership in beneficiary groups. E.g., in 1953 the Government declared that 
upon conversion to Hinduism, those born in another religion (whose parents were converts) 
should not be treated as members of the Scheduled Castes "whatever may be their original 
family connections" (RCSCST 1953: 132). Reconverts born as Hindus, on the other hand, have 
been treated as Scheduled Castes. 

244. RCSCST 1967­1968: 33. 
­338­ 
departmental advice seems to accept what amounts to delegation to a single individual of power to 
confer membership ­­ often, one assumes, an individual who is consulting his own interest rather 
than group interests. 
Like Rajagopal, this advice reflects the heavy sway of the notion of unitary group membership, 
according to which an individual is a member of the group either for all purposes or for none at all. 
Such a notion is not only a deceptive guide to the social realities; it is an obstacle to rational policy­
making. Why should the group's decision for its purposes (much less an individual Scheduled Caste 
member's for his) necessarily bind the government in its administration of its policies of preferential 
treatment? The government should not be allowed to attach conditions to inclusion which are not 
relevant to the carrying out of these policies.  245.   But surely it may enquire whether inclusion of a 
new member would effectuate the particular policy involved. Where dealing with such collective 
preferences as political representation (or with Scheduled Tribes, where group autonomy is an aim 
of government policy) group acceptance might well prevail. On the other hand, in dealing with such 
benefits as scholarships, housing, etc., the government might well enquire whether the new member 
had such an affinity of conditions and resources that it would carry out the aims of the particular 
program to include him. 
The Mysore High Court employed such a policy analysis to fashion an incisive solution to the 
adoption problem in Shantha Kumar v. State of Mysore,  246.   a case which reminds us that 
exploitative manipulation is at least as possible with "neutral" categories as with conmmunal ones. 
Mysore had adopted an "income cum occupation test" for Backward Class status. Medical 
admissions were reserved for candidates whose parents were in manual occupations and had 
incomes below a specified amount.  247.   Shantha Kumar, the son of a supervisor in the National 
Extension Service, was given in adoption to a coolie (his mother's sister's husband?). The adoption 
supposedly took place three years earlier when the boy was sixteen, but the deed of adoption was 
executed just three months before the application to medical college. The Selection Committee 
refused to regard him as a member of the Backward 
____________________ 
245E.g., the "birth test" for reconverts described in n. 243 above. Since reconverts born as Hindus 
. are treated as Scheduled Castes, Hindu birth is made a condition for governmental recognition of 
membership in a Scheduled Caste. 
2461971 (1) Mys. L. J. 21. Another bench of the Mysore court came to a similar result in Nataraja v. 
. Selection Committee, 1972(I) Mys. L. J. 226 [NIC]. There a candidate for medical college, the 
son of a health inspector earning Rs. 2,000 per year, had been adopted by his mother's sister, a 
tailor with an annual income of Rs. 780. 
247. The Mysore income­cum­occupation test is disussed at chap. 8, §G, above, and in §G, below. 
­339­ 
Classes, and he petitioned the High Court to overturn that decision. The court deliberately 
suppressed its doubts about the genuineness of the adoption and decided to "proceed on the 
assumption that there was a genuine and valid adoption of the petitioner"  248.   at the age of sixteen. 
The court observed that the rationale for reservations for Backward Classes was to offset the 
unfavorable environmental conditions of such persons. But the applicant had not suffered any 
environmental disadvantage while living with his natural father for his first sixteen years, and the 
environmental conditions of the last three years "cannot be said to destroy or nullify the advantage" 
 249.   conferred by his earlier upbringing. In the case of a late adoptee who "had all the while imbibed 
the . . . environmental advantage of his natural father's income and occupation,"  250.   it is the income 
and occupation of the natural father rather than of the adoptive father that is relevant in determining 
eligibility for benefits for the Backward Classes. Otherwise, adoption could be used to divert 
benefits from the intended beneficiaries. 
This line of analysis produces a similar result even where the affiliation with a beneficiary group is 
in undoubted good faith. Thus in Urmila Ginda v. Union of India, a woman of high caste origin 
married to a Scheduled Caste man applied for a reserved post and claimed that by the Hindu theory 
of marriage she had become his Sapinda and caste fellow.  251.   The court put aside these 
"spiritual . . . considerations" and held that as one who was not personally handicapped she could 
not take advantage of the special provisions contemplated by Articles 15(4) and 16(4) "because she 
is not one of them ­­ she is a high caste Hindu who was not subject to any such backwardness either 
socially or educationally."  252.   To permit her to compete might defeat the purpose of such provisions 
by diverting benefits away from legitimate beneficiaries and open possibilities of abuse. The Court 
distinguishes Jahan Ara as a case of assimilation into tribal group.  253.   Urmila Ginda is devoid of 
any assertion of assimilation, affinity, of conditions, or acceptance by the husband's group. 
____________________ 
248. 1971 (1) Mys. L. J. at 22. 
249. Id., at 23. 
250. Id., at 23. 
251. A. I. R. 1975 Del. 115. During the Lok Sabha debate on the revision of the Scheduled Castes 
Order, one M.P. opined that it would be redundant to legislate Scheduled Caste status for 
intermarrying wives, since this would "happen automatically without anybody's intervention" 
because women in India took on their husband's social identities ( Shri G. Kuchelar, LSD [4th 
Series], Vol. 45, No. 9 col. 270 [ 20 Nov. 1970]). 
252. Id., at 116. An attempt to reverse this result by statute failed. See n. 52 above. 
253. See the discussion of Jahan Ara in §C, above. The court here implies that the words "tribal 
communities" in Art 340 provide more leeway for assessments of membership in cases of 
Scheduled Tribes than with Scheduled Castes (id., at 116). 
­340­ 
Subsequently, the Punjab and Haryana High Court encountered a case in which the missing element 
of caste acceptance was supplied. Not only had a medical college applicant been adopted by a 
member of the Scheduled Castes, but the Baradari [Assembly] of the caste had indicated its 
acceptance of him as a member. The court declined to read the Guntur Medical College case as 
extending the acceptance test to those with no historic connection with the community, for to do so 
would "open the floodgates for all sorts of undesirable practices to gain narrow ends."  254.   The 
floodgates argument is less persuasive in the Urmila Ginda situation than in the adoption case, since 
it is established public policy to foster intermarriages with Scheduled Castes.  255.   Although 
Government and Parliament declined to set a rule of automatic inclusion of wives,  256.   this does not 
imply that they are automatically excluded. Instead, it leaves a question of fact, to be decided in each 
instance according to the tests of intention, group acceptance, and the policy of the particular 
scheme of preference. One might imagine policies (like health care, drinking water, provision of 
agricultural loans) in which the inclusion of such wives would clearly be appropriate. 
G. NON­COMMUNAL CRITERIA: INCOME AND OCCUPATION TESTS 
Since membership problems of the kinds discussed above are a by­product of the use of communal 
units to identify the beneficiaries of preferences, it is sometimes anticipated that the replacement of 
com­ 
____________________ 
254. This case (Civil Writ Petn. No. 3970 of 1976) is described and all excerpt from the judgment 
reprinted at RCSCST 1975­1977: 19. 
255. For examples of various state programs of gifts and grants to couples entering such marriages, 
see RCSCST 1973­74: 184; cf. the account of the public reception ill Ahmedabad at which Prime 
Minister Desai presided over the honoring of eighteen couples entering mixed (Harijan­caste 
Hindu) marriages. The state gave each couple all award of Rs. 5,000 and undertook to give 
accorded various benefits to the offspring of these marriages ( Overseas Hindustan Times, 8 
June 1978). 
256. The Joint Parliamentary Committee set up in 1967 to revise the Scheduled Castes and Tribes 
Order recommended inclusion of wives. The provision in the 1970 bill: "Notwithstanding 
anything contained in any law . . . or in any custom on usage to the contrary, a woman who 
marries a person belonging to a caste specified in any part of the Schedule shall be deemed to 
belong to the Scheduled Caste to which her husband belongs," was greeted by a Government 
spokesman to be a "heartily welcome . . .encouragement for Savarna girls to marry Harijans" 
( Shri K. Hanumanthaiya, LSD [4th Series], Vol. 65, No. 6: 322 [17 Nov. 1970]). When the bill 
reappeared in 1976, this provision was removed from the Government version, and efforts to 
restore it from the floor were unsuccessful. 
­341­ 
munal criteria by economic and educational tests of backwardness would eliminate such disputes. 
But even criteria like income, occupation, and place of residence, which may seem subject to rather 
mechanical determination, raise a host of problems that are quite analogous to the membership 
disputes that we have reviewed. Movement to noncommunal criteria will change the nature of the 
disputes, but it will not eliminate disputes about who is included and who is not. 
Income tests immediately raise the question of whose income is be measured. Typically, the 
requirement is in terms of family income­­the income not of the applicant himself but of some 
larger unit. But what is that unit? those in certain biological relationships? those who are legally a 
joint family? those who share a common pocket? those who live under one roop In Jayaswal v. 
Principal,  257.   an applicant for medical college admission declined to supply information about his 
father's income and instead filed his own income certificate, showing that he was separated from his 
father as a result of a suit and that his annual income was below the maximum of Rs. 3,000. The 
Patna High Court upheld the rejection of his application because of omission of a certificate of his 
father's income.  258.   
In Shetty v. State of Mysore,  259.   the High Court held that even more than the father's income might 
be taken into account. Medical college seats were reserved for those who met an occupational test 
and whose annual income was less than Rs. 1,200. A candidate listed his father's income as Rs. 
348.20 ­­ a one­thirteenth share of joint family income. The court held that "therefore his family 
income was Rs. 4,526.60, which exceeds Rs. 1,200 ­­ the maximum prescribed by the Order." 260. 
The court does not explain why it regards all members of an undivided joint family as a "family" for 
purposes of the Order, but it hints that it is because "the social status of a member of a joint Hindu 
family"  261.   is related to the income of the joint family. But surely not all who share a common status 
are members of the same family. The Order itself seems to visualize somewhat smaller units than 
the family involved here (13 co­parceners and presumably a comparable contingent of females), 
when it explains that the Rs. 1,200 income ceiling was determined on the gound that the annual per 
capita income in the state in 1961 was Rs. 266 and that the average family consists of five members. 
It was clearly a different family that they had in mind!  262.   
____________________ 
257. A. I. R. 1968 Pat. 504. 
258. The survey cited in n.265, below, indicated that false declarations of separation from a joint 
family were about a ninth of all false certificates. 
259. A. I. R. 1969 Mys. 48. 
260. Id., at 50. 
261. Id., at 51. 
262. The absence of any sliding scale for family size tends to favor small families, perhaps a laudable 
policy but one never avowedly pursued in this setting. 
­342­ 
In practice, the conceptual problems of identifying family income are overshadowed by the 
formidable problems of accurate assessment of income. The difficulties offered by joint families, 
income in kind rather than cash, and so forth are compounded by the widespread inclination to 
understate income. The absence of income and tax data on most families had led the government to 
ascertain income by requiring production of a certificate signed by a government officer. The 
officers vary from place to place and scheme to scheme,  263.   but generally they are highly placed 
and invariably have no way of ascertaining the income of the person for whom they sign.  264.   There 
is no way of knowing how many of these certificates are false. In a 1964 sample investigation of 
15,438 beneficiaries in one state, it was found that 4,491 (29.1%) had given false statements of 
income. False income statements ran as high as 58.9% in one of the four districts sampled; none had 
less than 5.6% false statements.  265.   
A few years after the survey just mentioned, the Tamil Nadu Backward Classes Commission 
remarked of that state's experience with income certificates: 
[I]t is impossible to check the income declaration . . . of people who derive their income 
from land or small business; and those who are self­employed succeed generally in 
filing satisfactory income certificates. Only those whose opulence is known . . . do not 
attempt to secure scholarships for their children by questionable means.  266.   
____________________ 
263. Many such provisions are like that of Gujarat, which required a certificate from "a Revenue 
Officer not below the rank of Deputy Tahsildar" ( Gujarat, Administration Report of the Social 
Welfare Department [Backward Class Section] 1963­64: 75) Some states accept certificates from 
a wider range of officials. Thus in Madras the certificate must be "issued by a gazette officer, an 
Inspector of Police or a non­gazetted officer of the Revenue Department not lower than a Deputy 
Tahsildar or SubMagistrate" ( Govt. of Madras, 1962: 29). The anonymous study referred to 
below found that certification by revenue officials tended to be more accurate than that of 
gazetted officers, while certificates issue by honorary magistrates were least accurate. 
264. Thus in Nataraja v. Selection Committee, 1972(1) Mys. L. J. 226, 227 [NIC], the court noted that 
the machinery for determining income and occupation "has given ample scope for 
manipulations. We have noticed that certificates of backwardness have been issued by Tahsildars 
indiscriminately without proper investigation into the bonafides of the claims." 
265. I am not at liberty to disclose the provenance of this study, but the state concerned enjoys a 
reputation of having one of the most efficient administrations in India. Cf. the observations of 
the Punjab Evaluation Committee ( Welfare Department 1966: 141) that "with the adoption of 
economic criterion, there is a tendency among the people to get themselves falsely declared as 
backward." 
266. Tamil Nadu Backward Classes Commission 1971: 77. Irschick ( 1977) notes that from the time 
poverty certificates were introduced in Madras in 1921, they provided abundant "room for local 
manipulation." 
­343­ 
Such practices were so prevalent that 
[i]t is generally conceded that a very large percentage of people definitely above the 
income limit fixed are able to obtain scholarships by understating their income.  267.   
But of course opportunities for such dissimulation are not spread evenly. They are least available to 
those like schoolteachers and inferior government servants who receive fixed and known salaries. 
 268.   The Tamil Nadu Commission argued in favor of special treatment of lowpaid government 
officers, noting that among other problems they "suffer from the handicap . . . that they cannot . . . 
suppress their monthly salary and produce false income certificates."  269.   
Where income is ascertainable and accurately reported, there may be the additional problem of 
fluctuating eligibility, dramatically displayed in Shameem v. Medical College.  270.   The State of 
Kerala had established an income cut­off of Rs. 6,000 on membership in the Backward Classes for 
purposes of medical admissions. The income of applicant's father, a small businessman, was Rs. 
7,020 for the most recent year and Rs. 3,710 for the previous year. Although the average for the two 
years was Rs. 5,365, comfortably below the cut­off line, she was ruled not eligible. Here 
fluctuations, surely not unusual, give the whole arrangement a curious lottery­like quality. A student 
might be excluded in a singularly good year; another, might gain inclusion as a windfall in a 
singularly poor year. These problems could obviously be addressed by provisions for averaging, at 
the expense of more record­keeping and more possiblities for manipulation. 
But even if information was accurate and the fluctuation problem was solved, income tests have 
another peculiar arbitrary quality when applied to a non­divisible benefit like a medical admission. 
When applied to a divisible good like a stipend, the incremental nature of income makes it a 
measure that has the advantage of allowing benefits to be calibrated to need or desert. Benefits can 
increase as income falls. But income has a very different character when used to distribute a non­
divisible good like medical admissions or reserved posts. An eligibility measure that is continuous 
and incremental is subject to a single cutting point. The difference between Rs. 5,999 and Rs. 6,000 
becomes the difference between inclusion and exclusion. One might, of course, 
____________________ 
267. Tamil Nadu Backward Classes Commisson 1971: 160. 
268. The anonymous study referred to above found that false statements of income were more 
prevalent among agriculturists and the self­employed than among employed persons. 
269. Tamil Nadu Backward Classes Commission 1971: 77. 
270. A. I. R. 1975 Ker. 131, rev'd A. I. R. 1976 Ker. 54. 
­344­ 
import the incremental virtue of income as a measure into the selection process by according 
graduated weightage in proportion to low income. But so long as eligibility for benefits is an all or 
none affair, income tests create border lines that are difficult to justify as well as to ascertain. Like 
communal tests, income tests import their own brand of arbitrariness into the selection process. 
Occupation tests also may give rise to many membership problems. A series of Mysore cases 
illustrate some of the complexities. The Mysore order specified the following occupations as 
contributing to social backwardness: "(1) Actual cultivator; (2) Artisan; (3) Petty businessman; (4) 
Inferior service (i.e., Class IV in Government service and corresponding class or service in private 
employment) including casual labour, and (5) Any other occupation involving manual labour".  271.   
First, there is the simple problem of the definition of these broad categories. For example, who is an 
"actual cultivator" or an "artisan" is not always self­evident.  272.   What is an "occupation involving 
manual labor"? An applicant for admission to a medical college was rejected on the occupation test 
when she reported her father's occupation as purohit (i.e., family priest), who performed rituals 
manually.  273.   The court held that this was not an occupation "involving manual labor." It was 
necessary to look at the predominant character of the occupation: 
Though a purohit may use his hands in performing certain rituals and ceremonies the 
predominant character of his occupation is that it requires study and knowledge of 
scriptures and of the body of traditions and the performance of his work involves mainly 
chanting or recitation of "mantras" or scriptures.  274.   
Second, there is the problem of the composite occupation where there are several sources of 
income, some from listed occupations and some not. In Sridhara v. Director of Technical Education, 
 275.   the applicant's father earned Rs. 800 per year, Rs. 660 as a part time clerk, and Rs. 140 by 
dealing in pencils and paper. He claimed that he was a "petty businessman" who "supplemented" his 
business income by working part time as a clerk. In view of the figures, the court might have easily 
employed a rule of preponderance. But it chose to formulate a general policy against inclusion of 
any person who engaged part time in unlisted occupations. It finds the rationale of the Order in the 
notion that the environmental conditions which attend certain occupations contribute to 
backwardness, while others do not. If a person follows both, 
____________________ 
271. The order is reproduced in full in Shetty v. State of Mysore, A. I. R. 1969 Mys. 48. 
272. A "mechanic" is an "artisan" ( Pasha v. State of Mysore, 1969(1) Mys. L. J. 79). 
273. Sudha v. Selection Committee, A.I.R. 1967 Mys. 221 
274. Id., at 222. 
275. 1966(2) Mys. L. J. 774. 
­345­ 
it cannot be said that the advantage of following that latter category of occupations is destroyed or 
nullified by the same person following, in addition, an occupation of the former category. Having 
regard to the objects sought to be achieved by this classification, namely to help those who are at the 
bottom rung of the ladder of social and economic development . . . it is only reasonable to construe 
the Government Order as providing that only those who follow exclusively the listed occupations 
come within the Backward Classes. Any other interpretation will bring into the Backward Classes a 
large number of persons who do not suffer from serious environmental disadvantages. . . .  276.   
While this rule has the virtue of definiteness, it is difficult to see that the object is effectuated, for 
example, in a situation the reverse of the one in the Shridhara case. Suppose the father made Rs. 
660 from selling pencils and added Rs. 140 from part­time clerking. Does any income from a 
prestigious occupation automatically carry with it the disqualifying environmental advantages? 
Such a rule, if strictly applied, would be a disincentive to take up part­time employment which 
might deprive one's children of enhanced educational opportunities. 
Third, there is the problem of change over time.  277.   In Shetty v. State of Mysore,  278.   the applicant's 
father was a retired schoolteacher who was now engaged in agriculture. The High Court held that 
the occupations contemplated by the order 
are not casual or temporary occupations, but the habitual occupations of families, 
otherwise a person may claim to be treated as backward one year and forward in another 
as suits his advantage.  279.   
It was common knowledge, said the court, that a schoolteacher even after retirement was held in the 
same esteem, so he could not be regarded as a member of a Backward Class. 
Fourth and finally, there is the problem of whose occupation is to be measured. As in the case of 
income, it is generally assumed that it will he the father's. But what if there is a joint family and 
some members are engaged in higher­status occupations? In the Shetty case, the court insists that the 
family is a unit for purposes of backwardness. "It would not be possible to envision a situation in 
which one is treated as backward and another is not."  280.   But suppose you have a joint family 
____________________ 
276. Id., at 777. 
277. In what must be regarded as the reductio ad absurdum of the mechanical application of the 
occupation test, a student whose mother was a coolie was rejected because his mother was ill 
and therefore had no occupation. This was reversed by the Court in Viswanath v. Govemment of 
Mysore, A. I. R. 1964 Mys. 132 at 144. 
278. A. I. R. 1969 Mys. 48. 
279. Id., at 50­51. 
280. Id., at 51. 
­346­ 
as in Shetty with thirteen co­parceners. Are all the members of a family disqualified from being 
backward if one of the thirteen earns some money in a higher­status occupation? With family 
boundaries and members' occupations as variables, the possibilities for producing the indicia of 
eligibility are not negligible.  281.   Unlike the manipulation of communal identity, these expedients 
may be resorted to without excit ing reproach by self or others. 
What is striking in these cases is that, contrary to what might have been anticipated, judgments 
about status seem at least as tangible and determinate to the courts as do facts about economic 
standing. Where the economic facts appear doubtful or subject to interpretation, the court resolves 
the ambiguity by turning to assessments of status, and it experienced no difficulty in judging, e.g., 
the status of a clerk, a purohit, a schoolteacher, or a joint family with Rs. 4,000 income. The 
economic tests tend to be resolved into selection on the basis of status.  282.   But if this is so, it is not 
clear that reported income/occupation is really a better measure of status than communal 
identification. 
Courts have been relatively indifferent to the disincentive effects of these rules, or more precisely 
the choice they pose between disincentive to mobility and positive incentive to dissimulation. A 
family whose son aspired to be a doctor or who anticipated any other tangible boon from 
governmental preference would of course be well advised not only to keep reported income from 
rising above, e.g., Rs. 1,200, but even at lower levels of income to avoid employment that would 
jeopardize their occupational standing. But these are largely fictional. In reality, the difficulties have 
been of overinclusion rather than disincentives. 
Of course, these problems in administering economic tests have their parallel in the administration 
of communal tests. Apart from cases of dissimulative identification, there were a host of borderline 
cases in which issues of group membership were litigated when a challenger with a heavy personal 
stake sought to wrest a valuable prize by establishing his eligibility or disqualifying a competitor. 
Dispute's about economic tests are different in several ways: compared with community, income is 
not likely to be a matter of general knowledge; compared with community there is little 
compunction about or stigma attached to dissimulation of income; and finally the categories, 
malleable and fluid, not just at the border lines, are less established in common experience. 
The cases that reach the courts are not unusual contests about dis­ 
____________________ 
281. For an unsuccessful but instructive attempt, see Shantha Kumar v. State of Mysore, 1971 Mys. L. 
J. 21, discussed in §F, above. 
282. Cf. the observations of Sabberwal ( 1972: 74) on the tendency of administration to employ status 
as an indicator and rationing device in place of determinations which would require resources 
and competence to make. 
­347­ 
puted border lines, but cases of ordinary applicants testing the margins of routine administration. 
Were there serious attempts to monitor economic tests, there would be many more cases of this kind 
than there have been in the communal group membership area. On the other hand, there is a greater 
possibility of avoiding such litigation by more careful, refined, and definite formulation of the 
administrative rules. This would seem to be an area in which courts could contribute to solutions on 
the administrative level by pressuring the administrators to eliminate some of the gross inequities 
that pervade the rules. However, so far no litigant has found a way to provide the courts with a 
handle for exerting such control. 
H. CONCLUSION: PRAGMATIC AND FORMAL SOLUTIONS 
The judicial treatment of group membership problems displays two divergent tendencies, each with 
some support from the Supreme Court. The first, which might be called the empirical or pragmatic 
approach, is represented in the Jasani, Shyamsundar, Kartik Oraon, Wilson Reade, and Jahan Ara 
cases.  283.   The second, which might be called the formal or fictional approach is represented in the 
Dora case and in the cases dealing with the "Hinduism" test for Scheduled Castes.  284.   
Both approaches recognize the "compartmental" nature of Indian society. But the empirical 
approach is willing to give recognition to the areas of blurring and overlap that are found within it. 
The formal approach, emphasizing theoretical symmetry, tends to picture the society as one of 
mutually exclusive and hierarchically ranked compartments; where in fact individuals straddle 
compartments, the Court sees its task as assigning them to one or the other. The empirical approach 
does not share this notion of resolving the ambiguity into a single identification but is congenial to 
multiple and overlapping affiliations; it addresses itself to whether, in the light of the policy of the 
particular legislation involved, the individual can be said to be a member of the group concerned. 
The formal approach concentrates on the theoretical consequences of certain acts: one who attains 
caste status loses his tribal affiliation; one who declares himself a member of a non­Hindu religion 
loses caste membership. The empirical approach pays less attention to such theoretical 
incompatibility and gives greater weight to the facts of intention and acceptance. Thus sect members 
can retain their caste membership, an Anglo­Indian or a Tamil Christian can become a tribal, a 
convert can remain a tribal. 
____________________ 
283. See §B and C, above. 
284. See §D and E, above. Cf. the Saha case, §A, above. 
­348­ 
The empirical and formal labels are not entirely satisfactory, for both approaches are formal in the 
sense of being selective and abstract, and both are empirical in the sense of rendering the most 
accurate picture of Indian social reality as it is experienced. But that reality is so vast and complex, 
and any individual experience of it so partial and fragmentary, that there is no single correct way to 
render it. The two approaches I have distinguished differ in which aspects of that experience they 
find most salient. What I have called the empirical approach emphasizes the group affiliations as 
experienced by (at least some of) the participants at the particular social locations in question; the 
formal approach tends to rely more heavily on perceptions of these affiliations by observers at 
distant (often hierarchically superior) locations in the society. These perceptions, mediated by 
categories of hierarchy and exclusiveness, are more evident in dealing with the Hindu caste order 
than in dealing with tribal groups. Problems of acquiring or losing caste status tend to trigger 
bewildering and often unpersuasive conceptual acrobatics. In cases of membership in tribal groups, 
the courts seem to experience less difficulty in aligning entitlements with the factual contours as 
experienced by the participants. As the Court in Jahan Ara says: 
When a person in the course of time has been assimilated in the community it is 
somewhat difficult to comprehend how that person can be denied the rights and 
privileges which may be conferred on that community . . . by constitutional provisions. 
 285.   
Is one of these approaches preferable or constitutionally incumbent upon the courts? The empirical 
approach is clearly more consonant with current knowledge of Indian society. It sheds the 
encumbrance of theories of caste which reflect perceptions of Indian society current several 
generations ago. Modern students of Indian society have modified older notions of the caste system, 
which emphasized varna and inflexibility, and have achieved a new understanding of its complexity, 
which includes its local variability, the ambiguity of caste­ranking, the existence of mobility, and 
the limitations of varna theory.  286.   The courts have long recognized the deficiencies of the older 
notions.  287.   Now that 
____________________ 
285. A. I. R. 1972 S. C. 1840 at 1850. The Court notes some additional textual support for this 
pragmatic approach in the tribal setting when it commends the view of the High Court that Art. 
342's mention of "tribal communities" as well as "tribes" evidences an awareness that those who 
are not members of a tribe in the strictest sense may by assimilation become members of the 
tribal community. 
286. See Mandelbaum 1970; Fox 1969; Singer and Cohn, eds., 1968 Rudolph and Rudolph 1967; 
Srinivas 1962 and 1966; Karve 1961; Marriott 1959. 
287. On the limitations of the varna model, see §D, above. For instructive discussion of the 
complexity of the caste system, see Muthusami v. Masilamani, I. L. R. 33 Mad. 342 (1909). 
­349­ 
courts are no longer hampered by the necessity of giving legal effect to a picture of Hindu society 
consisting of the four varnas,  288.   they may appropriately employ the new perspectives to confront 
the actualities of Indian society in order to implement the principles of the Constitution. 
It is submitted that the empirical approach is also to be preferred because of its consonance with the 
ideals and principles of the Constitution, while the formal approach contravenes these and should be 
discarded. The Constitution sets forth a general program for the reconstruction of Indian society. In 
spite of its length, it is surprisingly undetailed in its treatment of the institution of caste and existing 
group structure in Indian society. But it clearly sets out to secure to individuals equality of status 
and opportunity,  289.   to abolish invidious distinctions among groups,  290.   to protect the integrity of a 
variety of group ­religious, linguistic, and cultural,  291.   to give free play to voluntary associations, 
 292.   to confer the widest freedom of association to the individual,  293.   and generally the widest 
personal freedom consonant with the public good.  294   Without pursuing all of these in detail, it is 
clear that the following general principles are consistently in evidence: (1) a commitment to the 
replacement of ascribed status by voluntary affiliations; (2) an emphasis on the integrity and 
autonomy of groups within society; (3) a withdrawal of governmental recognition of rank ordering 
among groups. 
In all of these, the empirical approach seems more congenial to the constitutional design. Its 
emphasis on the actual conduct of the individual and the actual acceptance by members of the group 
gives greater play to the voluntary principle. The control this gives the group over determination of 
its own membership seems implicit in the recognition of the integrity of the group. Finally, it avoids 
the necessity of giving official recognition to the ranking of groups. 
The formal approach, on the other hand, severely limits the voluntary principle and the autonomy of 
the group by giving conduct unintended consequences on theoretical grounds and by determining 
the question of membership without consulting the views of the relevant 
____________________ 
288. The Constitution and statute law forbid enforcement of invidious caste distinctions. The 
enactment of the various Hindu Code acts in 1955­56 virtually eliminates varna as an effective 
legal concept and discredits it as a source of analogy. Although its use as a distinct legal status is 
still permissible for limited purposes (see nn. 70 and 71 above), the Constitution surely provides 
a mandate to confine it within the narrowest limits. 
289. Preamble, Arts. 14, 15, 16, 17, 18, 23, 46, 
290. Arts. 14­17, 25­30. 
291. Arts. 25­30, 347, 350A, 350B. 
292. Arts. 19(1)c, 25, 26, 30. 
293. Id. 
294 See, generally, Parts III and IV of the Constitution. 
­350­ 
groups. It gives, if unwillingly, legal effectiveness to the notions of rank order among groups and 
mutual exclusiveness among them. The persistence of such an approach should be regarded as an 
anachronism, and one expects that before long the courts will refine and extend the empirical 
approach. 
An empirical, pragmatic approach, sensitive to the egalitarian and welfare objectives of 
compensatory discrimination, seems particularly appropriate in the group membership area. In 
handling group membership problems, the courts draw close to the central paradox of the 
compensatory discrimination­­the employment of traditional categories of social distinction for the 
purpose of overcoming these distinctions. It is the task of the courts, in visualizing groups and their 
boundaries, to employ those parts of the traditional view which are useful in carrying out the 
compensatory discrimination policy and which are compatible with the constitutional principles of 
equality and voluntarism while discarding those which tend to perpetuate and extend invidious 
distinctions. 
However, this task itself involves the courts in a dilemma. It propels the courts into subtle 
distinctions, refined analysis of policy, and a searching examination of facts. But the massive job of 
certifying crores of recipients of preferences is going to be done, not by the courts, but by 
administrators and clerks. The litigation about group membership so far has arisen primarily from 
election disputes, which are atypical in several ways: the courts are directly involved and the stakes 
are high enough to justify a thorough investigation of the facts and to provide sufficient incentive to 
expend resources to render them problematic. But allotment of scholarships, housing sites, or small 
loans will be conducted remote from judicial scrutiny and in circumstances where litigation will be 
rare. For the courts to guide policy, the subtleties and refinements will have to be translatable into 
workable rules that can be applied routinely and inexpensively at the operating level by 
administrators and clerks who have neither the training and capacity nor the fact­finding resources 
of the judges. Inevitably, there will be some slippage between judicial formulations and actual 
administration. Courts cannot solve or even anticipate the problems of administration, but they can 
strive to design their tests so as to minimize the inevitable discrepancy. 
­351­ 
10 Social Identity and Judicial Action 
WE BEGAN chapter 1 with the image of a compartmental society composed of innumerable 
mutually exclusive groups, arranged in interlocking clusters, layers, and strata. Even acknowledging 
occasional difficulty in locating groups (or individuals), the imagery conveyed a sense of 
determinate location on an elaborate grid. And it was implied that our knowledge­­subject to 
inevitable blurs and gaps­depicted that complex society as it really was. Our subsequent tour of the 
issues of group identity leads us to reflect on the limitations of both the compartmental imagery and 
the ideal of knowledge as photographic representation of reality. 
Social identities are not innate qualities that inhere in or emanate from the group or individual; 
instead they are asserted to, imposed by, and negotiated with various other components of the 
society. The definition, location, standing, and boundaries of a group (or an individual) may differ 
as they are viewed from different vantage points. Thus the identity of a caste may differ from the 
perspective of its members, fellow villagers, the regional marriage network, nearby citydwellers, 
religious dignitaries, politicians, and the educational authorities. Its members may use any of these 
attributions of identity as counters in trying to renegotiate their identities in any of these other 
arenas. What the groups are, what their boundaries are, and who is in them depends on the observer, 
on the frame of reference, and on efforts at self­presentation. 
This relativistic vision contrasts sharply with what we might call the "realist" view of Indian society. 
Realism and relativism are useful to mark extremes of ways of looking at Indian society. "Realism" 
proceeds as if there were bounded social units, corresponding to our intellectual categories, carrying 
fixed identities, and arranged in relations of mutual inclusion and exclusion. Thus Hinduism 
includes varnas, which are in turn made up of endogamous jati groups; other religions are mutually 
incompatible with Hinduism, varna, and caste; caste is exclusive of tribe, and so forth. In the realist 
view all of these elements can be ordered so that each group and individual has a determinate 
location. 
­352­ 
Thus with sufficient information we can mark the line between untouchables and touchables, 
forward and backward, tribes and castes, Hindus and non­Hindus, etc. 
We have moved from this naive realism to a relativist view which posits a pluralistic process of 
assigning and constructing identities, a process in which elements and fragments of several 
overlapping but competing symbol systems (e.g., religious purity, political dominance, "Western" 
advancement, etc.) are combined and recombined by individuals and groups. These elements are not 
fixed and inherent traits but are counters to persuade others in a continuing argument in which 
identities are constructed by assertion, negotiation, and imposition. 
Of course, stable relations of inclusion and exclusion are many and typical, so that the realist grid is 
a useful way to organize a multitude of observations. It is tempting to infer that these regularities, 
multiplied and extended, form a single coherent system, a set of orderly blocks interlaced with a few 
deviant and transient instances of conflict, ambiguity, and uncertainty. But, relativism counsels us, 
conflict, ambiguity, and uncertainty are not marginal; they are part of the core of the process. What 
we have called the realist view is a tendency to overestimate the fixed, uncontroverted, and regular 
parts of the process. The relativist view denies that these exist apart from the movement, ambiguity, 
ambivalence, indeterminacy, and conflict. Like foreground and background, they are inseparable; 
the moving parts are as integral as those at rest. 
Realist and relativist approaches do not necessarily reflect disagreement about the behavior that 
occurs in Indian society. They are different cognitive and expository strategies for understanding 
and discussing that behavior. Few observers owe exclusive allegiance to one or the other of these 
strategies. The same judge, for example, may combine them or switch from one to another as the 
context commends. (Similarly, in this book I found the compartmental imagery useful to present an 
initial sketch, but subsequently veered around to a relativist perspective to do justice to the 
complexities revealed in the cases.) What is a helpful map for the traveller is not necessarily the 
most useful picture of the landscape for the surveyor or the engineer. 
Group identities have a peculiar dual character. They are not natural facts, like geological formations 
or rainfall, impervious to fiat and uninfluenced by understanding. But neither are they artificial 
products of policy, like courts, legislatures, or corporations, that can be dismantled or altered by 
human design. Like language usages, they have an intermediate character, combining natural 
givenness with some malleability. 
The relation of government to group and individual identities is 
­353­ 
extremely complex, for government cannot promulgate such identities nor can it simply record them. 
It cannot mold social reality into the categories it promulgates, because whatever it does is super­
added to the existing symbolic array made available by the multiple folk and learned systems that 
circulate in local, regional, and all­India networks and are internalized within the actors themselves. 
Government is an important source of messages about personal and group identity. These messages 
may suggest new models and make available and salient symbolic constellations that were 
previously inaccessible. Government symbols have a special potency because they can be used to 
advance or defeat claims to which attach concrete rewards and determinate results­­hence the 
tendency for other identities to be deflected toward alignment with official categories. Government 
specifications become facts that people can use, in combination with other elements, as resources in 
constructing or changing their identities. Instead of promulgating a definitive ordering, the 
government introduces new incentives for change and new possibilities for symbolic enterprise. 
But it is not open to government to eschew influencing the process in favor of observing and 
recording identities as if they were natural phenomena. The experience of recording caste and caste 
standing on Censuses is, of course, the classic example of the impossibility of neutral measurement. 
What is recorded reflects collective and individual strategies of self­presentation. The official 
record not only becomes an instrument of official policy, but mints new symbolic currency. 
The government (including the courts), though it cannot definitively portray the labyrinth of group 
identities nor regulate it decisively, cannot escape having policies regarding it. In a curious way 
government presides over the process and contributes importantly to it. How can government pursue 
policies that turn on group identity in this relativistic world? In many instances, I have argued that 
the courts should take a relativist view, recognizing and encouraging ambiguity, indeterminacy, 
asymmetry, and variation. I submit that courts (and by implication, government) should incline this 
way as a general policy. 
I have argued that relativism (as embodied in what I called the empirical approach to group 
membership questions and in a flexible "elastic" position on the question of caste units) is more 
consonant with the constitutional themes of voluntarism, group autonomy, and nonrecognition of 
rank ordering. The relativist approach embodies a more comprehensive understanding of Indian 
society that enables us to include ambiguity, indeterminacy, and movement as part of the process of 
the continual construction and reconstruction of identities, a process that can be understood, if not 
exhaustively cataloged or mapped in detail. It enables us to apprehend change without concluding 
that the 
­354­ 
old is totally supplanted (for example, the decline of interdependence and the growth of 
competitiveness among castes; or the effacement among elite groups of the sense of caste hierarchy 
that was so palpable earlier in the century). The relativist approach alerts us to the presence of 
multiple and competing principles, whose specific and changing manifestations can only be 
ascertained by empirical observation. It therefore invites the courts to link up with (and stimulate) 
current social scientific scholarship. 
Relativism is more useful in pursuing the compensatory discrimination policies. It equips the courts 
with analytic tools to puncture exploitative assertions of identity (e.g., adoptions or conversions for 
1 Visualizing identity as composite and relative enables courts to 
purposes of securing benefits).    
readily distinguish those aspects of it relevant to compensatory discrimination policy, undistracted 
by the notion that it is unitary and inseparable. Similarly, it liberates courts from the notion that 
backwardness is a single trait or one that is either present or absent. It enables the courts to examine 
government action with a critical eye to insure that government has focussed on aspects of identity 
that are relevant to compensatory discrimination policy. 
The relativist perspective can provide government not only with critical response but with more 
options for policy. Policy alternatives are narrowed by the realist vision in which untouchability is 
coextensive with Hinduism, caste units are inseparable from caste standing, caste is lost upon 
conversion, etc. The relativist perspective emancipates thought from these fusions, enabling us to 
disaggregate our observations. We can organize them in terms of variation and continuity rather 
than in all­or­none terms of inclusion, exclusion, and implication. Thus government is afforded the 
option of modulating measures to correspond with the factual contours of specific situations. 
Of course, what the courts have to say is only part of a wider discourse about these policies that 
takes place in government, among political elites, in the media, and among wider publics. It is a 
discourse in which courts are accorded a leading role. And it is one that is characterized by 
____________________ 
1 Exploitative assertions of identity (for example, a Brahmin having himself adopted by a 
Scheduled Caste family in order to gain admission to medical college) often entail a dramatic 
inversion of symbols. It might be argued that such attempts should be welcomed as 
disestablishing traditional notions of hierarchy. But as we noted earlier, one of the paradoxes of 
compensatory discrimination policy is that it confers benefits on those among the beneficiary 
groups who are personally least needful. Much of the benefit conferred on those at the bottom is 
indirect­­i.e., through provision of models, establishment of networks, provision of representation 
and opportunities for vicarious participation. In whose success and advancement do the worse­
off participate in these ways? Group boundaries, as summations of these patterns of 
participation, are then an important tool of compensatory discrmination policy. 
­355­ 
considerable cognitive disarray. Consider, for example, the nomenclature used to describe the 
policies (reservations, special treatment, concessions, privileges) and their beneficiaries (backward 
classes, Other Backward Classes, Scheduled Castes, harijans, weaker sections, etc.). It is a 
landscape of euphemism and equivocation in which terms have multiple, ambiguous, overlapping, 
and shifting meanings. Precise discussion of policy is rendered difficult: even officials stumble and 
public discussion is confounded. 
For example, there is a persistent tendency to fuse all policies and their beneficiaries under such 
general headings as reservations for backward classes, so the resentment and dismay engendered by 
use of the Other Backward Classes category to stake out massive claims on behalf of peasant­
middle groups in some southern states are transferred to the more modest measures for Scheduled 
Castes. Atrocity stories about extensive reservations for undeserving beneficiaries, generated by 
OBC provisions at the state level, are readily generalized to the discredit of the Scheduled Castes 
nationally. 
Another persistent feature of this discourse is that it proceeds without much attention to the actual 
performance of the programs. Ignorance about performance is not accidental but is an inherent part 
of the administration of the policies. The policies are not, for the most part, programmatic­­i.e., they 
do not involve deliberate pursuit of specified goals. They are, as Aggarwal and Ashraf observe, 
2 Officials can point with pride to a list of accomplishments. 
"activity centred and not goal centred."    
But the relation between program performance and the goals to which the program is ostensibly 
aimed are left unstated and unexamined. Even if goals were made explicit, the bookkeeping is 
3 Actually, large 
generally such that it is impossible to separate out the net effects of the program.    
numbers of reservations for those highly cherished positions are not filled. Yet overestimation of the 
effect of reservation is pervasive. There is widespread feeling that enough (or too much) has been 
done for these people. It is time for them to stand on their own feet. This perception that so much is 
being done is fostered by the immense variety of programs (in housing, land, legal aid, water 
supplies, etc.). Many of the programs are very small, and it is onerous and cumber­ 
____________________ 
2 Aggarwal and Ashraf 1976: 166. 
3 For example, an increase in the number of Class I government servants from the Scheduled 
Castes does not tell us the net effect of the reservation. This enables the government to take 
credit for all the jobs obtained by the Scheduled Castes, not only those conferred by the 
reservation. (Credit may be claimed several times over. For some of these jobs may have been 
gained through a scholarship program which enabled some Scheduled Castes to gain places on 
merit.) And it contributes to the overestimation of the occurrence and effects of preferential 
programs and to their unpopularity. 
­356­ 
some to obtain any benefits. Much of their resources are consumed in administration with only a 
tiny scatter of benefits (e.g., legal aid). Worse, rather than supplementing benefits available through 
regular government programs, these special programs often offer an inferior substitute. And their 
existence becomes justification for exclusion from general programs of benefits.    4 

It is beyond the courts to rescue these policies from systematic cognitive distortion, for courts 
cannot control the way that various actors and audiences perceive judicial (and other) 
pronouncements through the lenses of their own preconceptions and use them in various arenas. 
Nevertheless, the courts are extremely significant contributors to discourse on these matters. The 
focussing of attention by conflicting claims dramatizes their pronouncements. Their decisions are 
given authority by producing tangible outcomes that can (in theory at least) be reproduced in similar 
cases. Although their interventions are episodic, the courts are the custodians of a continuing 
didactic tradition: they have the opportunity to reappraise and rework what they have said before, 
elaborating, adding emphasis, and pointing out how they should be read. 
The linkage of courts to the beliefs and the rhetorical stock of both elites and wider publics, 
although subject to noise and distortion, supplies another reason for forthright relativism. By 
permitting exceptions that blur boundaries, by accrediting variations that dislocate hierarchical 
markers, and by refusing to promulgate a single authoritative picture of the group contours of Indian 
society, the courts help to disestablish the traditional picture of that society, a picture which 
provided legitimacy for traditional relations of domination and hierarchic ordering. Of course, 
ignoring the symbols of domination and hierarchy will not, as some have naively contended, abolish 
the domination and hierarchy themselves. Indeed, avoidance may conceal the continuation of old 
disparities. Courts face the problem of keeping these categories available for carrying out policies 
of eliminating old inequalities without invigorating them as sources of symbolic sustenance for 
hierarchic patterns. 
But if a relativistic approach to questions of identity generally seems favorable to equalizing 
policies, it is not sufficient nor is it free of perplexities that make problematic the role of courts in 
carrying out these policies. Courts are passive and reactive institutions. They do not actively 
mobilize an agenda of cases. Yet in an indirect and subtle way they do foment and instigate 
5 When they indicate the scope 
litigation.    
____________________ 
4 See Aggarwal and Ashraf 1976:171. 
5 Cf. Mayhew 1975. 
­357­ 
of a right or locate standing to vindicate that right, they help to mobilize claimants. Their message is 
like a seed that requires fertile soil and ample water to nourish it. Some potential litigants are better 
able to perceive the message and supply the matching resources needed to make it flourish. The 
results brought about by such judicial messages, then, depend not on the courts themselves but on 

the distribution of capabilities among litigants.    
Judicial willingness to go behind conventional categories to assess their empirical "fit" and their 
policy relevance invites litigants to make arguments and bring forward evidence along these lines. 
But empirical arguments about complex states of fact­­e.g., the contours of group membership or of 
backwardness­­are more difficult and expensive to make than logical arguments about relations of 
inclusion and exclusion of categories. Rules that entail complex showings of fact and are open to 
subtle argument about policy accentuate the role of litigant capabilities and amplify the effect of 
disparities in those capabilities. 
Compensatory discrimination policies are designed to redistribute resources and opportunities to 
those who enjoy the fewest advantages. It seems plausible that the intended beneficiaries of these 
policies enjoy not only less wealth, education, and status but also less capability as users of legal 
7 The 
remedies. This supposition seems confirmed by the pattern of litigation in this area.    
beneficiaries of compensatory discrimination are less successful than their adversaries and have 
8 The 
been unable to use litigation affirmatively to improve the implementation of these policies.    
development of legal doctrine favorable to their interests does not of itself enlarge the capacities of 
these beneficiaries to use legal remedies. (Indeed, increases in the complexity and evidentiary 
demands of the law may amplify disparities in litigant capacities.) To deliver fully the benefits of 
these policies requires an enhancement of 
____________________ 
6 The notion of disputant capability includes such attributes as ability to perceive grievance, use 
information about the availability of remedies, articulate and manage claims competently, seek 
and utilize appropriate help, and bear the psychic and social costs of disputing. These may be 
attributes of individuals as well as of larger social units. Beyond this sort of personal competence 
are a related set of structural factors: the scale, continuity, and organization of the disputant that 
permit it to store information, utilize advance intelligence, develop expertise, establish and 
maintain credibility as a disputant, adopt long­term strategies, coordinate activities in different 
forums, etc. This party capability notion is elaborated in Galanter 1974, 1975, 1976. 
7 This pattern is analyzed in chap. 14, §C, below. 
8 In many cases the putative interests of the beneficiaries are represented by counsel for the 
government, defending a particular scheme. The statement in the text holds for such cases. They 
also raise the question of how much the positions asserted by the government in such cases 
diverge from those that would be asserted by advocates responsive only to the beneficiaries. 
­358­ 
the capability of the beneficiaries to elicit favorable judicial doctrine and to utilize it. Such an 
enlargement of their capacities would require the provision of more continuous and specialized legal 
services; and it would require better organization of the beneficiaries to coordinate strategy, monitor 
administration, aggregate claims, support the development of legal expertise, and do other things 
that would be beyond the capabilities of isolated individual litigants. 
The implementation of compensatory discrimination policies is entrusted to administrators and 
clerks, not to judges; judicial intervention is intermittent. The complexities and refinements of 
judicial doctrine have to be translated into workable rules that can be applied routinely and 
inexpensively. The individualized weighing of multiple factors that is attractive in the judicial 
setting is worrisome in the administrative setting. It may lead to errors due to misunderstanding and 
lack of information; it may confer discretion where that is not wanted; and it may make 
administration slow and costly. Further judicial resolution is possible in theory, but in practice 
informed and continuing monitoring by the judiciary is possible only if the beneficiaries are capable 
of bringing to the courts the cases and the data required for such control. The competence of courts 
to oversee the implementation of compensatory discrimination policy and to resolve the tensions 
between that policy and other commitments is both created and limited by the kind of institution 
9 Within those limits there is great scope for achievement, but it is achievement that 
that courts are.    
is very dependent on having capable parties that can elicit and utilize creative judicial action. 
____________________ 
9 On the variation in the roles and structures of courts as institutions, see Galanter (forthcoming); 
Friedman 1967; Damaska 1975; Shapiro 1975. 
­359­ 
Part Three COMPENSATORY DISCRIMINATION AND THE JUDICIAL PROCESS 
In Part Two the courts and the compensatory discrimination policies formed the frame within which 
questions of group identity were addressed. In Part Three the courts and compensatory 
discrimination policy move into the foreground. In roughly a quarter of a century the higher 
judiciary in India (Supreme Court and High Courts) erected a jurisprudence of compensatory 
discrimination. This branch of Indian constitutional law was built from scratch; it is very much a 
domestic product, produced with little guidance or borrowing from abroad. Although there have 
been some false starts and curious turnings, the major lines of this body of law can be traced. 
Chapter 11 discusses the constitutional setting of compensatory discrimination policy and doctrine. 
It traces the divergent but overlapping readings of the constitutional design that informed judicial 
response to specific problems until a remarkable 1975 decision ( N. M. Thomas v. State of Kerala) 
reopened what appeared to be settled, creating a new set of doctrinal possibilities and raising the 
question of how far they would ramify and how they would relate to earlier doctrine. 
Chapters 12 and 13 assemble two clusters of problems that arise in connection with compensatory 
discrimination programs. These chapters examine the judicial response to these problems and relate 
that response to the design and administration of these preferential 
­361­ 
programs. In these judicial encounters we get some glimmering of the problems of basic 
institutional architecture inherent in compensatory policies. And we see the problems and 
possibilities of having courts monitor these policies. 
Chapter 12 takes up a series of problems in which the judiciary is asked to respond to the question 
of "how much": how many places can be reserved, for how large a class of beneficiaries, how much 
variation from general standards can be tolerated, etc. 
Chapter 13 takes up a cluster of problems that center around the design of redistributive policies: 
how to distribute benefits so that all benefits are not monopolized by the better­off; how to build in 
flexibility while curtailing discretion that could defeat basic policy; how to articulate preferences for 
several groups of beneficiaries. 
The responses of courts to specific claims over a quarter of a century have erected a body of law 
which begins to wrestle with these questions. Neither a stagnant backwater nor the center of intense 
controversy between courts and government, it is an area in which we can readily observe Indian 
lawyers and judges at work. Chapters 14 and 15 turn from judicial grappling with specific problems 
to the judicial process as such. 
Chapter 14 portrays the higher judiciary, the lawyers, and the litigants who have shaped the law of 
compensatory discrimination. It tries to understand the ways in which the doctrinal product is 
shaped by the structure of courts, the recruitment and work habits of judges, the organization of 
legal services, the goals, resources, and capabilities of litigants and the legal culture that all of them 
share. 
Chapter 15 turns this around to ask what difference the doctrinal product makes, what impact it has 
on compensatory discrimination policies. Even in an area where government is attentive to the 
courts and there is no pattern of resistance, the actual effects of decisions are always problematic. 
How can courts optimize the effects of their decisions? What are the options that they have in the 
light of their changing sense of judicial work and of the re­emergence in the late 1970s of Scheduled 
Castes and Other Backward Classes as central policy issues? 
­362­ 
11 The Constitutional Scope of the Compen­ satory Discrimination Policy 
IN THE British period, arrangements for the distribution of power and benefits according to 
membership in communal groups were accepted methods for the adjustment of the political balance 
among communities as well as for the amelioration and protection of the lowly.    1. The Constitution 
of independent India drastically restricts the use of such arrangements. They are discarded as a 
general principle of governmental operation and may be used only for the purpose of promoting 
equality. Communal preferences and quotas for the purpose of political accommodation are 
outlawed.  2.   The protective (or compensatory) discrimination authorized by the Constitution is 
envisaged as an exceptional and temporary measure to be used only for the purpose of mitigating 
inequalities; it is designed to disappear along with these inequalities. The Constitution makers did 
3 But 
not intend it as a device to consolidate and protect the separate integrity of communal groups.    
while the preferential principle is thus confined in the purposes for which it may be used, it has been 
expanded in scope to cover a wide array of governmental schemes and programs. 
____________________ 
1. Quotas and preferences in recruitment for government service were not uncommon. Similar 
arrangements occurred in education, housing, revenue, and agricultural administration. The 
Government of India Act, 1919, provided for communal representation for Mohammedans, Sikhs, 
Anglo­Indians, Indian Christians, Depressed Classes, Aborigines, and other groups. Similar 
communal representation was provided by the Government of India Act, 1935. 
2. Proposals for reserved seats in legislatures, reserved cabinet posts, and reserved posts in public 
service for minorities (Muslims, Christians, Sikhs, Anglo­Indians, and Parsis, among others) 
were eliminated from the final draft of the Constitution. See XII CAD 229. Cf. the rejection of 
the proposed elimination of the qualifying word "backward" from Art. 16, which would have 
allowed provision of communal reservations in public services (VII CAD 682, 704). For a 
discussion of the Constituent Assembly's handling of communal representation, see Retzlaff 
1963; Austin 1966: 12 ff, 146 ff. 
3 There are constitutional provisions for the protection of the identity and integrity of groups: 
religious denominations have a Fundamental Right to manage their own religious affairs and to 
maintain and administer institutions (Art. 26); groups with a 
­363­ 
A. PREFERENTIAL TREATMENT IN THE CONTEXT OF FUNDAMENTAL RIGHTS 
The effort to secure equality by means of preferential treatment must be seen in its constitutional 
context. In the Constitution, the compensatory theme appears juxtaposed with the theme of formal 
equality. The provisions for compensatory preference appear as exceptions within a framework of 
enforceable fundamental rights which attempt to curtail the significance of ascriptive groups and to 
guarantee equal treatment to the individual. 
The Constitution of India confers on all citizens a fundamental right to be free of discrimination by 
the State on grounds of race,    4. religion, and caste. Governmental discrimination on these grounds is 
prohibited generally by Article 15(1). More specifically, discrimination is prohibited by Article 
16(2) in regard to State employment, by Article 23(2) in regard to compulsory public service, by 
Article 29(2) in regard to State­run and State­aided educational institutions, and by Article 325 in 
regard to electoral rolls. In specific contexts government is further forbidden to discriminate on 
5. residence,    
grounds of place of birth,     6. descent,    
7. class,    
8. language,    
9. and sex.  10.   

It is envisaged that government will not only refrain from discriminatory practices in the public 
sphere, but will actively seek to eliminate them in the private sphere. Untouchability is forbidden, 
and citizens are protected from certain discriminatory practices by private persons and institutions. 
 11.   Discrimination by private individuals in regard to use of facilities and accommodations open to 
the public is prohibited.  12.   Discrimination in private educational institutions is forbidden.  13.   
____________________ 
distinct language, script, or culture are guaranteed "the right to conserve the same", (Art. 29); 
religious and linguistic minorities have the right to establish and maintain educational institutions 
(Art. 30 (1)). No similar guarantees are extended to caste groups. (However, at least some caste 
groups qualify as "religious denominations." See Galanter 1968.) However, the governmental 
commitment to the Scheduled Tribes does include preservation of their distinctiveness. See chap. 
5, §C, above. 
10. Arts. 15(2), 16(2), and 325. 
11. Art. 17. For a discussion of the working of these provisions, see Galanter 1972. 
12. Art. 15(2). 
13. Art. 28(3), Art. 29(2). 
4. "Race" does not play a prominent part in post­Independence Indian jurisprudence. Its presence 
signifies an end to earlier colonial discrimination on racial grounds. (See n. 103, chap. 7, above). 
But it may have some residual conceptual role. Thus one court conceived of preferences for tribal 
people as preference "on racial ground" ( Mahendra Nath v. State, A. I. R. 1970 A. & N. 32, at 
34). And cf. the special provisions for Anglo­Indians, discussed at n. 50, chap. 9, above. 
5. Arts. 15(2) and 16(2). 
6. Art. 16(2). But cf Art. 16(3). 
7. Art. 16(2). 
8. Art. 23(2). 
9. Art. 29(2). Cf. Art. 30(2) and see Arts. 29(1), 30(1), and 350, 350­A, 350­B, which grant rights 
regarding language. 
­364­ 
But this attack on discrimination, public and private, is only one facet of the constitutional scheme 
to secure equality. The Constitution also directs the government to undertake special measures for 
the advancement of backward groups. In addition to provisions for reservation in government 
employment and legislative representation, the Constitution declares as a "Directive Principle of 
State Policy:" 
The State shall promote with special care the educational and economic interest of the 
weaker sections of the people, and, in particular, of the Scheduled Castes and the 
Scheduled Tribes, and shall protect them from social injustice and all forms of 
exploitation.  14.   
The Directive Principles of State Policy are not themselves justiciable, but the Constitution 
prescribes it as "the duty of the State to apply these principles in making laws."  15.   
The scope of governmental power to make special provision for the backward was soon put to the 
test. When the Constitution came into force, the State of Madras apportioned seats in its medical 
and engineering colleges in accordance with its Communal G.O.,  16.   which provided that for every 
fourteen seats the selection committee was to admit the following: 

Non­Brahmin Hindus  6 
Backward Hindus  2 
Brahmins  2 
Harijans  2 
Anglo­Indians and 
Indian Christians  1 
Muslims  1 
A Brahmin student who would have been admitted on purely academic qualifications charged that 
this was in violation of Article 29(2), which provided: 
No citizen shall be denied admission into any educational institution maintained by the 
State or receiving aid out of State funds on grounds only of religion, race, caste, 
language or any of them. 
In State of Madras v. Champakam Dorairajan,  17.   the Supreme Court upheld the claim of the 
protesting student. Finding that the denial to the student of a chance to compete for the other seats 
was only on the basis of her caste and violated Article 29(2), the Court advanced two crucially 
important propositions. First, it held that "the right to get 
____________________ 
14. Art. 46. 
15. Art. 37. 
16. On the background of the Madras Communal G. O., see Irschick 1969. 
17. A. I. R. 1951 S. C. 226, 1951 S. C. J. 313. 
­365­ 
admission into any educational institution of the kind [covered by Article 29(2)] is a right which an 
individual citizen has a citizen and not as a member of any community or class of citizens."  18.   Thus 
the Court refused to countenance any notion of group rights. Second, the Court rejected the 
argument that the Directive Principle of Article 46 by its own force established a principle of 
preference which might lawfully be embodied in legislation. "The Directive Principles of State 
Policy have to conform to and run as subsidiary to the Chapter of Fundamental Rights. . . ."  19.   The 
Court had an additional textual argument for this position: noting that Article 16 was explicitly 
qualified by an exception allowing reservations for the backward in the area of State employment, it 
suggests that "the omission of such an express provision from Article 29 cannot but be regarded as 
significant."  20.   The government's argument about the implications of Article 46 would render 
superfluous the exception in Article 16. 
It should be pointed out that this was by no means an inevitable or unavoidable reading.  21.   An 
earlier case in the Punjab High Court upheld reserved seats for Harijans in educational institutions 
on the ground that "Article 46 must be taken as an exception to Article 29(2) . . ." and thus the state 
could constitutionally promote the educational interests of Scheduled Castes by adopting a system 
of reservations.  22.   In later years, the Supreme Court, still maintaining that Directive Principles were 
subordinate to Fundamental Rights, allowed that "these directive principles are to be taken into 
account in determining the scope and application of Fundamental Rights."  23.   From the vantage 
point of 1967, the then Chief justice of the Supreme Court indicated that "any reasonable attempt to 
raise the status of the Backward Classes could have been upheld on the principle of classification." 
 24   One 
____________________ 
18. A. I. R. 1951 S. C. at 227. 
19. A. I. R. 1951 S. C. at 228. 
20. A. I. R. 1951 S. C. at 228. 
21. Alexandrowicz 1957: 62 observes that in Champakam the Supreme Court "did not sacrifice the 
letter of the Constitution to make any spectacularly progressive move in the spirit of the 
Constitution." Cf. the remark of Prime Minister Nehru in the subsequent parliamentary debate 
that the proposed amendment "might almost be considered redundant" ( Parliamentary Debates, 
Vol. XII­XIII [Part II], col. 8820 [ 16 May 1951]). 
22. Om Prakash v. State of Punjab, A. I. R. 1951 Punj. 93, 97. 
23. Cf. with Mhd. Hanif Qureshi v. State of Bihar, A. I. R. 1958 S. C. 731 at 732, 737, where the 
"subsidiary" interpretation of Champakam Dorairajan is commended and the Court observes 
that although the Directive Principles are fundamental, the state must implement them "in such a 
way that its laws do not take away or abridge the Fundamental Rights, for otherwise the 
protective provision of Chapter III will be 'a mere rope of sand.'" 
24 "To remove the effect of centuries of discriminatory treatment and to raise the downtrodden to an 
equal status cannot be regarded as discriminatory against any 
­366­ 
wonders how much the outcome in Dorairajan was affected by the contingency that the question 
was presented in the context of a fullblown scheme of communal quotas rather than as a measure 
specifically designed to advance backward classes. 
On the same day in Venkataramana v. State of Madras, the Court struck down part of a similar 
communal quota in regard to government posts.  25.   Article 16(4), it found, permitted reservations 
only for "backward classes." The Court concluded that of the groups on the Madras list, only 
Harijans and "backward Hindus" could be said to be backward classes. The reservations for these 
groups were allowed, the others struck down. One may imagine a result in Dorairajan, along the 
lines of the result in Venkataramana. The Court might have said that Article 46 directs the State to 
promote the educational and economic interests of the weaker sections­­and mentions the Scheduled 
Castes and Tribes in particular; these groups then are clearly "weaker sections"; therefore the 
classification is not on the basis of caste "only"; therefore reservation for these groups is permissible 
in spite of Articles 15 and 29. As to other groups, the State had the burden of showing that they are 
"weaker sections" and within the exception to Article 15 and 29 implied by the Directive Principle; 
since there was no such showing, the reservations for these should be struck down. 
In the event, the Court held that the government had no power to reserve seats for backward 
communities (or, a fortiori, for any communities) in educational institutions. The implication was 
that all special provisions for backward caste groups (outside the area of government employment) 
would be open to the same kind of attack. The inference was later drawn by the Bombay High 
Court, which held that it was not a legitimate public purpose for the state to provide a housing 
colony for Harijans, but unconstitutional caste discrimination forbidden by Article 15.  26.   
Governmental reaction to this doctrine was swift. Along with other Supreme Court decisions that 
had discomfited the government, it was reversed by constitutional amendment less than two months 
after it had 
____________________ 
one. . . . With all due respect [to Champakam Dorairajan] the question of discrimination hardly 
arose because in view of [Articles 46, 340, 16(4)] . . . any reasonable attempt to raise the status of 
the backward classes could have been upheld on the principle of classification" ( Golak Nath v. 
State of Punjab, A. I. R. 1967 S. C. 1643 at 1706). Cf. G. S. Sharma 1965: 183 ff., tracing the 
greater reliance on the Directive Principles in later years, who explains Champakam by 
postulating the necessity of further time for the Court to realize that "under the special conditions 
of the Indian society a special treatment to a particular minority for a specific period could be 
treated as a value comparable to the value of maintaining secular traditions of admissions for 
educational institutions." 
25. Venkataramana v. State of Madras, A. I. R. 1951 S. C. 229. 
26. Jagwant Kaur v. State of Bombay, A. I. R. 1952 Bom. 461. 
­367­ 
been delivered. Among the amendments was Article 15(4), which provides: 
Nothing in Article 15 or Article 29(2) . . . shall prevent the State from making any 
special provision for the advancement of any socially and educationally backward 
classes of citizens or for the Scheduled Castes and the Scheduled Tribes.  27.   
The Constituent Assembly had rejected the inclusion in the original Constitution of a provision 
similar to Article 15(4).  28.   Now the same assembly, in its incarnation as the provisional Parliament, 
passed it with little opposition,  29.   on the ground that it was needed to empower the State to carry 
out the Directive Principles by insuring that the Fundamental Rights guarantees of equality did not 
obstruct substantive equalization. Together with the provisions for special treatment that were in the 
original text, Article 15(4) forms the constitutional basis for the subsequent governmental policy of 
compensatory discrimination. Before proceeding to an examination of this policy, we shall briefly 
review the scope and relations of these constitutional provisions. 
B. THE SCOPE OF THE CONSTITUTIONAL PROVISIONS FOR SPECIAL 
TREATMENT 
The only specific authorizations for preferential treatment in the original Constitution were in the 
fields of government employment and legislative representation. Discrimination in government 
employment is outlawed by Article 16, but Article 16(4) permits the State to make 
any provision for the reservation of appointments or posts in favor of any backward class 
of citizens which, in the opinion of the State, is not adequately represented in the 
services under the State. 
In addition, Article 335 provides: 
The claims of the members of the Scheduled Castes and the Scheduled 
____________________ 
27. Constitution (First Amendment) Act, para. 2 (1951). 
28. Professor K. T. Shah's proposal to amend Art. 15 to allow the state to make special provision for 
the "advantage, safeguard, or betterment" of the Scheduled Castes and Scheduled Tribes was 
rejected by the Assembly (VII CAD 655, 664). Dr. Ambedkar, chairman of the Drafting 
Committee, opposed the amendment as providing an opening for segregated facilities (VII CAD 
661). It should be noted that the Shah amendment is narrower in scope than Art. 15(4), which 
applies to Backward Classes as well as Scheduled Castes and Scheduled Tribes. 
29. This clause was adopted by a majority of 243 to 5, a slightly higher margin than either of the 
other major amendments ( Parliamentary Debates, Vol. XII­XIII (part II), col. 9833). Cf cols. 
9885, 9927 ( 1 June 1951). 
­368­ 
Tribes shall be taken into consideration, consistently with the maintenance of efficiency of 
administration, in the making of appointments to services and posts in connection with the affairs of 
the Union or of a State. 
The relation between these provisions remains somewhat obscure. Article 335 is confined to 
Scheduled Castes and Scheduled Tribes, while Article 16(4) extends to all "backward classes." 
Article 335 seems to include any method of preference, while Article 16(4) is confined to 
reservations. Article 335 extends to all appointments "in connection with the affairs" of the State, 
which may be broader than the "services under the State" referred to by Article 16(4). Finally, 
Article 320(4) provides that public service commissions need not be "consulted as respects the 
manner in which any provision referred to in clause (4) of Article 16 may be made or as respects the 
manner in which effect may be given to the provisions of Article 335."  30.   
Both Article 15(4) and Article 16(4) permit special provisions for backward classes. Article 15(4) 
applies to the State in all of its dealings, while 16(4) is confined specifically to the field of 
government employment. It has been held that more general Fundamental Rights provisions are 
displaced in relation to a specific area by provisions specifically pertaining to that area.  31.   Thus, the 
area of employment, offices, and appointments under the State is controlled by Article 16 alone, and 
preferences in this area must be within the scope of 16(4). This includes judicial office  32.   as well as 
administrative posts, but it does not include elective office.  33.   
The authorization to establish preferences in employment is not confined to posts directly under the 
State. Where the State acts as an employment agency for State­aided schools, it may make such 
preference a condition of the aid.  34   The situation in regard to public sector 
____________________ 
30. An attempt to include the Other Backward Classes within Art. 335 was proposed by Dr. 
Deshmukh (see n. 44, chap. 6, above) (IX CAD 598). 
31. Dattatraya v. State of Bombay, A. I. R. 1953 Bom. 311 (Art. 16 displaces Art. 15 in the area of 
government employment); University of Madras v. Shantha Bal, A. I. R. 1954 Madras 67 (Art. 15 
displaced from area controlled by Art. 29[2]). See also State of Madras v. Champakam  
Dorairaian, A. I. R. 1951 S. C. 226. On the relations between the more general and the 
increasingly specific guarantees of equality in Arts. 14, 15, and 16, see Rangachari v. General 
Manager, A. I. R. 1962 S. C. 36; Shyam Behari Tewari v. Union of India, A. I. R. 1963 Ass. 94. 
32. Venkataramana v. State of Madras, A. I. R. 1951 S. C. 229; Kesava Iyengar v. State of Mysore, A. 
I. R. 1956 Mys. 20. 
33. Dattatraya v. State of Bombay, A. I. R. 1953 Bom. 311. 
34 The Supreme Court held that Art. 16(4) allowed the state to provide that its Public Service 
Commission give preference to Scheduled Caste teachers in filling vacancies in state­aided 
schools run by religious minorities. This provision of the Kerala Education Bill was particularly 
objected to on the ground that Scheduled Caste 
­369­ 
enterprises is far from clear. Government lawyers advised that the constitutional provisions for 
reservations were not applicable in the sense that public sector enterprises were not constitutionally 
obliged to have a policy of reservations.  35.   Yet it was open to them to make reservations, and it was 
open to the government to direct them to do so.  36.   
Article 16(4) covers not only preferences in initial recruitment into government service, but also 
preferences in promotions within the service. After some hesitation and with considerable 
reluctance, the Supreme Court held, in Rangachari v. General Manager (1961), that the "posts" 
referred to in Article 16(4) included promotions as well as initial appointments, but indicated that 
the preference permissible under Article 16(4) would not necessarily extend to other aspects of 
employment covered by Articles 16(1) and (2)­­e.g., salary, increment, pension, retirement age. 37. 
Such matte are "absolutely pro­ 
____________________ 
teachers might know nothing of the religion of the minority by whom the school was run. 
Denying that it violated Art. 29(2), the Supreme Court upheld the provision as a permissible 
condition for receiving grants in aid from the state ( In re Kerala Education Bill, 1957, A. I. R. 
1958 S. C. 956, 983). (This provision of the bill was repealed in 1960.) However, this conclusion 
is rendered doubtful by Sidhrajbhai Sabbaj v. State of Gujarat, A. I. R. 1963 S. C. 540, 547, 
where the Supreme Court referred to the rights of minority institutions as "absolute" and subject 
only to those reasonable restrictions that are "in the interests of the institution." 
35. Thus the Law Ministry is reported to have advised that Arts. 15, 16, and 335 were "not applicable 
to the public sector enterprises" ( CWSCST [ 4th] 2: 9 [ 1969). This is said to be based on the 
decision of the Supreme Court in Rajasthan State Electricity Board v. Mohan Lal, A. I. R. 1967 
S. C. 1857. But in that case it was held that the Electricity Board was clearly an authority to 
which the provisions of Part III are applicable. The Court there reached this conclusion by 
reasoning that duties imposed upon the "State" by Art. 46 imply that the "State" as defined in 
Art. 12 comprehends "bodies created for the purpose of promoting the educational and economic 
interests of the people" (id., 1862). And surely public sector undertakings are not free to 
discriminate among employees on caste or religious lines (forbidden the State by Arts. 15(1) and 
16(2)). Shah, J., concurring, pointed out that there is a further question of whether the body in 
question is invested with the sovereign power of the state to impose the restrictions on the 
Fundamental Rights that the state is authorized to do. Thus, there remains a question of whether 
it might act on its own to implement 16(4). 
36. CWSCST ( 4th) 2:9 ( 1969). 
37. General Manager v. Rangachari, A. I. R. 1962 S. C. 36. The question was left open in the earlier 
case of All­India Station Masters v. General Manager, A. I. R. 1960 S. C. 384. The High Court in 
the Rangachari case had concluded that Art. 16(4) is more limited in scope than Art. 16(2). The 
latter forbids caste discrimination in respect to "any employment or office under the state." Art. 
16(4) is an exception to this but only as regards "appointments or posts." The High Court held 
that promotion was "employment . . . under the State" but was not "appointment to a service or a 
post." Thus, caste discrimination in promotions would be barred by 16(2) and would not be saved 
by 16(4). The net result would be to permit reservations only at the stage of initial appointment to 
a service. Presumably, reservation would also be permissible if the higher post constituted a 
separate service (A. I. R. 1961 Mad. 35 at 41). 
­370­ 
tected by the doctrine of equality of opportunity and . . . do not form part of the subject matter of 
Article 16(4)."  38.   While the dissenting judges argued that reservation was limited to securing 
adequacy of quantitative representation of the favored group,  39.   the majority held that reservation 
could legitimately be used to secure representation in posts of higher grades.  40.   Thus, in 
establishing preference in government service, the State can aim at placing members of privileged 
groups in positions of authority, responsibility, and prestige. In the 1963 case of Balaji v. State of 
Mysore,  41.   the Supreme Court took the unusual course of explicitly adding a gloss to its earlier 
decision, pointing out that the Rangachari case "posed the bare question about the construction of 
Article 16(4). The propriety, the reasonableness or the wisdom of the impugned order was not 
questioned."  42.   In Rangachari, the Court had pointed to the language of Article 335, recuiring the 
State to take into account the efficiency of the services in making provisions for Scheduled Castes 
and Tribes, and declared the necessity of striking a reasonable balance between the claims of these 
classes and the efficiency of the services. In Balaji the Court emphasized that the public interest in 
the efficiency of government services set limits to reservation in promotions, putting outside the 
scope of Article 16(4) any "unreasonable, excessive, or extravagant reservation," for that would, by 
eliminating general competition in a large field and by creating widespread dissatisfaction among 
the employees, materially affect efficiency.  43.   Thus the Court gave notice that it would carefully 
scrutinize the reasonableness of reservations in the area of promotions. This came to pass in the 
Devadasan case, where the Court invalidated the reservation because of its unreasonable extent. 44. 
Both Rangachari and Devadasan involved promotion preferences for Scheduled Castes and Tribes. 
The authorization for preferences in promotions presumably includes the Other Backward Classes 
(OBC) as well.  45.   Since the courts have shown themselves somewhat more stringent in dealing with 
reservations for OBC, it might well be that such preference in promotions as is allowable for these 
most disadvantaged groups would be excessive when done for OBC. 
____________________ 
38. A. I. R. 1962 S. C. 36 at 42. 
39. Wanchoo, J., id. at 48; Rajagopala J. Ayyangar, id. at 49. The former relied on the interpretation 
of "posts" referred to in n. 54, chap. 12, below. The latter agreed with the High Court, see n. 37, 
above. 
40. The majority couples this with the puzzling observation that it is impermissible to make any 
reservation at the cost of efficiency of administration (A. I. R. 1962 S. C. at 44). This would 
seem to limit reservations to posts for which there are a number of equally qualified candidates. 
Or perhaps it implies that there is no relation between ordinary selection procedures and 
efficiency. 
41. A. I. R. 1963 S. C. 649. 
42. Id. at 664. 
43. Id. 
44. Devadasan v. Union of India, A. I. R. 1964 S. C. 179, discussed in chap. 12,§B, below. 
45. Triloki Nath Tiku v. State of Jammu and Kashmir, A. I. R. 1967 S. C. 1283, A. I. R. 1969 S. C. 1. 
­371­ 
The wording "any provision" in Article 16(4) and "any special provision" in Article 15(4) gives the 
State great leeway in prescribing the method of operation of schemes for preference. Governmental 
agencies have utilized a wide variety of devices for the purpose of conferring advantage on 
backward groups. Reservations have involved not merely setting aside reserved places, but also such 
preferential rules of recruitment as waiver of age requirements, of application fees, and of minimum 
educational qualifications; establishment of a lower minimum of qualifying marks on competitive 
examinations; special assistance and training in preparation for competitive examinations.  46.   
Are such practices covered by the authorization of Article 16(4) that the State can make "any 
provision for the reservation" of government posts? It has been suggested that fixing a lower 
minimum level of successful marks in a competitive examination is outside the power of the State, 
since it is not a "reservation in any sense of the term under Article 16(4)."  47.   However, such devices 
would appear to be included within the constitutional authorization.  48.   While Article 16(4) confines 
the State to the method of "reservations," it is clear that Article 16(4) empowers the State to 
determine the precise method to be adopted in effectuating the reservation.  49.   Unless it included 
permission for special 
____________________ 
46. RCSCST 1961­62: I, 128 ff. 
47. Somnath J. Iyer, in Chandra Sekhera v. State of Mysore, A. I. R. 1963 Mys. 292 at 299­300 
(dictum). In Kulkarni v. State of Mysore, A. I. R. 1963 Mys. 303, petitioner challenged the 
selection of munsiffs in which maximum age requirements had been raised and the requirement 
of minimum experience at the bar lowered for Scheduled Castes and Scheduled Tribes. The 
petition was dismissed without considering on the merits constitutional objections to these 
practices. In Devadasan v. Union of India, A. I. R. 1964 S. C. 179, the Court never reached 
petitioner's argument of the unconstitutionality of the double standard of minimum qualifying 
marks. 
48. Devadasan v. Union of India, A. I. R. 1964 S. C. 179 at 187. The position under Art. 16(4) would 
then be analogous to that under Art. 15(4), where it is "open to the State to . . . fix different 
minimum marks" so long as the difference is reasonable for purposes of Art. 14. In Sukhdev v. 
State of Andhra Pradesh, 1966 (1) Andh. W. R. 294 at 312, there was a waiver of minimum 
qualifying marks for medical college admission from 45% to 40%. Cf. General Manager v. 
Rangachari, A. I. R. 1962 S. C. 36 at 42, where the Court implies lower standards are 
permissible. See chap. 12, §C, below. 
49. This interpretation of Art. 16(4) is reinforced by the presence of Art. 335, which seems to 
contemplate preferences in recruitment which lie beyond "reservation" in the widest sense. See 
General Manager v. Rangachari, A. I. R. 1962 S.C. 36 on the relevance of Art. 335 to the 
interpretation of Art. 16(4). Art. 335, though, applies only to the Scheduled Castes and Tribes. If 
"reservation" in Art. 16(4) is narrowly construed, government would still be able to rely on Art. 
335 to authorize special recruitment procedures for the groups covered there, but it would be 
curtailed in using such procedures for Other Backward Classes. Cf. Art. 320(4), which 
contemplates wide governmental discretion in respect to the "manner" in which Art. 16(4) and 
Art. 335 are given effect. 
­372­ 
rules of recruitment, the provision for reservation would be meaningless, for it would merely mean 
that those who otherwise qualified would be entitled to reserved places­ but they are entitled to that 
anyway under Article 16(1).  50.   
Article 15(4) extends to all areas of governmental activity which are not controlled by a more 
specific provision.  51.   Thus, preferences in housing, education, welfare, government contracts, 52. 
and local political arrangements  53.   are authorized by, and must be within the scope of, Article 
15(4). Unlike Article 16(4), this Article does not limit the State to reservations as a method of 
preference. Reservations may be used to implement Article 15(4), but government may also employ 
methods such as fee concessions, scholarships, special facilities (housing, medical, etc.) which do 
not involve reservations at all.  54.   Special treatment need not take the form of diversion of resources. 
It may instead involve special protection, as for example, protection against sale of land in execution 
of a creditor's judgment. The scope of the "protection" authorized is not restricted by the qualifiers 
"socially and educationally," which describe the classes who may be benefitted, but may extend to 
protection from economic exploitation.  55.   
The power conferred in Article 15(4) does not seem confined to measures which confer benefits 
directly on members of these groups, but would seem to encompass indirect provision for their 
"advancement" which might otherwise offend Article 15(1). Suppose the State were to pay lawyers, 
doctors, or agricultural specialists to provide services to 
____________________ 
50. The striking down of the "carry­forward rule" in the Devadasan case­­especially the view that 
there must be some sort of equal opportunity for others to compete on every occasion of 
recruitment (A. I. R. 1964 S. C. at 187)­­might serve as a basis for attacks on these preferential 
rules of recruitment. However, it is clear that the vice there was the extent of the reservation 
rather than use of devices to broaden Scheduled Castes and Tribes' eligibility to compete. See 
discussion at chap. 12, §B, below. 
51. Thus, Art. 15 would also be displaced from the area covered by Arts. 29(2) (state­aided 
education) and 23(2) (compulsory public service). 
52. Afahendra Nath Pathak v. State of Assam, A. I. R. 1970 A. & N. 32. 
53. See n. 68, below. 
54. Moosa v. State of Kerala, A.I.R. 1960 Ker. 355 (housing for Harijans). Cf. Raju v. Chief 
Electoral Officer, A. I. R. 1976 Guj. 66, 71, where the court rejected a challenge to the 
Representation of the People Act's provision for lower electoral deposits from Scheduled Caste 
and Scheduled Tribe candidates: lower deposits "cannot be said to be discriminatory at all, as this 
concession has been granted in order to enable the members of the Scheduled Castes and Tribes 
who are economically backward and Financially incapable to contest elections." The court does 
not mention Art. 15(4), but it is responding to a broadside claim by petitioner that the provision 
violates Art. 14, a contention that is stated by the court "only for rejecting it." Although it is not 
mentioned in the opinion, the A. I. R. headnote writer lodged this aspect of the case under Art. 
15(4). 
55. Chait Ram v. Sikander, A. I. R. 1968 Pat. 337. 
­373­ 
backward classes? Or suppose the State were to allot scarce educational opportunities to persons on 
the basis of an undertaking to provide services to legitimate beneficiaries of Article 15(4)? 
Curiously, only one court seems to have encountered this kind of argument. In Pradip Tandon v. 
State of Uttar Pradesh, the state contended that its reservations for hill and rural areas were 
"made . . . to provide medical service . . . to people of those areas."  56.   A Full Bench of the 
Allahabad High Court thought this argument belied by a report of a committee on medical 
education which suggested an urban bias in medical education and called for various measures to 
make rural practice more attractive to medical graduates. The Full Bench observed: 
Medical graduates hailing from rural areas may also be disinclined to return to the 
villages for medical practice on account of poor facilities. . . The [state's] assertion . . . 
that the reservations . . . was with a view to feed the dispensaries of these areas appears 
to be pretentious and cannot be justifiable ground for making reservations. Such 
grounds of reservation of seats . . . will be wholly outside the ambit and scope of Article 
15(4). . . .  57.   
The court here appears to assert that reservations have to be based on the deservingness of the 
recipients rather than on some beneficial consequences for a backward class. The court here did not 
believe there was a legitimate backward class involved, nor, apparently, did it accord the argument 
much weight as a serious statement of government policy. Hence the issues that lie buried here were 
not addressed. If the State can make "any special provision" for the "advancement" of a backward 
class, may it not do so by conferring benefits on non­members of that class (e.g., those who 
undertake to provide services in undersupplied areas) through whom benefits will be delivered to 
the backward class? Since the immediate beneficiaries of such schemes would be selected on 
grounds of the service to be provided rather than on any of the grounds forbidden by Articles 15(1) 
or 29(2), it is not clear that Article 15(4) need come into play. But if it does, it is difficult to see any 
constitutional warrant for reading it so narrowly that it could not encompass such schemes.  58.   
____________________ 
56. A. I. R. 1975 All. 1, at 4. This decision was reversed in part by the Supreme Court, A. I. R. 1975 
S. C. 563. The Supreme Court did not take up this "distribution of medical services" justification 
of the scheme, but it did uphold a reservation for one of the areas on behalf of which the state 
made this argument. See discussion in chap. 8, §H. 
57. A. I. R. 1975 All. 1, at 7. The court dismissed as "irrelevant" to Art. 15(4) a suggestion that the 
U.P. government take a bond from medical graduates to serve in rural areas. 
58. Art. 16, on the other hand, specifically refers to reservations for members of the backward 
classes. But presumably, selection of non­backward­class employees on grounds of their capacity 
to benefit backward classes would not be on the basis of any 
­374­ 
Article 15(4) authorizes not only direct protections, but also provision for protection by political 
self­help. The Constitution specifically provides reserved seats in the lower house of Parliament 59. 
and in the lower houses of the state legislatures for the Scheduled Castes and Scheduled Tribes. 60. 
These are the only special reservations in legislatures; the Constituent Assembly definitely rejected 
political safeguards for religious and other minorities.  61.   All of the other provisions for preferences 
are merely authorizations empowering the State to make special provision for the backward. These 
legislative reservations are the only ones where a reservation is specifically enacted by the 
Constitution itself. 
Seats are reserved in the proportion of the population of Scheduled Castes and Tribes to the total 
population and are to be filled by joint election. These reservations do not involve "separate 
electorates" ­i.e., the representation of a particular group by legislators chosen by an electorate 
composed solely of members of that group.  62.   The seats are "reserved" in the sense that candidates 
who stand for them must belong to the privileged groups, but the entire electorate participates in 
choosing among candidates so qualified. Separate electorates for Parliament and the state 
legislatures are specifically outlawed by Article 325, which provides that no person shall be 
excluded from any electoral roll on grounds of religion, race, or caste. The Supreme Court has 
indicated that Article 15 prohibits communal electorates in local bodies.  63.   
Contrary to widespread misapprehension, these legislative reservations are the only ones that are 
subject to a constitutional time limit. It was originally provided that such reservations should expire 
ten years 
____________________ 
ground forbidden in Art. 16(2) and would thus stand or fall on general classification grounds 
unless Art. 16(4) were read restrictively as exhausting all governmental power to confer special 
treatment on backward classes in the area of government employment. 
59. Art. 330. 
60. Art. 332. 
61. XII CAD 299. 
62. By separate electorates is meant the representation of religious (or other) minorities by 
legislators chosen by an electorate composed only of members of that minority. The provision 
and extent of such representation was an extremely troublesome political issue in India during 
the forty years preceding Independence. The Govenment of India Act, 1909, gave separate 
electorates to Muslims; the Government of India Acts of 1919 and 1935 provided separate 
electorates for Muslims, Sikhs, Indian Christians, and other groups. Proposals to give separate 
electorates to the Scheduled Castes under the 1935 Act were withdrawn after adamant resistance 
by M.K. Gandhi (see Ambedkar 1946). They were given instead reserved seats in the legislatures 
­­i.e., only they could be candidates for these seats, but the whole general electorate chose among 
the candidates. Separate electorates were a constant focus of dispute between Hindus and 
Muslims (see Coupland 1944; Dalal 1940). Concern about separate electorates gave rise to 
extreme sensitivity to respective population figures of minority groups. 
63. Nain Sukh Das v. State of U.P., A. I. R. 1953 S. C. 384 
­375­ 
after the commencement of the Constitution.  64.   In 1959 these provisions were extended for another 
ten years,  65.   in 1969 for another ten.  66.   and in 1980 for yet another ten.  67.   
The Constitution contains no provision for representation for Other Backward Classes in legislatures 
or in other political bodies. However, such political safeguards may lie within the scope of Article 
15(4), which is apparently broad enough to cover the reservation of seats in elective bodies on a 
local level.  68.   Presumably, Article 15(4) would not authorize additional reservation of seats in 
Parliament or in the state legislatures, since representation in these bodies is controlled by Parts XV 
and XVI of the Constitution, which provide reservations only for Scheduled Castes and Tribes and 
only in the respective lower houses. But Article 15(4) allows reservation for Scheduled Castes and 
Tribes (and presumably for Other Backward Classes as well) in local bodies such as district and 
municipal boards, panchayats, etc.  69.   Ordinarily, the use of communal criteria in conducting such 
elections would run afoul of Article 15(1),  70.   but Article 15(4) acts as an exception in proper cases. 
It is not clear whether special representation sanctioned by Article 15(4) must follow the legislative 
model of reserved seats from joint electorates or whether Article 15(4) authorizes the use of 
separate electorates in filling these reservations.  71.   Article 325, the ban on communal electorates, is 
concerned not with local bodies but only with Parliament and the state legislatures. 
____________________ 
64. Art. 334. 
65. Constitution (Eighth Amendment) Act, 1959. 
66. Constitution (Twenty­third Amendment) Act, 1969. 
67. Constitution (Forty­fifth Amendment) Act, 1980. 
68. A. R. V. Achar v. State of Madras, Writ Petition No. 568, High Court at Madras, August 25, 1952, 
aff'd on other grounds A. I. R. 1954 Mad. 563, upholding reserved seats on municipal council for 
Scheduled Castes and women. Cf Dattatraya v. State of Bombay, A. I. R. 1953 Bom. 311, holding 
that Art. 15(3) authorizes such reservations for women. 
69. See Dulichand v. Union of India, 1974 M. P. L. J. 200, 203, where the court, ignoring Art. 15(4), 
holds that Art. 14 does not prevent the government from reserving places on a Cantonment Board 
for Scheduled Castes and Scheduled Tribes because they "form a class by themselves" and "a 
provision or reservation of seats for the backward classes in matters relating to election . . . 
makes a reasonable classification which has a necessary nexus with the object sought to be 
achieved." The use of the term "backward classes" here may be only as a synonym for Scheduled 
Castes and Tribes; clearly it is only provision for the latter groups that was before the court. The 
classification grounds that the court invokes might be applied to Other Backward Classes as well. 
70. Bhopalsingh v. State, A. I. R. 1958 Raj. 41; Nain Sukh Das v. State of U.P., 1953 S. C. 384. 
71. A. R. V. Achar v. State of Madras, n. 58. If separate electorates are permissible to insure 
representation for backward groups, it would presumably be equally permissible to accomplish 
the same result by delineation of wards on caste lines as is reported to be done in Orissa 
(RCSCST 1960­61: I, 20). 
­376­ 
C. THE EXCEPTIONAL CHARACTER OF COMPENSATORY DISCRIMINATION 
We have seen that in the text of the Constitution, the general principle of compensatory 
discrimination is established as a Directive Principle, but the specific provisions authorizing it are 
framed as exceptions to general Fundamental Rights. This arrangement expresses the basic tension 
between the broad purposes to be achieved and the commitment to confine the device and make it 
comport with other constitutional commitments, especially that to formal equality. 
The exceptional character of compensatory discrimination has frequently been noted by the courts. 
That Article 15(4) "has to be read as a proviso or an exception to Articles 15(1) and 29(2)" was 
evident from its history, according to the Supreme Court in Balaji v. State of Mysore.  72.   This has 
been the received characterization of Article 15(4). Thus in State of Andhra Pradesh v. Sagar, the 
Supreme Court emphasizes that as an exception, Article 15 (4) cannot be extended so as in effect to 
destroy the guarantee of equality in Article 15(1).  73.   
In Devadasan v. Union of India, the Supreme Court emphasized that Article 16(4) "is by way of a 
proviso or an exception" to Article 16(1) and "cannot be so interpreted to nullify or destroy the main 
provision." Thus its "over­riding effect" is only to permit a "reasonable number of reservations . . . 
in certain circumstances. That is all."  74.   The characterization as an exception was challenged by 
Subba Rao, J., dissenting in Devadasan. In his view, Article 16(4) "has not really carved out an 
exception, but has preserved a power untrammeled by the other provisions of the Article."  75.   An 
extension of this argument was put forward in C. A. Rajendran v. Union of India, where it was urged 
that Article 16(4) was not merely an exception engrafted on Article 16, but was itself a Fundamental 
Right granted to the Scheduled Castes and Tribes and untrammeled by any other provision of the 
Constitution.  76.   The Supreme Court's response was that Article 16(4) imposed no duty on the 
government to make reservations for these classes, but that "Article 16(4) is an enabling provision 
and confers a discretionary power on the State to make a reservation."  77.   
____________________ 
72. A. I. R. 1963 S. C. 649, 657. 
73. A. I. R. 1968 S. C. 1379, 1382. The characterization of Art. 15(4) as an exception reappears in 
State of A.P. v. Balaram, A. I. R. 1972 S. C. 1375, 1394, and Janki Prasad v. State of J. & K., A. I. 
R. 1973 S. C. 930, 937. 
74. A. I. R. 1964 S. C. 179, 187. 
75. Id., at 190. 
76. A. I. R. 1968 S. C. 507. The argument was based upon Subba Rao, J.'s dissent in Devadasan v. 
Union of India, A. I. R. 1964 S. C. 179. 
77. A. I. R. 1968 S. C. at 513. More than a decade later, when the issue of reservations had become 
fiercely controversial, a Working Group appointed by the Union Home 
­377­ 
But if the particular means are discretionary, the object is not: the Constitution explicitly declares it 
"the duty of the State" to promote the interests of the "weaker sections" and to protect them.  78.   And 
if these provisions are exceptions, they are exceptions of a peculiar sort. They do not merely carve 
out an area in which the general principle of equality is inapplicable. Rather, they are specifically 
designed to implement and fulfill the general principle.  79.   
Articles 15(4) and 16(4) are undoubtedly exceptions to the constitutional prohibition of State 
employment of the otherwise forbidden criteria of caste, religion, and so forth. But it does not 
follow that they are exceptions to the policy of equal treatment mandated by Articles 14, 15, and 16. 
In respect to the general policy of equality they represent an empowerment of the State to pursue 
substantive equality in regard to the disparities between backward classes and others. It might be 
argued that in a State generally committed to formal equality this commitment to reduce these 
disparities is an exception. Thus the State could pursue a policy of overcoming the inequalities of 
backward groups that would lie outside its mandate in dealing with the poor, the handicapped, 
veterans, or victims of disasters. But it seems clear that these disparities of circumstance and fortune 
may be addressed by policies utilizing reasonable classifications. So, what these clauses do is to 
insure State power to pursue substantive equality vis à vis certain historic formations in Indian 
society. 
It was the realization that mere provision of formal equality would not suffice to bring about the 
desired "EQUALITY of status and of opportunity" that led to the adoption of these provisions. As a 
Full Bench of the Kerala High Court observed: 
It has however been realized that in a country like India where large sections of the 
people are backward socially, economically, educationally and politically, these 
declarations and guarantees [of equality] would be meaningless unless provision is also 
made for the uplift of such backward classes who are in no position to compete with the 
more advanced classes. Thus to give meaning and content to the equality guaranteed by 
Articles 14, 15, 16, and 29, provision has been made in Articles 15 (4) and 16 (4) 
enabling preferential treatment in favour of the "weaker sections."  80.   
____________________ 
Ministry suggested that Art. 16(4) be replaced by a mandatory provision for reservation of posts 
in state and public sector employment and also in those areas of private employment where 
reservations might be enforced ( Overseas Hindustan Times, 21 June 1979). 
78. Arts. 37, 46. 
79. Cf. Gupta 1969:83, who notes that these provisions "are only to be viewed as defining the 
principle of equality as contained in Art. 14. They do not derogate from the principle of equality 
in any manner." 
80. Hariharan Pillai v. State of Kerala, A. I. R. 1968 Ker. 42 at 47­8. 
­378­ 
Indeed, as the Supreme Court has observed, guarantees of equality might by themselves aggravate 
existing inequalities. If taken literally, 
instead of giving equality of opportunity to all citizens, it will lead to glaring 
inequalities . . . [I]n order to give a real opportunity to [the backward to] compete with 
the better placed people . . . [Article 16(4) is included in the Constitution]. The 
predominant concept underlying [Article 161 is equality of opportunity in the matter of 
employment; and, without detriment to said concept, the State is enabled to make 
reservations in favour of backward classes to give a practical content to the concept of 
equality.  81.   
The tension between these commitments to non­discrimination and to substantive equalization was 
poignantly expressed by Prime Minister Nehru when he remarked in the course of the First 
Amendment debate that "we arrive at a peculiar tangle. We cannot have equality because in trying 
to attain equality we come up against some principles of equality."  82.   The textual juxtaposition of 
guarantees of equality and authorization of compensatory discrimination reflects a deeper conflict 
between different views of equality and divergent notions of the goal and scope of protective 
discrimination. While in practice these views tend to merge and overlap, it may be helpful here to 
isolate in pure form what we may conveniently label the horizontal and the vertical perspectives on 
equality and compensatory discrimination. 
In the horizontal view, the relevant time is the present. Equality is visualized as identical 
opportunities to compete for existing values among those differently endowed, regardless of 
structural determinants of the chances of success or of the consequences for the distribution of 
values. One of the dissenting judges in the Thomas case sums it up neatly when he cautions that 
Article 16(1) "speaks of equality of opportunity, not opportunity to achieve equality."  83.   In this 
view, preferential treatment is accepted as a marginal adjustment to be made where results of 
complete equality are unacceptable. Compensatory discrimination detracts from equality: it amounts 
to a kind of social handicapping to insure fair present distribution among relevant units. Thus in 
Devadasan v. Union of India, the Supreme Court emphasized that in combining these provisions it 
might strike a fair balance between the claims of the backward and the claims of other communities. 
 84.   
The relation of equality and compensatory discrimination is viewed 
____________________ 
81. Triloki Nath Tiku v. State of Jammu and Kashmir, A. I. R. 1967 S. C. 1283 at 1285. 
82. Parliamentary Debates, Vol. XII­XIII (part II), col. 9617 (29 May 1951). 
83. Gupta, J., dissenting in State of Kerala v. N. M. Thomas, A. I. R. 1976 S. C. 490 at 543. 
84. Devadasan v. Union of India, A. I. R. 1964 S. C. 179; cf. C. A. Rajendran v. Union of India, A. I. 
R. 1968 S. C. 507. 
­379­ 
very differently in what we may call the vertical perspective. In this view the present is seen as a 
transition from a past of inequality to a desired future of substantive equality; the purpose of 
compensatory discrimination is to promote equalization by offsetting historically accumulated 
inequalities. Thus, compensatory discrimination does not detract from equality in the interest of 
present fairness; rather, it is seen as requisite to the fulfillment of the nation's long­range goal of 
substintial redistribution and equalization. Not only present claims but historical deprivations and 
national aspirations are relevant. Such a view was given its most clear judicial expression in 
Viswanath v. Government of Mysore, where Hegde, J., repulsed the argument that reservations 
should be confined with the observation that counsel 
did not appear to be very much alive to the fact that there can be neither stability nor 
real progress if predominant sections of an awakened Nation live in primitive 
conditions, confined to unremunerative occupations and having no share in the good 
things of life, while power and wealth are confined in the hands of only a few and the 
same is used for the benefit of the sections of the community to which they belong. . . . 
[U]naided many sections of the people, who constitute the majority in this State cannot 
compete with the advanced sections of the people, who today have a monopoly of 
education and consequently have predominant representation in the Government 
services as well as in other important walks of life. It is cynical to suggest that the 
interest of the Nation is best served if the barber's son continues to be a barber and a 
shepherd's son continues to be a shepherd. . . . We have pledged ourselves to establish a 
welfare State. Social justice is an important ingredient of that concept. That goal cannot 
be reached if we overemphasize the "merit theory." 
Advantages secured due to historical reasons cannot be considered as fundamental 
rights guaranteed by the Constitution. The nation's interest will be best served­­taking a 
long­range view­­if the backward classes are helped to march forward and take their 
places in a line with the advanced sections of the people.  85.   
From this "long­range" perspective, justice Hegde later elaborated, the "immediate advantages of 
the Nation [in the effective utilization of talent] have to be harmonized with its long­range 
interests."  86.   
Rarely do these contrasting views of equality and compensatory discrimination appear with such 
purity and clarity; more common is 
____________________ 
85. A. I. R. 1964 Mys. 132 at 136. A more summary and less vivid version of these observations was 
delivered by Justice Hegde after his ascension to the Supreme Court in Periakaruppan v. State of 
Tamil Nadu. 
86. Periakaruppan v. State of Tamil Nadu, A. I. R. 1971 S. C. 2303 at 2309. 
­380­ 
the attempt to harmonize them. Thus in Balaji v. State of Mysore, Gajendragadkar, J., observes: 
It is obvious that unless the educational and economic interests on the weaker sections 
of the people are promoted quickly and liberally, the ideal of establishing social and 
economic equality will not be attained. . . .  87.   
Surely the State is authorized to take "adequate steps" toward that objective, but these special, 
exceptional provisions do not override the Fundamental Rights of others. Furthermore, there are 
other crucial national interests which have to be taken into account: 
The interests of weaker sections of society which are a first charge on the States and the 
Centre have to be adjusted with the interests of the community as a whole.  88.   
The inevitable weighing and balancing is rendered particularly difficult because the constitutional 
provisions set up another tension as to the relevant units whose interests are to be balanced or who 
are ultimately to be equalized, as the case may be. The Constitution confers Fundamental Rights on 
individual citizens in their personal capacity, not as members of communal groups. All citizens have 
a fundamental right that another, excepting a member of a backward class, shall not be preferred by 
the State on the basis of his membership in a particular group. However, the government is obliged 
to advance the weaker sections or backward classes. Therefore the scope of compensatory 
discrimination involves tension between individuals or groups as objects of State policy. 
The constitutional embrace of the antagonistic principles of equal treatment and compensatory 
discrimination, individual rights and group rights, confronts both government and courts with the 
problem of reconciling them in specific settings. The sweeping language of Articles 15(4) and 16(4) 
indicates that their framers relied primarily on the discretion of the politicians and administrators of 
the future, rather than on the courts, to effect such a reconciliation. But while these provisions give 
the executive and the legislatures broad discretion in their application, judicial intervention is not 
entirely excluded. Leverage for judicial oversight is supplied by the placement of these provisions as 
exceptions to the judicially enforceable Fundamental Rights. These rights can only be vindicated to 
the extent that the courts scrutinize the government's designation of backward classes to see that 
only the backward are included, that the extent or method of operaiton does not prejudice others 
unduly, that the schemes are designed and 
____________________ 
87. A. I. R. 1963. S. C. 649 at 661. 
88. Id., at 663. 
­381­ 
administered to work in favor of the intended beneficiaries and not to their detriment. Such review is 
necessary not only to vindicate individual rights but to effectuate the policy of these provisions ­­ by 
preventing unwarranted dilution of benefits (for the more unrestrained the inclusion of' 
beneficiaries, the less the assistance the intended beneficiaries will receive) and abuses that 
undermine public support for these measures. 
As we shall see, the courts have indeed played a major role in shaping policy in this area by defining 
the constitutional boundaries of preferential treatment. Recent developments in the constitutional 
doctrine of equality indirectly raise the quesiton of whether courts might play an even more central 
role in the design and implementation of compensatory policies. 
D. A NEW CONSTITUTIONAL VISTA? THOMAS AND AFTER 
For the first quarter­century of constitutional development, the approaches discussed in the previous 
section bounded the discourse about the way in which competing commitments to formal equality 
and compensatory discrimination might be combined. But much of the earlier understanding of 
constitutional policies of compensatory discrimination is cast into doubt by a remarkable 1975 
decision of the Supreme Court in State of Kerala v. N. M. Thomas.  89.   Since there will be many 
occasions to refer to this decision, I will set out the story here.  90.   As will be evident, it is far from 
clear what it means or what its reception will be. 
Employees of the Registration Department of the State of Kerala were divided into Lower Division 
Clerks and Upper Division Clerks. The former could be promoted to the higher position on a 
"Senioritycum­merit" basis. To qualify for promotion it was necessary to pass some tests ­­ 
"Account Test . . . Lower Kerala Registration Test and . . . Test in the manual of office procedures" 
­­ this was the merit 
____________________ 
89. A. I. R. 1976 S. C. 490. 
90. The version of the facts here is a composite drawn from the accounts provided at various places 
in the seven opinions. I noted no major discrepancies in these accounts. See A. I. R. 1976 S. C. at 
493­95, 496­97 to 501, 502­04, 520, 523. One factual question puzzled me: what would happen 
to these clerks at the expiration of the extension period if they had not passed the tests? My 
initial reading was that during the period that they were incumbents in the higher posts, they 
were eligible for promotion on seniority, and since they were exempt from passing the tests, they 
would be confirmed in the higher posts on grounds of seniority. But the judges are unanimous 
that this is not the case, that the extension is only temporary, and that the tests must be passed 
before they can be confirmed in the higher posts (Ray, C.J., at 502, Mathew, J., at 520, Beg,.J. at 
523, Fazl Ali, J. at 544). 
­382­ 
requirement. Among those who satisfied this prerequisite, the promotions went to the most senior 
Lower Division Clerks. The rules allowed for temporary appointments to the higher posts for a two­
year period during which the clerk would have to pass the required tests; the two years was extended 
to four years in the case of Scheduled Caste and Scheduled Tribe clerks. Nevertheless, a number of 
SC clerks had not satisfied the test qualifications within the extended period and were facing 
reversion to the lower posts. 
In 1972 the state government promulgated a new rule, 13AA: 
Notwithstanding anything contained in these rules, the Government may, by order, 
exempt for a specified period, any member or members, belonging to a Scheduled Caste 
or a Scheduled Tribe, and already in service, from passing the tests. 
On the same day the state promulgated an order granting Scheduled Castes and Scheduled Tribes 
already in service "temporary exemption . . . from passing all tests . . . for a period of two years." In 
1974 this was extended for a further period to insure each employee two chances to appear for the 
required tests. This time the government ordered that "these categories of employees will not be 
given any further extension of time to acquire the test qualifications." 
Because of earlier difficulties with the test barrier, there was a heavy concentration of SC Lower 
Division Clerks with high seniority. When the test barrier was removed temporarily, many of them 
were promoted to the higher posts. Thus in 1972, of 51 vacancies in the category of Upper Division 
Clerks, 34 were filled by SCs who had not passed the tests, and only 17 were filled by persons who 
had passed the tests. A writ petition was filed by N. M. Thomas, a Lower Division Clerk who did 
pass the test and would have been promoted but for the extensions authorized by Rule 13AA. 
A Division Bench of the Kerala High Court concluded: "What has been done is not to reserve . . . 
posts . . . . [R]eservations had already been made . . . . What has been attempted by Rule 13 AA is to 
exempt persons from possessing the necessary qualifications." Such exemption lies beyond the 
scope of Article 16(4)'s authorization of reservations, and on the scale it is done here, directly 
violates Article 335's directive that claims of Scheduled Castes and Tribes may be taken into 
account in government employment "consistently with the maintenance of efficiency of 
administration."  91.   On appeal to the Supreme Court, counsel for the state took an innovative talk 
and argued that the extension 
____________________ 
91. Thomas v. State of Kerala, I. L. R. 1974 (1) Ker. 549 at 556­57. The quantum of preference 
aspects of the case are discussed in chap. 12, § B, below. 
­383­ 
need not be subsumed under Article 16(4)'s provision for reservations, but could be justified as a 
reasonable classification under Article 16(1).  92.   A seven­judge bench decided five to two to reverse, 
issuing seven separate opinions.  93.   One of the majority judges thought that Article 16(4) rightly 
interpreted would authorize the state's provision; the other four accepted some version of the broad 
classification argument advanced by the state.  94.   
Chief Justice Ray's opinion for the majority sketches the outline of the classification argument: 
providing equal opportunity in government employment is a legitimate objective; Article 46 directs 
the State to promote the economic interests of Scheduled Castes and Tribes with special care; 
Article 335 directs the State to take into their consideration their claims regarding service under the 
State. Thus the classification of employees belonging to these groups to afford them an extended 
period to pass tests for promotion is "a just and reasonable classification having rational nexus to the 
object of promoting equal opportunity . . . relat­ 
____________________ 
92. In the High Court, counsel for the state had argued unsuccessfully that the state's action fell 
within the meaning of "reservation" in Art. 16(4). 1. L. R. Ker. 1974(1) at 561. Counsel had also 
advanced an ingenious textual argument that if the reservations were covered by Art. 16(4), then 
preference in government employment apart from reservations was authorized by Art. 15(4), but 
the court held that Art. 16(4) covered the employment area and its provisions exhausted the 
goverment's power to deviate from the guarantees of Arts. 16(1) and 16(2) (id., at 557). There is 
no indication that counsel urged that Arts. 16(1) or 16(2) themselves conferred power to make 
such arrangements. 
93. State of Kerala v. N. M. Thomas, A. I. R. 1976 S. C. 490. The appeal was argued just after the 
onset of Mrs. Gandhi's Emergency Rule, and the decision was announced on September 19, 1975, 
at the end of three months of emergency rule and at the height of optimism that it heralded an 
"egalitarian breakthrough." 
94. Although Thomas marks a sharp departure from earlier doctrinal analysis of the special 
provisions for backward classes, it was not without precursors and anticipations. Thus 
classification standards were used to permit reservations for various deserving groups in medical 
admissions. See Chanchala v. State of Mysore, A. I. R. 1971 S. C. 1762 and other cases discussed 
in chap. 12, § D, below. For example, Sardool Singh v. Medical College, A. I. R. 1970 J. & K. 45, 
applied general Art. 14 classification standards to permit reservations in medical admissions for 
children of defence personnel, noting that Art. 14 permits such classification quite apart from the 
specific provisions for reservations in Arts. 15 (4) and 16 (4). Accord: Kushma Joshi v. Pro­Vice 
Chancellor, A. I. R. 1969 J. & K. 136. Reservations of government posts for ex­servicemen were 
upheld on a similar ground in Daya Ram v. State of Hatyana, A. I. R. 1974 P. & H. 279. The 
classification argument might be made where the beneficiaries would come under Art. 15(4) or 
16(4), as in Raju v. Chief Electoral Officer, A. I. R. 1976 Guj. 66, decided nine months before 
Thomas. The High Court had no difficulty in justifying a lower electoral deposit for Scheduled 
Castes and Tribes on Art. 14 grounds. See discussion at n. 54, above. The extent to which 
classification principles were perceived as available to justify compensatory preference is 
suggested by the remarks of Chief Justice Hidayatullah in Golak Nath v. State of Punjab, A. I. R. 
1967 S. C. 1643 at 1706, discussed in n. 24, above. 
­384­ 
ing to [public] employment."  95.   The difference in condition of these groups justifies differential 
treatment; just as rational classification is permissible under the general equal protection provision 
of Article 14, so it is permissible to treat unequals unequally under Article 16.  96.   
The extent of the doctrinal innovation here can be appreciated by considering the opinion of Justice 
Beg, the only member of the majority who does not participate in the reconceptualization of Article 
16. He helpfully restates a conventional understanding of the constitutional provisions. In this view 
"the guarantee contained in Article 16(1) is not by itself aimed at removal of social backwardness 
due to socio­economic and educational disparities produced by past history of social oppression, 
exploitation, or degradation of a class of persons."  97.   Instead, "it was in fact intended to protect the 
claims of merit and efficiency . . . against incursions of extraneous considerations."  98.   And 
efficiency tests, in turn, "bring out and measure . . . existing inequalities in competency and capacity 
or potentialities so as to provide a fair and rational basis for justifiable discrimination between 
candidates."  99.   Thus provisions for equality of opportunity are meant to insure "fair competition" in 
securing government jobs; they are not directed to "removal of causes for unequal performances." 
 100.   But such provisions do not stand alone: they are juxtaposed with Articles 46 and 335, which 
imply "preferential treatment for the backward classes" to mitigate the rigor of equality in the same 
sense of strict application of uniform tests of competence. 
Article 16(4) was designed to reconcile the conflicting pulls of Art. 16(1), representing . 
. . justice conceived of as equality (in the conditions of competition) and of Articles 46 
and 335, embodying the duties of the State to promote the interests of the economically, 
educationally and socially backward, so as to release them from the clutches of social 
injustice.  101.   
Thus Article 16 (4) may be thought to "exhaust all exceptions made in favour of backward classes." 
 102.   Yet the effect of the Kerala promotion rules here is "a kind of reservation," for it is a temporary 
promotion that would be confirmed only if the appointee satisfied specified tests within a given 
time. These rules may be viewed as "implementation of a policy of qualified or partial or 
conditional reservations" which could "be justified under Article 16(4)."  103.   
Dismissing the interpretation of Article 16(4) as an exception to Article 16(1), justice Mathew 
articulates the view of equality that 
____________________ 
100. Id., at 522. 
101. Id., at 522. 
102. Id., at 522. 
103. Id., at 524. 
95. A. I. R. 1976 S. C. at 500. 
96. Id., at 502. 
97. Id., at 522. 
98. Id., at 522. 
99. Id., at 522. 
­385­ 
implies the doctrinal shift. The equality of opportunity guaranteed by the Constitution is not only 
formal equality with fair competition, but "equality of result."  104.   In order to assure the 
disadvantaged "their due share of representation in public services" the constitutional equality of 
opportunity was fashioned "wide enough to include . . . compensatory measures."  105.   Thus the 
guarantee of equality "implies differential treatment of persons who are unequal."  106.   Article 16(1) 
is "only a part of a comprehensive scheme to ensure equality in all spheres."  107.   It implies 
"affirmative action" by government to achieve equality ­that is, "compensatory state action to make 
people who are really unequal in their wealth, education or social environment. . . equal."  108.   
If equality of opportunity guaranteed under Article 16(1) means effective material 
equality, then Article 16(4) is not an exception to Article 16(1). It is only an emphatic 
way of putting the extent to which equality of opportunity could be carried viz, even up 
to the point of making reservation.  109.   
Thus 
[t]he state can adopt any measure which would ensure the adequate representation in 
public service of the members of the Scheduled Castes and Scheduled Tribes and justify 
it as a compensatory measure to ensure equality of opportunity provided the measure 
does not dispense with the acquisition of the minimum basic qualification necessary for 
the efficiency of administration.  110.   
Justice Krishna Iyer propounds a complex vision of the constitutional commitment to equality. 
Interpreting the Constitution by "a spacious, social science approach, not by pedantic, traditional 
legalism,"  111.   he proposes to erect a "general doctrine of backward classification" to pursue "real, 
not formal, equality."  112.   According to the doctrine of backward classification, the State may, for 
purposes of securing genuine equality of opportunity, treat unequals equally. Thus Article 16(4) 
serves not as an exception [to the strictures of Article 16(1) and (2)] but as an emphatic 
statement, one mode of reconciling the claims of backward people and the opportunity 
for free competition the forward sections are ordinarily entitled to . . . .Closely examined 
it is an illustration of a constitutionally sanctified classcification.  113.   
____________________ 
104. Id., at 518. 
105. Id., at 518. 
106. Id., at 516. 
107. Id., at 519. 
108. Id., at 516. The idea of affirmative action is attributed to the United States Supreme Court, citing 
cases involving the unconstitutionality of the poll tax and the right of criminal defendants to a 
free transcript and counsel on appeal. 
109. Id., at 519. 
110. Id., at 519. 
111. Id., at 525. 
112. Id., at 529. 
113. Id., at 535. 
­386­ 
So, in addition to reservations provided by Article 16(4), the State may also confer "lesser order[s] 
of advantage" on the principle of classification under Article 16(1).  114.   
At this point in the argument there is a crucial divergence between the views of Justice Krishna Iyer 
and those of Justice Mathew. For the latter, the compensatory measures authorized by Article 16(1) 
might be extended to "all members of the backward classes,"  115.   not only to the Scheduled Castes 
and Scheduled Tribes. But justice Krishna Iyer's more complex vision of Article 16 contains a 
second layer: the power of classification outside the boundaries of Article 16(4) for purposes of 
overcoming inequality may be used only on behalf of Scheduled Castes and Tribes. 
Article 16(4) covers all backward classes, but to earn the benefit of grouping under 
Article 16(1) based on Articles 46 and 335 . . . the twin considerations of terrible 
backwardness of the type harijans have to endure and maintenance of administrative 
efficiency must be satisfied.  116.   
"Not all caste backwardness is recognized" as a basis for differential treatment under Article 16(1). 
 117.   
The differentia . . . is the dismal social milieu of harijans. . . . The social disparity must 
be so grim and substantial as to serve as a foundation for benign discrimination. If we 
search . . . we cannot find any­large segment other than the Scheduled Castes and 
Scheduled Tribes. . . . [N]o class other than harijans can jump the gauntlet of "equal 
opportunity guarantee. Their only hope is in Article 16(4).  118.   
This is perplexing, for it appears that the stronger measures of reservation may be taken on behalf of 
all the backward, yet those who suffer the most terrible backwardness are the only ones entitled to 
measures of "a lesser order of advantage." 
Later he suggests that to allow the Other Backward Classes to participate in these benefits may be 
detrimental to those who are most deserving: 
[N]o caste, however seemingly backward . . . can be allowed to breach the dykes of 
equality of opportunity guaranteed to all citizens. To them the answer is that . . . equality 
is equality . . . . The heady upper berth occupants from backward classes do double 
injury. They beguile the broad 
____________________ 
114. Id., at 536. 
115. Id., at 519. 
116. Id., at 536. Although here and elsewhere in his opinion, justice Krishna Iyer uses the term 
"harijans," his explanations suggest that he is employing this as a shorthand term for both 
Scheduled Castes and Scheduled Tribes. 
117. Id., at 537. 
118. Id., at 537. 
­387­ 
community into believing that backwardness is being banished. They rob the need based bulk of the 
backward of the . . . advantages the nation proffers.  119.   
This distinction is justified because "the Constitution itself makes a super­classification between 
harijans and others, grounded on the fundamental disparity in our society and the imperative social 
urgency of raising the former's sunken status."  120.   From the provision of Articles 330, 332, 335, 
338, and others, we may deduce that "the Constitution itself demarcates harijans from others. . . . 
This is based on the stark backwardness of this bottom layer of the community."  121.   This 
constitutional differentiation of Harijans is specifically extended to the area of government 
employment as part of the State's obligation to "promote the economic interests of harijans and like 
backward classes." Articles 14 to 16 are, according to Justice Krishna Iyer, "the tool kit" to carry out 
the "testament" of Articles 46 and 335.  122.   
An attenuated form of this "super­classification" argument is found in each of the four majority 
judgments that embrace the classification argument. That argument posits a general authorization 
flowing from Article 16(1) to adopt reasonable classifications for purposes of securing equality of 
opportunity. But it is conceded that this does not include a power to employ those classifications 
specifically forbidden in Article 16(2). Therefore each of these four "classification" opinions argues 
that the Scheduled Castes and Tribes do not constitute a classification on the basis of "caste." 
Thus Justice Krishna Iyer argues that the ban of 16(2) doesn't arise in connection with measures for 
the Scheduled Castes and Scheduled Tribes, for they are "no[t] castes in the Hindu fold but an 
amalgam of castes, races, groups, tribes, communities or parts thereof found . . . to be the lowliest 
and in need of massive State aid and notified as such by the President."  123.   Article 16(4) allows use 
of these forbidden grounds to identify the backward classes that may be recipients of reservation. 
The four­judgments imply that the other kinds of compensatory treatment justified directly by 
Article 16(1) are available only to classes which avoid the classification forbidden by Article 16(2), 
including Scheduled Castes and Tribes, since they are not castes. But if the idea is to confine 
compensatory classification under Article 16(1) to SC and ST, this argument proves too much. For 
there are innumerable categories ­­ e.g., on the basis of income, occupation, physical hand­ 
____________________ 
119. Id., at 539. 
120. Id., at 532­33. 
121. Id., at 533. 
122. Id., at 533. 
123. Cf. Chief Justice Ray, id., at 501; Justice Mathew, id., at 519Justice Fazl Ali, id., at 549, 552. But 
cf. Justice Gupta, id., at 542. 
­388­ 
icap, etc. ­­ that are not based on the classifications forbidden in Article 16(2). The argument for SC 
and ST as a "super­classification" cannot be sustained on the basis of the structure of Article 16, but 
only on the basis of their special recognition in the Constitution. But if that is the argument for their 
distinctiveness, the fact that they are not "castes" within Article 16(2) is a distracting irrelevance. 
In this and other matters, the Thomas opinions leave open many perplexing questions. Are 
equalizing measures permissible even if they employ the categories forbidden by Article 16(2)? Or, 
conversely, is the power to adopt equalizing crassifications under Article 16(1) to be used 
exclusively to address disparities along the dimensions listed in Article 16(2)? How about 
differences in class? income? Is the doctrine of classification merely permissive, so that government 
may make such compensatory classifications, but need not? Suppose it fails to do so? Do 
classifications that do not take account of inequalities violate the Article 16(1) guarantee of equality 
of opportunity? 
Yet Thomas is welcome because it makes reflection on compensatory treatment unavoidable. By 
revealing the constitutional indeterminacy and doctrinal disarray in this field, it poses new 
challenges forjurisprudence and policy. Where courts could once rely without much thought on 
quotations from Balaji or Sagar, they must now articulate their choices. Actors, governmental and 
private, can no longer assume that the categories of compensatory policy are immutable. 
In terms of doctrinal housekeeping, there was little need for the Thomas reconceptualization of 
Article 16. "Reservation" in Article 16(4) could readily have been construed to accommodate the 
Kerala scheme (as it was by Justice Beg). Such a construction would usefully have clarified the 
status of the age­waivers, fee concessions, travel allowances, coaching schemes, lowering of 
minimum marks, and other provisions that typically accompany reservations and often exist apart 
from reservations per se. If the State enjoys the plenary power to reserve places for a beneficiary 
group, does this not imply a power to take lesser measures to help the members of that group obtain 
places? In order to join issue with the large questions, the Thomas majority (the four) had to accede 
to a very narrow definition of "reservation" so that all of these practices are now exposed to 
challenge and possibly to the necessity of securing justification on broad classification principles, a 
course which may be strewn with conceptual land mines. Thus the radical reconceptualization of 
Article 16 arrives on the agenda at the cost of an unimaginatively narrow reading of what common 
sense would regard as the most relevant constitutional provision, Article 16(4). 
A generous construction of Article 16(4) not only would have addressed the problem in Thomas, but 
would have been readily comprehensi­ 
­389­ 
ble to all the different groups concerned with reservations in government posts ­­ officials in charge 
of appointments and promotions, government servants, and aspirants to government posts. Thomas 
reaches the same immediate outcome by elaboration of ambiguous and recondite doctrine that is not 
readily accessible to officials who must design and manage programs of preference, to the 
beneficiaries of such programs, to disappointed non­beneficiaries ­­ and in large measure is not 
readily accessible even to lawyers and judges. 
Perhaps, though, the accomplishment of Thomas lies in its providing government with more ample 
means to pursue compensatory policies. The doctrinal obstacles that might have impeded 
governmental policies in this area stood for the most part on weak ground, unlikely to withstand 
sustained conceptual attack.  124.   Openings for favorable reinterpretation were plentiful. To bestow 
on government more ample authority to do something does not automatically mean that more of it 
will get done. Was the critical shortage one of doctrine favorable to such policies? or the will, 
energy, and competence to implement them? 
Consider the situation of the Kerala registration clerks in Thomas. The state's scheme here seems 
less a carefully calculated modification of job requirements than a desperate improvisation 
amounting to a confession of failure of its earlier policies. Extensions hadn't worked before, so 
another one is tried. There is no indication that any thought was given to (1) some way of helping 
the Scheduled Caste Lower Division Clerks prepare to pass these tests; nor to (2) modifying the 
tests to eliminate cultural biases or extraneous matters and measure qualities genuinely needed for 
the job; or (3) modifying the job to make it suitable to candidates with the qualifications of these 
applicants. The immediate thrust of the decision is to enlarge the State's authority to confer 
preferential treatment. But the failure of the earlier measures was due not mainly ­­ perhaps not at 
all ­­ to lack of State authority, but to lack of will or capability to make its schemes work. Enlarging 
State authority will not necessarily supply the lack. Indeed, it may allow the State to substitute easy 
shortcuts (like the one here) for thinking about how to do the job effectively. Since it may now in 
effect decree the result of keeping these clerks in the higher posts, it may have less incentive to 
devise ways of motivating or enabling them to grasp these opportunities and improve their 
performance, thus prolonging their lack of qualification and reinforcing their dependence. And, as 
the State's authority is more broadly defined, there will be fewer occasions for the courts to 
____________________ 
124. E.g., the quantitative restrictions imposed by Devadasan, discussed in chap. 12, § B, below, or 
the restrictions on the use of layers of preferential treatment, discussed in chap. 13, § C, below. 
­390­ 
observe state programs and to monitor and energize government performance. 
Of course, it is not only a question of the capability of the State, but also of the recipients of 
preferential treatment. The argument in favor of the extension of the Scheduled Caste clerks in the 
upper­division posts is that it will somehow motivate or equip them to pass the test eventually. And 
unless they are motivated and enabled to take the initiative and improve their own capacity to 
perform, the extension (and any compensatory measure for that matter) is only a temporary 
palliative. There is no indication that the State of Kerala in any way enhanced its ability to elicit a 
satisfactory performance from these clerks.  125.   
Thus Thomas does not offer any hope of breaking out of the pattern of patronage and dependence. 
Both majority and minority judgments visualize the Scheduled Castes as the passive recipients of 
governmental largesse, rather than as active participants in their own improvement. And for all its 
apparent radicalism, Thomas enlarges State power in a way that may jeopardize the future of 
compensatory preference for the backward classes. In two senses this enlargement is a false victory 
for the Scheduled Castes. 
First, the Court passed up the chance to make the modest contribution of usefully expanding the 
meaning of reservations to align doctrine with existing practice about age waivers, coaching, etc. In 
taking the more radical course, the Court attempts to make legal doctrine yield up benefits that it 
cannot yield. Doctrine can give only authority, not power. And the crucial shortage was not of State 
authority, but of the will and capacity of the State to deliver benefits (and of the recipients to utilize 
those opportunities to enlarge their capabilities). The result is a symbolic breakthrough in which 
Scheduled Castes, their wellwishers, and wider publics are beguiled into thinking that much (or too 
much) is being done for Scheduled Castes. 
Second, the new reading of equality may detract from the attention and priority accorded to the 
backward classes. The new equality doctrine is so ample that it sweeps in its path all that confines 
the commitment to compensatory treatment to specific historic groups. If Articles 14­16 proclaim a 
regime of substantive equality and if the State may employ classification to remedy any falling short 
of the equality thus mandated, the government's responsibility to confer compensatory preference is 
vastly larger than it has hitherto been understood. It is a responsibility that runs not only to 
Scheduled Castes and Tribes and 
____________________ 
125. I am indirectly informed that by June 1978, most of the Scheduled Caste clerks whose promotion 
sparked the Thomas case had been provisionally certified as Upper Division Clerks upon passing 
the departmental test. 
­391­ 
to backward classes more widely conceived, either in terms of social groups or the poor; it also 
includes those who sufer difficulties as a result of personal misfortune (disaster victims), accidents 
of personal history (the physically handicapped) or as a result of meritorious service to the nation 
(ex­servicemen or dislocated children of diplomats). 
In a setting of chronic shortage, an enlarged commitment to remedy all undeserved difficulties 
betokens a commendable generosity of spirit. But it also raises the question of priorities and of 
allocation of scarce resources, including attention. Government's authorization to pursue substantive 
equality is vastly greater than the resources that will conceivably be available to it. Among the 
claimants on its compensatory powers will be many who are better placed to press their claims on 
the attention and sympathies of government. Will not the commitment to the lowest social groups­­
especially where these are perceived to receive massive benefits­­be overwhelmed by governmental 
response to better­placed claimants on its compensatory attentions? 
Of course, the State could always take account of these difficulties under Article 14 in terms of 
reasonable classification. But the notion that the State has a general obligation to produce 
substantive equality means that the kinds of disadvantage that afflict the better­off­­diplomats, 
central government employees, retired army officers, physically handicapped children in well­to­do, 
educated families­­are now elevated, as far as compensatory responsibilities go, to parity with the 
government's commitment to overcome disparities associated with the traditional social hierarchy. 
The sense of a regime of formal equality qualified by a singular exception to alleviate disparities 
derived from position in the traditional social hierarchy is liquidated or dissolved into a general and 
unfulfillable commitment to substantive equality. 
The distributive potential of this new dispensation is dramatically realized in Jagdish Rai v. State of 
Haryana, where the state reserved a substantial portion of government posts for ex­servicemen on 
the ground that they were handicapped because "over the years [they] have lost opportunities for 
entering government service and have also lost contact with ordinary civilian life."  126.   The Full 
Bench sweeps aside the notion that reservations have to be justified by Article 16(4) as "a relic of 
the old way of thinking . . . . The old idea has now given way to the idea that [Article 15 (4) and 
Article 16 (4)] . . . are themselves aimed 
____________________ 
126. A. I. R. 1977 P. & H. 56, 61. Since India has no conscription, these were volunteers. Two 
separate reservations figured in this case: the Division Bench addressed a 50% reservation for 
"Ex­Emergency Commissioned Officers" as demonstrators in a Dental College (id., at 57). On 
reference to a Full Bench, attention shifted to a reservation of 28% of posts of sub­inspectors in 
the Food and Supplies Department for "released Army Personnel" (id., at 58). 
­392­ 
at achieving the very equality proclaimed and guaranteed by Article 14 and other clauses of Articles 
15 and 16."  127.   After extensive citation from Thomas, which it commends for having "got rid of the 
old sterility" and "introduced a new dynamism and a new dimension into the concept of . . . equality 
of opportunity," the court observes that "[i]t is no longer necessary to 'apologetically' explain laws 
aimed at achieving equality as permissible exceptions. It can now be boldly claimed that such laws 
are necessary incidents of equality."  128.   Reservation of posts for ex­servicemen is justified, for they 
suffer difficulties in competing with civilians for civilian jobs, and the state has an obligation to 
provide them employment. Thus the state is justified in classifying them separately as a source of 
recruitment.  129.   To secure a just proportion of posts to those who suffer a peculiar handicap in 
competition is "an extension of the principle of Article 16(4) to those that do not fall under Article 
16(4)."  130.   
In effect, other deserving groups are now entitled to reservations along with the backward classes 
specified by Article 16(4), and in effect all the kinds of preferential treatment which seemed to be 
allowed exclusively to the backward classes may now be bestowed on other groups regarded as 
deserving by the State. It is a fitting symbol that the losing petitioner in Jagdish Rai was herself a 
Scheduled Caste. 
Thus Thomas opens Pandora's box: compensatory classification is available­­perhaps incumbent­­to 
succor all the disadvantaged. The earlier notion that "the interests of the weaker sections are a first 
charge"  131.   on government is dissolved into a diffuse, and in the nature of things, largely symbolic, 
generalized egalitarianism. Justice Krishna lyer's doctrine of super­classification may be understood 
as an attempt to close the lid again, confining compensatory classification to Scheduled Castes and 
Tribes. Were this doctrine to gain acceptance, State authorization to confer benefits on these groups 
might end up not very far from where liberal reading of Articles 15(4) and 16(4) would have left it. 
Super­classification would re­establish the priority of Scheduled Caste claims by emphasizing a 
picture of Indian society as riven by an unbridgeable dichotomy between Scheduled Castes and 
Tribes and the rest of the population. This dichotomous picture, by drawing a rigid line between 
Scheduled Castes and Other Backward Classes, could impede program administration and obstruct 
the even­ 
____________________ 
127. Id., at 58. 
128. Id., at 60. 
129. Id., at 61. The "source" language is borrowed from Chanchala v. State of Mysore, A. I. R. 1971 S. 
C. 1762 (discussed in chap. 12, § D, below). This borrowing confirms the suspicion that the 
distinction between making reservations and specifying sources is a thin one indeed. 
130. Id., at 61. 
131. Balaji v. State of Mysore, A. I. R. 1963 S. C. 649 at 651. 
­393­ 
tual dismantling of preferential treatment by preventing merger of these categories. More 
immediately, it would further stigmatize Scheduled Castes and Tribes by portraying them as 
uniquely hapless and helpless specimens, a potential demonstrated in the overtones of 
condescension found in some of the Thomas opinions. 
But it would be shortsighted to judge Thomas only in terms of what it does to governmental power 
in this area. There is another side to the coin and one that is potentially of even greater significance. 
If Articles 14, 15, and 16 are read as mandating the pursuit of substantive equality, then to what 
extent is substantive equality­­or at least governmental efforts to promote it­­an enforceable 
Fundamental Right? Since those denied Fundamental Rights may resort to the courts to enforce 
them, would the potential beneficiaries of such equalizing measures have a right to resort to the 
courts to secure governmental compliance with this right of equalization? 
It would not be surprising if the courts would shrink from affirmative enforcement of these 
reconceptualized rights to equality. But imagine for a moment that they were willing to do so. 
Scheduled Castes and others would not have to wait for government preference in their behalf, but 
could take the initiative in the courts to secure the "enforcement" or implementation of their rights 
to substantively equalizing measures. Scheduled Caste movements could mobilize around these 
issues, generating the kind of political movement that would make government responsive to 
judicial proddings. 
We shall in a moment mention the doubts that surround this scenario. But to emphasize it once 
more, in this reading the ultimate significance of Thomas is not the enlargement of State authority 
to confer preferential treatment, but the acknowledgment of a Fundamental Right to substantive 
equality and the possibilities for affirmative litigation by disadvantaged groups to force the State to 
fulfill its responsibilities. This would amount to an ironic reversal of the meaning of Thomas. We 
started with the obvious view that Thomas loosened judicially imposed restraints on government, 
allowing it to patronize the least advantaged; we end by viewing Thomas as suggesting the 
imposition on government of a new and onerous accountability to these disadvantaged, an 
accountability mediated through the courts. 
This scenario is subject to a number of contingencies. First, do the litigants and their lawyers have 
the legal imagination to devise these claims? Second, do they have the capacity to organize to 
sustain such litigation and press for the implementation of favourable rulings? (Of course, the 
capacity for organization might be enhanced by creative deployment of legal services.) Third, are 
the judges likely to be responsive to such claims, claims that ask them to depart from deeply held 
­394­ 
notions about the judicial role? They would not only have to innovate original standards of what is 
appropriate movement toward equality, they would have to undertake a sustained activist monitoring 
of government that they may find both ideologically uncongenial and institutionally discomfiting. 
 132.   (Were they to shrink from this, while maintaining an interpretation of equality rights as 
substantive, Thomas would, in the guise of exalting the Directive Principles, have helped to demote 
these Fundamental Rights to the non­enforceable status that led to the disdain of the Directive 
Principles.) Fourth, even if claimants come forward and courts are responsive, do the realities of 
litigation in India the delay and maneuver and cost, make it possible for groups to improve their 
position through litigation against powerful adversaries? This is related to a fifth contingency: are 
the real barriers to improvement of the conditions of these groups reachable by even most well­
disposed and capable judiciary? To what extent is the notion of remedying these problems through 
litigation yet a further tempting illusion? We may be better able to address these questions after we 
have followed the course of the first quarter­century of litigation about compensatory 
discrimination. 
It is not clear yet in what ways and how deeply Thomas will transform earlier doctrine and practice. 
It was pronounced by a divided court; it expresses the heated symbolic egalitarianism that was both 
institutionalized and discredited during Emergency Rule. Although it provides no satisfying 
answers, its presence permits and requires some fresh thought about the principles underlying 
compensatory discrimination policies 
____________________ 
132. On the accomplishments, problems, and discomforts of the American counterpart of this kind of 
activist "public law" judging, see Chayes 1976; Horowitz 1977 and Galanter, Palen, and Thomas 
1979. On its divergence from the Indian judicial style, see chap. 14, §A, below. 
­395­ 
12 The Scope and Extent of Preferences: Constitutional Limits and Judicial Control 
A. HOW MUCH RESERVATION? THE MINIMUM 

ALTHOUGH there is a general constitutional duty to protect and advance the interests of the 
"weaker sections," the Constitution does not explicitly obligate the government to employ any 
particular method of doing this. There is no provision in the Constitution that the State need reserve 
any minimum number of posts in government service seats in educational institutions; nor divert 
any minimum part of its resources to benefits for backward groups. Except for reserved seats in the 
legislatures, the quantum of preferential treatment is unspecified. 
The courts have been unresponsive to claims attempting to establish some minimum quantum of 
compensatory discrimination. We saw earlier that courts were unwilling to admit as justiciable 
claims for inclusion by groups which claimed to be rightfully included within the Backward Classes 
1. Similarly, the courts have been unwilling to vindicate the claims of 
or Scheduled Castes.    
admittedly proper beneficiaries who complain that the provisions for preference made by the 
2. the 
government are insufficient in quantity and scope. In C. A. Rajendran v. Union of India,    
central government had confined reservations in promotions to Class III and Class IV posts. A 
Scheduled Castes employee in Class III claimed that the limitation to inferior posts violated the 
guarantee afforded by Article 16(4). He argued that Article 16(4) was not merely an exception 
engrafted on Article 16, but was itself a fundamental right, granted to Scheduled Castes and Tribes, 
which places on the government the duty of making reservations for these groups. The Supreme 
Court repelled this contention with the observation that 
Article 16(4) does not confer any right on the petitioner and there is no constitutional 
duty imposed on the Government to make a reservation for 
____________________ 
1. See chap. 8, §I. 
2. A. I. R. 1968 S. C. 507. 
­396­ 
Scheduled Castes and Scheduled Tribes, either at the initial stage of recruitment or at the stage of 
promotion. In other words, Article 16(4) is an enabling provision and confers a discretionary power 
3. 
on the State to make a reservation. . . .    
Other courts have concurred that Articles 15(4) and 16(4) do not confer on backward groups any 
Fundamental Right to such arrangements.     4. These provisions represent an exception to the rights 
that others would enjoy in their absence to complain of such arrangements as violative of the 
Fundamental Rights granted in Articles 15, 16, and 29; they thereby empower the government to 
confer preferential treatment, but they do not require it to do so. Government may constitutionally 
5. 
omit to make any particular scheme of preferences.    
The courts seem to infer that since there is no duty to confer any particular sort or amount of 
preferential treatment, there is no duty to confer any at all. In effect, they hold that a discretion 
sufficiently broad to allow a zero response to any individual claim is taken to imply a discretion to 
make a zero response to every claim. But this is somewhat paradoxical in view of the clear and 
explicit constitutional duty to make some special provision to advance the interests of the weaker 
sections. Article 46 directs: 
The State shall promote with special care the educational and economic interests of the 
weaker sections of the people, and, in particular, of the Scheduled Castes and Scheduled 
Tribes, and shall protect them from social injustice and all forms of exploitation. 
While this Directive Principle is "not enforceable in any court . . . it shall be the duty of the State to 
6. The 
apply [it] . . . in making laws."    
____________________ 
3. C. A. Rajendran v. Union of India, A. I. R. 1968 S. C. 507 at 513. This is a holding in regard to 
promotions dictum in regard to initial appointments. In regard to initial recruitment, this point 
was explicitly decided in Dasa Rayudu v. Andhra Pradesh Public Service Commission, A. I. R. 
1967 A.P. 353, 363, where the court rejected a petition seeking to require the Andhra Pradesh 
Government to restore reservations for backward classes. 
4. Arts. 15(4) and 16(4) are enabling provisions which "do . . . not impose an obligation, but merely 
leave . . . it to the discretion of the appropriate Government to take suitable action, if necessary" ( 
Balaji v. State of Mysore, A. I. R. 1963 S. C. 649 at 664). 
5. "[I]t must be remembered that Art. 15(4) and Art. 16(4) are not mandatory, but are merely 
enabling provisions. The State Government therefore may or may not make any rule or issue any 
executive direction for the purposes of reserving some posts for extending other concessions to 
backward classes . . . when the State Government can in its discretion make such rule or issue 
such administrative instructions, it can also cancel or modify it. It was not incumbent upon the 
State Government to prepare any list of backward classes, nor is it incumbent upon it now to 
prepare any list" ( Dasa Rayudu v. Andhra Pradesh Public Service Commission, A. I. R. 1967 
A.P. 353 at 363). 
6. Art. 37. 
­397­ 
commitment to preferential treatment is even more specific in the employment area. Article 335 
states: 
The claims of the members of the Scheduled Castes and of the Scheduled Tribes shall 
be taken into consideration, consistently with the maintenance of efficiency of 
administration, in the making of appointments to services and posts in connection with 
the affairs of the Union or a State. 
The statement in C. A. Rajendran that "there is no constitutional duty" seems to involve an 
unspoken equation between "constitutional duties" and duties "enforceable by a court" and a finding 
that the latter is not present. It is not clear whether the Court is suggesting that the government's 
duty is unenforceable in courts under any circumstances or whether the government's provision here 
is so comfortably above the constitutional floor that it is clearly under no enforceable duty to do 
more. This would leave open the situation where government response was absent or patently 
inadequate. Thus, for the time being at least, there is a complete absence of judicial limits of any 
kind over minimum preferential treatment. 
When government does confer preferences, the courts have been disinclined to exercise scrutiny 
over their scope or coverage. The government need not include among the beneficiaries every group 
which might conceivably qualify as backward. It may make reservations in some services and not 
others, in some grades and not in others. Thus, in C. A. Rajendran, the discrepancies in the 
treatment of one stream of promotion and another, of higher and lower grades, of one service 
differently from others, were found to be justifiable classifications which did not offend equal 
protection. The Court did not require from the government any showing of the reasonableness of 
drawing the lines where it had, finding it "obvious" that an order limiting reservations to Class III 
and IV posts was a reasonable classfication in view of the higher requirement of efficiency and 
7. 
greater responsibility of upper echelons.    
This unwillingness to oversee the coverage of preferences excludes not only beneficiaries 
complaining of their insufficiency, but also nonbackward­class employees who claim that 
preferences are unfairly concentrated to their disadvantage. In Devadasan v. Union of India, the 
Supreme Court held that employees who complained of reservation in promotions in the Central 
Secretariat whereas other agencies had no such arrangement had no complaint "merely because 
8. 
reservation is not made in every kind of service under the State."    
____________________ 
7. A. I. R. 1968 S. C.507. 
8. Devedasan v. Union of India, A. I. R. 1964 S. C. 179 at 185. 
­398­ 
B. HOW MUCH RESERVATION? THE MAXIMUM 
The Constitution does not explicitly provide any maximum or limitation on the quantum or extent of 
preferences. Article 15(4) authorizes "any special provision for the advancement" of the backward; 
Article 16(4) empowers the State to make "any provision for the reservation of appointments or 
9. In spite of this broad and unrestricted 
posts" in favour of any under­represented backward class.    
language, a doctrine of constitutional limitation has emerged in dealing with those preferences 
which take the form of percentage reservations of government posts and places in educational 
institutions. 
In Venkataramana v. State of Madras, where the question of extent was not in issue, the Supreme 
Court upheld reservations for Harijans and Backward Hindus of 19.7% of posts filled by 
competitive examination.  10.   In Kesava Iyengar v. State of Mysore, the first court to consider 
explicitly the question of extent upheld a reservation of seven out of ten posts.  11.   The Mysore High 
Court there conceded that "reservation" in Article 16(4) signified a "small fraction" or "small 
portion of the main" (presumably less than one­half). However, on the ground that "each backward 
class is an independent class whose claim for appointment can be sustained under Article 16(4)," 12. 
the court suggested that the state could reserve such a portion for each backward community and 
could, on this basis, give all the available posts to backward groups. This expansive view, which 
permits total reservations to exceed 50% and would allow reservation of all posts, is now 
discredited. 
In Triloki Nath Tiku v. State of Jammu and Kashmir [II],  13.   the Supreme Court rejected the state's 
policy of giving 50% of government posts to Kashmir Muslims, 40% to Jammu Hindus, and 10% to 
Kashmir Hindus, on the ground that "[i]n effect the State policy . . . was a policy not of reservation 
of some appointments or posts; it was a scheme of distribution of all the posts communitywise." 14. 
Such distribution was contrary to Article 16(1) and (2) and was not saved by Article 16(4). 
The chief draftsman of the Constitution, defending Article 16(4) 
____________________ 
10. A. I. R. 1951 S. C. 229. The State had specified that of83 vacancies as munsiffs, 12 be filled by 
members of the Subordinate Judicial Service and 71 from the bar, assistant public prosecutors, 
etc. on a basis of communal quotas. The quotas originally set for Harijans and Backward Hindus 
amounted to 40.8% of appointments from the bar (36.1% of all posts). The appointments actually 
made included 14 members of these groups among the 71 appointed from the bar (19.7% of this 
group, 16.8% of the total). 
11. A. I. R. 1956 Mys. 20. 
12. Id., at 24. 
13. A. I. R. 1969 S. C. 1. 
14. Id., at 4. 
9. Cf. the limitation in Art. 335 to claims consistent "with the maintenance of efficiency of 
administration." But Art. 335, unlike Art. 16(4), is not limited to "reservation" as a means of 
preference. 
­399­ 
before the Constituent Assembly, indicated that the reservation authorized was "a minority of seats" 
and gave the example of an aggregate reservation of 70% of posts as falling outside the power 
bestowed by the clause.  15.   Although the draftsman visualized judicial review of the extent of 
reservation, the text of Article 16(4) itself supplies no apparent warrant for it, unless the word 
"reservation" is construed necessarily to imply a minority (less than 50%) of posts ­­ not merely a 
minority for any single backward group, as the court in the Kesava Iyengar case construed it, but for 
the aggregate of all backward groups. 
In Balaji v. State of Mysore, the Supreme Court has attempted to supply a constitutional limit to the 
extent of preference, not on narrow technical construction of "reservation," which was not available 
to it ­­ since the reservation was under Article 15(4) rather than Article 16(4) ­­ but on broader 
grounds of policy.  16.   The Mysore scheme reserved a total of 68% of places in engineering and 
medical colleges: 15% for Scheduled Castes (SC); 3% for Scheduled Tribes (ST); and 50% for 
Backward Classes. Proceeding from the notion that Articles 15(4) and 16(4) are provisions of a 
special and exceptional character, rather than principles that can be given general operation, the 
Court finds that these provisions should not be read as completely excluding or ignoring the 
fundamental rights of all citizens. It characterized the purpose of these special provisions not merely 
as conferring special privileges on the backward, but as serving the interests of the whole society by 
promoting the advancement of its weakest elements. Thus, 
____________________ 
15. Dr. B. R. Ambedkar, VII CAD 701. 
16. A. I. R. 1963 S. C. 649. So long as the argument for the percentage limitation rested on the word 
"reservation" in Art. 16(4), it presumably would not apply to benefits under Art. 15 (4) ­­ 
including "reservations" in educational institutions. Such a distinction was apparently inferred by 
the chairman of the Backward Classes Commission. Contrast his recommendation of maximum 
49% reservation of government posts with this recommendation of a 70% reservation in existing 
educational institutions (I BCC ix, xii). Cf. Ramakrishna Singh v. State of Mysore, A. I. R. 1960 
Mys. 338, where the High Court struck down a scheme of reservation in professionl colleges by 
which Scheduled Castes and Tribes enjoyed a 20% reservation and another 45% was distributed 
among various groups designated as Backward Classes. While subjecting this scheme to 
withering criticism, the court did not see fit to object to the extent of reservations even though 
the total reservation comprised well over half of the seats. Unlike the Kesava Iyengar and 
Venkataramana cases, above, nn. 11 and 10 respectively, this case involved a scheme of benefits 
under Art. 15(4), rather than reservation of government posts under Art. 16(4). Were such a 
distinction recognized the state would have been able to reserve only 50% of the government 
posts for members of a preferred group but able to reserve for them a larger percentage of the 
educational opportunities necessary to qualify for these posts. A percentage limitation applicable 
to reservation of seats in educational institutions and on local political bodies had to rest on some 
other ground than interpretation of "reservation." 
­400­ 
any scheme for preference can be weighed in terms of the interests of the whole society. In 
considering reservations in institutions of higher education, it is necessary, then, to balance the 
policy of advancing the backward against the national interest in the full utilization of talent. 
Lowering of the quality of graduates, says the Court, is the "inevitable consequence of reservation." 
 17.   But the need for technical, scientific, and academic personnel "is so great that it would cause 
grave prejudice to national interest if considerations of merit are completely excluded by wholesale 
reservation of seats in all technical, medical, or engineering colleges or institutions of that kind." 18. 
Thus, special provisions for the backward must be within "reasonable limits": "the interests of the 
weaker sections of society which are a first charge on the States and the Centre have to be adjusted 
with the interests of the community as a whole."  19.   
Although recognizing the difficulty of adjusting these competing claims, the Court experiences no 
difficulty in finding that reservation of "practically all the seats available in all the colleges . . . 
clearly would be subverting the object of Article 15(4)."  20.   While reluctant to set definite limits, the 
Court indicates that "speaking generally and in a broad way, a special provision should be less than 
50%; how much less than 50% would depend upon the relevant prevailing circumstances in each 
case."  21.   The reservation of 68% here it finds plainly inconsistent with Article 15(4). This clear 
inconsistency with the concept of the special provision, combined with sole reliance on caste as a 
criterion for selecting these classes, was such a covert or latent transgression of the constitutional 
power as to amount to a "fraud on the Constitution."  22.   
How much less than 50% need the reservation be? The Court mentions in passing the 
recommendation of the central government, incorporating the recommendations of the All­India 
Council for Technical Education, that the total reservation (for SC and ST as well as Other 
Backward Classes) may be up to 25%, with marginal adjustments not exceeding 10% in some 
exceptional cases. The central government had specifically suggested to the Mysore Government 
that in all nongovernmental institutions the reservations should not in any case exceed 35%.  23.   In 
discussing reasonableness, the Court mentions that the University Education Commission observed 
that reservations should not exceed a third of the total number of seats.  24.   Reservations of 25% or 
35% are identified as acceptable, and reservations over 50% as clearly unreasonable, leaving an area 
of doubt between 35% and 50%. Of course, even a reservation in the acceptable range would not 
auto­ 
____________________ 
17. A. I. R. 1963 S. C. at 662. 
18. Id. 
19. Id., at 663. 
20. Id. 
21. Id. 
22. Id., at 664. 
23. Id., at 656. 
24. Id., at 662. 
­401­ 
matically be "reasonable"; it might still be attacked as unreasonable under the circumstances. The 
limits suggested by the Court include the share reserved for SC and ST. Since it seems to be 
generally accepted that these groups may be given reservations in the range of their ratio to the total 
population (about 15% and 6% respectively on an all­India basis), this would seem to leave a very 
limited scope for reservations in favor of Other Backward Classes. 
Although Balaji directs attention to the lower percentages that might, under the circumstances, be 
reasonable, the state governments and the high courts have ignored the qualifications and have 
responded by taking the 50% that Balaji mentioned as the farthest permissible limit as a flat 
maximum. The first High Court to apply the Balaji views on the permissible extent of reservations 
upheld a reservation of 30% of places in engineering and pre­medical colleges for Backward 
Classes.  25.   This reservation was in addition to reservations of 15% for SC and 3% for ST. The 
Mysore High Court there found that the reservation of 30% for Backward Classes was not 
"excessive." Backward Classes were defined in terms of income and occupation; the state contended 
that, as so defined, they are "very much more than 50% of the [state's] population."  26.   The High 
Court accepted the 50% limit mentioned in Balaji as an absolute maximum and took no notice of 
the lower estimates of reasonableness mentioned in Balaji.  27.   Seven years later the Supreme Court 
found Mysore's reservation of 48% of medical school places for backward classes (including SC and 
ST) "apparently . . . on the high side" but not shown to be "unreasonably excessive."  28.   
Similarly, the Kerala High Court later upheld a scheme reserving 50% of munsif's posts, 10% for SC 
and ST and 40% for various Backward Classes. The court noted that 50% was the upper limit but 
that within this limit "it is for the State Government to decide."  29.   Thus, while Balaji has brought 
about the end of those expansive schemes of reservation which transfromed the whole competition 
into a set of communal quotas, it has not had as restrictive an effect on the extent of reservation as 
might have been expected from an examination of the 
____________________ 
25. D. G. Vishwanath v. Government of Mysore, A. I. R. 1964 Mys. 132. 
26. Id., at 137. 
27. The Supreme Court, upholding the scheme against challenge on other grounds, did not comment 
on the quantum of reservations ( Chitralekha v. State Of Mysore, A. I. R. 1964 S. C. 1823). 
Mudholkar, J., dissenting, felt that the extent of the reservations was one of several factors which 
reduced the consideration of merit to an impermissible level. 
28. Chanchala v. State of Mysore, A. I. R. 1971 S. C. 1762 at 1770. 
29. Hariharan Pillai v. State of Kerala, A. I. R. 1968 Ker. 42 at 51. 
­402­ 
decision. The qualifications have fallen away, leaving a flat 50% rule.  30.   
In the Balaji case, the extent of the reservation was only one of several objectionable features, and 
the Court indicated that its remarks about a 50% limit were not to be taken as a precise formula. 
However, less than a year later another bench of the Supreme Court  31.   carried the limitation on 
reservations much further and made it more rigid. In Devadasan v. Union of India, the central 
government provided reservations of 12 ½% for SC and 5% for ST for promotions from Grade IV to 
Grade III posts in the Central Secretariat Service to be filled by competitive examination. The 
government applied a "carry forward rule" under which unfilled reservations from the previous two 
years were added to the reservation for the current year. A total of 45 appointments were made, of 
which 29 (64.4%) were from the SC and ST.  32.   Finding that the carry­forward rule permitted 
reservations of over 50% (in the third year), the Supreme Court, one judge dissenting, found the 
scheme unconstitutional. The majority found that the holding of the Balaji case was that any 
reservation greater than 50% was unconstitutional  33.   But their application of this "holding" to a 
very different set of facts greatly curtailed the government's power to make reservations and its 
discretion in implementing them. 
Article 16(4), says the Devadasan majority, is "a proviso or an exception" to Article 16(1) and 
cannot be interpreted to nullify or destroy the main provision. "To hold that unlimited reservation 
could be made under cl. (4) would in effect efface the guarantee contained in cl. (1) or at best make 
it illusory."  34.   Article 16(4), then, must be read to avoid obliterating the rights granted in the rest of 
the article. Thus, "the overriding effect of cl. (4) on cls. (1) and (2) could only extend to the 
____________________ 
30. As one judge baldly states it: "It is settled law that 50 percent posts have to be thrown open to the 
general public" ( Daya Ram v. State of Haryana, A. I. R. 1974 P. & H. 279). However, this 
requires a qualification, for distribution restrictions may reduce the open competition portion 
further. See below at §D of this chapter. 
31. It should be noted that although Devadasan was decided only eleven months after Balaji, the five 
judges sitting in the former included none of the five who had sat in the latter. Sixteen months 
later, in B. N. Tewari v. Union of India, A. I. R. 1965 S. C. 1430, the Supreme Court indicated 
that the Devadasan decision did not affect the current reservations for Scheduled Tribes or 
Scheduled Castes, nor the validity of the provision for "carry­over" of reserved seats from one to 
another of these groups. The bench in the latter case included three members of the Balaji court 
and a single member of the Devadasan court. 
32. Apparently, the three years' reservations of 17½% of current appointments could add up to a 
reservation of 64.4%, since there were fewer new vacancies to be filled during the third year. In 
fact, the government had reserved a total of 32 out of 48 vacancies (66.6%), but it filled only 29 
of the 32 (A. I. R. 1964 S. C. 179 at 180). 
33. Id., at 186. 
34. A. I. R. 1964 S. C. 179 at 187. 
­403­ 
 making of a reasonable number of reservation [ sic
   ] of appointments and posts in certain
    
35.
circumstances."      
A "reasonable" number is one which strikes a "reasonable balance" between the claims of the 
backward classes and those of other citizens. The notion of a "reasonable balance" between 
conflicting interests echos the Balaji case, were the Court contemplated that "steps for the 
advancement of the weaker elements" must be balanced with "the requirements of the community at 
large."  36.   However, the notion of reasonableness in Devadasan is both novel and exceedingly 
restrictive. To the Devadasan majority the purpose of reservations under Article 16(4) is to insure 
that the backward "are not unduly handicapped in securing government employment."  37.   
Where a rule makes reasonable allowance for the backwardness of members of a class . . 
. what the state would in fact be doing would provide the members of the backward 
classes with an opportunity equal to that of the members of the more advanced classes 
in the matter of appointment to public services. If the reservation is so excessive that it 
practically denies a reasonable opportunity for employment to members of other 
communities . . . it would be open . . . for a member of a more advanced class to 
complain that he had been denied equality by the State.  38.   
Thus reservations are legitimate to the extent that they provide backward classes with an 
"opportunity equal to that of members of the more advanced classes"; apparently they are excessive 
if they do more than provide equal opportunity. But such a formula raises some puzzling questions. 
What does equal opportunity mean in such a setting? Let us assume that in a service of 100 
employees there are 20 members of Backward Group X. Let us further assume that the X population 
contains the same distribution of native talent, effort, and incentive as the population as a whole, but 
that it differs in background, education, and training to an extent that in a test of objective 
performance the Xs will score substantially lower than the rest of the population. Let us further 
assume that if there are 10 promotions available in this service and "equal opportunity" to compete 
for them, that none of the 20 Xs would qualify in open "merit" competition. 
If a scheme of preference or "compensatory discrimination" is introduced into such a situation, it 
cannot be with the purpose of giving the Xs equal opportunity to compete, for this they enjoy 
already, but for the 
____________________ 
35. Id. 
36. A. I. R. 1963 S. C. at 663. Cf. 662. I am unable to find anywhere in the text of the Balaji decision 
any reference to "claims of other employees" such as is adverted to by the Devadasan majority at 
A. I. R. 1964 S. C. 187. 
37. A. I. R. 1964 S. C. at 183. 
38. Id., at 185. 
­404­ 
purpose of giving them greater opportunity to achieve success in the competition. The scheme is 
justified precisely on the ground that the deficiencies of the Xs in background and education make 
them unable to compete on terms of equality. If they have merely equal opportunity to compete for 
all posts, they will have less than an equal probability of succeeding. A scheme of reservations, 
then, is designed to increase their probability of success and it does so by curtailing the opportunity 
of others to succeed­­by denying these others the opportunity to compete for some portion of the 
available places. 
What the majority in Devadasan seems to suggest is that the maximum permissible reservation for 
the Xs is that under which both they and the others would enjoy "equal opportunity." But equal here 
cannot mean equal opportunity to enter the competition, for no reservation is needed to accomplish 
that and no reservation is compatible with that. Does it mean equal probability of success? Let us 
assume that all of our 20 Xs and 80 others compete for the promotions. If 5 of the 10 higher posts 
are reserved for the Xs, it is clear that the statistical probability of a given X attaining one of the 
higher posts is greater (25%) than the chances of one of the others (6.25%). Does the fact that the 
chanes of success are, at least theoretically, far higher for the Xs make their opportunity unequal? 
Must these opportunities be equalized by confining the reservation to one in which the successful 
candidates are likely to stand in the same proportion to those who compete? This would limit the 
reservation to the percentage of Xs already eligible to compete­­in this case 20%, or 2 of the posts­­
which would give both the Xs and the others a 10% statistical chance of success. But since the 
probability of succeeding also depends on ability to obtain the minimum qualifying marks, and this 
does not necessarily bear any relation to the extent of reservation, such equalizing of statistical 
probabilities would be a barren exercise with no relation to the actual chances of the various 
competitors.  39.   But even if it were possible to do so, it is clear that preferences are 
____________________ 
39. The discrepancy between reserved percentages and posts obtained nearly always is to the 
disadvantage of the backward groups. For example, in the year ending 31 Mar. 1962 
appointments to the Grade III of the Central Secretariat Service (Permanent Strength) were as 
follows:­ 
Scheduled
Castes 
Scheduled
Tribes 
Others 
% of Scheduled
Castes & Tribes 
By direct
recruitment 


18 

By Promotion 
10 

496 
2.17 
Total 
10 

514 
2.09 
These are the same posts involved in the Devadasan case. Figures from RCSCST 1961­62: 135. 

­405­ 
not intended to "equalize" opportunities in this fashion. The arrangement of reservations as fixed 
quotas, by making them in effect separate competitions, has been definitively rejected by the courts. 
 40.   
Another meaning of "equal opportunity" might be the notion that the reservation should be of such a 
size that equal "effort" should be rewarded by equal success, in spite of lower performance due to 
deficiencies of background and training. This kind of intuitive fairness is probably closest to the 
notion of equal opportunity. But although such fairness may be one factor taken into account in 
deciding on the extent of preferences, it is patently impossible for either the government or the 
courts to administer any rule that involves the measurement of effort or other subjective factors 
which make some as deserving as others. 
The Court adds further complexity to the meaning of equality by noting that "the reservation for 
backward communities should not be so excessive as to create a monopoly or unduly disturb the 
legitimate claims of other communities."  41.   Other communities, of course, have no claims as 
communities, nor do their members have any constitutional claims by virtue of their membership in 
such communities. Even the backward classes have no claims as communities. Except in the case of 
political safeguards, members of backward classes are entitled to special treatment as individuals; 
their membership in certain communities is only used to identify them as deserving beneficiaries. 
The right to equality and freedom from discrimination is an individual one.  42.   The only "claim" 
that members of other communities have is a right to compete as individuals in open merit 
competition; they have no "claim" for "reasonable" representation for members of their respective 
communities.  43.   The backward classes enjoy the same right to compete in open merit competiton 
and in addition a right­­where it is provided by the State­­to a selection biased in their favor. Neither 
of these competing "claims" vests in any communal group; the rights of individual members of 
backward and advanced groups cannot be affected by the number of others who succeed or fail. 
____________________ 
40. Dippala Sura Dora v. V. V. Giri, A. I. R. 1958 A.P. 724; Raghuramulu v. State of Andhra 
Pradesh, A. I. R. 1958 A.P. 129; S. A. Partha v. State of Mysore, A. I. R. 1961 Mys. 220 at 235; 
Shimoga Durga Bovi v. Returning Officer, 1962 Doabia's Election Cases 31 High Court, 1960); 
Digambar Rao Bindu v. Dev Rao Kamble, 15 E.L.R. 187 (Bombay High Court, 1958). See 
discussion in chap. 13, §B. 
41. A. I. R. 1964 S. C. at 187. 
42. Venkataramana v. State of Madras, A. I. R. 1951 S. C. 229; State of Madras v. Champakam  
Dorairajan, A. I. R. 1951 S. C. 226. 
43. Cf V. V. Giri v. D. Sura Dora, A. I. R. 1959 S. C. 1318, where a combination of reservation and 
"merit" gave both seats in a double­member constituency to Scheduled Tribe candidates. The 
Supreme Court upheld this result by construing the reservation of one seat as a guaranteed 
minimum, not a compartment. Obviously, a concern for the "legitimate claims of other 
communities" did not enter into the Court's calculations. 
­406­ 
How then are the right of all to open merit competition and the right of some to a biased or 
weighted selection to be reconciled? There is no mechanical rule that can achieve such a balance. 
Presumably, what the Court is saying is that in striking the balance there must be enough unreserved 
places to make the opportunity of non­backward individuals to gain a place on merit a substantial 
one. But it is clear that how substantial this right is for individuals bears no necessary relation to the 
percentage of places reserved; it depends at least as much on the number of places available, the 
absolute number of competitors, and the absolute number of places reserved. And, of course, it 
differs from one individual to another depending on his own ability, training, and preparation and 
that of his rivals. There is no mechanical formula by which it can be determined that this 
opportunity is "equal" between different individuals. Thus in reconciling these conflicting 
commitments more factors than mere "equality" must be taken into consideration. These might 
include judgments about the morale and incentive of employees, the full utilization of talent and 
training, the efficiency of the service, the intuitive fairness of having equal effort rewarded by equal 
success, the intangible effects of reservation on the backward classes in terms of their economic 
well­being, incentive to education, feelings of self­esteem, and political effectiveness, and the effect 
of their presence on government dealings with the disadvantaged. Striking a balance among these 
competing and elusive interests of the backward classes, the other employees and the public at large 
is a judgment of policy. The Constitution has not specified any formula for striking it. What impels 
the Court to deem this a case requiring judicial overturning of the balance struck by the government 
when it applied the "carry­forward rule"? 
Once a percentage limitation on reservations is inferred from the Constitution, some peculiar 
problems arise when a time dimension is added. Presumably, such a limit applies to the total 
number of posts or seats reserved. But, typically, the percentage of backward classes among present 
incumbents of these posts is much lower than the reserved percentage. Does the limit also apply to 
the permissible reservation among appointments currently being made? For example, if a state 
reserved 30% of Grade I posts for members of group X, may 60% of 100% or current appointments 
be reserved for the Xs until the 30% is reached? Temporary reservations of up to 100% have been 
employed by some states.  44.   Again, can the state place its 30% reservation for the Xs on a long­
term basis, so that a paucity of X candidates in earlier competitions can be redressed in later ones? 
This is, in the abstract, the problem of the "carry­forward rule." 
____________________ 
44. See RCSCST 1953: 134; RCSCST 1954: 171; RCSCST 1961­62: 139: 
­407­ 
 Justice Subba Rao, dissenting in the  Devadasan
    case, suggests that the reasonableness of a
    
reservation is to be measured by its relation to total cadre strength: a reservation is constitutional so 
long as the "privileged" groups do not make up a disproportionate part of the total complement of 
employees. For him the words "any provision" in Article 16(4) are broad enough to cover schemes 
of reservation which operate over several years and in relation to the entire strength of the cadre; 
reasonableness is not to be determined on the basis of the percentage of seats reserved on one 
particular occasion. The state may use any method, including reservation of 100% of posts until the 
prescribed level is reached.  45   But for the majority the guarantee of equality in Article 16(1) 
means that on every occaison for recruitment the State should see that all citizens are 
treated equally. . . [E]very citizen who is seeking employment . . . under the State is 
entitled to . . . an opportunity for seeking such employment . . . whenever it is intended 
to be filled. In order to effectuate this guarantee each year of recruitment will have to be 
considered by itself and the reservation for backward communities should not be so 
excessive as to disturb the legitimate claims of other communities.  46.   
It is not clear here whether the majority means "every occasion" or "each year." They are obviously 
not the same. The former standard would rule out such devices as rosters of rotation for recurrent 
vacancies  47.   and special examinations open only to the members of the backward class,  48.   even if 
designed only to fill the reservation for the given year. Subsequent courts seem to have fastened on 
the "every year" standard and to have ignored the "every occasion" language.  49.   But whichever it is, 
the decision seems to make mandatory a temporal 
____________________ 
45 A. I. R. 1964 S. C. at 192. 
46. Id., at 187. 
47. Prescription of rotational rosters (in which sequential vacancies are arranged in cycles of, e.g., 20 
and certain ones in the cycle are specified as reserved for certain groups) is a common device for 
implementing reservations in hiring. For a particularly elaborate example, see Government of 
Kerala 1960: 66. Rotational systems (apart from backward classes matters) were held not to 
violate Art. 16(1) ( Kelkar v. Chief Controller of Import and Exports, A. I. R. 1967 S. C. 839). 
Recently the Supreme Court has described the "rotational or roster system [as] a common 
adopted and easily understood method" of administering entry quotas, but held that it was not the 
only allowable method ( Chauhan v. State of Gujarat, A. I. R. 1977 S. C. 251, 261). 
48. Cf. the special examination in the Central Secretariat Stenographer's Service in Dec., 1960 
(RCSCST 1961­62: 135). Devadasan involved promotions. If the same rule were to apply to 
original recruitment, presumably special examinations to fill vacancies for Scheduled Castes and 
Scheduled Tribes could not be held­­unless government servants have interests of greater weight 
than those attempting to enter service. 
49. Thus in C. A. Rajendran v. Union of India, A. I. R. 1968 S. C. 507 at 513, the Supreme Court 
reads Devadasan as saying that each year must be considered by itself. 
­408­ 
segmentation of reservation.  50.   Each year's reservation is to be judged reasonable or unreasonable 
by itself, regardless of the extent to which reservations have gone unfilled in the past. The majority 
does not specify that "carry­forward" provisions are invalid only when they raise a current 
reservation over 50%. Presumably, such a rule would be impermissible whenever its operation 
deprived other employees of opportunities equal to those of the backward classes.  51.   
It appears, too, that this equality must prevail in every separate category of competition within the 
year's recruitment.  52.   For in this case, posts to be filled by competitive examination constituted only 
30% of the appointments for the year. Another 40% were to be filled by direct recruitment from 
those who had obtained lower ranks on the I.A.S., etc., examinations, and 30% were to be filled by 
promotions from Grade IV on the basis of "seniority­cum­fitness." Even if we assume that these 
other methods of selection were accompanied by the reservation applicable to "direct recruitment" 
(16 2/3% for SC and 5% for ST), the actual percentage of posts reserved during the year may have 
been only 37.56% (assuming there were no carry­forward provisions applied to those categories of 
recruitment). Actually, the Court does not consider any of the recruitment done under the other 
headings 
____________________ 
50. But cf. General Manager v. Rangachari, A.I.R. 1962 S.C. 36, at 39/45. The order there providing 
reservations in promotions was issued in 1959 to be effective from Jan. 4, 1957. The Court there 
held that the reservation could operate retrospectively as well as prospectively. Presumably, then, 
it is permissible for the State to cumulate reservations in retrospect but not to do it prospectively 
by a "carry­forward rule." 
51. The majority does not indicate whether it is still open to the government to leave unfilled the 
reserved places which are not filled by backward classes in a given year. There is nothing in the 
majority opinion to indicate that other employees have any constitutional right to compete for the 
reserved places which cannot presently be filled with qualified members of the backward classes. 
By making no appointments at all for these places, government could increase the number of 
vacant appointments­­and the number of reserved places­­at the next competition when there 
might be more candidates who could qualify for the reserved places. Since there would seem to 
be no constitutional objection to such a practice, government would be able to achieve at least 
part of the object of the "carry­forward" rule. However it could only do so by sacrificing the 
value of a full roster of employees during the interim. What the majority decision does is to 
remove from government the option of both filling its roster now from the available candidates 
and preserving the reservation until qualified members of the backward classes are available. 
52. Thus government may not club together otherwise independent cadres of posts solely for 
purposes of administering a reservation where that unduly impinges on the chances of non­
backward employees. When grouping promotions of Railway Mail Service Inspectors and Postal 
Inspectors led to two of three vacancies in the former going to Scheduled Castes, it was 
"excessive reservation" which "unreasonably denied this opportunity" to a high ranking non­
backward employee ( Director General of Posts and Telegraphs v. Natarajan, 1971 [2] M. L. J. 
79, 80, aff'g A. I. R. 1970 Mad. 458). 
­409­ 
during the same year, so we have no indication of the precise extent of the effective reservation of 
posts during that year.  53.   
In order to appreciate the extent to which the Devadasan majority departs from the Balaji case and 
extends judicial policy­making in this area, it is helpful to consider some of the salient differences 
between the two cases. First, the factual situation in colleges, where there is a complete turnover of 
students every 2 or 3 years, is quite different from that of a government service where the 
incumbents stay from 20 to 30 years. In schools a smaller reservation may suffice to bring the 
proportion of the backward in the school population up to the reserved strength in a few years; in the 
service such a process would take a long time. For example, under a reservation of 17 ½ % for a 
service where the average tenure of employment is 30 years, it would take 30 years before the 
reserved percentage was reached (assuming that the full number of backward class members 
succeeded every year).  54.   
Second, the higher percentage of reservation created by the operation of the carry­forward rule is 
quite different from a straight reservation of a similar percentage. For the carry­forward rule 
depends upon the fact that others had increased success in the past; the chances of success for the 
backward this year are increased commensurately with their lack of success in previous years. 
Again, the carry­forward provision is selfliquidating and non­recurring; to the extent that the higher 
reservation is effective, the reservation returns to its original lower level. Thus arguments based on 
the public interest in utilizing talent and insuring the efficiency of the services have less application 
to the carry­forward rule than they do to a situation like that in Balaji, for over the long run the 
reservation is not increased above an acceptable level. 
The whole policy of preferences introduces a factor of historical balance or compensation;  55.   the 
backward are to be given a more than equal chance this year in order to overcome the deprivations 
and inequalities of the past. But according to the majority, the government, 
____________________ 
53. Calculation based on data at A. I. R. 1964 S. C. at 182 ff. 
54. The inordinate length of time required for such reservation to become effective inspired 
Wanchoo, J., dissenting in General Manager v. Rangachari, A.I.R. 1962 S. C. 36 at 46­48, to 
propose that Art. 16(4) allowed two kinds of reservation: reservation of a percentage of current 
appointments and also reservation of a certain portion of "post" in the whole service. Under the 
second type all appointments could be reserved to the backward classes until the reservation was 
filled, so long as candidates met the minimum qualifications. This was based on a textual 
argument about the meaning of the term "posts" in Art. 16(4), which was rejected by the majority 
there. The Devadasan majority reads Balaji as forbidding even a temporary reservation about 
50% (A. I. R. 1964 S. C. at 188). 
55. Cf. Gajendragadkar, J., for the majority in General Manager v. Rangachari, A. I. R. 1962 S. C. 
36 at 42. 
­410­ 
which is empowered to depart from equality with a view to this long historical time­perspective, is 
not authorized to devise its measures with an eye to results in the two previous years. 
Third, the Devadasan situation concerns a scheme of reservation by the central government, rather 
than that of a state government as in Balaji. In Balaji the Court had national policy available to it as 
a guide in reconciling competing interests in interpreting Article 15(4).  56.   But no such guide is 
available to the Court when it undertakes to sit in judgment over the policy formulated by an agency 
directly responsive to the highest levels of national policy­making. Yet the majority confines the 
discretion of the central authorities more strictly than Balaji did the state's discretion. 
Fourth, connected with this is the fact that the beneficiaries in the Devadasan case were Scheduled 
Castes and Tribes, who are specified under special constitutional authority and to whom there is a 
firm constitutional commitment of preferential treatment. In the Balaji case, on the other hand, the 
beneficiaries were backward classes created ad hoc by the state.  57.   
Fifth, in considering the various interests which must be balanced in determining the reasonableness 
of the reservation, the Devadasan case takes a narrower view of the interests involved. On the one 
hand are the public interest in efficient operation of the services and the "legitimate claims" of the 
other employees; on the other, the right of the backward to adequate representation in the services. 
But as the Court in Balaji pointed out, special treatment is not solely for the purpose of benefitting 
the backward in the short run; there is a long run interest of the society at large in promoting the 
advancement of the backward elements in the population.  58.   The claims of the backward classes are 
not necessarily in derogation of the interests of the wider public; they may be the instrument for 
forwarding the public interest in an egalitarian society. 
Finally, the Devadasan majority takes a much more assertive view of the Court's role in sitting in 
review of the government's determination of policy. In General Manager v. Rangachari, the 
Supreme Court had 
____________________ 
56. See A. I. R. 1963 S. C. at 662. 
57. It should be noted that Devadasan was the first case in which any court has struck down as 
unconstitutional a central scheme for preferential treatment. It was also the first time since the 
addition of Art. 15(4) to the Constitution in 1951 that any court has held unconstitutional any 
arrangement, central or state, for the benefit of the Scheduled Castes or Tribes. All the previous 
cases in which arrangements had been held unconstitutional had dealt with state schemes for the 
other backward classes. 
58. A. I. R. 1963 S. C. at 662. For an eloquent rejection of the facile equation of the "merit" claims 
of the advanced with the national interest, see Vishwanath v. Government of Mysore, A. I. R. 
1964 Mys. 132 at 136. Cf. General Manager v. Rangachari, A. I. R. 1962 S. C. 36 at 42. 
­411­ 
declined to undertake any supervision of the government's policy in making the reservation 
retroactive. Conceding that it might be wise not to give it retrospective effect, the Court noted that 
"with the propriety or the wisdom of the policy underlying the circulars we are not directly 
concerned".  59.   In exercising its power the State "must attempt . . . to strike a reasonable balance 
between the claims of backward classes and the claims of other employees as well as the important 
consideration of the efficiency of administration."  60.   The striking of the balance is left to the 
government with the Court merely admonishing it that such factors must be considered in every 
case. 
In Balaji v. State of Mysore, the Court speaks of "adjusting" the "interests of the weaker sections of 
society" to "the interests of the community as a whole," and declares that "a formula must be 
evolved which would strike a reasonable balance between the several relevant considerations." 61. 
But it is clear that it is the government which must do the required balancing and adjusting. All the 
Court is asserting is its right to step in when the government's final "adjustment" is "plainly 
inconsistent" with the constitutional authorization for preference.  62.   It seems clear that both the 
Balaji and Rangachari cases concede that the balance between conflicting interests is to be struck 
by the State in the first instance. There is a recognition that there is a wide range of possible 
arrangements that might be considered reasonable adjustments of the complex factors that must be 
taken into account. The work of the courts is clearly confined to setting constitutional boundaries by 
insuring that government has in fact taken the relevant interests into account, but stops far short of 
substituting their judgment of how they should be reconciled. 
But in Devadasan, the Court seems to go much further in addressing itself to the task of striking a 
reasonable balance of conflicting interests. It not only sets constitutional boundaries of 
reasonableness but goes on to specify in some detail what reason requires. The difference is 
particularly striking because in Balaji the Court concerned itself not with detailed balancing of 
Article 15 (4) and 15 (1) but with broad consideration of interests to determine whether the state's 
action was justified under Article 15(4).  63.   In determining this, it had an abundance of concrete 
data about the state's selection procedure for backward classes, and the 68% reservation was of such 
magnitude and obvious impact that no delicate consideration of marginal utilities was required. 
However, Devadasan involves the government's decision for arranging over time an admittedly 
modest and proper reservation. The Court had before it no concrete data about the basis upon which 
government 
____________________ 
59. Id., at 433. 
60. Id. 
61. A. I. R. 1963 S. C. at 663. 
62. Id. 
63. Cf. id., at 657, 659. 
­412­ 
decided on this scheme or how it had worked in practice.  64.    In effect, then,  Devadasan
    represents a
    
shift from a "rule of reasonableness" under which state action is tested to see whether the balancing 
of interests by the state falls within broad constitutional limits, to a "balance of interests" rule by 
which the Court undertakes directly to determine the constitutionality of the government's action by 
weighing and balancing competing claims. 
The readiness of the Devadasan majority to step in to oversee the government's balancing of 
interests in such a marginal and ambiguous matter raises the question of the role of the courts in 
reviewing government action under Articles 15(4) and 16(4). The dissenting judge in Devadasan 
adopts a line of severe judicial restraint. To him, Article 16(4) "has not really carved out an 
exception, but has preserved a power untrammeled by the other provisions . . . of Article [16]." 65. 
What provisions are to be made in exercising this power is, for him, entirely within the discretion of 
the State, so long as they fall under the mandate of that clause. Judicial review would be confined to 
seeing whether government action properly fell within that mandate­­i.e., whether the arrangement 
was in fact "a reservation of appointments or posts" and whether in fact the beneficiaries were a 
backward class of citizens. 
There is some support in the text of the Constitution for such a restrained view of the judicial 
function in these matters.  66.   It is clear that the Constitution puts the primary responsibility with 
government for setting policy in this field. All the cases before Devadasan qualify this view only to 
indicate that they will consider broad lines of policy as laid down in the Constitution in determining 
whether the State's action is a valid exercise of these powers. The courts find themselves in the 
difficult position of needing sufficient power to check abuses and see that constitutional principles 
are protected while leaving to the government sufficient flexibility to implement effectively the 
policy of preferences. The Devadasan opinion was the first exercise of judicial review to involve 
serious curtailment of the government's discretion. 
____________________ 
64. Again, the Balaji court could find some constitutional guidelines in the fact that the state had 
utilized constitutionally forbidden criteria. Even if the presumption of reasonableness which 
ordinarily accompanies governmental classifications was not entirely dissipated by the flagrant 
use of criteria forbidden by Art. 15(1), it added a concrete element of constitutional 
disreputability which was absent in Devadasan. 
65. A. I. R. 1964 S. C. at 190. 
66. Cf. Art. 320(4), which exempts implementation of Arts. 16(4) and 335 from even the necessity of 
consultation with Public Service Commissions. While the absence of this ordinary guarantee 
might be thought to impose on the courts a special responsibility to check abuses, it is difficult to 
imagine the Constitution makers excluding the Public Service Commissions and simultaneously 
anticipating a detailed power of review in the courts. 
­413­ 
The paradoxes of expansive judicial intervention combined with a narrow view of the interests 
 involved are dramatically displayed in  Abdul Latif v. State of Bihar
     67.   In 1952 the Bihar 
 . 
Government had decided to give preference to members of Scheduled Castes and Scheduled Tribes 
in the settlement of new excise shops. In 1958, finding that the implementation of this policy was 
"not . . . very effective," the government decided to take further measures. It gave special publicity 
about proposed settlements to ST and SC. It stipulated that if there was a single suitable candidate 
from one of these groups, he should have the shop. If more than one suitable candidate from these 
groups applied, the shop was to given to one of them chosen by lot. 
In 1963 and excise shop was to be settled in Ranchi and the 39 applicants included 7 from the Sc 
and St. The Superintendent of Excise settled the shop by drawing lots among the 7. Abdul Latif, 
who had petitioned to be included in the drawing, obtained a writ from the High Court quashing the 
Superintendent's order. The Patna High Court found that an arrangement which "completely 
excludes the rest of the society . . . is clearly outside the scope of Article 15(4)."  68.   Citing Balaji on 
the need for reasonable limits on reservations and for adjusting the interests of the weaker sections 
with those of the whole society, the court found that the order violated Article 15(1). The vice of the 
arrangements was that no others were allowed to compete. But of course this is precisely what every 
reservation does­­it sets aside some portion of the available prizes and says that members of the 
non­preferred groups may not compete for them. The question, then, is whether the portion set aside 
is within the limits permitted by the Constitution. 
In order to answer this question, we must know what is the relevant unit or group of prizes of which 
we are measuring a portion. If we assume, as did the Patna High Court, that each single bhang shop 
is a relevant unit, then the portion reserved is 100%, which is clearly outside the limit suggested by 
Balaji's case. If, on the other hand, we are willing to take an overall of long­run view of excise shops 
as a whole, then this appears as a smaller reservation. Balaji tells us that reservations must be 
reasonable. It is clear the wherever, as in Abdul Latif, selection is in small increments, an acceptable 
reservation can be devised only if we take the long­run view of the whole class of separate 
competitions. But if Devadasan is correct in saying that the reservation in each separate competition 
must be reasonable, then it is impossible to have a reservation were or in anything that is distributed 
in small 
____________________ 
67. A.I.R. 1964 Pay. 393. (Decided 27 March, 1964. although Devadasan, which was decided 29 
August 1963, was already reported, the Patna court took no notice of it.) 
68. Id., at 395. 
­414­ 
increments­­e.g., it would be unconstitutional to provide a roster for reservation of every tenth future 
vacancy, since that vacancy would be 100% reserved when it came up. 
Government presumably has a mandate to provide that some reasonable number of its licensees are 
from the Scheduled Castes and Scheduled Tribes. Since selection of licensees is done one at a time, 
a percentage reservation is of no utility. Must government reorganize its system of licensing so that 
shops are settled in groups, which would allow for a percentage reservation? Or is there another way 
to achieve the desired reservation? Might not the Abdul Latif situation be thought of as involving 
two stages of reservation­­(1) a reservation of a certain percentage of liquor shop licenses and (2) a 
further reservation of opportunities to enter into the competition to acquire them? If we think of 
opportunities to enter the competition as analogous to government posts or places in colleges, then 
government might require that a certain number of places in the competition or drawing for liquor 
shops be reserved for Scheduled Castes and Tribes. Balaji by itself would seem to require only that 
the reservation of chances to enter the competition be a reasonable one. But what does reasonable 
mean in this context? Devadasan suggests that it means substantially equal opportunity for the non­
backward in each selection­­i.e., that a number of non­backward Ys be allowed in the competition 
so that their chances are substantially equal to that of the preferred Xs. The fallacy here is that the 
chances of each individual competitor in a drawing are equal; his chance is affected only by the 
absolute number of others in the competition and not at all by the relative number of Xs and Ys. As 
this example points out, the equal­opportunity test suggested by Devadasan does not apply to 
individuals at all, but applies only to the collective chances of the Ys. But while the Constitution 
empowers the government to be concerned with the percentage of Xs who enjoy such benefits, it 
does not give to individual Ys any justiciable rights in the collective chances of the Ys. 
Abdul Latif presents in miniature the crucial problems of policy involved in the tension between 
"compensatory discrimination" on the one hand and "equal opportunity" on the other hand. The first 
suggests reservation, which in the nature of the situation must be 100%; the second points to the 
desirability of avoiding 100% reservations, even temporarily, because of their upsetting and 
discouraging effects. The choice is unavoidable. But is it to be made by the courts or by the 
government?  69   In Balaji the Supreme Court indicated that courts might 
____________________ 
69 The Constitution itself provides no final formula for when intervention by the judiciary is proper. 
The sweeping phrases "Nothing in this Article" and "any provision" suggest an absolute 
derogation from the coverage of Art. 15. But, per contra,Art. 15(4) speaks of "special" provision 
only, and special can be said to militate against the 
­415­ 
enquire whether what the government did was within the scope of the purpose of Article 15(4)­­and 
this scope was to be ascertained by consideration of a variety of publiċ interests. In regard to 
 admission to medical colleges  Balaji
    found that excessive reservation neglected and undermined
    
important national interests in the full utilization of talent. In regard to government posts, 
 Devadasan  found that excessive reservation threatened the public interest in morale and incentive in
    
 national services. But  Abdul Latif
    is notable for the absence of any consideration of public policy. It
    
is difficult to see what the analogous public interest is in this case. Can any national interest in 
 motivation for talent or morale in services be said to be present in the case of a  bhang
    70.   May 
  shop? 
 
not government decide that the encouragement of enterprise in this field is outweighed by the 
benefit to society from channelling these opportunities to disadvantaged groups?  71.   
Does the public have any real interest in assuring opportunities to invest in bhang shops to members 
of non­backward communities? If the courts are to undertake a balancing of interests, surely the 
public interest is not to be equated with the interests of the sections of the population that do not 
enjoy reservations. As many courts have noted, special treatment for backward groups is authorized 
precisely because there is a strong national interest in the advancement of the weaker sections of 
society.  72.   In the case of women, no one doubts that special provisions for the advancement of 
women are justifiable, even when men are thereby put at a temporary disadvantage.  73.   It is 
presumed that what helps women to assume a productive place in the society is good for the society. 
The interests of the society cannot be equated with the interests of the men who are temporarily 
discomfited by the special treatment of women. Similarly, there is no justification for equating the 
interests of the advanced classes with those of the nation.  74.   The 
____________________ 
absoluteness of the derogation. Plainly, the power of review here flows, as Upendra Baxi has 
noted, not from the text of the articles but from the structure of Part III of the Constitution­­from 
Arts. 12, 13 and 32. 
70. It is particularly difficult to find an element of efficiency or incentive where, as here, the 
selection is by lot. Concern for enterprise in this area is especially ironic when we note that the 
prohibition policy commits the government to eventually taking away all rights to operate such 
shops. 
71. May government not decide that extensive reservation might be desirable in promoting a 
particular development scheme? 
72. Balaji v. State of Mysore, A. I. R. 1963 S.C. 649 at 662. Cf. Vishwanath v. Government of 
Mysore, A. I. R. 1964 Mys. 132 at 136; Hariharan Pillai v. State of Kerala, A. I. R. 1968 Ker. 42. 
73. For a few examples, see Dattatraya v. State of Bombay, A. I. R. 1953 Bom. 311; Ramchandra 
Mahton v. State of Bihar, A. I. R. 1966 Pat. 214. 
74. See text at nn. 36, 58, above. 
­416­ 
Constitution makers had the foresight to realize that the overcoming of vast social and economic 
disparities was a requisite for having a modern nation. While they entrusted this task primarily to 
the government, there is some room for judicial review, to see that the power is used only for the 
public interest. But it must be recalled that one of the public interests in the balance is the nation's 
long­run interest in the spread of opportunities and resources into the lower strata of the population. 
This interest may sometimes be outweighed by other pressing interests. But it is difficult to see how 
it failed to prevail against an interest as indefinite, diffuse, and questionable as that involved in 
 Abdul Latif . 
  
In the delicate task of balancing the merit principle with other interests, a flat percentage limitation 
on the extent of reservations is of less use than it might appear. It has the virtue of blocking schemes 
of reservation so expansive as to transform the competition into a set of communal quotas. But the 
results of Devadasan and Abdul Latif provide eloquent testimony of the possible dangers of the 
mechanical application of such a standard. The percentage of reservation may be a poor indication 
of its impact where reservations go largely unfilled. Beyond this, the flat percentage limit contains 
other dangers of deception. Even where a reservation is filled, the size of a reservation is not 
necessarily directly related to the amount of distortion of the merit principle in a given competition. 
For example, a reservation of 50% of seats for a group of backward­class candidates who would 
obtain 40% of the seats on merit is actually less of a distortion of merit than a 25% reservation for a 
group of backward classes who would obtain only 5% of the seats on merit. In the first case, only 
10% of the successful candidates would not have prevailed in open merit competition; in the second 
case, 20% would not have been selected on merit. The effective impact of the reservation in the 
second case is twice as great as in the first case, although the reservation in the second case is only 
half as large. The real impact of a scheme on the chances of others and on the merit principle cannot 
be known from the percentage of places reserved. 
The effective impact of a reservation, then, depends not only on its size, but on a variety of other 
factors: whether it operates as a guaranteed minimum or on an "over and above merit" basis; 
whether there is a minimum floor of qualification that must be met; how many of the group for 
which reservation is provided would succeed on merit; how the reservation is divided into 
compartments; how the unfilled reserved places are disposed of We shall be taking up several of 
these matters below. In the meantime, it should be clear that the impact of a reservation on the 
chances of others or on the merit principle cannot be determined from the nominal size of the 
reservation. What is needed is a way of estimating the net effect of resrvations. One simple way of 
doing 
­417­ 
this would be to ask how many those selected by a particular scheme of selection would have been 
selected on a pure merit basis.  75.   It would then be possible to compare the effective extent of 
reservation, rather than its nominal extent. 
Since Devadasan the Supreme Court has never returned to full­dress consideration of how much 
preferential treatment is permissible, but it has obliquely undermined its earlier learning on the 
subject. When the Court encountered the quantity­of­reservation question a decade later in State of 
Kerala v. N.M. Thomas,  76.   it was entangled with much broader issues about the scope of the 
government's power to pursue policies of compensatory discrimination. The Kerala service rules 
specified a period during which government servants temporarily appointed to higher posts might 
pass the tests required for permanent appointment in those posts. A series of government orders 
granting Scheduled Castes and Tribes employees a temporary waiver from required examinations 
enabled a number of Scheduled Castes Lower Division Clerks to be promoted to Upper Division 
Clerks on the basis of seniority in their higher posts. During 1972, out of 51 vacancies in the 
category of Upper Division Clerks, 34 were filled by Scheduled Castes members who had not 
passed the tests; only 17 were given to persons who had passed the test.  77.   Thus the beneficiaries of 
preference constituted two­thirds of that year's appointments; this preponderance was one of the 
grounds upon which the scheme was struck down by the High Court.  78.   
A seven­judge bench of the Supreme Court reversed, five to two. Since it overruled the High Court 
on this, the Devadasan rule of 50% in any year seems seriously undermined. However, there are 
suggestions in the five majority opinions that the scheme in question was not a reservation regulated 
by Article 16(4) but a preferential device empowered directly by Article 16(1) or (2). If so, it may be 
that the Devadasan rule is still applicable to reservations even though it does not 
____________________ 
75. Data on the extent to which backward classes would have obtained seats on merit are scarce. The 
Kumara Pillai Commission in Kerala ( 1965: 134) presented some data of this kind. For example, 
in medical school admissions in 1962­63, 40% of the seats were reserved, but in fact 71% of the 
successful candidates would have been selected on the basis of merit. The effective reservation, 
then, was 29%. Less systematic data come to light in Sagar v. State of Andhra Pradesh, A. I. R.. 
1968 A.P. 165 at 189. In a medical college selection, 123 of the 189 backward classes candidates 
admitted would have secured places on merit. To the extent that reservations are not effective, 
they are unnecessary and should be abolished. But the point here is that they may be ineffective 
and at the same time deceptive as to the amount of special treatment actually being conferred. 
76. State of Kerala v. N. M. Thomas, A. I. R. 1976 S. C. 490. The broader issues in this case are 
discussed at chap. 11, §D, above. 
77. Id., at 495. 
78. I. L. R. 1974 (1) Ker. 549. 
­418­ 
apply to other preferential measures regarding government employment. Of the five majority judges, 
three address the quantity of preference question in their opinions. 
Chief Justice Ray in the majority judgment finds the High Court mistaken in condemning the 34/51 
as excessive. 
The promotions made in the services as a whole are nowhere near 50 per cent of the 
total number of posts. The Scheduled Castes and Scheduled Tribes constitute 10 per cent 
of the State's population. Their share in the gazetted services of the State is said to be 2 
per cent, namely 184 out of 8,780. Their share in the non­gazetted appointments is only 
7 per cent, namely 11,437 out of 1,62,784 . . . therefore . . . the orders are meant to 
implement not only . . . Article 335 but also the Directive Principle under Article 46.  79.   
Apparently, the 50% limit is to be measured by the total number of posts in the services rather than 
the number of promotions. Again it seems to be suggested that higher than 50% is permissible until 
such point as the protected groups have places in service proportionate to their portion in the 
population. 
Fazl Ali, J., concurring, frankly repudiates Devadasan and suggests that such a rule may defeat 
provisions for preference in public employment.  80.   The 50% limit "no doubt laid down" by earlier 
decisions of the Supreme Court is to be taken as "a rule of caution" which "does not exhaust all 
categories."  81.   JusticeKrishna Iyer expresses agreement with Fazl Ali's reading of Article 16(4). 82. 
This is somewhat confusing, since these judges found that the provision in question is justified 
directly under Article 16(1) and does not involve Article 16(4). Krishna Iyer, J., too would not agree 
with an arithmetical limit of 50% in any one year. "Overall strength in a department does not 
depend on recruitment in a particular year, but the total strength of a cadre."  83.   
By overturning the High Court's application of Devadasan the Supreme Court reopens the question 
of the quantity of preference that is constitutionally permissible. Unfortunately, the question is not 
well framed by the Thomas facts. Kerala's award of two­thirds of a year's promotions to Scheduled 
Castes appears less a deliberate application of policy than a temporary (and possibly inadvertent) 
result of a series of earlier unsuccessful measures designed to facilitate SC promotions. The Court's 
response plunges the whole question into obscurity. It upholds at least temporary provision for more 
than 50% of current appointments.  84   It is unclear, though, whether it is abandoning the Devadasan
____________________ 
79. A. I. R. 1976 S. C. at 501. 
80. Id., at 555. 
81. Id., at 554­55. 
82. Id., at 537. 
83. Id., at 537. 
84 In response to the Thomas case the Government of India in 1977 abandoned its earlier 
instructions (embodying the Devadasan 50% limit) and ordered that even 100% 
­419­ 
50% rule as to all reservations under Article 16 (4)­­as a later bench implies.  85.   And, beyond that, it 
 is unclear whether the entire  Balaji
    limitation on the extent of preferences and the extent of
    
preferred groups is unsettled. Justice Fazl Ali, the only member of the majority who is explicit about 
this, makes it clear that he would reopen the whole question of the scope of preference policy. 
Suppose, he asks, a state had a population of which 80% were backward classes and the government 
reserved 80% of the jobs in its services for them. And he answers: it could not be said that such a 
provision "violates the permissible limits" of Article 16(4).  86.   
C. THE INTENSITY OF PREFERENCE 
Connected with the amount of preferential treatment is a set of problems that may be thought of as 
questions about its intensity: just how far may the State deviate from the standards it would 
otherwise employ? May it waive minimum standards altogether? Just how much may an individual 
beneficiary be helped and how much may those around him be discomfited? May the State deviate 
as much in distributing outcomes (services to citizens or desirable posts) as in distributing 
opportunities (admission, etc.) or improving access to them (e.g., by coaching or waiver of 
application fees, etc.)? 
It is accepted that the State may lower barriers to opportunity exclusively for backward classes: 
provisions for waiver or reduction of filing fees, application fees, and so forth are widespread and 
have attracted few challenges.  87.   The Representation of the People Act's provision for lower election 
deposits for Scheduled Castes and Tribes was held not to be discriminatory, for it enabled members 
of these groups to contest elections.  88.   Clearly, preferences need not take the form of reservations, 
but may be to enable members of these groups to enter the competition by waivers of fees, age 
limits, and so forth. In line with the general notion that special measures to enable them to compete 
at 
____________________ 
of available posts could be reserved where reservations had gone unfilled in previous years 
(Department of Personnel and A.R.O.M. No. 16.3.1973 ESD [SCT] dated 27.12­1977). 
85. Thirteen months after Thomas, in Chauhan v. State of Gujarat, A. I. R. 1977 S. C. 251 at 263, the 
Court (per Krishna Iyer, J.), after distinguishing the Devadasan carryforward situation on the 
facts, remarks that "it is doubtful whether Devadasan's case survives State of Kerala v. N. M. 
Thomas. . . ." 
86. A. I. R. 1976 S. C. 490 at 555. This inference may gain persuasiveness in the setting provided by 
the new primarily tribal states in northeastern India. In Nagaland there are indeed 80% 
reservations for Scheduled Tribes. 
87. See cases cited in chap. 11, n. 47. 
88. Raju v. State, A. I. R. 1976 Guj. 66. 
­420­ 
intake are unobjectionable, special arrangements have been made to prepare members of these 
groups for examinations for government service.  89.   
It is established that the State may, to insure the inclusion of these groups, employ some lower 
standard of selection at the intake point. For example, SC and ST applicants for Class III posts in 
central services are judged by "relaxed standards."  90.   It is generally accepted that reservations may 
entail the setting of a lower floor of minimum qualifying marks for members of the preferred 
groups. The common practice  91.   of most medical colleges to relax the minimum qualifying marks 
for SC and ST by 5% has been upheld; it is "open to the State to . . . fix different minimum marks" 
so long as the difference is reasonable for purposes of Article 14.  92.   
But courts are divided on how far this can go. In Dilip Kumar v. Government of U.P., a Division 
Bench of the Allahabad High Court held that it was allowable to reduce the minimum qualifying 
marks "to any reasonable figure below 30 per cent," but that to specify "that there shall be no 
minimum qualifying marks for the candidates from Uttarkhand Division" was "unreasonable and 
arbitrary" because it defeats "the object of securing suitable candidates."  93.   In a different case, the 
Supeme Court eventually upheld the state's designation of residents of Uttarkhand as backward, 
without taking up the "no minimum marks" question.  94.   
However, in Gupta v. Union of India,  95.   another bench of the Al­ 
____________________ 
89. See the account of coaching centers for preparing members of Scheduled Castes and Tribes for 
high level examinations discussed in chap. 4, §B, above. 
90. In non­technical and quasi­technical Class III posts, if the relaxed standards fail to admit enough 
candidates to fill the reservation the best candidates who fill the minimum educational 
qualifications are to be appointed and given in­service training. In practice, appointing 
authorities often deviate from these instructions (RCSCST 1975­77: I, 50). 
91. RCSCST 1975­77, II36­37. 
92. Sukhdev v. State of Andhra Pradesh, 1966(1) Andh. W. R. 294 at 312. That such lower standards 
are also permissibile in the government employment area is implied in General Manager v. 
Rangachari, A. I. R. 1962 S. C. 36 at 42; Devadasan v. Union of India, A. I. R. 1964 S. C. 179 at 
187. 
93. Dilip Kumar v. Government of U.P., A. I. R. 1975 All. 592 at 596. The no­minimummarks feature 
produced a very high variance between reserved and general candidates. In Pradip Tandon v. 
State of U.P., A. I. R. 1975 All. 1 at 2, it is reported that in a subsequent U.P. medical admissions 
competition the last admitted general candidate obtained a test score of 271 marks and the lowest 
score on the waiting list (of 100 candidates) was 266. The lowest successful candidate in the 
reserved category had 128 marks. 
94. State of U.P. v. Pradip Tandon, A. I. R. 1975 S. C. 563. 
95. A. I. R. 1974 All. 288. 
­421­ 
lahabad Court not long after dealt with a case of admission to the elite Indian Institute of 
Technology. The normal minimum qualifying marks were 40%, and it was alleged that persons 
receiving as little as 10% had been admitted to reserved seats. Once a reservation has been made, 
says the court, "the method of filling . . . reserved seats may be different. . . . The qualifying marks 
for this class may also be different or there may be no qualifying marks at all."  96.   The court asserts 
that such a waiver cannot lead to the dire result of lowering the quality of engineers, prophesied by 
the petitioner, for "it is only at the stage of admission that concession is granted. . . . No concession 
is being granted them in passing the final examination: unless they qualify themselves through the 
education given them at the Institute they will not pass and the country will not get technologists 
and engineers who are not fully qualified."  97.   Thus the court visualizes a combination of maximum 
flexibility at the intake stage with inflexible standards at the output. But it is not clear that this neat 
pattern is constitutionally mandated or readily realizable.  98.   
The question has never been addressed directly. Where it was obliquely present in several recent 
cases, the judicial response is so disparate that it is hard to predict what would happen if it were 
encountered directly. In Janki Prasad v. State of Jammu and Kashmir,  99.   the Supreme Court struck 
down the use of the selection committee in the promotion of teachers to headmasters, because the 
procedure was one in which relevant information was disregarded and the standards were set so low 
as to reflect abandonment of concern with quality. But the setting of absurdly low minimal scores 
made it a "travesty of selection."  100.   Although reservations for Backward Classes were involved in 
the selection, it was not a waiver of quality for them that was objected to, but the setting of a general 
floor for all candidates. Also, this selection was contaminated by other constitutional shortcomings: 
there was no use of essential information about performance on the job, and the Backward Classes 
reservations were defective in various ways. Again, the establishment of these inadequate selection 
standards was seen as part of a prolonged effort to effectuate a scheme of unconstitutional 
communal 
____________________ 
100. Id., at 936. 
96. Id., at 289. 
97. Id. 
98. The system approved in Gupta proved unsatisfactory, for many of those recruited "even after 
special coaching . . . could not withstand the strain of the five year course." A new arrangement 
admitted only those Scheduled Caste and Scheduled Tribe students who obtained two­thirds of 
the marks obtained by the lowest ranking successful general candidate, with a special system of 
recruitment to fill the remaining reserved places (RCSCST 1975­77: I, 100). For a sensitive 
account of the working and effects of the preferences at I.T.T. Bombay, see Kirpal 1978. 
99. Janki Prasad v. State of J. & K., A. I. R. 1973 S. C. 930. 
­422­ 
quotas, an effort marked by shifting stratagems and defiant evasion of judicial decrees.  101.    If  Janki
    
 Prasad  indicates that the courts will scrutinize the setting of minimal standards, it tells us little
    
about how much a state may vary minimal scores in a selection where relevant information is 
utilized, and where there is a reservation for legitimate beneficiaries that is legitimate in scope. 
The intensity of preference that is allowable may vary, depending on what kind of opportunities are 
at stake. Is it primarily a matter of distribution of individual benefits, or is the quality of public 
service directly affected? Is the matter in question the consumption of benefits or the production of 
them for the public? Thus in Janki Prasad, the Court observes that 
where appointments and promotions to responsible public offices are made, greater 
circumspection would be required in making reservations for the benefit of any 
Backward Class because efficiency and public interest must always remain paramount. 
It is implicit in the idea of reservation that a less meritorious person is to be preferred to 
another who is more meritorious.  102.   
If reservation implies inferior service to the public and the public's interest is paramount, the scope 
for reservation might be severely restricted­­at least in regard to "responsible public offices" like 
headmasters. 
The "intensity" issue surfaces briefly in the Thomas case. Scheduled Caste clerks were kept in the 
posts of Upper Division Clerks for many years by a series of orders exempting them temporarily 
from the requirement of passing various departmental examinations. Justice Khanna, dissenting, 
holds that in regard to government posts "it is not permissible to waive the requirement of minimal 
educational qualification and other standards essential for the maintenance of efficiency of service." 
 103.   The majority judges concede that Article 335 inserts into any scheme for preference in 
government employment a requirement th t be consistent with the efficiency of administration. 104. 
Yet none of them appear troubled by the possibility that efficiency might be com­ 
____________________ 
101. The headmaster posts at issue in Janki Prasad are the very same ones that were involved in 
Triloki Nath Tikku v. State of J. & K., A. I. R. 1967 S. C. 1283 and A. I. R. 1969 S. C. 1. Janki 
Prasad was the fourth appearance of this dispute in the Supreme Court. In Makhan Lal v. State 
of J. & K., A. I. R. 1971 S. C. 2206, 2209, 2210, the state was admonished for responding to 
Triloki Nath with "absolute non­compliance" with its own rules and "complete defiance of the 
law laid down by this Court." 
102. A. I. R. 1973 S. C. at 939. 
103. A. I. R. 1976 S. C. 490 at 506. 
104. Even compensatory measures to ensure equality of opportunity justified by a broad classification 
argument may "not dispense with the acquisition of the minimum basic qualification necessary 
for the efficiency of administration" Justice Mathew, id., at 519). 
­423­ 
promised by the waiver of these examinations.  105.   Some are reassured by other evidence of 
competence. Justice Fazl Ali notes that the Scheduled Caste clerks are "senior . . . and have gained 
more experience."  106.   JusticeBeg notes that these employees must be "otherwise qualified and 
suitable for appointment" and were "presumably quite satisfactorily discharging their duties in the 
higher grade."  107.   Others indicate some skepticism about whether the test really measures 
qualification for the job. Thus, Justice Mathew finds the existence of lengthy exemptions for all 
employees an indication that "passing of the test is not absolutely essential for holding the post." 108. 
And Justice Krishna Iyer introduces an elaborate "trichotomy" to distinguish "basic qualifications 
for eligibility to a post," "secondary qualifications . . . useful to discharge the duties of the post," and 
"virtues which will make the employee ultra efficient."  109.    The tests in  Thomas
    are "demanded for
    
better performance, not basic proficiency." Since "their absence [is] not . . . fatal to efficiency," 
relaxation of standards is allowable "in suitable . . . cases."  110.   Suitability in this case is established 
by the Olympian observation that 
After all, here he is a pen­pushing clerk, not a Magistrate, accounts officer, forest 
officer, sub­registrar, space scientist, or top administrator or one on whose initiative the 
wheels of a department speed up or slow down.  111.   
Summing up, Justice Krishna Iyer suggests that "to relax on basic qualification is to compromise 
with minimum administrative efficiency" and is presumably barred by Article 335. But "to relent 
for a time on additional test qualifications is to take a calculated but controlled risk, assured of a 
basic standard of performance."  112.   Presumably, then, such additional qualifications can only be 
waived in "suitable cases," "for a time," only where there is some assurance of a "basic standard of 
____________________ 
105. Curiously, the majority judges in Thomas erect a temporal counterpart to the notion that there 
cannot be a waiver of minimum marks. Thus they emphasize that the state could not dispense 
with the tests forever, for to promote a Scheduled Caste clerk "without passing any test at all . . . 
[would] destroy the concept of equality" (Fazl Ali, J., A. I. R. 1976 S. C. 490, at 551). In spite of 
their emphatic embrace of substantive rather than formal equality, a close reading of the Thomas 
majority opinions suggests that the substantive equality pursued by compensatory classification 
is not the only equality to be found in Art. 16. 
106. Id., at 552. 
107. Id., at 523. 
108. Id., at 520. At one point Justice Krishna Iyer seems about to take up the question of the way in 
which efficiency is defined: "efficiency means . . . not marks in examinations only, but 
responsible and responsive service to the people." But he veers away without developing the 
point (id., at 534). 
109. Id., at 527. 
110. Id., at 527. 
111. Id., at 527. See his further observations on "alert quilldriving," id., at 536. 
112. Id., at 527. 
­424­ 
 performance." Thus this acceptance of broad waiver in  Thomas
    carries the suggestion that lowering
    
of standards for compensatory purposes is limited and hints at principles for judicial definition of 
those limits. 
Thomas suggests that the source of those limits is the competing commitment to efficient 
administration, not the frustration of other employees. Reservations that are otherwise valid are not 
invalidated by the fact that they endow beneficiaries with very extensive benefits, to the 
discomfiture of those surrounding them. Thus, reservations in promotion that entail the promotion 
of junior officers over their seniors are allowable, notwithstanding the vexation of the latter, because 
of long­run benefits from erasing group disparities.  113.   
Although the State may make such generous waivers, it is not required to do so. When the State of 
Orissa conducted a combined examination for several of its services, 133 candidates were 
successful­­i.e., had test scores high enough to support appointment. Although 18% of the places in 
each of the two services were reserved for Scheduled Castes, there was just one successful SC 
candidate, who had scored 105th on the examination. Of the successful candidates, 33 were 
appointed to the elite "A class" posts and 53 to "B class" posts. The sole SC appointee challenged 
the Public Service Commission's recommendation of him to a "B class" post and sought an order 
that the reservation be implemented by appointing him to an "A class" post. The High Court 
declined to intervene, for such an appointment depended on whether the PSC thought him suitable, 
"having regard to the maintaining of the efficiency of the administration."  114.   Thus the 
government's preference policy is mediated through judgments involving competing goals of policy. 
This is most clear in the government service area, where Article 335 requires an assurance that 
efficiency will be maintained. But similar considerations would seem to be present wherever service 
to the public rather than distribution of purely individual benefits is involved. 
D. THE WIDER USE OF THE RESERVATION DEVICE 
The reservation device has been increasingly employed to assure places to other categories of 
persons, in addition to the backward classes, who seem especially desirable or deserving. 
Government jobs have occasionally been reserved for political sufferers and exservicemen. But 
professional and especially medical education is the place where reservations have proliferated. 
From time to time, states have provided reservations in the professional colleges for such disparate 
categories of persons as overseas Indians; refugees; nominees of 
____________________ 
113. State of Punjab v. Hira Lal, A. I. R. 1971 S. C. 1777. 
114. Sudharkar v. State, A. I. R. 1974 Or. 224. 
­425­ 
neighboring states or of the central government; students who have distinguished themselves in 
sports, extra­curricular activities, National Cadet Corps, or President's Scouts and Guides; children 
of medical practitioners, armed services personnel, university employees, displaced goldsmiths, and 
"political sufferers." In addition, there are sizable reservations for women. State governments have 
attempted to distribute the available seats by residence in districts or regions and to allocate seats 
among those who qualified by several alternative examinations. 
None of these categories are "backward classes": the provisions of Article 15(4) are not applicable. 
But the allotment of places under these headings has affected provisions for backward classes in 
several ways. Before exploring the connections, it is necessary to sketch the judicial response to 
these other categories of reservation. 
Reservations for women are upheld simply on the ground that Article 15(3) authorizes the State to 
make special provisions for women.  115.   The other reservations and distributions, when challenged, 
have to meet the general requirements of reasonable classification of Article 14:  116.   they must be 
based upon an intelligible differentia and bear a rational relation to the object being pursued. The 
results have been varied. 
District and regional residence distributions were upheld in several states  117.   before the Supreme 
Court held that "district­wise"  118.   and [regional] "unit­wise"  119.   selections are contrary to Article 
14. Distribution of places among candidates qualifying by different examinations was struck down, 
 120.   but distribution among the graduates of different universities has been upheld.  121.   
____________________ 
115. Sagar v. State of Andhra Pradesh, A. I. R. 1968 A. P. 165; Ramchandra Vishnu v. State of 
Madhya Pradesh, A. I. R. 1961 M.P. 247. Since sex is not one of the criteria forbidden by Art. 
29(2), which deals specifically with admissions to educational institutions, such reservations 
might be valid even without consideration of Art. 15(3); cf. University of Madras v. Shantha Bal, 
A. I. R. 1954 Mad. 67; Anjali v. State of West Bengal, A. I. R. 1952 Cal. 825. 
116. P. Rajendran v. State of Madras, A. I. R. 1968 S. C. 1012; Subhashini v. State of Mysore, A. I. R. 
1966 Mys. 40. Contra: Ramchandra Vishnu v. State of Madhya Pradesh, A. I. R. 1961 M.P. 247. 
117. Sagar v. State of Andhra Pradesh, A. I. R. 1968 A. P. 165; Lalita Shuri Tikku v. State of Jammu 
and Kashmir, A. I. R. 1966 J. & K. 101. 
118. P. Rajendran v. State of Madras, A. I. R. 1968 S. C. 1012. This result was anticipated in State of 
Kerala v. Jacob Mathew, I. L. R. 1964(2) Ker. 53. 
119. Periakaruppan v. State of Tamil Nadu, A. I. R. 1971 S. C. 2303. 
120. State of Andhra Pradesh v. Balaram, A. I. R. 1972 S. C. 1765. An earlier version was upheld in 
Sagar v. State of Andhra Pradesh, A. I. R. 1968 A.P. 165; an even earlier one was struck down on 
the ground that the shares were disproportionate, Sukhdev v. Government of Andhra Pradesh, A. 
I. R. 1966(1) Andh. W. R. 94 (1963); and a yet earlier one upheld in Gullapalli Nageswara Rao 
v. Principal, A. I. R. 1962 A.P. 212. 
121. Chanchala v. State of Mysore, A. I. R. 1971 S. C. 1762. Universities can favor 
­426­ 
Courts have struck down reservations for President's Scouts and Guides,  122.   and for children of 
medical practitioners,  123.   university employees,  124.   legislators,  125.   and displaced goldsmiths, 126. 
on the ground that they have no relation to the object of medical education. 
On the other hand, judges have uniformly endorsed reservations for outstanding sportsmen on the 
grounds that they exhibited qualities which were desirable in medical practitioners.  127.   Reservations 
for children of defense personnel and ex­servicemen were thought to be clearly related to defense 
policy.  128.   Similarly, children of political sufferers may be disadvantaged by the unsettled life and 
economic sacrifices of parents; a careful and detailed provision which makes it possible to identify 
them has been held to have "a reasonable nexus with the object of the rules which can be nothing 
else than a fair and just distribution of seats."  129.   Looser provisions have been struck down  130.   or 
have provoked judicial condemnation.  131   Reservations for childen of 
____________________ 
applicants who qualify by their own examinations­­i.e., are their own students. Since student 
bodies are largely local, this has some of the same effect as the invalid districtwise distribution. 
See Katzenstein 1979, chap. 9. 
122
Sagar v. State of Andhra Pradesh, A. I. R. 1968 A. P. 165. 

123Jacob Mathew v. State of Kerala, A. I. R. 1964 Ker. 39 at 68 (Aff'd. on this point by I. L. R. 1964 
. [2] Ker. 53); Surendrakumar v. State of Rajasthan, A. I. R. 1969 Raj. 182. 
124
Umesh Chandra Sinha v. V. N. Singh, A. I. R. 1968 Pat. 8. 

125
Surendrakumar v. State of Rajasthan, A. I. R. 1969 Raj. 182. 

126
Sagar v. State of Andhra Pradesh, A. I. R. 1968 A. P. 165. 

127. State of Kerala v. Jacob Mathew, I. L. R. 1964(2) Ker. 53 (rev'g on this point A. I. R. 1964 Ker. 
39); Subhashini . v. State of Mysore, A. I. R. 1966 Mys. 40 at 46 ("a good sportsman cannot 
afford to be a book worm . . . he is likely to be a better doctor or engineer than his competitor 
who knows only books but not men and matters"); Sukhdev v. Government of Andhra Pradesh, 
1966 (1) Andh. W. R. 294 (1963). 
128Subhashint v. State Of Mysore, A. I. R. 1966 Mys. 40; Surendrakumar v. State of Rajasthan, A. I. 
. R. 1969 Raj. 182; Sardool Singh v. Medical College, A. I. R. 1970 J. & K. 45; Kushma Joshi v. 
Pro­Vice Chancellor, A. I. R. 1969 J. & K. 136. Similarly, reservation of posts for ex­servicemen 
has been held permissible to encourage young men to join the forces and expose themselves not 
only to the dangers of that profession but to the additional hazard of compulsory early retirement 
( Daya Ram v. State of Haryana, A. I. R. 1974 P. & H. 279). 
129
Chanchala v. State of Mysore, A. I. R. 1971 S. C. 1762 at 1776. 

130. Surendrakumar v. State of Rajasthan, A. I. R. 1969 Raj. 182. 
131 In Ramchandra Vishnu v. State of Madhya Pradesh, A. I. R. 1961 M.P. 247; Vinor Sagar Sood v. 
State of Madhya Pradesh, A. I. R. 1967 M.P. 182. The Madhya Pradesh Court, while 
condemning these reservations as unrelated to the object of promoting efficiency in the medical 
profession, said that if it were a statutory provision it would not hesitate to strike it down as 
violative of Art. 14. But since it was merely an executive instruction and not a law, it did not 
attract Art. 15. The hope expressed in the first case that government would respond favorably to 
judicial animadversions was proved fruitless in the latter. 

In the reductio ad absurdum of such reservations, the State of Himachal Pradesh reserved 
medical school places for wards of "political sufferers and persons of Himachal 

­427­ 
Indian diplomats and for various categories of out­of­state students (overseas Indians, Colombo 
Plan scholars, refugees from Burma, Asian and African students, Goans, residents of Union 
Territories and of other states) were held to be reasonably related to the State's policies of breaking 
down barriers and creating an integrated society.  132.   
Whatever plausibility one accords the supposed link between athletic ability and qualifications for 
medical practice, the claim on behalf of children of political sufferers or of diplomats is of a 
different kind. It is not a claim of unrecognized qualification, but a claim for fairness to offset 
disadvantage. Government might, in addition, assert an interest in the composition and morale of 
the diplomatic corps or in integrating the society by including residents of other states and so forth. 
What are the objects that government may legitimately pursue when it distributes, for example, 
medical admissions? Presumably it aims to select those best qualified to practice medicine, and 
presumably, it is not confined to prior academic performance in judging such qualifications. Indeed, 
if the goal of the exercise is to enhance the provision of health services to the population, there may 
be an obligation to enlarge the notion of qualification accordingly (to include, for example, 
inclination to provide service in under­serviced areas and perhaps disinclination to emigrate). 133. 
But, may government take into account other goals, unrelated to provision of health services? In one 
area, the answer is clearly yes ­­ the State is empowered to promote inclusion of the backward 
classes while distributing places.  134.   May it pursue other goals that are extraneous to health 
services? 
____________________ 
Pradesh with outstanding social background." A Division Bench struck this down as violative of 
Art. 14 ( Sukhvinder Kaur v. State of Himachal Pradesh, A. I. R. 1974 H.P. 35). A few months 
later, when the son of an I.N.A. veteran sought to be admitted under this heading, another 
Division Bench (including one of the same judges) went to heroic lengths to produce an 
acceptable construction, finding that [t] he phrase 'outstanding social background' necessarily 
refers to outstanding performance . . . in the political field, coupled with . . . backwardness in the 
matter of education" ( Registrar v. Shashi Pal Singh, A. I. R. 1974 H.P. 41 at 45). 
132. Chitra Ghosh v. Union of India, A. I. R. 1970 S. C. 35; Subhashini v. State of Mysore, A. I. R. 
1966 Mys. 40; Sagar v. State of Andhra Pradesh, A. I. R. 1968 A. P. 165. 
133. A substantial proportion of India's medical graduates do emigrate. One report estimates that 
from 1961 to 1964, 18% of medical graduates emigrated annually, leading to a permanent loss 
each year of about 7% of all medical doctors ( World Bank 1975: 83). Indeed, M. R. Balaji, 
petitioner in the leading case, was reported in 1979 to be practising medicine in the United States 
(interview with S. K. Venkata Ranga Iyengar by Robert Hayden in Bangalore, Oct., 1979). 
Likelihood of emigration might be thought relevant if measurable (and if the administration of 
such measures were feasible). But cf. the judicial response to a post hoc claim of selection on the 
basis of providing service to rural areas (discussed above in chap. 11, at nn. 56 ff.). 
134. Inclusion is used here to comprehend the whole range of goals (acceptance, integration, fairness, 
representation, redistribution, incubation) which may be present in measures favoring backward 
classes. For a sketch of these, see chap. 3, §D, below. 
­428­ 
This question has not been addressed so clearly as by a divided 1955 Full Bench of the Patna High 
 Court in  Sukhnandan Thakur v. State of Bihar
     135.   A retrenched employee complained that in 
 . 
appointing and retaining supply inspectors, the state had favored political sufferers and displaced 
persons over other employees with greater seniority.  136.   Chief Justice Das, dissenting, asks: 
What . . . are the objects of employment . . . under the State. Undoubtedly, one object is 
to get the work of the office done efficiently. I am . . . unable to agree that this is the 
sole object of state employment. . . . State employment can be for other purposes as well 
. . . here the directive principles in Part IV of the Constitution . . . are relevant.  137.   
Reading Article 39's directive about adequate means of livelihood and Article 41's about the right to 
work, he concludes that 
the object or purpose of state employment may [also] . . . be to give public assistance in 
cases of unemployment or any other cases of undeserved want.  138.   
Displaced persons and political sufferers are instances of undeserved want, and the preference for 
them is justified. Where a large number of persons possess qualification for a job, a selection must 
be made "based on some principle which is related, not so much to the required or minimum 
qualification for the office, but to something . . . which is peculiar to the candidate himself. . . ."  139.   
If, under conditions of oversupply, the selection process could be used to pursue other policies 
without significant sacrifice of "merit," it did not strike the other Judges as constitutionally 
permissible. It was not open to government to set qualifications unrelated to "the demands and 
needs of the office."  140.   The Government's test was not rational, for it is "based on extraneous or 
collateral considerations"  141.   By addressing these irrelevant considerations, the appointing authority 
has violated Article 16(1). "[T]he essential point is that the qualities or qualifications required of the 
candidate must have some relevance to the efficient performance of the duties and obligations or the 
particular office."  142.   
Although it was never again addressed so explicitly, there has been a clear shift away from this view. 
Chitra Ghosh v. Union of India  143.   concerned 
____________________ 
135. A. I. R. 1957 Pat. 617. 
136. Of the 169 Supply inspectors appointed, 52 were political sufferers and 10 were displaced 
persons (id., at 628). 
137. Id., at 624. 
138. Id., at 624. 
139. Id., at 626. 
140. Ahmed, J., id., at 621. The judge notes that Art. 16(4) and Arts. 16(3) and 16(5) represent an 
exception to this and that their presence argues that "no other exception was intended" (id., at 
622). 
141. Ramaswami, J., id., at 629. 
142. Ramaswami, J., Id., at 632. 
143. A. I. R. 1970 S. C. 35. 
­429­ 
admissions to a medical college in Delhi. In addition to reservation of 15% for SC, 5% for ST, and 
25% for women, there was provision for 23 of the 125 seats to be reserved for the following 
categories and filled by nominees of the central government: 
c.  Sons/daughters of residents of [specified] Union Territories . . . 
d.  Sons/daughters of Central Government servants posted in Indian Missions abroad 
e.  Cultural Scholars 
f.  Colombo Plan Scholars 
g.  Thailand Scholars 
h.  Jammu and Kashmir State Scholars  144.   
Applying the tests of permissible classification under Article 14, the Court notes that since it is not 
possible to throw open admission to students from all over the country, "the government cannot be 
denied the right to decide from what sources the admission will be made."  145.   This is a matter of 
policy depending on "an overall assessment . . . of the requirements of residents of particular 
territories and other categories of persons for whom it is essential to provide facilities for medical 
education."  146.   
Have the sources been properly classified? The Union Territories and Jammu and Kashmir 
categories are responsive to special shortages of medical education facilities. The Cultural, 
Colombo Plan, and Thailand scholars are justified by "reciprocal agreements of an educational and 
cultural nature."  147.   As to the children of diplomats: "it is . . . well known that due to exigencies of 
their service these people are faced with a lot of difficulties in the matter of education."  148.   With 
the sources classified in this manner, "it is difficult to see how that classification has no rational 
nexus with the object of imparting medical education and also of selection for the purpose." 149. 
This is a revealing locution. It is evident that some of these sources are selected on grounds of 
fairness, avoidance of hardship, indirect contribution to cultural enrichment, or the morale of the 
overseas employees and are not directly related to the specified object of securing "the best possible 
material" for medical training.  150.   The single­minded "relevant qualifications" notion of 
Sukhnandan is at least enlarged in some unexplained way and possibly displaced entirely. This 
suggests large possibilities for the use of the 
____________________ 
144. Id., at 37. Since the reservations for women are to be made after the Government of India 
nominations are filled, the maximum reservation could run as high as 58%. But in fact it was 
less. The Government of India nominated only 9 persons against the 23 allowed for categories 
(c)­(h). And very likely the net effect of the reservation for women was much less than the 25% 
specified. 
145. Id., at 39. 
146. Id., at 39. 
147. Id., at 38. 
148. Id., at 38. 
149. Id., at 39. 
150. Id., at 38. 
­430­ 
reservation device to pursue all kinds of governmental policies. We shall return to the question of 
possible conceptual limits after considering whether there are fixed quantitative limits on its use. 
A new sort of "how much" problem arises when reservations under these other headings are 
combined with the reservation for the backward classes under Article 15(4) arid the total is above 
50%. Does the Balaji­Devadasan 50% rule apply to all reservations ­­ i.e., does it concern the extent 
of the opportunities for merit ­­ or is it merely a limitation on what the State can do for the 
backward classes? In Subhashini v. State of Mysore,  151.   16.1% of the 750 seats were reserved for 
miscellaneous categories (out­of­state, foreign students, sportsmen, etc.), and then 48% of the 
remainder were reserved for the Backward Classes, Scheduled Castes, and Scheduled Tribes. Thus a 
total of 56.3% of the seats were reserved and only 43.6% remained open to merit competition. To 
the argument that Balaji ruled out reservations which reached higher than 50%, the court replied 
that that case was concerned only with the permissible extent of reservations under Article 15(4) 
and did not include other reservations.  152.   Since each of the reservations here was reasonable, the 
Constitution was not offended by the fact that they totalled more than 50%.  153.   
The Punjab Court has taken the opposite view. In Ramesh Chander Garg v. State of Punjab, 154 
medical college seats were subject to the following reservations: 
Nominees of Government of India 
Nominees of Government of 
State of Jammu and Kashmir  6% 
Scheduled Castes and Scheduled Tribes 20% 
Backward classes 2%
Sportsmen 2%
Students from recognized backward areas
of Punjab unable to get admission on merit  10%
Women  20%
Total Reservations   60%
____________________ 
151. A. I. R. 1966 Mys. 40. 
152. Where the state had reserved 90% of the places in a condensed M.B.B.S. course for licentiates in 
state service, the Mysore Court held that this was not a reservation but an allotment and that Art. 
15 standards did not apply ( Narayana Swami v. State of Mysore, A. I. R. 1968 Mys. 189). 
153. Sagar v. State of Andhra Pradeh, A. I. R. 1968 A.P. 165, similarly read Balaji as dealing only 
with reservations under Art. 15(4) and as standing for the proposition that Art. 15(4) 
reservations should not exceed 50%. Reservations for women were justified separately under 
Art. 15(3), and the miscellaneous ones (sportsmen, etc.) did not offend either Art. 15(1) or 
29(2). 
154 A. I. R. 1996 Punj. 476. The scheme became even more elaborate in succeeding 
­431­ 
The Punjab­High Court took the view that a reservation of 60% was "clearly hit by the test laid 
 down by [  Balaji
      155.   "[S]pecial provision is allowed to be made for some kinds of minorities and 
 ]." 
this permission should not be used by State authorities . . . so as to reduce the seats available to 
candidates competing on an open merit basis to less than even 50 per cent."  156.   In view of the 
interests of the meritorious candidates and the national interest in utilizing talent, the Court held 
"that making reservation for the weaker sections of the society under any label whatsoever should 
not in the aggregate exceed 50 per cent of the total available seats . . . ."  157.   The 60% reservation is 
"invalid and violative of the rule of equal protection of laws and amount[s] to a fraud on Article 
15(4) of the Constitution."  158.   
Neither of these views is entirely satisfactory. The Subhashini view ignores the concern expressed in 
Balaji and other cases to give scope to talent. The constriction of merit operates to the detriment of 
the public and the general applicant as much where the reservation is for sportsmen as where it is 
for backward classes. Of course, it could be argued that these are admissions on "merit," with the 
ordinary measure of merit broadened to include athletic ability, etc. But if these are desirable 
qualities to be encouraged among the student body, some way might be found of measuring it in all 
students, not just in a few, and measuring it in increments rather than on an all­or­none basis. 
On the other hand, the Punjab position is not convincing insofar as the categories are not truly 
additive but are cross­cutting. It was the reservation for women that sent the total over the 50% 
mark. It is not clear why this should be counted. Special provisions for women occupy a unique 
constitutional position. If the state had two medical colleges of equal size, one only for women and 
one only for men, would it be foreclosed from making any reservation for Scheduled Castes in these 
colleges? Why should the result be different if the two medical colleges are co­educational and 
places in them are reserved for women? Unlike the other reservations, those for women cut across 
the other categories and are not additive. This was in fact the situation in Garg. More than 20% of 
the successful merit candidates were women, and the reservation was not utilized at all. In fact, 251 
of 400 candidates there were admitted on the basis of merit; the effective reservation was less than 
____________________ 
years, but the reserved total dropped to 50%. Several categories (candidates from Border areas, 
children of defense personnel, etc.) were added, but the reservation for women was reduced from 
20% to 1%. See Gurinder Pal Singh v. State of Punjab, A. I. R. 1974 P. & H. 125 at 126. 
155. Id., at 478. 
156. Id., at 478. This position was recently taken (without discussion) by the court in Chhotev Lal v. 
State of U.P., A. I. R. 1979 All. 135, 152 [NIC]. 
157. Id., at 478. 
158. Id., at 478. 
­432­ 
38%. So the court's objection was highly formalistic and completely unrelated to the extent to which 
merit considerations were actually diminished in the selection. 
Both positions reveal the inadequacy of applying a strict percentage rule without consideration of 
underlying policy and the factual implications of the arrangements. The size of the reservation does 
not necessarily reflect its impact on the merit principle. The broader the categories for which 
reservation is made, the more likely they are to pick up students who would be selected on merit. 
Thus large categories like "women" either may have no effect on the selection or may have a 
marginal effect rather than seriously detracting from the merit principle. The more the categories 
cut across one another, the less will a simple addition of percentages reveal the effective extent of 
reservations. The problem of composite reservations emphasizes the need to go beyond nominal 
reservations to a measurement of effective reservation. 
The conflict among High Courts pointed to a set of questions which remain unanswered after the 
Supreme Court's purported resolution of the matter. In Chanchala v. State of Mysore,  159.   the Court 
encountered a version of the Mysore admissions policy that had achieved a Byzantine complexity. 
The four medical colleges run by the state in connection with its three universities had places for 
765 entering students. (In addition, the state controlled 59 places in the four private medical colleges 
located in the state.)  160.   Of the total of 824 places, 60 were set aside for the following categories: 
students from Union territories and States where there are not medical colleges, students 
from relatively less developed Commonwealth countries, cultural scholars and students 
under T.C.S. of the Colombo Plan and Special Commonwealth Assistance Plan, students 
from Nepal, repatriates from Burma, Ceylon, Mozambique, children of Defence 
Personnel and Ex­Defence Personnel, students who have passed L.A.M.S. and L.U.W.S., 
lady students taking family planning programme, children of political sufferers, and . . . 
students from Goa.  161.   
Of the remaining places,  162.   80% were open to those who passed the Pre­University Course ("the 
common qualification for eligibility" to 
____________________ 
159. Chanchala v. State of Mysore, A. I. R. 1971 S. C. 1762. 
160. At A. I. R. 1971 S. C. 1764 the Court refers to government control of 10% of 480 seats, but at pp. 
1765 and 1767 the figure mentioned is 59 seats. It is not clear to what extent the distribution or 
reservation provisions applied to these 59 seats. 
161. Id., at 1765. 
162. Not 705 but 720, since the Court reports that "any one of the lady candidates can take up" one of 
the 15 seats set aside for the Family Planning Programme (A. I. R. 1971 S. C. at 1770). 
­433­ 
medical courses) or an equivalent examination; 20% were open to graduates who had taken the 
required subjects. Forty­eight percent of these same places were reserved for backward classes (15% 
for SC; 3% for ST; and 30% for Other Backward Classes). Finally, each of the three universities was 
supposed to allot at least 80% of its seats to students passing from its own colleges. By my 
calculation less than 46% of the seats remained open to "general" applicants, assuming that the 
backward classes and miscellaneous reservations did not overlap very substantially in practice. 
The Court distinguished the "university­wise" distribution from the district and area distributions 
previously struck down, finding that it related to legitimate educational objectives. It is open to the 
government "to decide the sources from which admission would be made, provided of course such 
classification is not arbitrary and has a rational basis and a reasonable connection with the object of 
the rules."  163.   
When it turns to the 60 seats for the various categories, the Court again invokes the notion that this 
is not a reservation but a specification of sources. 
[T]he Government is entitled to lay down sources from which selection for admission 
would be made. A provision laying down such sources is . . . not a reservation as 
understood by Article 15 against which objection can be taken on the ground that it is 
excessive.  164.   
Apparently, the 50% limit is specifically related to reservations for Backward Classes under Article 
15(4). 
Setting apart 60 seats [for various other categories] is . . . not a reservation, but laying 
down sources for selection necessitated by certain overriding considerations such as 
obligations toward those who serve the interests of the country's security, certain 
reciprocal obligations and the like.  165.   
Since analytically a reservation is precisely that ­­ a specification of sources from which a selection 
will be made ­­ the distinction is not a very illuminating one. A few pages further on, the Court 
takes up one of these "sources" (the provision for admission of "Children of Political Sufferers"). In 
the course of its discussion, the Court identifies the question as whether there may be "valid 
reservations, apart from those permissible under Article 15(4)."  166.   It finds the "political sufferers" 
provision justified by extending the principle of Article 15(4) to others who are handicapped in 
competition, in this case by the disruption and economic stress entailed by their parents' 
involvement in the national 
____________________ 
163. Id., at 1769. The same notion would seem to fit the PUC vs. graduates distribution, as well. 
164. Id., at 1770. The "sources" language comes in with Chitra Ghosh, discussed at nn. 143 ff, above. 
165. Id., at 1770. 
166. Id., at 1775. 
­434­ 
struggle. The Court's portrayal is persuasive, but it is clear that the question is one of special 
entitlements for candidates whose personal characteristics make them especially deserving; it is not 
a question of arranging sources of recruitment to achieve any medical or educational objectives. 
"[S]uch a classification has a reasonable nexus with the object of the rules which can be nothing 
else than a fair and just distribution of seats."  167   The notion of single­minded devotion to the 
substantive policy at hand (here, medical education) has now given way to recognition that fair 
distribution of medical admissions is itself a legitimate object of governmental policy.  168.   But this 
raises a host of questions, which are only obscured by the formula that "a provision laying down a 
source is not a reservation."  169.   
Two questions emerge of profound import for policies of compensatory discrimination for the 
backward classes. First, what are the other legitimate grounds for government preferment (other 
than performance related to the matter at hand)? That is, what conditions of inequality and 
unfairness may government attempt to remedy? Is there any quantitative limit on the amount of such 
preferment? Specifically, when such preferential treatment takes the form of reservations in scarce 
facilities, is there any quantitative limit on the cumulative total of preferential treatment? 
The Full Bench decision in Jagdish Rai dramatizes the enlarged doctrinal possibilities for 
preferential treatment supplied by Thomas and Chanchala. Reservation of 28% of posts for ex­
servicemen was upheld on the grounds that Thomas had introduced a new dynamic concept of 
equality of opportunity which permitted application of general principles of classification under 
Article 16(1). Just as Chanchala permitted preferential treatment for children of political sufferers 
who are handicapped but did not fall under Article 15(4), so it is "fair and equitable that a just 
proportion of posts" be given to ex­servicemen who suffer peculiar handicaps in competing for jobs. 
 170.   For the State to "classify 
____________________ 
167.Id., at 1776. 
168. The narrower view is still alive. Thus the Court in Pradip Tandon says the only object of 
selection for medical college is to choose the "best possible student" (A. I. R. 1975 S. C. 563 at 
570). And Justice Khanna, dissenting in Thomas, A. I. R. 1976 S. C. 490 at 508, adheres to the 
view that the only legitimate object of rules regarding recruitment is "to ensure efficiency of 
service." 
169. State of U.P. v. Pradip Tandon, A. I. R. 1975 S. C. 563 at 571. Provision of a source is not a 
reservation. What is and what is not a "source" remains unfathomed, but some categories are 
not. In Pradip Tandon the Court finds that "rural area in Uttar Pradesh cannot be said to be a 
source for reservation of the type in Chitra Ghosh's case" (A. I. R. 1975 S. C. 563, 570). 
170. Jagdish Raj v. State of Haryana, A. I. R. 1977 P & H 56, 61. That the shift is 180° is confirmed 
by the judgment's words of commendation for the minority view in Sukhnandan (id. at 60). 
­435­ 
them separately as a source of recruitment and reserving posts for them" is "an extension of the 
principle of Art. 16(4) to those that do not fall under Article 16(4)."  171.   Supplied with such 
abundant doctrinal resources for justifying preferential treatment of virtually any object of 
governmental solicitude, is there any limit to the expansion of the compensatory principle? 
Expansion of the compensatory principle to a sovereign remedy for any falling short of the 
substantive equality promised by Articles 14­16 equips the government with an authorization so vast 
that the commitment to the lowest social groups represented by Articles 15(4), 16(4), 46, 335, etc., 
is in danger of being overwhelmed by competing commitments. In effect, any other groups thought 
to be the victims of "undeserved want" are now entitled to have the reservation device used on their 
behalf, too. New competitors for special reatment will proliferate,  172.   and public patience with the 
use of reservations may be exhausted. 
These new categories cannot use the classifications forbidden by Articles 15(1) and 16(2), but the 
beneficiaries under these various headings are overwhelmingly drawn from relatively advantaged 
strata of society. They are the temporarily or situationally disadvantaged members of those sections 
of society which enjoy social, educational, and economic advantages. Now all the kinds of 
disadvantage which afflict the better­off­­diplomats, central government employees, retired army 
officers, physically handicapped children in well­to­do families­­are candidates for preferential 
treatment along with the backward classes. It is doubtful that, in a setting of chronic shortage, the 
backward classes will fare as well where government is thought to have a general obligation to 
produce substantive equality as in a regime of formal equality, qualified by a single major exception 
for measures to overcome the disparities associated with the traditional social hierarchy.  173.   
Justice Krishna Iyer, in the course of his contribution to enlarging the notion of equality of 
opportunity in Thomas, discerned the potential for diffusion of the commitment to the backward 
classes and attempted to provide a formula for circumscribing the compensatory principle: 
____________________ 
171. Id., at 61. 
172. Some of the categories commonly employed can be seen as attempts to exempt favored groups 
from the effects of restrictive eligibility requirements, adopted to appease localist sentiments, 
without discarding those requirements. As local restrictiveness grows (see Katzenstein 1979) so 
do provisions for out­of­state categories. 
173. Before we can tell the net effect of these other categories of preference, we have to know 
whether they will operate as guaranteed minimums or on an "over and above" basis (see chap. 6, 
§B). If they are guaranteed minimums and the beneficiaries are from more or less advantaged 
groups, there may be little net effect; these reservations would be little more than a device for 
allowing authorities to take credit for admissions that would have been made anyway. 
­436­ 
The distinction would seem to be between handicaps imposed accidentally by nature and those 
resulting from societal arrangements such as caste structures and group suppression. Society being, 
in the broad sense, responsible for these latter conditions, it also has the duty to regard them as 
relevant differences among men and to compensate for them whenever they operate to prevent equal 
access to basic, minimal advantages enjoyed by other citizens.  174.   
The initial attractiveness of this formula is soon dissipated by reflection. Putting aside the question 
of whether medical admissions or high government posts should be distributed according to the 
same principles as "basic minimal advantages," how sturdy is the asserted distinction between 
natural and social handicaps? Natural and social categories overlap: which, for example, is 
"woman"? or "handicapped"? Beyond this, why is compensation thought appropriate to offset 
conditions "resulting from social arrangements" but not those "imposed . . . by nature"? 
What does it mean to be "responsible" for former conditions? It was not the existing State that 
imposed them. Is it responsible for conditions imposed, supported, or tolerated by its predecessors? 
Or because it has power to alter them and undo their results? But the State has the power to 
overcome "natural" conditions as well. Does it not have comparable responsibilities for failing to do 
so?  175.   For example, is an accidental injury (or a form of blindness.) that could have been prevented 
by known measures any less a matter of societal responsibility than "caste structures"? The former 
may be considerably more tractable to government action. If this attempt to erect a coherent theory 
of compensation proves wanting, it should not obscure the fact that the Constitution and a quarter­
century of attendant policy­making do embody a sustained commitment to offset a specific historic 
constellation of inequalities implicated with the traditional social hierarchy in India. To generalize 
from this commitment may diffuse rather than strengthen it. 
E. THE NUMBER OF BENEFICIARIES 
Judicial control over the extent of reservations leads back to the question of whom the State may 
designate as backward. Does a limit on the percentage of places that may be reserved imply a limit 
on the percentage of the population for whom they may be reserved? There is no explicit 
constitutional limitation on the number of persons that may constitute the backward classes. 
However, independently of any limita­ 
____________________ 
174. State of Kerala v. N. M. Thomas, A. I. R. 1976 S. C. 490 at 538. 
175. Cf. Calabresi and Bobbitt 1978. 
­437­ 
tion on the percentage of places reserved, the courts have inferred some quantitative limit on the 
 percentage of the population designated as beneficiaries. In the  Kesava Iyengar
    case, the court did
    
not object to the inclusion among the backward of over 95% of the population of the state.  176. 
 However, in  Ramakrishna Singh v. State of Mysore
   , another bench of the same court held invalid a
    
scheme for the same list of beneficiaries on the ground that designation of 95% of the population as 
Backward Classes "is a fraud on the Constitution."  177.   Even in the absence of any fraudulent intent, 
the use of the power bestowed by Article 15(4) in order to benefit others than the backward was 
invalid. Finding that the Constitution did not intend that this power be used to benefit all but a few, 
the court held the scheme an unconstitutional discrimination against the excluded 5%. Similarly, 
fifteen years later, a Full Bench of the Allahabad High Court thought that to call backward the 80% 
of Uttar Pradesh's population residing in rural areas "would . . . be an appalling opinion, staggering 
to one's imagination."  178.   The Supreme Court, affirming on this point, thought this amounted to "a 
case of discrimination in favour of the majority of rural population."  179.   In both of these instances 
the putative beneficiaries were found not to be legitimate objects of preferential treatment, so the 
courts did not reach the question of numerical limits on the number of legitimate beneficiaries. 
Although the courts have recoiled from 80 and 90% figures, reservations for groups ranging from 
over 30 to over 50% have passed muster on several occasions without exciting judicial hostility. 180. 
When some 20% for Scheduled Castes and Tribes are added, this means that beneficiaries of 
preferential treatment are a clear majority of the population. No limit corresponding to the widely 
accepted 50% ceiling on reservations has even been hinted at, although existing doctrine provides 
some starting points for such an argument. 
In the Balaji case, the Supreme Court did not address itself explicitly to the number of beneficiaries 
of the Mysore scheme. However, it is clear that the Court's strictures against preferences for those 
who were merely "less advanced" or only "relatively backward" indicate that the number of 
beneficiaries was an important point. Several of the Court's 
____________________ 
176. A. I. R. 1956 Mys. 20. 
177. A. I. R. 1960 Mys. 338 at 348. 
178. A. I. R. 1975 All. 1, at 7. 
179. A. I. R. 1975 S. C. 563, at 571. 
180. Thus in Balaram the Court upheld Backward Classes lists estimated to consist of 38% of the 
population of Andhra Pradesh (A. I. R. 1972 S. C. 1375). In Janki Prasad, the Court approved of 
almost all of a list which apparently consisted of 46% of the state's population (A. I. R. 1973 S. 
C. 930). In Periakuruppan, the Court upheld a Backward Classes list supposed to contain 51% of 
the population of Tamil Nadu (A. I. R. 1971 S. C. 2303). 
­438­ 
tests of backwardness imply severe quantitative limitations. The Court holds that "only communities 
 which are  well below the state average
   . . . can properly be regarded as educationally backward
    
181.
classes of citizens."      The "state average" in question is the mean, and presumably considerably 
less than half of the population could qualify as "well below" the mean. Second, the Court limits 
those who may legitimately be designated as backward to those whose "backwardness [is] 
comparable to Scheduled Castes and Scheduled Tribes."  182.   Such a standard, if applied to any 
tangible criteria of literacy, income, etc., would severely restrict the number of possible 
 beneficiaries. (The  Balaji
    Court elsewhere endorses an economic test of backwardness, but it is not
    
clear whether it means the "well below the state average" or the "comparable to the Scheduled 
Tribes and Castes" standards to specify the level at which an economic test must be set.) 
The relationship between the percentage of seats reserved and the percentage of the population who 
enjoy the reservation is a puzzling one. It seems to be generally assumed that a reservation should 
be at about the same level as the representation of that group in the population­­thus, for example, 
seats or posts are normally reserved for the Scheduled Castes and Scheduled Tribes roughly in 
accord with their percentage in the population. In Pandit v. State,  183.   the Bombay High Court found 
the beneficiaries percentage of the state's population "an objective and just test for determining the 
proportion of seats to be reserved in the medical colleges."  184.   There is "Perhaps . . . no better basis 
for such reservation than the proportion of the population of the Backward Classes to the whole 
population of the State."  185.   Such 
____________________ 
181. A. I. R. 1963 S. C. 649 at 660 (emphasis added). 
182. Id., at 668. However, in the same paragraph the Court observed that "[i]t was realized that . . . 
there were other classes of citizens who were equally, or maybe somewhat less, backward than 
the Scheduled Castes and Scheduled Tribes and it was thought that some special provision ought 
to be made even for them" (emphasis added). 
183. A. I. R. 1972 Bom. 242. 
184. Id., at 250. The reservation in question operated as a guaranteed minimum. The obvious 
"representative" thrust of the Court's argument may not hold in cases where the reservation is of 
places over and above those acquired by merit. See chap. 13, §B, below. 
185. Id., at 251. This proportionality theme provides the basis for Justice Fazl Ali's argument that the 
50% limit on reservations is not a definite limit but merely "a rule of caution": if the government 
of a state with "backward classes of citizens which constitute 80% of the population . . . reserves 
80% of the jobs for them, can it be said that the percentage of reservation . . . violates the 
permissible limits of . . . [Art. 16(4)]? The answer must necessarily be in the negative," since the 
object of this provision is to secure adequacy of representation ( State of Kerala v. N. M. 
Thomas, A. I. R. 1976 S. C. 490 at 555). 
­439­ 
statewide ratios may be used even when the population in the catchment area of a particular scheme 
of benefits may vary from it. 
Many reservations for the Backward Classes are much smaller than the population percentage of 
those classes. Courts have responded differently to this discrepency. In Jacob Mathew the fact that 
the reservation was smaller than the population of the group impressed the judge as evidence that 
the selection was a prize of caste politics rather than a genuine attempt to assist the backward. 186. 
On the other hand, in Vishwanath the Mysore court found the fact that 30% of the seats were 
reserved for 50% of the population a warrant of the reasonableness and moderation of the measure. 
 187.   Since, as we have seen, there is no necessary relationship between the nominal size of a 
reservation and its effect, it might be doubted whether either presence or absence of 
commensurateness to population ratios is in itself an indication of reasonableness. 
Furthermore, quite apart from the guesswork by which they are estimated, population figures for 
backward classes may be misleading characterizations of the number of beneficiaries. The 
population of backward classes and the beneficiaries of any given scheme cannot be equated. The 
relationship between them is problematical. If we are dealing with a scheme for providing health 
care or free books and slates to First Form schoolchildren, we might take the percentage of the 
population in the backward classes as a rough estimate of the percentage that might benefit from 
this scheme. But in what we have called higher­echelon benefits, those which require previous 
training and accomplishment, there may in fact be few members of the backward classes in the pool 
of persons eligible to compete for a benefit such as a high government post or a medical college 
admission. Even if 60% of the population are members of the backward classes, it may be that only 
30% of the government officers eligible for promotion are from these groups. 
Symmetry between the percentage of places reserved and the percentage of the population who are 
backward may produce some curious distortions. When effective competitors from these groups are 
few, to confine the backward population to the percentage of places reserved might diminish the 
pool of competitors. Conversely, if because of population figures, the reservation is set high relative 
to the number effectively eligible, there may be a windfall pile­up of benefits on a few fortunate 
individuals.  188.   
But the abandonment of symmetry exposes the real problems of a 
____________________ 
186. A. I. R. 1964 Ker. 39 at 61. 
187. A. I. R. 1964 Mys. 132 at 137. 
188. Cf. the situation described in Harpartap Singh v. Union of India, 1970 S. L. R. 40; see also State 
of Punjab v. Hira Lal, A. I. R. 1971 S. C. 1777. 
­440­ 
limit on the number of beneficiaries. First, as the beneficiary class approaches the total population 
there arises the question of whether the administration of eligibility requirements is justifiable. Is it 
worth its cost in money, energy, and attention? And, as a scheme becomes more inclusive, at some 
point it takes on the character of singling out the omitted groups for unfavorable treatment rather 
than conferring special treatment on those who are included. Schemes conferring benefits on 
majorities may be hard to terminate or reduce. 
The non­included are not the only victims of expansive lists of Backward Classes. As the backward 
category expands, benefits may be diluted. Since many benefits are distributed among the backward 
on a merit basis inter se, the least prepared and advantaged will suffer in merit competition with 
more favored groups. Large numbers of Backward Classes will provide political support for 
measures for these groups, so that the most backward may benefit from enlargement of this 
category. But unless there is some internal sub­division, the most backward groups are likely to be 
squeezed out by the superior "merit" of the less backward as well as by their superior access to 
officials and programs. 
Who is hurt by enlargement of the Backward Classes category will depend in part on whether 
preferences are distributed on a guaranteed minimum or an "over and above" basis.  189.   If a 
reservation is merely a guaranteed minimum, enlarging the Backward Classes category will 
presumably shift places from the most backward groups to the best prepared among the Backward 
Classes. On the other hand, if reservations are places over and above those gained by merit, 
enlargement of the backward category will increase the effect of reservation at the expense of non­
beneficiaries as well as squeezing out the most backward. In either case, the most backward lose. 
F. BENEFITS OTHER THAN RESERVATIONS 
The State is not limited to "reservations" as a method of preference. Only in the area of government 
employment is "reservation" constitutionally specified, but even there reservations may be 
accompanied by other preferential devices­­fee concessions, waiver of age limit, etc. 
The Balaji case disposed of the notion that a limitation on the extent of reservations must rest upon 
the construction of the word "reservations" in Article 16(4) and established that such limits are 
applicable to 
____________________ 
189. See chap. 13, §B, below. 
­441­ 
reservations under Article 15(4) as well. The question arises whether there are similar constitutional 
limitations on the quantity of other kinds of preferences. Presumably, "reservations" would be 
covered wherever used. But the national interests which made the limitation imperative in the 
professional and technical colleges or in promotions in government service may not be present to 
require a similar limitation on the reservation device in other schemes. 
A percentage limitation of the sort applied to reservations would have little meaning with respect to 
other kinds of benefits authorized by Article 15(4), such as expenditures for housing and education 
or protection against foreclosure for debt.  190.   The State is entitled to provide such things as housing 
facilities for the exclusive use of Scheduled Castes.  191.   Can it not do the same for other backward 
groups and with other kinds of facilities? May it not have a health scheme for Scheduled Castes only 
or an agricultural program for Scheduled Tribes only? Presumably, the State may do any of these 
things. Is the State then limited in the percentage of expenditures for e.g., housing, that it may 
devote to such purposes? Perhaps the notion that provisions for the advancement of the backward 
must be balanced against broader national interests might be applied here. But the broad national 
interests in the limitation of expenditure for the backward are more difficult to identify. In the case 
of government employment or admissions to medical colleges, there is a public interest in some 
share of these scarce resources going to the "haves" because, presumably, they will be more 
productive and efficient, bestowing more benefit on the public. In addition there are the guaranteed 
and enforceable fundamental rights of the merit competitors in these areas. But expenditures for 
such matters as housing and health do not involve specific and enforceable rights of others. Nor is it 
clear that there are parallel public benefits in the rewarding of the advantaged. There is not 
necessarily any accretion of public benefit from giving housing subsidies to those who already have 
good housing. 
If the Devadasan majority's notion that opportunity must be equalized annually were imported into 
this area, it would make impossible large­scale diversion of resources to meet the problems of 
backward groups. Unlike legislative seats, government posts, and medical college admissions, these 
welfare areas do not involve values of which government is the exclusive or prime dispenser. Valued 
resources in health, 
____________________ 
190. Of course, it may prove a useful instrument of policy to set percentage goals for expenditures. 
For example, the Karnataka Government recently instructed municipal bodies to earmark 18% of 
their annual budgets for programs for welfare of the Scheduled Castes and other weaker sections 
( Times of India, Aug. 22, 1977, p. 6). 
191. Moosa v. State of Kerala, A. I. R. 1960 Ker. 335 Cf. Jagwant Kaur v. State of Bombay, A.I.R. 
1952 Bom. 461. 
­442­ 
housing, land, etc., lie largely in private hands. Government's role in these fields is not that of the 
major distributor of scarce opportunities, but rather that of a countervailing re­distributor. There 
would seem to be no constitutional bar to prevent government from channelling all of its 
expenditure on, e.g., housing, to the backward classes. 
The case for judicial control over the extent of such preferences is far less clear than over 
reservations. For schemes of reservation directly involve the guaranteed and enforceable 
fundamental rights of other citizens. But in the case of expenditures for such matters as housing and 
health, there are no tangible rights of others involved, and the courts would have only a vague 
national interest to guide them­­or at most the non­justiciable rights of other citizens contained in 
the Directive Principles of State Policy. It seems clear that in making these nonjusticiable the 
purpose was to keep the details of policy involved in development out of the courts. Any detailed 
judicial review would inject the judiciary into an area in which both its authority and its competence 
are not clear. 
­443­ 
13 The Design and Operation of Compensatory Discrimination Programs: Constitutional 
Limits and Judicial Control 
A. RULEFULNESS AND DISCRETION 

WE HAVE seen that preferential treatment may be initiated by Centre or states, by the executive as 
well as by legislative bodies. Courts have been vigilant to insure that reservations and other special 
provisions are made in a ruleful manner and by competent authorities.     1. A state may not employ a 
scheme of reservations where there has been no government order specifying the backward classes 
2. Nor may a state reserve places in medical 
or fixing the proportion of places reserved for them.    
school admissions where, at the time of selection, no special provision for this had been made by 
3. Nor may a state appoint Scheduled Castes or Back­ 
either legislative enactment or executive order.    
____________________ 
1. The recent decision in Sharda v. Visveswariah Engineering College, A. 1. R. 1978 Kar. 111 
[NIC], departs strikingly from the requirement of rulefulness which imbues the cases cited 
below. The university had permitted a Backward Class student to transfer from an electrical 
engineering to an electronics program in preference to other would­be transferees with higher 
marks. The single judge held that in the absence of any express order permitting such 
preferences, the university's course might have been illegal, but since it was fair and substantially 
just, this was not a fit case for the exercise of the court's extraordinary power to issue a writ. 
2. Triloki Nath Tiku v. State of Jammu and Kashmir [1], A. I. R. 1967 S. C. 1283. 
3. Lalita Shuri Tikku v. State of Jammu and Kashmir, A. I. R. 1966 J. & K. 101: "there must be 
some order or notification to show that state has applied its mind" (Id., at 104). Some formality 
must be observed, although the Supreme Court has left open the question whether formal 
publication is required to make an order effective ( Triloki Nath Tiku v. State of Jammu and 
Kashmir [11], A. I. R. 1969 S. C. 1 at 3). Where a valid list exists, a particular community may 
not be omitted because of administrative oversight without express modification of the order "by 
a proper decision of the government" ( Susila v. State of Madras, A. I. R. 1970 Mad. 399, 400). 
­444­ 
ward Classes to posts in contravention of its own rules of recruitment, by invoking a broad power of 
the Public Service Commission to consider the case of any candidate with the prescribed 
4. 
qualifications, to the disadvantage of more­qualified candidates.    
Similarly, preferential treatment, when properly authorized, must be implemented rulefully. Notice 
must be given to applicants that a selection is subject to reservations: implementation of an 
otherwise valid reservation (for women, in this case) was set aside where the announcement of the 
Selection Board led applicants to rely on a merit competition unencumbered by reservations.     5. A 
principal of a medical college may not decide that unfilled seats reserved for Scheduled Castes 
should be filled with members of Backward Classes rather than revert to the general pool, where 
6. Nor may the State overlook 
this "carry­over" arrangement is not authorized by statute or rule.    
merit within its specified reservations: the State may not select Scheduled Castes candidates ranking 
lower in the competition over those who rank higher.    7. 

The Devadasan notion of limiting the future effects of unfilled reservations has been extended to a 
8. on the interests 
more general principle that reservations may not operate with retrospective effect    
9. Thus it has been held that the State might not 
of individual (non­beneficiary) appointees.    
withdraw a post from an employee already appointed in order to give it to a Scheduled Caste­­even 
where the seat was originally reserved and was given to the non­Scheduled Caste employee because 
there was no applicant eligible for the reservation.  10.   
____________________ 
10. In Sudama Prasad v. Divisional Superintendent. A. I. R. 1965 Raj. 109. a railroad employee 
appointed as Chief Clerk obtained an order setting aside the order of the railway authorities who 
had demoted him in order to replace him with a less senior 
4. Channabasaviah v. State of Mysore, A. I. R. 1965 S. C. 1293. 
5. Abodha Kumar v. State of Orissa, A. I. R. 1969 Or. 80. 
6. Abdul Aziz v. State of Maharashtra, High Court at Bombay, Special Civil Application No. 1363 
of 1962, K. K. Desai and Chandrachud, J. J., decided 12 Oct. 1962; Mohd. Abdul Mabood Khan 
v. State of Maharashtra, High Court at Bombay, Special Civil Application Nos. 1802 and 1829 of 
1962, Tambe and Palekar,.J. J. decided 19 Feb. 1963. 
7. Lalita Shuri Tikku v. State of Jammu and Kashmir, A. I. R. 1966 J. & K. 101 at 105; Kushma 
Joshi v. Pro­Vice­Chancellor, A. I. R. 1969J. & K. 136 at 139. 
8. Similarly, a debtor whose caste was added to the Scheduled Caste list by the 1979 amendment of 
the lists was not entitled to benefit from debt relief provisions applicable to loans subsisting at an 
earlier point in time ( Mangilal v. Mittilal­Radhaylal Rastogi, A. I R. 1978 M.P. 160 [NIC]). 
9. The vast corpus of case law on service matters contains indications that it is open to government 
to use the seniority provisions to offset temporary delays in filling reservations. Where 
reservations go unfilled, the government may assign seniority according to its rotational system, 
so those who belatedly take the reserved places take seniority over those who were hired (or 
promoted) because of the temporary lack of persons to fill the reserved places ( Chauhan v. State 
of Gujarat, A. I. R. 1977 S. C. 251, 263). 
­445­ 
But beneficiaries of reservations who are disadvantaged by retrospective changes in reservation 
policy have had little success in securing judicial relief A Scheduled Caste typist who lost his 
chance for promotion when the basis for computing seniority was changed was held to have no 
vested right to promotion on the original basis.  11.   A member of a Backward Class promoted to a 
temporary post in accordance with a state order reserving promotion posts might be reverted to his 
original post when the state government reversed its policy and withdrew reservations from 
Backward Classes. This case involved a temporary post, and the original order was unconstitutional 
and therefore void.  12.   It is not clear whether the rule would apply to permanent appointments or to 
appointments pursuant to orders which are constitutionally valid. But it has been applied to a 
government servant who passed up a desirable selection post on "merit" to take a more desirable 
promotion post reserved for Backward Classes. When the government reversed its policy, he was 
reverted out of the promotion post; the other post had in the meantime been filled and he lost out on 
both.  13.   
While the courts have required regularity in the designation of backward classes and the 
authorization of reservations, they have moved slowly in controlling broad personal discretion at the 
later stages of the selection process. The operation of selection committees provides a vivid 
example. In several states, the portion of the competitive score for college admissions allotted to 
interview by a selection committee is so substantial that it creates the potential for what is in effect a 
covert reservation. 
____________________ 
employee who had belatedly established that he was a member of a Scheduled Caste. But cf. 
Mahendra Nath Pathak v. State of Assam, A. I. R. 1970 A. & N. 32. The original settlement was 
restored when the putative Scheduled Tribe member failed to establish his tribal membership. 
11. Pramanick v. Union of India, A. I. R. 1969 Cal. 576. The basis for computing seniority was 
changed from date of comfirmation to "merit as determined by Public Service Commission or 
results of training tests." Scheduled Castes had gotten confirmation earlier than other employees. 
Even though the court says there was no right to complain of a change, it repels the claim of 
discrimination by observing that "no reservation or special provision in favor of members of the 
Scheduled Castes can be carried to the length of impairing the efficiency of the administration" 
(id., 577). It is not clear whether the same result would have been reached had the change of 
bases been in the opposite direction. 
12. Dasa Rayudu v. Andhra Pradesh Public Service Commission, A. I. R. 1967 A.P. 353. 
13. G. Satyanandhan and others v. State of Andhra Pradesh, etc. Writ Petition Nos. 278, 1352, etc., 
of 1964. High Court of Andhra Pradesh, Basi Reddy and Gopal Rao Ekbote, JJ., 8 March 1965. 
Such cases as these were numerous in Andhra when the government withdrew its Backward 
Class list and reservations in promotions. Some departments permitted those promoted when the 
reservation was in effect to retain their new posts, others reverted them. The Supreme Court 
denied special leave to appeal on this point. 
­446­ 
Selection committees were a favorite device in pre­Independence Madras for assuring the proper 
communal balance.  14.   When the Madras system of outright communal quotas was struck down by 
the Supreme Court in 1951, the new modest reservations that replaced them were accompanied by a 
revival of selection committees. Committees consisting of local notables, presided over by the Vice­
Chancellors, were empowered to take into consideration physical fitness, sports activities, special 
aptitude, and the backwardness of the applicant's family in the matter of higher education. In their 
original incarnation these procedures "had the intended effect of reducing drastically the proportion 
of Brahmin students."  15.   An enthusiastic defender of selection committees leaves little doubt that 
this remains a major element in their purpose: 
The sole criterion of highest merit based on marks, so beneficial to Brahmins, has gone. 
Physical fitness, special aptitude, and educational backwardness of the family, will all 
operate against the Brahmins. The test of physical fitness for once will eliminate puny 
creatures with no personality from becoming engineers and doctors in the future, and 
this is a step in the right direction.  16.   
The potential for communal bias in the operation of the Madras selection committees is greatly 
enhanced by the maintenance of an entry for caste in the Secondary School Leaving Certificate 
books which must be supplied by applicants.  17.   
The combination of modest reservations and use of selection committees in Madras had the effect of 
discouraging litigation, so that the Madras system of selection long remained unchallenged. When 
writ petitions were finally brought in the late 1960s, the first one to challenge the use of selection 
committees was P. Rajendran v. State of Madras.  18.   Three­man selection committees were to award 
up to 75 marks (out of 300) for extra­curricular activities under five heads. The three members did 
not sit together, but each interviewed each candidate separately 
____________________ 
14. "When the Justice Party was in power they appointed Selection Committees for admitting 
students into the Colleges, not being able to trust the Principals, who were mostly Brahmins to 
do justice" ( Dharmalingam 1951: 79). The committees flourished from 1922 until they were 
abolished in 1938 ( Irschick, 1977). See Irschick 1969: 238 ff. Selection boards are credited by 
knowledgeable observers with effecting the virtual exclusion of Brahmins from schoolteaching in 
Maharashtra. (Interview, Poona, 1966.) 
15. Rama Rao 1967: 74. 
16. Dharmalingam 1951:82. 
17. Rama Rao ( 1967: 74) reports that after Independence the Madras Government "has felt bound to 
abandon requirement of information of caste in application forms, [but] the S. S. L. C. [Secondly 
School Leaving Certificate] books are being maintained as previously, offering an unfailing clue 
about the caste of all candidates" 
18. A. I. R. 1968 S. C. 1012. 
­447­ 
and gave up to 25 of the total of 75 marks. After the selection was over the mark­sheets were 
destroyed. The Supreme Court, striking down the scheme on other grounds, found both the 
individual interviews and the destruction of the mark­sheets decidedly "odd."  19.   Nevertheless, the 
Court found the rule "indicates what matters have to be taken into consideration" and was 
not prepared to accept that the Committee did not follow the criterion indicated in Rule 
10(d) in allotting the marks provided in that rule. . . . We do not think that it is possible 
to provide any further guidance in the matter and the rest must be left to the Selection 
Committee.  20.   
The selection committee problem had been more fully considered in earlier Mysore litigation.  21.   In 
1963, when the Mysore Government abandoned its community­wise list of backward classes and 
switched to an occupational­cum­means test, it suddenly exhibited an enhanced concern for the 
extra­curricular accomplishments of applicants to professional colleges.  22.   The State decided that 
the interview marks should comprise 25% (75 marks) of the maximum marks on the examination in 
the optional subjects.  23.   The two selection committees, each operating over half the state, were 
authorized to allot marks for interview­having regard to . . . 
1.  General knowledge. 
2.  Aptitude and personality. 
3.  Previous academic career, including special distinctions. etc. 
4.  N.C.C. [National Cadet Corps], A.C.C. [Auxillary Cadet Corps], etc. 
5.  Extra curricular activities including sports, social service, debating, dramatics etc.  24.   
According to the Committee's chairman, each "interview" generally took two or three minutes: 
generally 3 to 4 questions were put to each candidate to test his general knowledge, 
aptitude, the practical application of scientific knowledge to 
____________________ 
19. Id., at 1017. 
20. Id. 
21. Chitralekha v. State of Mysore, A. I. R. 1964 S. C. 1823, aff'g Viswanath v. Government of 
Mysore, A. I. R. 1964 Mys. 132. 
22. Previously, selections committees who interviewed candidates were to award a maximum of 20 
marks (out of 300): "A maximum of 20 marks is fixed as interview marks to be awarded by the 
selection committee for assessing the suitability of the candidate for engineering profession, 
having regard to the following factors: (i) aptitude, (ii) general ability and physical fitness, (iii) 
sports or N. C. C. activities or scouting or other similar activities, (iv) extra curricular activities 
such as social service etc" (quoted at A. I. R. 1964 Mys. 142). 
23. A. I. R. 1964 Mys. at 141. 
24. Id. 
­448­ 
everyday life, and mental alertness and ability to express himself clearly, the last two factors being 
relevant for the criterion of personality. While each candidate was interviewed, one of the members 
of the Selection Committee was examining the Certificates produced by candidates in support of 
their extra­curricular activities like N.C.C. or A.C.C., sports, debating, dramatics and social 
services.  25.   
The committee's allotment of marks on this basis was attacked as an arbitrary, unguided, and 
uncontrolled power. The High Court in Viswanath v. Government of Mysore denied that is was 
unguided power, for in the absence of specific allocation it must be presumed that the government 
intended that "each one of those heads should carry 1/5th of the 'Interview' marks."  26.   Since the 
Committee had "acted under the impression that 25% of the marks were left in their hands to be 
given in any matter they considered best," they did not deal with the several heads separately and 
gave only consolidated marks. 
Instances were brought to our notice where . . . applicants who were proved to possess at 
least 3 out of the 5 tests . . . have secured . . . 10 marks or even less out of 75 marks.  27.   
This failure to use "objective standards" was a misuse of the power conferred on them.  28.   The court 
quashed the interview and directed the selection committees to interview petitioners afresh in 
accordance with the directions. 
On appeal to the Supreme Court it was argued that selection by interview is repugnant to the 
doctrine of equality embodied in Article 14, because it introduces a strong subjective factor into 
selection, giving ample room for discrimination and manipulation.  29.   The Court conceded that 
manipulation was possible in interviews as in other selection procedures, but 
that it is capable of abuse is not a ground for quashing it. So long as the orderlays down 
relevant objective criteria and entrusts the business of selection to qualified persons, this 
Court cannot obviously have any say in the matter. In this case the criteria laid down by 
the Government are certainly relevant . . . .  30.   
____________________ 
25. Id., at 143. 
26. Id., at 144. 
27. Id., at 144. 
28. Id. 
29. The argument was also made that giving weight to the interview was changing or raising the 
qualifications and therefore outside the power of the State, since the co­ordination and 
determination of standards of a University is a subject reserved to the Union by VII Schedule, 
List 1, Entry 66 of the Constitution. The courts have indicated that the fixing of standards was 
pre­empted from State authority only where the State's action would make impossible or difficult 
the exercise of Parliament's power. Here, there was no such encroachment (A. I. R. 1964 S. C. at 
1839­40). 
30. A. I. R. 1964 S. C. at 1831. 
­449­ 
The Court seemed to leave the door open for a further look when it concluded: 
We cannot, therefore, hold without better and more scientific material placed before us 
that selection by interview in addition to . . . the written examination is itself bad as 
offending Article 14. . . .  31.   
But the Court took no notice of the disparity between the criteria and the time in which qualified 
persons could possibly reach such complex judgments. In a later Mysore case, counsel convinced 
the High Court that the interview methods employed were "highly rudimentary," but the very 
disparity between the immensity of the committee's task and the paucity of its resources convinced 
the court that under the circumstances the interviews were satisfactory.  32.   
Subsequently, in Jammu and Kashmir, a Full Bench encountered an even more flagrant use of 
interviews.  33.   Medical school admissions were distributed entirely on the basis of marks given by a 
committee (consisting of two members of the Public Service Commission) which interviewed each 
student for two to three minutes. The committee interviewed only those who had obtained aggregate 
marks of 45%. It gave no weight to the academic accomplishment of those who came over this 
threshold. The committee avowed that it allotted marks as follows: physical fitness, 10; aptitude, 10; 
general knowledge, 40; personality, 40. In what is surely the ultimate tribute to the political primacy 
of medical school admissions, the committee's recommendations were submitted to the Chief 
Minister, who issued the order of admission. 
The High Court found that it was well­nigh impossible for the committee to assess merit objectively 
within such a short time; that its failure to take academic merit into consideration made the 
selection patently unreasonable, arbitrary, and subjective; that ignoring merit was "inherently 
abhorrent to the notions of a civilized and progressive state of Society."  34.   The court politely 
concluded that what the committee did was not what government intended it to do (i.e., to give due 
weight to academic merit along with interview marks) and struck down the admissions.  35.   
____________________ 
31. Id., at 1832. 
32. Subhashini v. State of Mysore, A. I. R. 1966 Mys. 40 at 47. 
33. Subhash Mohan Jalali v. Principal, Medical College, A. I. R. 1967 J. & K. 106. 
34. Id., at 110. 
35. The Full Bench spelled out the acceptable ways in which such a selection might be made. These 
included (1) use of aggregate marks at the first qualifying examination, with or without an 
interview; (2) adding interview marks to the marks in the last qualifying examination; (3) 
interviewing only candidates with high scores at the last qualifying examination and allotting 
marks at the interview, giving due weight and consideration to "the academic merit of the 
candidates as revealed by their perform­ 
­450­ 
 When the selection committee issue returned to the Supreme Court in  Periakaruppan v. State of
    
   36.   the bench included the judge who had written the High Court judgment in 
 Tamil Nadu , 
 Viswanath . The state had set up regional selection committees to interview applicants to medical
    
college. Of the total of 275 marks, 75 were to be awarded on the basis of the following tests: 
1.  Sports or National Cadet Corps Activities; 
2.  Extra Curricular special services; 
3.  General Physical condition and endurance; 
4.  General ability and 
5.  Aptitude. 
The committees were composed mostly of medical practitioners, and the interviews lasted roughly 
three minutes for each candidate. The Court reflected that 75 out of 275 marks seemed excessive for 
a threeminute interview, but felt itself unable to hold it beyond the power of the government to 
provide such a high portion.  37.   Although there was "disturbing" discrepancy between interview 
marks and other measurements of performance, the Court declined to conclude that either the 
committees or the government were manipulating the interview marks.  38   Nor was it fatal that no 
objective criteria were established: the 
____________________ 
ances in the last qualifying examination" (id., at 109­10). When it was later charged that these 
acceptable modes of selection had not been followed, the High Court called for all the relevant 
records of the Selection Committee and found that "by and large candidates have been selected 
purely in order of merit by adding the marks secured by them at the interview to the marks 
fetched by them at the last qualifying examination." The documents revealed that "the Selection 
Committee did not try to plump up marks of a particular candidate so as to tilt the scale of one 
candidate in favour of another. The marks awarded . . . appear to be fair and reasonable and are 
fully consistent with the aggregate fetched by the candidates in the last qualifying examination" ( 
Sardool Singh v. Medical College, A. I. R. 1970 J. & K. 45 at 47). it appears that once tamed into 
rulefulness, selection committees with only a few minutes to spend can do little that makes the 
outcomes very different than those produced by the examinations. 
36. A. I. R. 1971 S. C. 2303. 
37. Some courts have been unperturbed by the perfunctory character of interviews. In Daya Ram v.  
State of Hatyana, A. I. R. 1974 P. & H. 279, 280, petitioners claimed that in the selection of sub­
inspectors the interview was an empty formality: "The candidates entered through one door and 
were made to get out from the other. No question whatsoever was put to them." The selection 
board denied this and said that "only 253 candidates were interviewed on the said date between 7: 
00 A.M. and 3: 00 P.M." Assuming that the board worked continuously, this would leave 1.7 
minutes per candidate. 
38 A. I. R. 1971 S. C., at 2307­8. "[N]umerous students whose performance in the University 
examinations was none too satisfactory nor their past records creditable . . . secured very high 
marks at the interview. It is also true that a large number of students who had secured very high 
marks in the University examination and who performed very well in the earlier classes had 
secured low marks at the interview" (id., at 2307). 
­451­ 
Court found the specification of the five headings "sufficiently objective in character."  39.   However, 
even though the rule did not prescribe separate marks for separate heads, this could not be read as 
permitting the committee to allot marks as it pleased. The Court found the failure to divide the 
interview marks under the various heads and giving them "in a lump" was clearly illegal.  40.   
In addition, the committee departed from the rules by taking into consideration irrelevant matters 
and failing to consider those required by the rules. Thus under N.C.C training the committee judged 
candidates not on their records as cadets but on the response to questions about the role of N.C.C. in 
national life. And physical condition and endurance were not judged by medical examination: "It 
was in the nature of things not possible for the selections committee, though composed of eminent 
doctors, to find out the physical condition and endurance by a mere look at the candidate."  41.   From 
the facts available, the Court concludes that "it is clear that the candidates were not interviewed in 
accordance with the rules governing the interview."  42.   
Pursuant to the Court's direction, the state constituted a new selection committee to fill the 
remaining 24 unfilled seats, but again the petitioner failed to be selected, because of low interview 
marks. Although the Court was obviously sympathetic to petitioner's claim that he was the target of 
retaliatory hostility, it found that he failed to carry the burden of proving mala fides on the part of 
the committee. But, it found that the procedures followed by the selection committee contravened 
the Court's earlier judgment by departing from a strict allotment of interview marks among the five 
heads and by proceeding on "wholly wrong premises while granting marks under the head 'National 
Cadet Corps activities.'"  43.   The selections were quashed, and the government was ordered to 
appoint yet another fresh selection committee. 
Subsequently, in Janki Prasad v. State of Jammu and Kashmir,  44.   another attempt of the Jammu 
and Kashmir, Government to use the selection committee device to salvage a system of communal 
quotas was struck down by a constitution bench of the Supreme Court. This was the 
____________________ 
39. Id., at 2308. 
40. In Gudigar v. State of Mysore, 1972 (2) Mys. L. J. 202 [NIC], the court declined to apply this 
mechanical allotment requirement where the interview was designed to test suitability rather than 
to examine for knowledge. Mention of desired traits was "by way of guidance" to the selection 
committee, and assurance against arbitrariness was to be had by ensuring that the selection was 
entrusted "to persons of known experience, integrity and capacity" (id., at 215). 
41. A. I. R. 1971 S.C. at 2308. 
42. Id., at 2309. 
43. A. Periakaruppan Chettiar v. State of Tamil Nadu, A. I. R. 1971 S. C. 2085, 2089 [NIC]. 
44. A. I. R. 1973 S. C. 930. 
­452­ 
fourth round of Supreme Court litigation in a protracted and bitter struggle over the communal 
quotas within the Jammu and Kashmir schools.  45.   Promotions to the post of headmaster had been 
found invalid on the ground that they were distributed on the basis of community. To comply with a 
Supreme Court order to make new selections, the state reverted all of those earlier promoted and 
required all of the 1,100 teachers who were eligible for promotion to appear for an interview by a 
Departmental Promotion Committee. (Four hundred Kashmiri Pandits boycotted the interview 
because they "did not have any faith in this system of selection.")  46.   The previous appointments 
(invalidly made on a communal basis) had given 178 posts to Muslims and 134 to Jammu Hindus. 
Selecting anew on merit, the selection committee chose 177 Muslims and 134 Hindus. But finding 
some substance to the state's claim that things might have been different if the Kashmiri Pandit 
teachers had participated, the Court declined to "go merely by the coincidence."  47.   
But the interview was in two other ways "thoroughly unsatisfacory". First, the selection was based 
entirely on the interview, without any reference to the service records, confidential reports, or any 
other information about the candidates. The Court thought that in such a case where there are 
relevant records that should be consulted, "the interview cannot be made the sole test."  48.   The 
state's failure to avail of this other information was improper. 
Second, the committee fixed a "cutting score" or minimum that was in the Court's view absurdly 
low: "those who got more than 30% marks and more than 20% marks from the [outside] expert 
were declared eligible for selection. This is indeed a travesty of selection . . . a selection made on 
such a poor basis cannot be called a real selection at all."  49.   The whole selection process, then, was 
"wrong and unsatisfactory and must be set aside."  50.   The Janki Prasad court takes judicial 
monitoring of 
____________________ 
45. In Triloki Nath Tiku v. State of J. & K. [I], A.I.R. 1967 S.C:. 1283, the Supreme Court found that 
it had insufficient material to decide if the reservations of 50% for Muslims and 40% for Jammu 
Hindus were indeed for Backward Classes and directed the High Court of Jammu and Kashmir to 
gather the necessary material. In Triloki Nath Tiku v. State of J & K [II], A.I.R. 1969 S. C. 1, the 
Supreme Court found the distribution of posts not within Art. 16(4), but an unconstitutional 
distribution on the basis of community and residence. The state's initial evasive response was set 
aside in Makhan Lal Waza v. State of J. & K., A.I.R. 1971 S. C. 2206. 
46. Id., at 934. 
47. Id., at 934. 
48. Id., at 934. The Court was clearly of the view that the government's failure to avail of these 
records was an exercise of bad faith, finding it "impossible to conceive" that the reports were not 
available as the government somewhat lamely claimed. If a few were missing, said the Court, to 
decline to use any of them "would be ridiculous" (id., at 935). 
49. Id., at 936. 
50. Id., at 936. 
­453­ 
 selection an important step further than  Viswanath
    and 
   Periakaruppan
   . In those cases it was the
    
failure to conform to the rules prescribed for the selection that was the focal point of 
unconstitutionality. Here the rules themselves are held to fall short of minimal constitutional 
standards by disregard of relevant information and adopting standards that abandon minimal 
concern with quality. So rulefulness is not enough where crucial substantive commitments are 
subverted. The history of this litigation does little to dispel the conclusion that the selection 
committee was here yoked to the unconstitutional purpose of communal quotas. It is then difficult to 
be sure how much the State might vary the sources of information or minimal scores in a scheme of 
preference that is legitimate in its scope and in its intended beneficiaries. 
In spite of its dubious history, the selection committee does not lack proponents. Doubts about the 
relevance of grades as the predominant basis for selection are compounded by concern for the social 
bias incorporated in those grades.  51.   If in principle intellectual and leadership qualities can be better 
ascertained in personal encounter than by written examination, it remains unclear that this promise 
could be fulfilled without an investment of resources far beyond that evident from the cases. 
As the cases suggest, the selection committee is a device with an enormous potential for abuse. It is 
abuse that is especially difficult to correct because it is difficult to prove or even to discern. The 
potential for distortion of merit by selection committees is at least as great as that introduced by 
reservations. 
It is now firmly established that the standards of overtness and rulefulness applied to other phases of 
the selection process are also applicable to selection committees. Just as courts have looked beneath 
the formal shell of expansive reservations to find that schemes amounted to "fraud on the 
constitution," so they have begun to look beyond the forms of the selection to assess the operative 
realities.  52.   
____________________ 
51. Thus the Havanur report proposed that interviews be alloted weight equal to the written 
examination: "The merit of a candidate can be assessed better by his alertness, intelligence, 
intellectual outlook, his presence of mind and nervous equipoise. Emphasis on the interview will 
help offset the deficiencies of the candidates coming from the rural areas, and the peasant, 
artisan and low castes who do not have adequate opportunities to prove their 'talents' at the 
written examination. Better performance at the University or qualifying written examination is 
due to urban biased education. The present system of open competition has tended to deprive the 
backward communities, drawn mostly from the rural parts, of their adequate share in the 
administration, and the administration is not representative of the society at large" (Government 
of Karnataka 1973: I, 212­13). 
52. The courts are not the only influence operating to curtail capriciousness, as is illustrated by the 
situation in Mysore. By 1967 the scoring of the interview had been 
­454­ 
Judicial scrutiny of the adequacy of selection procedures may contribute in an important way to the 
implementation of reservations policies. For interviews and selections committees can be abused, 
not only to create what are in effect illegitimate reservations, but to frustrate and undermine 
legitimate ones.  53.   So far, though, the persistent complaints of beneficiaries that they are unfairly 
eliminated in interviews have not been brought before the courts. 
B. RESERVATIONS AS GUARANTEED MINIMUMS 
Any scheme of preference must be administered so as to operate in favor of the backward class. In 
Raghuramulu v. State of Andhra Pradesh,  54.   admission to a medical college was denied Backward 
Class candidates, although their scores were higher than candidates admitted in the "general 
competition," on the ground that the 15% reservation for Backward Classes had been filled. The 
interpretation of the reservation as a ceiling on admissions "instead of advancing the cause of the 
backward classes prevented some members of those classes from getting seats which they would 
have otherwise gotten if all the seats were brought under a common pool."  55.   Such rules, said the 
Andhra Court, were to be confined in their operation to cases in which the backward classes were 
benefitted, and their application in cases where they did not operate for the advancement of the 
backward classes violated the fundamental rights of the latter. Article 15 (4) authorized the 
abridgment of fundamental rights in order to make special provision for backward classes. Thus, if 
backward classes secure more seats by merit than are reserved for them, the reservation cannot be 
used to limit them to a prescribed number. 
The Supreme Court has upheld the position that reservations are not to be construed as 
compartments. A Scheduled Tribe candidate who had made the special declaration required to stand 
for the reserved seat in a double­member parliamentary constituency was held not disqualified from 
being elected to the general seat. The Supreme Court found that the reservation (authorized by 
Article 330) was to be construed as a 
____________________ 
abandoned and the interview reduced to a check on the documents which support the application. 
Informants suggested that this was because the scoring provided little additional separation of 
applicants, but exposed the committee to unwelcome pressures and charges of corruption (letter 
from Glynn Wood, dated Bangalore 21 Dec. 1967). 
53. Cf. the persistent complaints of Scheduled Caste candidates that they are unfairly eliminated in 
interviews and personality tests in competing for higher posts in the services. See chap. 4, §B, 
above. 
54. A. I. R. 1958 A.P. 129. 
55. Id., at 131 
­455­ 
guaranteed minimum and not as a barrier in election to additional unreserved seats.  56.   This 
principle that reservations are guaranteed minimums, rather than maximums, ceilings, or closed 
compartments, seems equally applicable to reservation of posts under Article 16(4), which 
empowers the State to make reservations "in favour of" backward classes.  57.   
On the other hand, members of backward classes are not constitutionally entitled to a system of 
selection which transforms a reservation into an award of a guaranteed number of seats over and 
above those obtained by merit.  58.   In Puppala Sudarsan v. State of Andhra Pradesh,  59.   an 
unsuccessful Backward Class candidate for admission to medical college contended that a higher 
ranking Backward Class candidate should be given a seat from the general pool rather than a 
reserved seat, thus enabling petitioner to obtain a reserved seat. The High Court, however, said that 
the selection should not be made in separate compartments in such a way as to allow Backward 
Class candidates to compete for both general and reserved seats to the detriment of others. 
Reservations are to operate as a guaranteed minimum. Backward Class candidates, it was held, may 
obtain additional seats by merit but cannot require that reservations operate as a guarantee of seats 
over and above those acquired by merit. 
In Chamaraja v. State of Mysore,  60.   a narrowly unsuccessful Backward Class candidate for a 
medical college seat argued that the reservations provided under Article 15(4) were distinct from, 
and in addition to, the right of successful merit candidates, whose right to admission was 
guaranteed by Article 29(2). The court held that Article 15(4) only empowered reservations but did 
not require them, leaving to the State broad discretion about how it uses this power, including an 
option to mesh its reservation with the guarantees of Article 29(2), as it had done 
____________________ 
56. V. V. Giri v. D. Sura Dora, A. I. R. 1959 S. C. 1318, aff'g. A. I. R. 1958 A.P. 724. Double­member 
constituencies were abolished in 1961. This abolition seems to have been inspired principally by 
the concern of politicians with the greater difficulty and expense of campaigning in a double­
sized district. But this was compounded by resentment of the infrequent but well­publicized 
instances in which the unreserved "general" seat as well as the reserved seat was won by 
Scheduled Caste or Scheduled Tribe candidates. Dushkin ( 1972: 192) says the Dora decision 
triggered the move for abolition. 
57. Harpartap Singh v. Union of India, 1970 S. L. R. 40, 43: an objection to reservations on the 
ground that Scheduled Castes already held a higher percentage of the promotion posts was based 
on "complete misunderstanding of the situation." 
58. This "over and above" arrangement was employed in Mysore until the decision in the Partha 
case, discussed below at n. 61. In the 1960 professinal college admissions, Backward Classes who 
enjoyed a 22% reservation secured 40% of the seats (Government of Mysore 1961: 23). 
59. A. I. R. 1958 A.P. 569. 
60. A. I. R. 1967 Mys. 21. 
­456­ 
here. In effect, the State might make reservations in the form of guaranteed minimums. 
In Partha v. State of Mysore,  61.   it was held that a reservation could not be used to set up two 
separate competitions; it must operate as a guaranteed minimum within a general competition for all 
seats. Allowing members of the backward group to compete separately for reserved seats and for 
general (i.e., unreserved) seats was not permissible, since it abridged the fundamental rights of the 
others. The only permissible method, said the court, was to prepare a single consolidated list of 
candidates in order of their merit, to draw a tentative line under the number of places open, to count 
the number of members of "privileged" groups above the line. In the event that some privileged 
group had a number of successful candidates which was less than the number of seats reserved, 
additional members of that group were to be selected in order of merit and an equal number of 
others deleted in inverse order of merit.  62.   
The very different results that different methods may lead to may be seen by a simplified tabular 
representation ( table 23 ). Let us take the simplest case of a reservation: a double­member 
constituency in which one of the seats is reserved for member of group R and the other seat may be 
contested by any member of the general public (G).  63.   If two Rs and two Gs contest the two seats, 
we may observe that under some circumstances different methods of declaring winners lead to 
different results. 
Methods A and B lead to identical results. If there is a single uncompartmented competition it does 
not matter whether the reserved place is declared first and then the rest of the places in order of 
merit; or whether all the places are distributed in order of merit and the reservation is employed 
only in the event that the reservation is not filled with merit candidates (the Partha method). By 
either procedure, the Rs will obtain the number of reserved places or the number of merit places, 
whichever is higher; the Gs will get all the places earned on merit except for the number of reserved 
places unfilled by merit. 
The other methods, C and D, lead to very different outcomes. Each of them involves a division into 
separate competitions. In method C, the selection is treated as two separate and simultaneous 
competitions, one for a reserved seat (in which only Rs compete) and one for the general seat (in 
which only Gs compete). In case 4 this procedure leads to the 
____________________ 
61. A. I. R. 1961 Mys. 220. 
62. Id., at 235. 
63. Although double­member constituencies are no longer used for parliamentary and legislative 
elections, multi­member constituencies are in use for local and panchayat elections, and the same 
problems of method are very much alive. Cf Shanshikalabai v. Election Officer, A. I. R. 1968 
Bom. 445, for some anomalous results. 
­457­ 
anomalous result that the higher ranking R2 is rejected in favor of G1. This procedure has been 
rejected by the Supreme Court and several High Courts, because it penalizes rather than helps the 
backward classes.  64.   
Method D also treats the selection as two separate competitions, not simultaneous but successive, 
with the general to be decided first, in which the Rs are eligible in both. In cases 5 and 6, this leads 
to the anomalous selection of R2 rather than the higher ranking G1. This is what we have referred to 
as the "over and above merit" method of reservations. 
TABLE 23 POSSIBLE METHODS OF DECLARING WINNERS IN A DOUBLE­MEMBER 
CONSTITUENCY 
Case  Case  Case  Case  Case  Case 
1  2  3  4  5  6 
Order of candidates  1st  G1  G1  G1  R1  R1  R1 
2nd  G2  R1  R1  R2  G1  G1 
3rd  R1  G2  R2  G1  R2  G2 
4th  R2  R2  G2  G2  G2  R2 
Method A 
single  R1  R1  R1  R1  R1  R1 
Case  Case  Case  Case  Case  Case 
1  2  3  4  5  6 
competition after  G1  G1  G1  R2  G1  G1 
filling reservation 
Method B 
single  G1  G1  G1  R1  R1  R1 
competition  R1  R1  R1  R2  G1  G1 
with reservation as floor 
Method C 
separate and  R1  R1  R1  R1  R1  R1 
simultaneous  G1  G1  G1  (G1)  G1  G1 
competitions 
Method D 
separate successive  G1  G1  G1  R1  R1  R1 
competitions, with  R1  R1  R1  R2  (R2)  (R2) 
general decided 
first, Rs in both 
____________________ 
64. V. V. Giri v. D. Sura Dora, A. I. R. 1959 S. C. 1318 (double­member parliamentary constituency); 
Digambar Rao Bindu v. Dev Rao Kamble, 15 E. L. R. 187 (Bombay High Court, 1958); 
Raghuramulu v. State of Andhra Pradesh, A.I.R. 1958 A.P. 129 (reserved places in medical 
college). 
­458­ 
Some important schemes, notably the central government's job reservations for Scheduled Castes 
and Tribes, operate by the guaranteedminimum method. But a large number of reservations by the 
states are of the "over and above" type. Thus we find "over and above" arrangements in admissions 
in Andhra Pradesh,  65.   Punjab,  66.   Tamil Nadu,  67.   and U.P.  68.   "Over and above" type of reservations 
of government posts are found in Andhra Pradesh,  69.   Kerala,  70.   Punjab,  71.   and Tamil Nadu,  72.   also 
in Maharashtra,  73.   where it is accompanied by a ceiling. Merit may expand recruitment from 
specified classes above the 24% reserved for them, up to 40% of the total vacancies. At this point, 
presumably, the guaranteed­minimum principle would reassert itself. This is in effect a combination 
of the two methods: it is a 24% reservation on an "over and above" basis, but as the success in 
general and reserved pools reaches 40%, it is transformed into a guaranteed­minimum reservation of 
40%. "Over and above" schemes have been urged on the central government as a way of augmenting 
its reservations, but have not been accpeted.  74.   
In recruitment of Scheduled Castes and Tribes, such provisions until now have had little effect in 
most areas. But where combined with an expansive definition of Backward Classes, "over and 
above" reservation for Backward Classes may have major distributive consequences. 
____________________ 
65. The Andhra Pradesh Backward Classes Commission ( 1970: 102) recommended restoration of 
the "over and above" method that had been abandoned a decade earlier. The new arrangement is 
found in Balaram v. State of Andhra Pradesh, A. I. R. 1972 S. C. 1375. 
66. In Ramesh Chander Garg v. State of Punjab, A. I. R. 1966 Punj. 476, we find a reservation for 
"students from recognized backward areas of Punjab unable to get admission on merit." 
67. Tamil Nadu Backward Classes Commission 1971: II, 196­97. 
68. E.g., the U.P. medical admissions scheme challenged in Dilip Kumar v. Government of U.P., A. I. 
R. 1973 All. 592, 594 specified that "candidates belonging to the reserved categories who qualify 
. . . for . . . general seats on the basis of merit will . . . be treated as general and admissions 
against the reserved seats . . . will be made only from amongst those candidates of reserved 
categories who do not qualify . . . for admission against the general seats." (The reservation for 
women was excepted from this provision.) This aspect of the admissions scheme was not 
challenged in Dilip Kumar or the other cases which overturned it on other grounds. 
69. Andhra Pradesh Backward Classes Commission 1970: 115. 
70. Rule 14(b) of Kerala State and Subordinate Service Rules 1958 (as amended to 1970), App. IX in 
Kerala( Backward Classes Reservation Commission) 1971: II, 169. This was present but not 
challenged in Hairharan Pillai v. State of Kerala, A.I.R. 1968 Ker. 42. 
71. RCSCST 1973­74: 99. 
72. Tamil Nadu Backward Classes Commission 1971: I, 32. The Commission notes (id., at 98) that 
"it is expected that the better . . . aspirants will get into the unreserved quota of Government 
service on their own merits without the protection of the reservations." 
73. RCSCST 1973­1974: 99. 
74. Id. 
­459­ 
 Thus in  Periakaruppan
   , the Supreme Court observed that "candidates of Backward Classes have
    
75.
secured about 50% of the seats in the general pool. . . ."      In a guaranteed­minimum method, this 
would have liquidated the 25% reservation for Backward Classes, but under the "over and above" 
method, it had no such effect. 
Returning to our schema, we may summarize the constitutional acceptability of the alternate 
methods as follows: separate compartments (method C) are an unconstitutional method of 
administering reservations; the guaranteed­minimum methods (A and B) have been commended; 
the "over and above" method (D) has been held not to be constitutionally required. One court found 
the over­and­above scheme to be constitutionally prohibited, but since that 1961 case this kind of 
scheme has come before courts on several occasions without exciting judicial condemnation.  76.   In 
Periakaruppan, the Supreme Court's reaction to heavy success by Backward Classes in the general 
pool was that this showed that "the time has come for a de novo comprehensive examination of the 
question [of who are the Backward Classes]."  77.   
The matter was raised even more pointedly in State of Andhra Pradesh v. Balaram, where it was 
argued that since more than the reserved quota of Backward Class candidates had secured seats on 
merit, there could be no further selection against the reserved quota.  78.   The Court declined to 
address this as a question of constitutional stature: "we can only state [it responded] that it is the 
duty of Government to review the question of further reservations of seats for such groups."  79.   In 
both Periakaruppan and Balaram, it was pointed out to the state government that the question of 
designation was open to judicial review, but neither court seemed inclined to take up the method of 
administering reservations as a constitutional question. In view of the broad authorization of Article 
15(4), which empowers the State to make "any special provision" for the advancement of the 
backward, and of Article 16 (4), which permits "any provision" for the reservation of posts for the 
backward, it does not appear that the State is constitutionally debarred from employing an 
____________________ 
75. A. I. R. 1971 S. C. 2303, at 2311. 
76. In Jacob Mathew v. State of Kerala, A. I. R. 1964 Ker. 39 at 63, there is dictum to the effect that 
reserved places may be calculated on such an "over and above merit" basis. It appears that this 
conclusion is based on a misreading of the Andhra cases discussed in nn. 54 and 59 above. 
77. Periakaruppan v. State of Tamil Nadu, A. I. R. 1971 S. C. 2303, at 2311. This judgment was 
delivered while the Tamil Nadu Backward Classes Commission was at work. The Commission 
submitted its report two months later, but its terms of reference did not include a mandate to 
address the question of who should be included among the Backward Classes ( Tamil Nadu 
Backward Classes Commission 1971: I, 2). 
78. A.I.R. 1972 S. C. 1395 at, 1400. 
79. Id., at 1400. 
­460­ 
"over and above" merit scheme of reservation. The Constitution itself employs a variant of the "over 
and above" mechanism in the case of reservations of legislative seats.  80.   
The "guaranteed minimum" type of reservation has several advantages. First, it insures that the 
amount of effective reservation is somehow commensurate with the backwardness that inspired it. 
That is, the less able the members of the backward group are to compete on merit, the more 
preference they get as a result of a given reservation. If there are reservations of 20% for backward 
group A and 20% for backward group B and the As win 5% of the places on merit and the Bs 15%, 
the net effect of the reservation (assuming that sufficient candidates secure the minimum acceptable 
marks) is 15% in the case of the As and 5% in the case of the Bs. The more backward the group, the 
greater the effect of the reservation. More important, such a reservation is selfliquidating. As 
increased educational opportunities lead to better performance by members of Backward Class A, 
the effect of the reservation declines as they increase their success in merit competition. The 
disadvantage of such "guaranteed minimums" is that they may overstate the effective amount of 
reservation, leading to complications with the 50% limit in the courts and having adverse effects on 
the morale of the non­backward, who may be led to underestimate their own opportunities to 
compete on merit, and of the backward whose entrance on merit is labelled as a reservation. 
An "over and above" reservation may be easier to administer than a guaranteed minimum in 
situations like, e.g., elective office from singlemember constituencies. In recruitment to services 
where a roster of rotation is used to designate reserved turns, it is obviously simpler to be able to 
ignore the success of backward class candidates on the unreserved turns. 
Another advantage of the "over and above" method is substantive as well as administrative. A 
guaranteed minimum requires that everyone in the competition be identified as belonging to the 
preferred group or not, so that all members of that group who succeed on merit can be subtracted 
from the reservation. Thus, it is necessary to collect identity data for everyone, instead of only for 
those members of the preferred group who wish to be considered for the reserved places. The "over 
and above" method, in contrast, is compatible with the privacy of nonbeneficiaries and those 
backward class members who prefer not to disclose their identity. And the "over and above" method 
may afford some welcome flexibility in policy in providing preferential treatment to a sizable group 
who are already getting a number of places on merit. 
____________________ 
80. I.e., there is no diminution of reserved seats as a result of additional seats won in open 
competition in unreserved constituencies. 
­461­ 
Suppose a state provides a reservation of 20% for backward class A which makes up 40% of the 
state's population at a time when As are winning only 5% of the places on merit. In the course of 
time, improvement of their condition leads to a situation in which As are winning 20%. Now the net 
effect of the reservation is nil. If the state still considers the As backward and deserving of 
preference, it has two choices: it may raise the reservation ­­ say to 30%, which would mean that the 
net preference would be 10% ­­ with all of the problems of general morale that this would entail. Or 
it could substitute a lower reservation that was not a guaranteed minimum, but of the "over and 
above" type ­­ 10% again. This 10% reservation would have the same net effect as the 30% 
guaranteed minimum. It has the advantage that the public knows just how much preference the As 
are getting, and protects against adverse effects on public morale of exaggerated notions of the 
amount of reservation. But the "over and above" method has the decided disadvantage of rigidity. 
What may be reasonable one year may very shortly be obsolete. This method does not have the 
selfadjusting and self­liquidating features of the "guaranteed minimum" and therefore requires 
closer judicial scrutiny to prevent abuses. 
But the principal argument  81.   advanced by the proponents of "over and above" reservations is that it 
would insure that some benefits go to the really backward.  82.   This is a particularly acute problem 
when the reservation is uncompartmented and the beneficiaries include some groups or strata that 
are highly qualified relative to other groups or 
____________________ 
81. A curious argument by Kuppuswamy ( 1978: 111) that guaranteed minimum type reservations 
will lower the motivation to excel ("each group will then perpetuate itself as being backward and 
depend wholly on reservation for its prospects") is unconvincing in its assumption of a high 
degree of group management over the performance of individual candidates. 
82. See, e.g., the plea for the "over and above" merit method (D) contained in the resolutions of the 
1959 meeting of Andhra Pradesh Backward Classes Association: "It is highly regrettable that the 
Government are still following the procedure . . . of selecting candidates first for the reserved 
seats and then for general seats [Method A above] . . . . This procedure strikes at the very root of 
the principle of reservation. The principle of reservation is meant for enabling candidates 
belonging to the Scheduled Castes, Scheduled Tribes and other Backward Classes, who are 
otherwise qualified but are unable to compete for general seats, to come up educationally and 
enter Government services. This does not prevent candidates belonging to these classes from 
competing for general seats. The procedure now followed by the Government in this respect is 
against all canons of fairness and justice. So, this . . . Body earnestly urge upon the Government 
to restore the old, procedure of selecting candidates first for general pool and then for reserved 
seats." ( Andhra Pradesh Backward Classes Association 1959: 2.). A decade later the Andhra 
Pradesh Backward Classes Commission ( 1970: 102) agreed. 
­462­ 
strata. "Over and above" reservations prevent the success of the former on merit from liquidating the 
reservation. But this protection of the most backward from the effect of disparities among the 
backward classes will obtain only in the unlikely event that the relatively advanced groups and strata 
succeed on merit but do not also cluster at the top of the list for the reserved places. If the less 
advanced groups within the backward classes are to be protected from these disparities, it can be 
more efficiently accomplished in the long run by more careful designation of the backward classes 
and in the short run by arrangement of compartments within the reservation to assure that the top 
groups within the backward classes do not obtain all the benefits. 
C. COMPARTMENTS AND LAYERS 
The requirement that preferences must operate for the benefit of the backward raises the puzzling 
problem of differentiated treatment of the backward­­a problem that dramatizes the difficulty of 
implementing the principle of compensatory discrimination. This problem arises in several forms. 
One is the permissibility of compartmental treatment. Is the State confined to one aggregate 
reservation for all whom it designates as backward classes? Or may it make separate reservations for 
component parts of the total backward group? It also arises in the form of layers or slabs: must the 
State bestow the same preferential treatment on all of the backward? Or may it designate some of 
the backward to receive more preferences or to have first call on limited preferences? 
Opinion is divided about the desirability of these devices. A breakdown of reservations has been 
specifically recommended on the ground that a single uncompartmented reservation would help 
only the relatively advanced groups among the backward classes and would be disadvantageous to 
the most backward.  83.   On the other hand, arrangements for compartments or layers of backward 
classes have been officially opposed on the ground that such distinctions promote caste feeling.  84.   
____________________ 
83. I BCC xxi, 141. Cf. the recommendation of the Lokur Committee that "the various castes and 
tribes in the lists should be administratively . . . catagorized so as to give higher priority in 
planning and development to the more needy, and lower priority for the comparatively advanced" 
( Dept. of Social Security 1965: 8). 
84. RCSCST 1957­58: I, 9. This opinion was apparently shared by the Ministry of Home Affairs. 
Yet, the commissioner agrees that it is necessary to compartmentalize benefits for Backward 
Classes on the one hand and Scheduled Castes and Tribes on the other, lest the former benefit 
from the preferences intended for the latter. RCSCST 1952: 70. 
­463­ 
A number of schemes involving compartments are in use,  85.   and some of these have come before 
the courts. It is clear that differentiated treatment of the Scheduled Castes, the Scheduled Tribes, 
and the Other Backward Classes is permissible, and perhaps it is even constitutionally required. 86. 
So far the courts have not settled conclusively the permissibility of the use of compartments within 
the Scheduled Castes or within the Backward Classes. Several early decisions suggested that any 
such compartmentalization of reservations is unconstitutional. 
The first court to address itself explicitly to this issue was in Ramakrishna Singh v. State of Mysore. 
 87.   Twenty percent of seats in the state's professional and technical colleges were reserved for 
members of the Scheduled Castes and Scheduled Tribes; 45% were reserved for Other Backward 
Classes. One hundred and sixty­four communities (castes, sub­castes, and religious groups) were 
listed as Backward Classes. These communities were arranged in 14 groups (ranging from 2 to more 
than 100 communities) and each group was assigned a portion of the reservation (from 1.2% to 
8.5% of the total seats). 
Among the features of this scheme that the High Court found to be invalid are two which concern 
the validity of compartments. First, the arrangement of the groups so that "each group of backward 
classes includes one relatively forward class" made it possible that "the limited percentage of seats 
reserved" for each group "would be captured by those communities who are more forward . . . 
leaving the really backward classes with no chances of getting any seats. . . ."  88.   The court held that 
such an arrangement, by diverting benefits from "really backward" to "relatively forward" groups, 
failed to meet the requirement of the Raghuramulu case  89.   that the arrangement be for the benefit 
of the backward classes. Here, the court's holding is based on the conclusion that the favored groups 
are not legitimate recipients of preferences at all. But the court's reasoning would seem to invalidate 
any scheme of compartments which combined, in a single compartment, groups of disparate 
backwardness­­even where both groups were legitimate recipients of benefits. 
This argument does not dispose of the question of compartments, for 
____________________ 
85. E.g., the communal rotation among various Backward Classes in Kerala. In Jacob Mathew v. 
State of Kerala, A. I. R. 1964 Ker. 39, a "sub­rotation" among various Backward Classes was 
struck down because the compartments were drawn wholly on lines of caste and were not 
commensurate with the relative backwardness of the respective groups. The rotation was later 
restored by the Division Bench, when it upheld the classification ( State of Kerala & Jacob 
Mathew, I. L. R. 1964 [2] Ker. 53). Sub­rotation among the major groups of Other Backward 
Classes was not challenged in Hariharan Pillai v. State of Kerala. A. I. R. 1968 Ker. 12. 
86. See below, §D. 
87. A. I. R. 1960 Mys. 338. 
88. Id., at 351. 
89. A. I. R. 1958 A.P. 569. 
­464­ 
it is really only an objection to the arrangement of the groups. But the court has a second and more 
basic argument against compartments. Under the Mysore scheme, members of each Backward Class 
"can only compete for the seats . . . reserved for that group and are not eligible for the remaining 
seats reserved for the backward classes. . . . [T]hey are debarred from capturing the said remaining 
seats in open competition amongst the members of backward classes . . . ."  90.   Extending the 
 principle of the  Raghuramulu
    case that the reservation must operate to the benefit of the backward,
    
the court holds that a backward class must be able to compete not only for the unreserved seats (as 
 in the  Raghuramulu
    91.   That is, while the fundamental 
  case) but also for all of the reserved seats. 
 
rights of the general public may be abridged by State power to permit preferences for members of 
Backward Class A, the court finds that Article 15(4) does not authorize abridgement of the 
fundamental rights of Backward Class B in order to permit preferences for Backward Class A. Since 
compartments restrict the rights of the Bs to compete for the seats reserved for the As, they are 
unconstitutional. 
The two arguments against compartments, both inferred from the principle that reservations must be 
for the benefit of the backward, are ultimately inconsistent. The first objection­­the unfairness of 
competition between members of groups of unequal backwardness­­could be remedied only by more 
precise and equitable arrangement of compartments; the second objection­­that all of the backward, 
whatever disparities exist among them, must be free to compete for every reserved seat­­can be met 
only by complete elimination of compartments. 
But the notion that compartmental arrangements are impermissible because members of the 
backward classes must be free to compete for all the reserved seats encounters both practical and 
theoretical problems. Does this principle make it unconstitutional for the state to provide, for 
example, that some preferential scheme applies only to backward classes in a given geographical 
area? Would backward classes throughout the state be automatically eligible to participate? Would it 
invalidate the separate treatment of Scheduled Castes, Scheduled Tribes, and Backward Classes­­
giving each of these a right to participate in any schemes for the benefit of another of them?  92.   
____________________ 
90. A. I. R. 1960 Mys. at 351. 
91. Cf. Partha v. State of Mysore, A. I. R. 1961 Mys. 220 at 235. 
92. An argument for separate compartments for Scheduled Castes and Tribes might be found in the 
wording of Art. 15(4) itself, which mentions them separately from and alternately to backward 
classes. Art. 16(4), which does not mention Scheduled Tribes and Castes separately, does not 
supply any such textual ground for a distinction in the government employment field. But 
presumably such a distinction could be founded upon Arts. 46 and 335. 
­465­ 
Beyond these, this conclusion lacks any apparent constitutional foundation. There is nothing in 
Articles 15 or 16 to indicate that State power to make provisions for backward classes falls short of 
power to limit preferences to one backward class in order to benefit another. There is no indication 
that persons who are designated as backward do not share with other citizens the possibility of 
having their fundamental rights limited to the extent that it is for the purpose of benefitting other 
backward groups. 
The Supreme Court also intimated some doubts about the constitutionality of compartments. In 
Balaji v. State of Mysore, the Supreme Court rejected a scheme under which separate reservations 
were provided for lists of "Backward Classes" and of "More Backward Classes." Article 15(4), the 
Court held, authorized special provisions only for the "really backward classes." Finding that the 
Mysore scheme benefitted not only these classes but also classes that were merely "less advanced, 
compared to the most advanced classes in the State," the Court held that "the classification of the 
two categories . . . [was] not warranted by Article 15(4)."  93.   Since the Court clearly indicates that 
the scheme would have been equally bad had all its beneficiaries (i.e., both legitimate and 
illegitimate recipients of preference) been grouped together, its rejection of the compartments is 
only dicta. The Court indicated that one of the objectionable features of the use of "such relative 
tests" is that they would entail "several layers or strata of Backward Classes . . . each one [of which 
might] claim to be included under Article 15 (4)."  94.   But it does not explain why provision for a 
plurality of backward classes is thought to lie outside the scope of Article 15(4). 
A decade later, compartments did not excite the same concern. The Andhra Pradesh Backward 
Classes Commission ( 1970) recommended parcelling out reservations among four groups in 
proportion to their estimated populations: 7% to 37 aboriginal and nomadic tribes; 1% to Harijan 
converts; 13% to 21 "vocational groups;" and 9% to 33 "other classes."  95.   The Commission never 
explains the recommendation for separate reservations beyond stating that it would afford them 
"equal opportunities" and would be "useful for watching the progress made by each group."  96.   In 
Balaram the state's apportionment of the reservation along these lines was attacked as an instance of 
the division (into more and less backward) condemned in Balaji. The Supreme Court repelled this 
attack by distinguishing this "distribution of seats among the reserved classes in proportion to their 
population."  97.   
____________________ 
93. A. I. R. 1963 S. C. 649 at 661. 
94. Id., at 658. 
95. Andhra Pradesh Backward Classes Commission 1970:103. 
96. Id. 
97. A. I. R. 1972 S. C. 1375 at 1399. A distribution among Scheduled Castes (and converts to 
Buddhism), Scheduled Tribes, Denotified Tribes, and Other Backward 
­466­ 
Compartments may be made sufficiently fine­grained that they apportion seats among communities. 
Thus in Kerala the reservations for Backward Classes in professional colleges were apportioned 
among Ezhavas (9%); Muslims (8%); Latin Catholics (2%); Other Backward Christians (1%); and 
Other Backward Hindus (5%).  98.   This feature has not attracted any challenge in court. The 
Backward Classes Reservation Commission proposed an even more elaborate sub­rotation of 
reserved posts among nine groups.  99.   
Presumably, compartments require some justification beyond apportionment of seats among 
communities, but so far the courts have not produced an analysis of the circumstances under which 
they are permissible. But that they are sometimes permissible seems not to be in doubt. 
But if groups can be sorted into compartments for purpose of apportioning reservations, can they 
not be sorted into "layers" or "slabs," some of which receive more preferential treatment than 
others?  100.   Such differentiated or layered preferences have long been in use. For example, Uttar 
Pradesh at one time had two lists of Backward Classes: the first received educational concessions 
only, the second received concessions in both education and state employment.  101.   Uttar Pradesh 
also introduced economic "layers" of Scheduled Castes for purposes of certain welfare measures. 
 102.   The central government has special schemes for sweepers and scavengers, who are the most 
needy of the Scheduled Castes.  103.   (And, of course, schemes based on an economic 
____________________ 
Classes was challenged and upheld in Pandit v. State of Maharashtra, A. I. R. 1972 Bom. 242. 
100. One, recent High Court decision says they may not. In Chhotey Lal v. State of U.P. [NIC], A. I. 
R. 1979 All. 135, 154 [NIC], the court declares flatly that "making classification of the backward 
classes into two sections . . . backward classes and . . . most backward classes is itself 
unconstitutional." This rule is announced here in the context of a finding that the state has failed 
to establish the backwardness of the groups in question. It is purportedly drawn from Balaji but 
ignores the tensions in Balaji as well as intervening judicial development of the point. 
101. RCSCST 1957­58: I,9. 
102. For the purpose of assuring that the most needy receive the most benefit from welfare schemes, 
the Scheduled Castes have for this purpose been divided into three groups: those with monthly 
incomes under Rs. 100; those with monthly incomes under Rs. 250; those with high incomes. 
Members of the last group are not eligible for benefits from welfare schemes; members of the 
second group may receive such benefits only after all the members of the first group have been 
provided for (RCSCST 1961­62: I, 28). 
103. Id., at 98 ff. Special preference for these groups and for the most backward among the Scheduled 
Tribes has been suggested by the Commissioner (RCSCST 1960­61: 148). 
98. Kerala ( Kumara Pillai Commission) 1966: 76. 
99. Kerala ( Backward Classes Reservation Commission) 1971: I, 124 ff. 
­467­ 
test often sort out the recipients into layers­­as has the central government's post­matriculation 
scholarship scheme for Other Backward Classes since 1963.)  104.   Layers, like compartments, may of 
course be used for purposes other than protecting the most backward. The Kerala Backward Classes 
Reservation Commission, proposing a series of successively higher income cut­offs in connection 
 with reservations for higher posts, expressed concern that the  Balaji
    pronouncements about layers
    
105.
might cause difficulties.      Of course, the layers it proposed are not for the purpose of preserving 
some share of benefits for the least well­off. Quite the opposite! They are to enable benefits to be 
extended to highly placed government officers who are doing well by any allIndia standard.  106.   
But, for the most part, schemes for preferences are unaccompanied by provision for compartments, 
and in many cases there is a massive "creaming" of benefits by the more advanced groups within the 
backward category.  107.   Thus in Tamil Nadu, a group of Most Backward Classes were from 1957 
listed separately for purposes of some educational fee concessions but not treated separately for 
purposes of admissions or government posts.  108.   These castes (including such occupations as 
barbers, washermen, basket­makers, stone­cutters, and fishermen) were estimated to make up about 
44% of the population of Backward Classes.  109.   In 1970 they formed 9.4% of the Backward Classes 
students admitted to engineering colleges and 11.0% of those studying in medical colleges.  110.   They 
made up only 13.9% of Backward Classes candidates selected for Class II government service. 111. 
If their extreme backwardness prevented them from competing for such high­echelon benefits, it 
might be thought that the benefits for them would be concentrated at earlier­stage preferences. But 
in 1970 they formed only 18.9% of Backward Classes students appearing in the Secondary School 
Leaving Certificate examinations.  112.   And they accounted for only 22.4% of the expenditures on 
scholarships for Backward Classes. In fact, the per capita expenditure on scholarships for them was 
only 78 
____________________ 
104. See chap. 6, §C, above. 
105. Kerala ( Backward Classes Reservation Commission) 1971: I, 128 ff. 
106. A more detailed account of this proposal is discussed in chap. 8, §G, above. 
107. Cf the observations of Justice Krishna Iyer at N. M. Thomas v. State of Kerala, A. I. R. 1976 S. 
C. 490 at 531­32. 
108. Tamil Nadu Backward Classes Commission 1971: I, 18. The report says that these groups were 
selected because they were "almost as backward as Scheduled Castes . . . [but] did not satisfy the 
criterion of untouchability." But a dissenting member of the Commission suggests that the list 
was copied from the list of "Most Backward" groups compiled by the Kalelkar commission ( I, 
241). 
109. Id., at 176. 
110. Id., at 160. 
111. Id., at 150. 
112. Id., at 160. 
­468­ 
paise, compared with Rs. 2.10 for the remainder of the Backward Classes.  113.   
As one focusses on smaller groups, this creaming effect and the corresponding exclusion of the 
most backward is even more pronounced. Table 24 summarizes data supplied by the Commission 
comparing the flow of benefits to nine castes who are prospering under the protections for the 
Backward Classes and a group of seven numerous castes (all but one from the Most Backward 
Classes list) who are not doing nearly as well. 
TABLE 24 A COMPARISON OF UPPER AND LOWER STRATA OF BACKWARD 
CLASSES IN TAMIL NADU 
Nine Castes  a 
Seven Numerous Castes    
Percentage of Backward Classes 
11.3  12.1 
Population 
Students admitted, Engineering, 
44.3  c  1.8c 
1970 
Studying first­year MBBS 1969­70 
(one college omitted)  47.3c  0.4c 
Students appearing in SSLC 
examination, March 1970  30.3c  3.5c 
Candidates selected for Group II 
Services by TNPSC 1964­1969  44.0c  1.8c 
Non­Gazetted Posts in all 
37.3c  1.9c 
departments 
Gazetted Posts in all departments  48.2c  0.9c 
Per capita expenditure on  Rs. 4.63  b 
[Rs.0.78]    
scholarships  Source: Tamil Nadu Backward Classes 
Commission, Vol. I. 
Notes: a Six from Most Backward Classes 
list. 
b This figure is for the entire Most 
Backward Classes group ( I, 156). The 
Commission does not provide a figure for 
its seven numerous castes, but the figures 
it gives ( I, 157) for five of the seven 
(0.64, 0.28, 0.50, 0.45, 0.14) suggests that 
Nine Castes  Seven Numerous Castes    a 

an aggregate figure would be lower than 
the MBC average. 
c Percentages of all Backward Classes 
present. 

____________________ 
113. Id., at 156. 
­469­ 
These disparities point in two directions. First, they point back to the question of designation. Are 
groups and strata that are succeeding in such ample measure properly included among the Backward 
Classes? Should certain groups be removed? Or can income limits be employed to eliminate the 
"upper crusts" within a group? Thus the Tamil Nadu Backward Classes Commission suggests the 
adoption of an income cut­off to exclude the "upper crusts" within the Backward Classes.  114.   But 
there is a second problem. Even if the most affluent and successful are removed, the Backward 
Classes category will still span groups with a wide disparity of competence and resources. Some 
segments of the Backward Classes may need training or other special measures to increase their 
ability to utilize benefits. But even if such disparities are reduced, it is unlikely that they will be 
eliminated. One way of providing that some of the benefits flow to the least advantaged sections of 
the backward is to provide a special sub­reservation for them. Thus the Tamil Nadu Commission 
proposed that almost half of the reserved portion of admissions and posts be set aside separately for 
the Most Backward Classes.  115.   
There is no definitive pronouncement by the courts on such schemes of layers. In the Balaji case, as 
we have seen, the Supreme Court indicated that it found objectionable the division of the recipients 
into two categories of "backward" and "more backward." But there was no attempt in the Mysore 
scheme to apportion benefits in accordance with the need of the several classes. Since the scheme 
was found unconstitutional because it covered those who were merely relatively "less advanced" as 
well as the "really backward classes,"  116.   the permissibility of the layers as such was not before the 
Court. But the Court clearly indicates its objection to "several layers or strata of Backward 
Classes . . . each one of which might claim to be included under Article 15(4)."  117.   
The Supreme Court's objection to compartments, along with its insistence that beneficiaries of 
preferences be both socially and educationally backward,  118.   amounts to a proposal that the 
Backward Classes are some single, undivided, and unvarying group who are to be treated uniformly. 
But backwardness is not a single isolable trait; it exists in kinds and degrees. The backward are not a 
uniform group. By whatever standards the "Backward Classes" are selected, some of them will be 
more backward than others; and, ergo, some will be less backward. Whatever arrangement is 
proposed for the distribution of preferences among them, it will work to the advantage of some and 
the detriment of 
____________________ 
114. Id., at 152. 
115. Id., at 21. 
116. A. I. R. 1963 S.C. at 661. 
117. Id., at 658. 
118. See chap. 8, §C, above. 
­470­ 
others; e.g., the "most backward" would be disadvantaged by a single competition, while the less 
backward would be disadvantaged by layers and compartments. The whole system of compensatory 
discrimination rests on the notion that because of disparities in resources and background, the 
backward must be protected against open competition with the general public. But a single 
uncompartmented reservation for all of the backward tends to reproduce within that group the same 
kind of .unfairness that protective discrimination is designed to eliminate. If the backward are to be 
protected against open merit competition with the general public, why cannot the most backward be 
protected against such competition with other sections of the backward? Again compensatory 
discrimination is aimed at channelling benefits according to need. But uncompartmented 
competitions may well give the least benefit to those with the greatest need. 
The ultimate objection to compartments and layers seems to rest on the assumption that there is 
some constitutionally pre­ordained group of "really backward" persons who may be the 
beneficiaries of preference but may never be disadvantaged by preference for others. But this has no 
evident constitutional foundation. Articles 15 and 16 seem to anticipate a plurality of backward 
classes. The wording of the articles does not indicate that every class which is found backward for 
one purpose must be treated as backward for every purpose or that all classes found backward for a 
particular purpose must be treated uniformly. As the High Court in the Partha case points out, "the 
classification of . . . [backward classes] may vary from time to time and with reference to the nature 
of . . . [the] backwardness which is sought to be remedied by special provisions made in respect of 
it."  119.   Thus there appears to be no constitutional bar on the use of the compartments and layers by 
the State. Such arrangements seem appropriate once it is recognized that among the backward, some 
are more backward than others in particular respects and are more necessitous and/or deserving of 
help of certain kinds. The Constitution seems to envisage that preferences for Scheduled Castes and 
Tribes will be more extensive than those for Backward Classes, a distinction corresponding to their 
relative lack of resources and opportunities. It would be strange were the government prohibited 
from distinguishing among Scheduled Castes  120.   and among Backward Classes in order to make 
preferences commensurate with the backwardness of the recipients. The broad discretion vested in 
the State by Articles 15(4) and 16(4) would seem to cover such arrangements. It 
____________________ 
119. A. I. R. 1961 Mys. at 235. 
120. Preferential treatment for the most backward sub­castes within the Scheduled Castes had been 
recommended, for example, by the Central Advisory Board for Harijan Welfare ( Times of India, 
15 Apr. 1959). 
­471­ 
may be asserted that the use of compartments, even if constitutionally permissible, has the 
undesirable feature of providing what are in effect guaranteed communal quotas. To some extent, 
this is inherent in any scheme of preferences in which the recipients are designated along communal 
lines, and it cannot be eliminated by merely eschewing compartments but only by substituting other 
criteria of backwardness. 
The main danger in the use of compartments and layers is that their arrangement will reflect, not the 
relative need of the various backward groups, but their relative political power to secure 
arrangements favorable to their interests.  121.   Again, this can be remedied, not by eliminating 
differentiated treatment, but by insuring that the compartmental arrangement corresponds in some 
measurable way to the relative backwardness of the groups. 
If compartments or layers of preference are constitutionally permissible, courts may be faced with 
another kind of attack on these arrangements. It may be claimed that a particular scheme of 
compartments or layers actually operates to the disadvantage of a "more backward" group while 
benefitting the less backward. In the Ramakrishna Singh and Balaji cases, the courts indicated that 
schemes which benefitted the "relatively less backward" at the expense of the "really backward" 
were invalid. The implication was that the former were not legitimate recipients of preferences at all. 
However, it is easy to visualize the case in which both groups are "really backward" but one is 
clearly more backward than the other. In such a case, are the courts to intervene to see that the most 
backward of the groups are not disadvantaged by the scheme? For example, may a group complain 
that its relative backwardness entitles it to inclusion among the groups receiving the greater 
quantum of preferences? Such questions could be decided only if the courts were willing to apply 
some objective standards of backwardness. But the willingness of the court in Ramakrishna Singh 
to invalidate a scheme of preference on the ground that unequals were treated equally indicates that 
the same standards might be applied to invalidate schemes where the quantum of preference is 
incommensurate with relative backwardness. 
D. RELATIONS BETWEEN MULTIPLE RESERVATIONS 
The constitutional provisions for special treatment envision a plurality of backward groups. When 
there are reservations for more than one 
____________________ 
121. As was so dramatically demonstrated in Ramakrishna Singh v. State of Mysore, A. I. R. 1960 
Mys. 338. 
­472­ 
category of backward classes, the question arises as to whether the reservations are entirely 
independent of one another or whether they are articulated. One problem of articulation is that of 
immunity from "bumping." Is a member of one backward group immune from being "bumped" in 
favor of a member of another backward group? Suppose there is a competition for 10 seats among 
20 candidates and that there are two backward groups, A and B, for each of whom one seat is 
reserved. Suppose, further, that in open merit competition, two As score respectively first (A­1) and 
tenth (A­10) in order of merit, and that the highest B is eleventh (B­11). Since reservation is not a 
ceiling, the reservation cannot be used to limit the number of As. Since the Bs have not obtained 
their quota of reserved seats by open competition, B­22 is to be given a seat. This requires that one 
of the successful candidates be eliminated. Is A­10 immune from elimination? Should he be passed 
over and the lowest of the successful non­backward candidates be eliminated in his stead? To make 
this more concrete, it is easy to imagine that the As are the Scheduled Castes and the Bs the 
Scheduled Tribes or Backward Classes. So far, there is no indication of how this dilemma is to be 
resolved. If the Constitution supplies no answer to this dilemma, it is difficult to see that the State's 
power to make special provision for members of a backward group does not encompass a power to 
make seats that members of this group gain on merit immune from "bumping" by members of other 
backward groups. At least one state has a rule to this effect: the Mysore rule provides that deletion 
be of "applicants not belonging to any of these catagories."  122.   
One may imagine a further elaboration of the "bumping" problem. Suppose an income cut­off 
provision makes the most affluent members of Backward Class X ineligible for reservations, and 
suppose further that an affluent X is the last person on the merit list and that an insufficient number 
of backward Xs gain places on merit. To fulfil the reservations for the Xs it will be necessary to 
eliminate one of those on the merit list. Should the reservation for a Backward Class X be filled by 
taking the place that would have gone to the affluent X? This is ordinarily the way that a reservation 
as guaranteed minimum would work. May the State immunize affluent Xs who secure seats on 
merit against being displaced by Backward Class candidates for reserved places? Absent some 
finding of their backwardness, to choose between the affluent Xs and other successful competitors 
might involve forbidden discrimination. May affluent Xs be granted a kind of intermediate status in 
which they are sufficiently backward to be exempt from being bumped but not sufficiently backward 
to entitle them to share in the reservations? Such 
____________________ 
122. Chamarala v. State of Mysore, A. I. R. 1967 Mys. 21, at 22. 
­473­ 
an arrangement would seem consonant with the recognition of the relative and variable nature of 
backwardness.  123.   It might also be regarded as a measure for the advancement of the backward, 
even if the beneficiaries are not themselves backward, since the affluent Xs attainment of valued 
positions might, through ties of kinship and group loyalty, redound to the benefit of the less­affluent 
backward Xs.  124.   
The reverse side of the bumping problem is whether, if the reservation operates as a guaranteed 
minimum, the affluent Xs should be taken into account in calculating whether the reservation is 
filled. Literally, one would think not, for they are not members of the Backward Class that is entitled 
to the reservation. They would then be counted in for immunity from bumping but out for purposes 
of calculating the reserved seats to be awarded. 
It is quite common for reservations to go unfilled, especially those for Scheduled Tribes and 
Scheduled Castes. States not uncommonly provide for a "carry­over" in which the seats unfilled by 
one backward class are to be filled by another and only revert to the general pool if no candidates 
from the latter are available. In Partha, a "carry­over" arrangement, whereby any places reserved for 
Scheduled Castes and Scheduled Tribes and not filled by them were to be filled with additional 
members of the Backward Classes, was held unconstitutional, since these were clearly 
distinguishable classes and the giving to the Backward Classes of the extra seats other than by open 
competition was an unreasonable restraint on the fundamental rights of other citizens.  125.   
In Garg 8 seats (2%) were reserved for Backward Classes and 80 (20%) for the Scheduled Castes 
and Tribes, but only 10 Scheduled Caste and Tribe candidates were eligible. Under a "carry­over" 
provision, the 70 surplus seats were given to Backward Classes, raising their total to 78 (19.5%). 
The High Court struck this down as arbitrary, and "contrary to the spirit and ratio of [the Balaji] 
judgment."  126.   In its view, specified reservations should represent the maximum limit to which 
members of given class might be admitted on considerations other than merit. 
In Sukhdev,  127.   the selection was first divided into regional ones, and 
____________________ 
123. See the discussion at §C, above. 
124. On the broad interpretation of "advancement," see chap. 11, §B. 
125. A. I. R. 1961 Mys. at 235. But some kind of "pourover" arrangement seems to have been present 
in the situation described in Venkataramana v. State of Madras, A. I. R. 1951 S. C. 229 at 229, 
where Harijans received fewer, and Backward Hindus more, posts than originally were reserved 
for them. No objection was raised on this ground. 
126. Ramesh Chander Garg v. State of Punjab, A. I. R. 1966 Punj. 476, at 478. 
127. Sukhdev v. Government of Andhra Pradesh, 1966 (1) Andhra W.R. 294 (1963). 
­474­ 
16% within each regional selection was reserved for the Scheduled Castes and Tribes and 25% for 
the Other Backward Classes. It was then provided that where there were insufficient candidates in 
one of these groups in a region, those seats should not revert to the general pool in that region, but 
should carry over to the corresponding group in another region. Besides the vagueness and 
arbitrariness of this provision, it was held to militate against the principle of the regional scheme. 
The court held it to be discriminatory and a violation of Article 14. 
Ordinarily, the government has treated Scheduled Castes, Scheduled Tribes, and Backward Classes 
separately, making separate provision for each. Courts have looked favorably on this.  128.   In 
Janardhan Surraraya v. State of Mysore,  129.   the Supreme Court indicated that its decision in the 
Balaji case did not invalidate Mysore's reservations for Scheduled Castes and Scheduled Tribes, 
which were not in any manner irregular or unjustified by Article 15 (4). Reservation for these 
groups, the Court indicated, "is distinct and separate from, and independent of the other reservation 
which was challenged."  130.   The High Court in the Partha case found the "carry­over" arrangement 
objectionable precisely on the ground that it lumped together as a single category the Scheduled 
Castes, Scheduled Tribes, and Other Backward Classes. Such treatment struck the court as opposed 
to the clear constitutional differentiation among them.  131.   
The opposite tack was taken by the one court to squarely uphold a "carry­over" provision. In S. G. 
Pandit v. State of Maharashtram,  132.   the state had reserved 34% of the seats for four groups: 
Scheduled Castes and Buddhist converts from them, 13%; Scheduled Tribes, 7%; Denotified and 
Nomadic Tribes, 4%; Other Backward Classes, 10%. It was provided that "reserved seats remaining 
vacant in any of the above . . . groups for want of students in that group should go to the other 
groups . . . .  133.   The court found this unobjectionable because "[a]ll the four groups form one 
category of socially and educationally backward citizens. . . . The sub­division into the four groups 
is made obviously only to allocate the reservation . . . so that the comparatively brighter students in 
one group may not keep out the students of the other groups."  134.   In effect the court reads this 
scheme as a single reservation compartmented to protect the less educationally robust among the 
beneficiary groups. But the desirability of compartments rests on differ­ 
____________________ 
128Ramakrishna Singh v. State of Mysore, A. I. R. 1960 Mys. 338; Dasu Rayudu v. Andhra Pradesh 
. Public Service Commission, A. I. R. 1967, A. P. 353. 
129

A. I. R. 1963 S. C. 702. 
130. Id., at 703. 
131. A. I. R. 1961 Mys. 235. 
132. A. I. R. 1972 Bom. 242. 
133. Id., at 244. 
134. Id., at 251. 
­475­ 
ences among groups of beneficiaries, not on their similarities. And it requires some further 
argument to justify a carry­over from one to another. Is there some affinity among the groups so that 
one benefits from places gained by the other? Are the carry­overs in fact reciprocated? Are the 
numbers so small that there is a need to protect the reservation from temporary fluctuation in the 
number of qualified candidates in each of the groups? Where this is not the case a carry­over rule 
creates the possibility that some beneficiary groups will enjoy reservations of vastly greater extent 
 than could be authorized for them. (Thus in the  Garg
    case, the carry­over rule amplified the
    
reservation for Backward Classes from 2% to 19.5%.) Carry­over provisions can help to deliver 
preferences without undue fluctuations. But whether they are likely to do so in a given instance 
depends on the size of groups and the distribution of qualified candidates among them (which in 
turn may depend on the kind of benefit that is at issue). It can hardly be deduced in advance from 
notions of the conceptual difference or affinity between the various beneficiary groups, since they 
of course not only are different but are also joined by the circumstance of being beneficiaries of 
compensatory discrimination policies. 
Some states have provided composite reservations for Scheduled Castes and Scheduled Tribes  135.   ­­
a practice which has never been found objectionable; nor have provisions for reciprocal "carry­over" 
of unfilled reservations between these two groups, which are used both by states and by the Centre. 
 136.   
Although it is clearly open to the State to recognize such a differentiation between Scheduled 
Castes, Scheduled Tribes, and Other Backward Classes, it may be wondered whether the 
Constitution requires that it do so. In the Balaji case, the Supreme Court stresses that "the Backward 
Classes . . . contemplated by Article 15 (4) are [those whose backwardness is] comparable to 
Scheduled Castes and Scheduled Tribes."  137.   If this is the case, presumably it would be reasonable 
for the State to treat them alike. And, since preferences are, it is hoped, temporary expedients, these 
provisions should be construed with an eye to preventing impediments to ending them. It may well 
be that the amalgamation of these separate categories will be a step toward the eventual dismantling 
of the system of compensatory discrimination. There would seem to be no constitutional warrant to 
rule out such a step. 
____________________ 
135. E.g., in Kerala as in Jacob Mathew and Hariharan. 
136. E.g., the "carry­over" provision in the Central Secretariat was left standing after the Devadasan 
case ( Tewari v. Union of India, A. I. R. 1965 S. C. 1430). 
137. A. I. R. 1963 S. C. at 658. 
­476­ 
14 The Setting and Incidence of Judicial Intervention 
AFTER OUTLINING the background and contours of the preference policy, we have examined in 
some detail the working out of the compensatory discrimination principle (and its articulation to 
other principles) within one institution­­the courts­­and by one means ­litigation. We shall proceed 
to assess the impact of the courts on the preference policy and to reflect more generally on the 
capabilities of the common­law constitutional court to handle such tasks and forward such policies. 
But first it is necessary to examine the setting, institutional and cultural, in which this judicial 
activity takes place. 
A. COURTS AND JUDGES 
The judicial encounter with compensatory discrimination occurs almost entirely within the higher 
1. At the apex of India's unified, hierarchical judiciary is the Supreme 
reaches of the judicial system.    
2. The 
Court, consisting in 1980 of a Chief Justice and thirteen puisne judges.    
____________________ 
1. Except for a few cases involving protections (debt relief, etc.) and election cases, which until 
1966 were initiated in Election Tribunals, all cases involving compensatory discrimination 
originated in the High Courts or the Supreme Court. In 1966 the High Courts were made the 
courts of first instance in election complaints ( Representation of the People (Amendment) Act, 
1966, § 38). It was this arrangement that provided the setting for Mrs. Gandhi's election case, 
which provoked the 39th Amendment. That Amendment enacted, inter alia, a new Article 329 A, 
which prohibited the calling into question in any court the election of someone who was at the 
time of the election Prime Minister or Speaker of the Lok Sabha, or who was appointed to either 
of those positions subsequently. The Amendment and the Election Laws (Amendment) Act 1975 
(Act 40 of 1975), which empowered the President of India to decide whether someone convicted 
of corrupt campaign practices should be disqualified from an election, were upheld by the 
Supreme Court in Raj Narain v. Indira Nehru Gandhi, A. I. R. 1975 S. C. 2299. 
2. It has grown from 7 judges at its formation in 1950 to 11 in 1956 to 14 since 1960. 
­477­ 
Court sits in benches of two, three, five (in constitutional cases), and (extraordinarily) seven or more 
judges.  3.   
The Supreme Court is the highest appellate court in all matters of criminal, civil and constitutional 
4. Its appellate jurisdiction extends to the 
law. Its authority extends to state as well as central law.    
High Courts and through them to all inferior courts and to all quasi­judicial tribunals.     5. It has a 

limited original jurisdiction, notably to issue writs to prevent violations of Fundamental Rights.  6.   In 
7. 
addition it has an advisory jurisdiction.    
8. Each state has a single judicial hierarchy with a High Court at 
There is no lower central judiciary.    
its head.  9.   The High Court exercises an 
____________________ 
3. Sitting in benches is sanctioned by Art. 145, but Art. 145(3) requires that a minimum of five 
judges sit to decide "any case involving a substantial question of law as to the interpretation of 
this Constitution" or to exercise the Court's advisory jurisdiction. In 1976, the Forty­second 
Amendment inserted the requirement that only a bench of seven or more should determine the 
constitutional validity of any central or state law and that a law should not be declared 
constitutionally invalid by less than two­thirds of the judges sitting. This provision (Art. 144A) 
was repealed before it came into effect by the Constitution (43rd Amendment) Act, 1977. §5. 
4. The 42nd Amendment added to the Constitution Art. 32A which withdrew the jurisdiction of the 
Supreme Court to hear writs under Article 32 for determining the constitutionality of state 
legislation. This provision was repealed before it came into effect by the Constitution (43rd 
Amendment) Act, 1977, §3. 
5. Articles 132, 133, 134, 136. These broad grants of jurisdiction are subject to some explicit 
limitations in regard to compulsory acquisition of property (Art. 31[6]) and delimitation 
controversies (Art. 329). 
6. Art. 32, Art. 131. 
7. Art. 143. The advisory jurisdiction has not been involved in any of the cases concerning 
compensatory discrimination, with the exception of In re Kerala Education Bill, 1957, A. I. R. 
1958 S. C. 956, 982­83, which touches on the power of the State to make reservations in non­
governmental schools. This case is not counted in the various computations in this chapter. 
8. There are no lower central courts, but there are various tribunals which are part of the central 
government: income tax, labor, industrial, etc., tribunals. Beyond this, the High Courts 
themselves occupy a mixed status: the Constitution discussed them under the heading of "The 
States," but the judges are paid from the Consolidated Fund of India; appointment is by the 
President; removal by impeachment is by the central Parliament; judges may be transferred from 
one High Court to another by the central Ministry of Home Affairs. On the other hand, staff are 
appointed and administrative expenses borne by the state government. 
9. There are 18 High Courts, one for each state, except that Assam and Nagaland share a High 
Court (since the latter's formation in 1962) as do Punjab and Haryana (since the latter's formation 
in 1966). Delhi has had its own High Court since 1966. Three of the Union Territories have 
courts of Judicial Commissioner with analogous power; the other Union Territories are included 
within the jurisdiction of a High Court. In four states the High Courts have permanent benches 
sitting in more than one location. The High Courts of Bombay and Calcutta also exercise an 
extensive original jurisdiction in those cities, which is not of concern here. ( H. P. Dubey 1968: 
386 ff.). 
­478­ 
extensive jurisdiction to review the lower courts  10.   and other tribunals, and it shares with the 
Supreme Court an original jurisdiction to issue writs against violations of Fundamental Rights 11. 
(although since 1976 it cannot hold unconstitutional a central law). The High Courts, which vary in 
size from two judges (in Sikkim) to forty (in Uttar Pradesh), sit as single judges, Division Benches 
of two judges, or, occasionally, as Full Benches of three, five, or even seven or more judges. In some 
instances an appeal or a reference on a point of law is possible from a smaller to a larger bench of a 
High Court. 
Judges of the High Courts are recruited directly from the bar (usually the High Court bar) and by 
promotion from the subordinate judiciary.  12.   Although the Constitution provides for direct 
appointment of lawyers and of "jurists" to the Supreme Court, so far all appointments to that Court 
have, with but three exceptions, been by promotion from the High Courts.  13.   Seniority has played a 
prominent role in selection.  14.   Just over half the appointees have been Chief Justices of their High 
Courts. The average age of appointees is fifty­seven. The preponderant role of seniority in making 
these appointments and the compulsory retirement age of sixty­five combine to produce a relatively 
rapid turnover on the Supreme Court.  15.   The average service of judges on that 
____________________ 
10. Below the High Courts are the Subordinate Courts, usually consisting of two somewhat 
overlapping levels. In criminal matters, there are various grades of magistrates, graded by the 
scope of their penal powers, and above them Sessions Courts, which hear appeals from 
magistrates and which are trial courts for more serious offenses. From the Sessions Courts there 
are appeals, revision petitions, and references to the High Court. On the civil side, there are 
various classes of Subordinate Judges; appeal from them and original jurisdiction over larger 
suits lies with the District Judge. There is appeal from the court of District Judge to the High 
Court. There are also a variety of Small Cause Courts, Revenue Courts, judicial panchayats, and 
a rapidly increasing number of administrative and quasi­judicial tribunals ( Dubey 1968 ; Law 
Commission 1958; Khosla 1949 ). 
11. Art. 226. Cf. Art. 227. Petitioners must elect to pursue their writ remedy in either the High Court 
or the Supreme Court. A petitioner who obtains a decision on the merits in an Art. 226 
proceeding cannot then file an Art. 32 petition in the Supreme Court ( Daryao v. State of U.P., 
1962 [1] S. C. R. 574). 
12. Judges in the lower courts are members of the state judicial service, chosen either by competitive 
examination among recent law graduates or holders of other government posts, or by selection 
from members of the bar of three to six years standing. 
13. See Gadbois 1968­69: 326. 
14. On the preponderant role of seniority in these appointments, see Law Commission 1958: I, 209. 
15. The retirement age in the High Courts is sixty­two, which insures a comparable turnover there. 
Compulsory retirement has been accounted a political threat to judicial independence, since 
many judges look to government for post­retirement reemployment and appointments. See S. 
Dhavan 1964: 344; Seervai 1968: 1007. Cf. Gadbois 1968­69:330; and R. Dhavan 1977: 29 ff., on 
the post­retirement occupations of Supreme Court judges. 
­479­ 
Court has been 67 months, and the turnover seems to be increasingly rapid: judges appointed from 
1950 to 1964 served an average of 82 months; those appointed after 1964 served only 52 months. 16. 
Judges of the High Courts typically serve ten years or more.  17.   
Appointments to the higher judiciary are formally by the President after consultation with the Chief 
Justice of India and, in the case of the High Courts, with the Chief Justice of the High Court and the 
Governor of the State.  18.   No legislative approval is required. In practice, the President acts on the 
recommendation of the Home Ministry, the government department in charge of services. Chief 
Justices of the High Courts have exercised initiative in appointments, but this is subject to the 
concurrence of the Chief Minister of the State, who has effective control over appointment to the 
High Courts.  19.   To the dismay of at least the upper echelons of the profession, there is in effect 
some admixture of considerations of regional and communal balance in the selection of High Court 
judges. The emphasis on seniority and regional balance transmits a similar profile to the Supreme 
Court. Once appointed, a judge of a High Court or the Supreme Court enjoys complete security of 
tenure and can be removed only for cause and only by a special parliamentary charge.  20.   
The Chief Justices of the Supreme Court and the High Courts occupy positions of strategic 
importance, because they are consulted on appointments and because they distribute work among 
their colleagues, deciding who should sit on which panels. A convention under which Chief Justices 
are appointed in order of strict seniority on the Court was broken by Prime Minister Gandhi in 1973 
and re­established with the 
____________________ 
16. The average is calculated for all of the 44 judges appointed after the court was formed in Jan. 
1950 who had retired, resigned, or died in office through Feb. 1978. (If we set aside the 4 judges 
who resigned in the supercession controversies, the average tenure was 66.7 months.) 
17. This assertion is based on the following computation: From the ages and dates of appointment of 
High Court judges sitting in 1977, it was possible to ascertain the expected length of their 
service, unless shortened by resignation, promotion to the Supreme Court, or untimely death. 
Such length of expected service was computed for each of the four High Courts with ten or more 
cases in our universe of compensatory discrimination cases. (See n. 45 below.) The expected 
length of service on these High Courts was Karnatak, 15 years; Punjab and Haryana, 14.5 years; 
Patna, 12.8 years; Andhra Pradesh, 12.6 years ( Ministry of Law, Justice and Company Affairs 
1977). I know of no reason to believe that these courts are unrepresentative. 
18. On the formalities of selection, see Art. 124 (Supreme Court) and Art. 217 (High Courts). 
19. On the practice of selection, see M. Gupta 1964: 368; Gadbois 1968­9: 335; Law Commission 
1958: I, 34, 69. 
20. Art. 124(4). Art. 218 makes this provision applicable to the High Courts. 
­480­ 
appointment of Chief Justice Chandrachud in 1978.  21.   In the Supreme Court, this has led to a rapid 
turnover of Chief Justices. From 1950 to early 1978, there were 14 occupants of the Chief justice's 
seat­­an average tenure of just two years.  22.   
Although judges and lawyers share a common legal education and a common professional milieu, 
members of the higher judiciary possess some distinctive characteristics. They tend to be drawn 
from prosperous and high status families,  23.   to have been educated in elite institutions,  24.   to have 
had distinguished professional careers and, with few exceptions, very little political involvement. 
Indian judges, especially in the superior courts, have a tradition of relative isolation from political 
life and from governmental connections.  25.   Since Independence there has been some decline in this 
"tradition of isolation and alooffness."  26.   
____________________ 
21. The chief justiceship of the Supreme Court came up for appointment in April 1973, shortly after 
a momentous decision holding that Parliament's power to amend the Constitution did not include 
a right to alter its "basic structure" aggravated the government's simmering displeasure with the 
judiciary ( Keshavananda v. State of Kerala, A. I. R. 1973 S. C. 1461). Departing from the 
seniority convention, the government passed over Justices Shelat, Hegde, and Grover, all of 
whom had opposed the government's position in Keshavananda, to appoint A. N. Ray the Chief 
Justice of India. The three "superseded" judges promptly resigned in protest. A fierce controversy 
ensued and spawned a vast polemical literature­­e.g., Palkhivala 1973. In January 1977, Justice 
Beg was appointed Chief Justice, in spite of the greater seniority of Justice Khanna, who 
resigned in March 1977. After the demise of Mrs. Gandhi's Emergency Rule, the Janata 
Government in 1978 pointedly adhered to the convention of appointing the seniormost judge, 
notwithstanding grumblings that the judge in question had not been sufficiently stalwart in his 
opposition to the misdeeds of the Emergency. 
22. Had the seniority convention not been broken by Mrs. Gandhi, the number of Chief Justices 
during this period would have been 16 (barring untimely deaths). 
23. Scheduled Castes and Scheduled Tribes were virtually absent from the higher judiciary during 
the period discussed in this book. Of the 352 high court judges sitting in 1977 there were only 
four members of Scheduled Castes and not a single member of the Scheduled Tribes ( RCSCST 
1975­77: I, 11). 
24. For example, they are much more likely to have had legal training in England. Fully one­third of 
the judges appointed to the Supreme Court until 1967 had been called to the bar in England. On 
the careers of Supreme Court judges, see Gadbois 1968­69 ; R. Dhavan 1977. 
25. Gadbois 1968­69: 330. Tripathi ( 1966: 28­29) attributed the "divergence between national 
aspirations and judicial pronouncements" to the recruitment of judges from among those who 
"generally speaking kept themselves aloof from this historic upheaval [the independence 
movement) and continued to function in the imperial and princely governmental institutions as 
administrators, lawyers, judges and civil servants." 
26. Judges are relatively little given to extra­judicial expression through speeches, lectures, and 
books­­except for inspirational talks to bar meetings. Chief Justice Subba Rao shocked 
professional opinion by his active candidacy for the presidency in 1967. More recently, some 
judges have earned a reputation as advocates of (pro­government) causes off the bench. 
­481­ 
As arbiters of the constitutionality of governmental action, equipped with broad powers of 
intervention, the judiciary was inevitably embroiled in battles between contending political forces. 
In applying the broad generalities of the Constitution to the actions of the government, the judiciary 
emerged as critic and monitor of government­­a role enhanced by the fact that other organs of 
criticism and redress were often feeble or overburdened. So the higher judiciary served not only as 
vindicator of Fundamental Rights but as the single most persistent and significant agency of 
criticism of government operation of terms of constitutional norms.  27.   The courts have not been 
noted for a marked solicitude for government, which according to Gadbois' calculation came out the 
loser in about 40% of the cases in which it was a party. Nor did the courts shrink from employing 
their power to strike down offensive legislation. For example, from 1950 to 1967, the validity of 
legislation was attacked in 487 Supreme Court cases; the Court struck down all or part of the 
impugned enactment in 128 cases.  28.   
The higher judiciary in India enjoys enormous respect. Judges are seen as honest, independent, and 
immune to the partiality and narrowness that are attributed to most actors in public life. The 
judiciary had to undertake these expanded responsibilities as critic and monitor of government at a 
time when its own prestige and that of the bar was undergoing a relative decline. The emphasis on 
economic development and the burgeoning of elective politics enhanced the power and prestige of 
other callings, accelerating the long­term drift away from the bar as a focus of ambition for the 
educated. Judges and lawyers were often seen as mere word­mongers who had little to contribute to 
national development­­when they were not obstructing it by overzealous protection of vested 
interests. The judiciary was severely criticized for thwarting governmental social and economic 
policy, for being too solicitous of the claims of property, for being too conservative in labor matters 
(and, in other quarters, for being too ready to bear with the executive).  29.   Sensitive to criticism, 
judges have from time to time employed their broad contempt powers to curb prominent detractors. 
In the early 1970s, critics aligned with the government mounted a campaign for a "committed 
judiciary"­­generally taken to mean 
____________________ 
27. Gadbois 1977. 
28. Gadbois 1970a: 3. 
29. See M. Gupta: 1964: 363 ff. The impatience with the bar and bench in the early years of the 
Constitution was shared by Prime Minister Nehru, who in the course of the debate over the First 
Amendment observed to Parliament: "Somehow we have found that this magnificent 
Constitution that we have framed was later kidnapped and purloined by the lawyers" ( 
Parliamentary Debates, Vol. XII­XIII [part II], col. 8832 [ 17 May 1951]). 
­482­ 
judges favorably disposed toward the "socialist" policies of the government. 
In spite of outspoken complaints about the formalism of the law and the conservatism of judges, and 
against a background of some mistrust of courts and lawyers, attitudes toward the higher judiciary 
remain remarkably respectful. Among those who deplored particular decisions, there has been little 
talk of outright defiance; even talk of evasion is not respectable in Indian public life. During 
Emergency rule, the government took care to curtail the authority of the courts by formal legal 
means, rather than by breaching the fabric of legal authority. Although the Supreme Court 
disappointed many critics of the Emergency, the High Courts emerged from Emergency rule with 
their reputation enhanced as guardians of citizens' rights. The extraordinary regard in which the 
higher judiciary is held is evidenced by their frequent use for extra­curricular "integrity jobs." 
Hardly any situation in which there is political stalemate or loss of confidence in other authorities 
arises without aggrieved parties calling for­­and often obtaining­­an inquiry by a High Court or 
Supreme Court judge.  30.   
On the bench, Indian judges tend to take an active, sometimes assertive, role in the taking of 
evidence and the controlling of argument of counsel. (It is my impression that less initiative in the 
management of the case is left to the lawyers than is the situation in the United States.) Indian 
judges feel free to introduce arguments and to consult sources which are not put before them by the 
lawyers.  31.   But such initiative is limited by their enormous work load. There is an emphasis on 
verbal skills; opinions are sometimes delivered orally and otherwise are dictated to stenographers. 
 32.   The higher courts are heavily overworked, and often judges have little time for editing the 
dictated version. The predominant style is formalistiC.  33.   Social phenomena are addressed by fitting 
them into a framework of pre­existing legal categories. The Indian version of formalism emphasizes 
the autonomy and the determinative character of legal learning. Judges expatiate on the inherent 
meaning of words; they elicit definitive answers from textual passages; they portray themselves as 
controlled by inexorable rules of procedure and precedent. Judges apply pre­existing rules; that this 
____________________ 
30. E.g., four of the commissions appointed by state governments to review the controversial matter 
of identifying the Backward Classes were chaired by retired High Court judges: the Kumara 
Pillai Commission in Kerala, the Jammu and Kashmir Commission, the Andhra Pradesh 
Commission, and the Gujarat Commission. 
31. See, e.g., n. 98, below. 
32. Since the early 1970s the Supreme Court has had a few research officers. But the judges do not 
have clerks (in the American sense) or other professional assistance. 
33. On legal formalism, see Kennedy 1973 ; Friedman 1966. 
­483­ 
involves choices is rarely emphasized, for it is assumed that such choice will be guided by formal 
principles (e.g., canons of statutory interpretation) rather than by assessment of consequences. 
Therefore, making choices is a matter of professional knowledge and skill rather than of personal 
values or political commitments. 
The prevailing style of judicial work has impressed many foreign observers as less than optimal. The 
Indian judicial process, one distinguished comparativist put it, "is in an interesting, though not 
satisfactory, position."  34.   Speaking of law in India generally, he observed that "the difficulty is after 
all one of intellectual attitude and conception­­the failure to approach legal problems functionally 
and its corollary, a static conception of law."  35.   And a close student of the Indian Supreme Court 
recently characterized the Indian legal culture as one of "procedural nitpicking," "hair splitting 
legalisms," "literal interpretation," "narrow, technical, and mechanical," expressing "concern for 
form, not for policy or substance."  36.   One can recognize in this unflattering portrait some of the 
pervasive stylistic features of the judicial process in India. But our compensatory discrimination 
cases suggest that the formalistic style should not automatically be equated with an absence of 
policy concern on the part of the judges. We find instances of judges smoothly blending the idiom 
of the formal style with reflective pursuit of substantive policy.  37.   We even find instances of tacit 
but incisive pursuit of policy by strict invocation of the canons of formalism.  38.   (And, as 
everywhere, judges may use the canons of formalism to torpedo unworthy or unappealing claims.) 
 39.   Just as purposive forays are possible within the formal idiom, abandonment of that idiom does 
not guarantee that chosen policies will be effectuated.  40.   
Although the formalistic idiom undoubtedly constricts the range of 
____________________ 
34. Von Mehren 1963a: 279. 
35. Von Mehren 1963b: 15. 
36. Gadbois 1977. 
37. E.g., Chatturbhuj Vithaldas Jasani v. Moreshuar Parashram, 1954 S.C.R. 817; Shantha Kumar v. 
State of Mysore, 1971 Mys. L. J. 21; Urmila Ginda v. Union of India, A. I. R. 1975 Del. 115. 
38. Periakaruppan v. State of Tamil Nadu A.I.R. 1971 S.C. 2303, discussed in chap. 13, §A, above. 
39. Consider Ramalingam v. Boddu Abraham, A. I. R. 1970 S. C. 741, 743, where the losing 
candidate failed to prove his allegation that the winner had converted to Christianity (and was 
therefore not a Hindu), because it was the winner's parents who had converted. Where "because 
of clumsy blundering the petitioner undertook a much greater burden than the law required him 
to take" (i.e., proving winner's conversion rather than that he was a non­Hindu), he lost because 
of "the vagaries of litigation which was to be carried on according to rules." 
40. Cf. the analysis of the imponderable consequences of the "purposive" decision in the Thomas 
case, chap. 11, §D, above. 
­484­ 
policy considerations that judges take into account, it does not exclude considerations of substance 
or policy entirely. The restrictive effect of judicial ideology cannot be separated from the other 
features of the judges' situation that militate against purposive concern with policy. The enormous 
work load that faces them, the fact that they encounter isolated pieces of complex situations, the 
lack of information or analysis on which to base a decision, and the imponderable effects of many 
decisions surely impose limits at least as formidable as attachment to the formal style. The isolation 
of the judiciary from the purposive effectuation of policy has sources deeper than judicial style. But 
this style has important effects as a cultural model, if not as a structural barrier. By educating 
lawyers about what are acceptable arguments (and ultimately educating claimants about what are 
justiciable claims), formalism stabilizes the role of the judiciary: it elicits claims and arguments that 
courts respond to in the formal idiom. 
Judges and lawyers, whether trained in India or abroad, have been exposed to legal education which 
stressed formal analytic treatment of legal doctrine, rather than considerations of policy. They have 
virtually no exposure to social inquiry in or out of law college. There is little in their professional 
life to offset this emphasis. India has not had a flourishing legal scholarship. As one reflective High 
Court judge observed, "[t]he one great weakness of the Indian judicial system today [ 1964] is that 
[it] lacks . . . theoretical nourishment."  41.   Since then a critical literature has begun to emerge,  42.   but 
there is little to indicate that it has reached many of the lawyers and judges. Most judges are not 
voracious readers, and if they read any legal periodical, it is more likely to be an English or 
American than an Indian one. 
In their judgments, the judges do not often venture outside the boundaries of case law. Standard 
legal works are sometimes cited, as are government reports. But reference to current scholarly 
comment is rare. In the litigation about compensatory discrimination there has been only a single 
instance in which a judgment referred to scholarly comment on this topic.  43.   In some of the group 
membership cases, judges have referred to ethnographic works for evidence of tribal usage. But 
where judges venture beyond the boundaries of Indian case law, the sources are likely to be Anglo­
American rather than Indian and conceptual rather than descriptive. In our compensatory 
discrimination cases, the most elaborate citation of the external sources is found in the various 
judgements delivered in the Thomas case. Although the judges 
____________________ 
41. S. Dhavan 1964: 348. 
42. As reported ( R. Dhavan 1977: 71) and exemplified by that judge's son. 
43. In Sagar v. State of Andhra Pradesh, A. I. R. 1968 A.P. 165 at 189, Jaganmohan Reddy, J., cites 
Radhakrishnan 1965 on the interpretation of Chitralekha. 
­485­ 
spend most of their time on the Indian case law, several refer copiously to American cases as well. 
 Justice Ray cites a  California Law Review
    article by Tussman and ten Broek; Justice Mathew cites
    
 Plato, Aristotle, Harold Laski, Bernard Williams, R. H. Tawney, John Rees, a  Harvard Law Review
    
comment, and the Moynihan Report. Justice Krishna Iyer draws upon Charles Black, Wolfgang 
 Friedmann, Anthony Lester in the  Modern Law Review
   , Ralph Ellison, and the off­thebench
    
writings of former Chief Justices Hidayatullah and Gajendragadkar, and he provides what appears to 
be the single reference to any work analyzing, Indian social phenomena.  44.   Clearly, the connection 
between judicial decision­making and systematic empirical learning about Indian society remains to 
be established. 
Overall the courts have been harsh critics of governmental arrangements for compensatory 
discrimination. A search of the law reports turned up a total of 113 reported litigations about 
compensatory discrimination from 1950 to 1977.  45.   Putting to one side the 32 electoral disputes, 
consider the outcome in the other 81 cases. The High Courts 
____________________ 
44. His statement, A. I. R. 1976 S. C. 490 at 531, that "research conducted by the A. N. Sinha 
Institute of Social Studies, Patna, has revealed a dual society among Harijans . . ." apparently 
refers to the research published as Sachchidananda ( 1977). 
45. For purpose of this and subsequent computations, the universe consists of all reported cases in 
the High Courts and the Supreme Court in which a substantial claim was asserted as to the 
existence, method, or extent of a policy of compensatory discrimination, or the designation or 
membership of the recipients of preference. It does not include attempts to extend doctrine from 
this area to other kinds of fact situations. Reported cases seems a reasonable place to draw the 
line because they are part of the corpus of authority and because someone with a grasp of the 
Indian legal scene made the decision that they were worth reporting. Thus, I have omitted some 
unreported cases which clearly fit the description above, along with some election cases which 
apparently did not proceed beyond the Election Tribunal and a few cases in which the claim in 
regard to preferences seems marginal. I suspect that a few cases, reported only in local law 
reporters, have escaped this enumeration, but there is no reason to believe that they would in any 
way alter the general profile. The same is true for the few reported cases discussed in this book 
but discovered too late for inclusion in these computations and marked [NIC]. 

Obviously, there is no assurance that this sample is not biased because of unreported cases. 
Although much trivia is reported, in a group this large over this period of time, I think it is safe to 
assume that the reported cases fairly represent what the courts have done. A partial check was 
permitted by examining the 6 unreported cases in my possession and by examination (in 2 cases), 
and indirect reconstruction (in the others) of 16 unreported judgments in litigations where later 
stages were reported. I could detect no important difference in subject or disposition, except for a 
greater preponderance of election disputes among the unreported cases. 

Although for convenience they are referred to as "cases," it should be noted that the unit here is 
not reported case as such but a litigation­­i.e., a dispute involving any number of parties and 
going through one or more courts. Otherwise, the accident of a case being reported two or even 
three times could bias our computations. 

­486­ 
struck down governmental action in whole or in part in 27 cases while upholding it in 40. State 
rather than central arrangements drew the most critical response from the High Courts: in 57 
litigations concerning preferential treatment conferred by the state governments, they ruled against 
the state in 24. But in ten cases involving central preferences, they ruled against the Centre in only 
three. While all of the central preferences were for Scheduled Castes and Tribes, 16 of the 24 invalid 
state arrangements were for the Backward Classes. 
Central schemes and schemes for Scheduled Castes and Tribes fared better in the Supreme Court as 
well as in the High Courts. Putting aside the electoral cases, the Supreme Court gave decisions in 24 
litigations, in which it upheld the government's preferential scheme in only 8 cases and struck it 
down in whole or part in 16 instances. Only 4 of these cases involved central schemes and they were 
upheld in 3 of 4 cases. But state schemes were struck down in 15 of 20 cases. The Supreme Court 
was particularly censorious of government in the exercise of its original writ jurisdiction. Eleven 
Article 32 petitions were brought against state government schemes; they were struck down in 10 
cases.  46.   The Supreme Court was more permissive when it sat on appeals from the High Courts. In 
10 appeals it reversed the High Courts six times.  47.   In 5 of these, the Supreme Court held that the 
preference conferred by the state was permissible where the High Courts had said that it was not. 48. 
Three of these 5 concerned Scheduled Castes­­including the 2 major cases enlarging government 
power to confer preferences on them.  49.   
Cases involving compensatory discrimination have, at times, placed some strain on the judges 
because of the intense political and social controversy involved. This is an area in which judges' 
values, ideals, and perceptions of their society cannot but affect their reactions. But the relative 
infrequency of such litigation, the system of sitting in 
____________________ 
46. Perhaps this reflects the Court's ability to pre­screen Article 32 cases at the admissions stage. 
47. The six cases are General Manager v. Rangachari, A. I. R. 1962 S. C. 36; Sarda v. Mizo District 
Council, A. I. R. 1967 S. C. 829; State of Punjab v. Hira Lal, A. I. R. 1971 S.C. 1777; State of 
A.P. v. Balaram, A. I. R. 1972 S. C. 1375; State of U. P. v. Pradip Tandon, A. I. R. 1975 S. C. 563; 
State of Kerala v. N. M. Thomas, A. I. R. 1976 S. C. 489. There were also three reversals in 
electoral cases ( Punjabrao v. Meshram, A.I.R. 1965 S. C. 1179; Abhoy Pada Saha v. Sudhir  
Kumar Mondal, A. I. R. 1967 S. C. 115; Laxman Siddappa Naik v. Chandappa, A. I. R. 1968 S. 
C. 929), in each of which the High Court had held that a successful candidate was not a 
Scheduled Caste, but the Supreme Court found that he was. 
48. Only in Sarda is the governmental action not upheld, and the preference element there is not 
clear. In Pradip Tandon the Court upholds the more permissive of the High Court responses to 
the state scheme and overrules an earlier, more restrictive, High Court case. 
49. I.e., Rangachari and Thomas. 
­487­ 
benches, and rapid turnover on the courts have allowed few judges to sit in enough cases involving 
such schemes to articulate in their judgements an explicit and comprehensive point of view on 
compensatory discrimination. For example, if we take the Supreme Court, we can identify 39 cases 
dealing with the compensatory discrimination policy from 1951 through early 1977 ­­ an average of 
just one and a half cases per year. Of the 59 judges who served on the Supreme Court from 1950 
through early 1977, 50 sat in one or more of these cases. Because of benches of varying sizes, this 
amounted to 160 sittings by these 50 judges ­­ an average of only 3.2 sittings per judge. Fifty­two 
judgments were delivered ­­ just more than 1 per judge. (If we exclude the 14 cases which involved 
only questions of group membership, we find 44 different judges sitting 113 times in 25 cases ­­ an 
average of 2.6 sittings and 0.84 written opinions per judge.) 
While some pronounced differences in approach are evident, there are not enough instances to make 
any quantitative comparisons. No judge has emerged as the champion of compensatory 
discrimination; Justice (later Chief Justice) Gajendragadkar probably comes closest by virtue of his 
opinions in Dora and Rangachari, two of the small number of Supreme Court cases in which the 
scope of preferential treatment was given a broadening, expansive interpretation. A few other judges 
sat often enough to display a consistent tendency to curtail its operation. Overall quantitative 
scoring is not very revealing here, for judges' views tend to vary from issue to issue: the same judge 
may take a very different stance in regard to different aspects of the preference policy. Thus Justice 
Subba Rao, the only judge who took a broad view of governmental power to make extensive 
reservations in Devadasan, took a very restrictive view of the State's power to employ the caste 
criterion in Chitralekha. Again, the unreliability of quantitative scoring is increased by the fact that 
many of these cases involve a whole bundle of issues, so that there is no necessary connection 
between the doctrine pronounced on any one point and the outcome of the case. In Chitralekha, 
animadversions against the caste basis were set forth at length, although no state action of this sort 
was before the Court. On the other hand, in P. Rajendran the Court was quite permissive on the use 
of caste by the state, but the scheme was struck down on a different ground. 
In the High Courts, too, few judges have had sufficient exposure to these issues to facilitate 
quantitative comparison. Even in Mysore, which had a disproportionately large share of these cases 
(19 in 18 years), the largest number of reported cases sat on by any judge was four and the median 
was two. Possibly, some High Court judges may have encountered many more such cases at the 
unreported admissions stage (but interviews with lawyers and judges did not uncover any lore about 
this). 
­488­ 
Styles or tendencies of some of the High Courts as a whole can be more readily discerned than 
those of individual judges. Thus in Kerala there was a reluctant willingness to go along with the 
state on the composition of the Backward Classes while expressing misgivings and pressing the 
government for reforms; in Andhra, on the same issues and during the same period, we find a less 
accommodating strictness on the part of the High Court. 
The homogeneity of judges' backgrounds and experience and strong pressures for unanimity are 
reflected in the infrequency of nonunanimous judgments. Dissenting and concurring opinions are 
considerably less frequent in India than in the United States. Gadbios reports that from 1950 to 
1967, 91% of the judgments in the Supreme Court were unanimous;  50.   in the court's first twenty 
years, Dhavan reports, there were dissents in only 7.4% of all cases derided.  51.   He points out that 
70% of the dissents are of a single judge and non­programmatic in character: 
Dissent is usually on a specific non­recurring matter, or of a token nature, where the 
judge merely expresses another point of view and having made his point, makes no 
attempt to try to get the others to adopt his opinion.  52.   
In the 39 cases involving compensatory discrimination in which judgments were delivered by the 
Supreme Court, 50 different judges sat ­­ a total of 160 sittings. Eight of these judges registered 
dissenting opinions in 6 (16%) of these cases ­­ which suggests that this might be an area of law 
somewhat more volatile and unsettled than most. This impression is confirmed if we exclude those 
14 cases involving only questions of disputed group membership. In the remaining 25 cases, which 
deal with the contours of the preference policy, we have non­unanimous courts in 6 out of 25 cases 
(24%) ­­ a rate three times the Supreme Court's overall rate. (Of the dissent, five were objections to 
the Court's "expansion" of preferences and a sixth was in the same spirit; two objected to undue 
restriction of governmental power to make preferences.) 
There is other evidence of division and disagreement. Dissents are even rarer on the High Courts. In 
part, this reflects the narrower opportunities ­­ a court that disagrees with an earlier court is bound 
either to follow it or to refer the matter to a larger bench; a divided Division Bench is required to 
refer the case to a larger bench. In the 7 preference cases derided by a Full Bench, we find 3 dissents 
(and 1 separate concurring opinion) ­­ but, of course, these are a very select 
____________________ 
50. Gadbois 1970a: 2. Non­unanimous decisions of the United States Supreme Court increased from 
11% in 1930 to 76% in 1957 ( ZoBell 1959: 205). Non­unanimous decisions (in fill opinion 
cases) were 66% in 1968, 72% in 1969, and 81% in 1970 ( Grossman and Wells 1972: 165). 
51. R. Dhavan 1977: 34. 
52. R. Dhavan 1977: 32. 
­489­ 
group of cases; the Full Bench was empanelled precisely because of their difficulty. In the 13 cases 
in which there was appeal from the decision of a single judge to a larger bench, we find 4 reversals. 
And in the 24 appeals from the High Courts to the Supreme Court, we find 9 reversals.  53.   
Strain or change in the law cannot be measured only by dissent or reversal. For structural features ­­ 
sitting in small and shifting benches by overworked judges whose short tenure provides few 
opportunities to encounter these issues and who enjoy few resources to effect a comprehensive 
synthesis of learning on these matters ­­ increase the possibility of shifts from case to case that are 
not reflected in these measures. Most change appears not in dissents but within majority opinions, 
 54.   occasionally by overruling of earlier cases,  55.   but more often by elaborating conflicting 
tendencies within the corpus of case law. The fluctuation of the courts with regard to the 
permissibility of caste units and the disparate strands in their handling of group membership 
problems are examples of the way that different approaches may flourish in spite of the absence of 
articulated disagreement. In spite of powerful unifying influences, the Indian judicial system 
contains features which seem to amplify the normal tendency of a case­law system to accommodate 
inharmonious strands of judicial doctrine. 
The opportunities for diversification without confrontation have increased over time. As the case 
loads of the courts have increased, the number of judges has grown. Thus, the Supreme Court's case 
load increased tenfold from 1951 to 1971,  56.   and the number of judges increased from 5 in 1950 to 
14 in 1980. Their tenure on the Court has decreased.  57.   After 1954, the Court started sitting in 
benches (at first the minimum membership in a bench was three; in 1966, the minimum was lowered 
to two).  58.   The average number of judges sitting in our compensatory discrimination cases has 
fallen from 5.4 in the 1950s to 4.1 in the 1960s to just 3.6 in the 1970s. 
B. AUTHORITY AND PRECEDENT: FORMAL STRUCTURE AND OPERATING NORMS 
Indian courts, like those of other common­law systems, are governed by the doctrine of precedent, 
or stare decisis. All courts in India are 
____________________ 
53. The 62.5% rate of affirmance is very close to the 60% rate for all appeals from High Courts from 
1950 to 1967, calculated by Gadbois 1970b: 38. 
54. Cf. R. Dhavan 1977: 32. 
55. Cf. R. Dhavan 1977: 44. But overruling is more difficult since the 42nd Amendment imposed a 
requirement that reversal be by a larger bench. 
56. R. Dhavan 1977: 116­17. 
57. See above at nn. 15­17. 
58. 1966 Supreme Court Rules O. VII, R. I. 
­490­ 
bound by the judgments of the Supreme Court, not only on questions of constitutional and central 
law, but on questions in every field of law, central or state, on which the Supreme Court has passed. 
 59.   The courts are bound not only by the narrow holding of the Supreme Court's decisions, but also 
 by the broader expressions and  obiter dicta
    of the Supreme Court. 
    60.   That Court, however, is not 
bound by its own prior decisions, but it usually treats them as authoritative.  61.   Where there are 
conflicting Supreme Court decisions, the decision of the larger bench controls.  62.   If the conflicting 
benches are of equal size, the later decision is controlling  63.   ­­ at least in theory. But as we have 
 seen in the handling of  Balaji
    and 
   Chitralekha
   , a later decision with a distinctively different stance
    
may be ignored or assimilated to the earlier decision. 
Neither the High Court in a state, nor the subordinate courts there,  64.   are bound by the decisions of 
the High Courts in other states, but usually these decisions are considered of persuasive authority 
unless their own High Court has ruled differently.  65.   Within a High Court, a 
____________________ 
59. Art. 141. The courts other than the Supreme Court are also bound by decisions of the predecessor 
Federal Court ( 1937­1950) and the Judicial Committee of the Privy Council delivered before 
Jan. 26, 1950 ­­ e.g., Kishan Chand v. Ram Babu, A. I. R. 1965 All. 65. This includes obiter dicta 
of the Privy Council as well as its holdings ­­ e.g., Dass Bank v. Kali Kumari, A.I.R. 1958 Cal. 
530. Judgments of the Privy Council delivered after that date are of merely persuasive force as 
authorities. Decisions of the Privy Council and the Federal Court are not binding on the Supreme 
Court, although they may carry persuasive force ( Srinivas v. Narayan A.I.R. 1954 S.C. 379; 
State of Bihar v. Abdul Majid, 17 S.C.J. 300 [1954]). See, e.g., Gurunath v. Kamalabai, A.I.R. 
1955 S.C. 206 (followed where settled character of law important). 
60. Municipal Committee v. Hazra Singh, A. I. R. 1975 S. C. 1087, 1088­89. A lengthy compendium 
of High Court citations may be found at Saxena 1963: 208 ff. This includes advisory opinions ( 
Ram Kishore v. Union of India, A.I.R. 1965 Cal. 283). 
61. E.g., see Income Tax Officer v. Nadar, A. I. R. 1968 S. C. 623; Superintendent v. Corporation of 
Calcutta, A. I. R. 1966 S. C. 997. 
62. State of U. P. Ram Chandra, A. I. R. 1976 S. C. 2547, 2556. 
63. Superintendent v. Ram Ajodhya, A. I. R. 1965 Cal. 348; Bhavan v. Commercial Tax Officer, A. I. 
R. 1961 Mys. 3. 
64. Subordinate courts in a state are bound by all judgments of the High Court of that state: 
"omission [to follow High Court precedent] is as much dereliction of duty as omission to refer to 
sections of the statute. The disregard of authority is, however, something still more objectionable. 
It amounts to an act of insubordination" ( Rex v. Ram Dayal, A. I. R. 1950 All. 134, 137). Courts 
are divided on whether subordinate courts are bound by the obiter dicta of their High Court. 
They are not, according to Vinayak Shamrao v. Moreshwar Ganesh, A. I. R. 1944 Nag. 44, 541; 
Baula v. Chief Executive Officer, A. I. R. 1955 Nag. 49, 63. But see, e.g., Basunta Kumar v. 
Chief Electrical Engineer, A. I. R. 1958 Cal. 657, 660. 
65. Emperor v. Dholraram, A. I. R. 1941 Sind 221. A court may not follow decisions of other High 
Courts when binding decisions of its own High Court are available ( Bisanchand Lalchand v. 
Union of India, A. I. R. 1961 M.P. 220; Ramaswami v. Chandra Kotayya, I. L. R. 48 Mad. 693 
[ 1924]). But a High Court is not bound by obiter dicta of its own court ( Trilochan Choudhury v. 
Dayanidhi Patra, A. I. R. 1961 Or. 158; Ram Narain v. Director of Consolidation, A. I. R. 1965 
All. 172). 
­491­ 
smaller bench is bound by the decision of a larger bench or a bench of equal strength. Thus a 
Division Bench is bound by a Full Bench or an earlier Division Bench; a single judge is bound by a 
Full Bench, a Division Bench, or an earlier single judge.  66.   If a Division Bench or a single judge 
disagrees with a prior decision of a court of equal strength, it should request that the case be 
referred to a larger bench.  67.   A Full Bench can be overruled only by another Full Bench ­­ possibly 
only by a larger one.  68.   Where there are conflicting decisions of courts of equal strength, it is 
unclear whether a subsequent court may choose or is bound to follow the earliest decision.  69.   Nor is 
it clear whether binding authority is retained by those points in overruled decisions which were not 
decided on by the overruling bench.  70.   
Usually there is a single written opinion in a case decided by a higher court.  71.   Where there are 
several opinions of equal strength, as where both judges of a Division Bench deriver separate 
opinions, they are presumably of equal authority. There is considerable uncertainity about the 
precedential value of concurring judgments of the Supreme Court.  72.   
____________________ 
66. Dhanki Mahajan v. Rana Chandubha, A.I.R. 1969 S.C. 69. 
67. Jai Kaur v. Sher Singh, A. I. R. 1960 S. C. 1118; Mahadeolal v. Administrator­General, A. I. R. 
1960 S. C. 936; Jaisri v. Rajdewan, A. I. R. 1962 S. C. 83. There are slips: within a month of 
each other two division benches of the Allahabad High Court came to opposite conclusion 
regarding the constitutionality of reservations for candidates from rural and hill areas. There is no 
indication that the second ( Dilip Kumar, A. I. R. 1973 All. 592, which struck them down) knew 
of the first ( Subhash Chandra, A. I. R. 1973 All. 295, which upheld them). In Pradip Tandon (A. 
I. R. 1975 S. C. 563, 565) the Supreme Court noted reprovingly: "It is desirable from the point of 
view of judicial propriety to refer to earlier decisions of the same High Court." 
68. Chandhlal v. Babulal, A. I. R. 1952 M. B. 171. However, in practice, it seems that full benches 
can be overruled by subsequent full benches ( Maktul v. Manbhari, A. I. R. 1958 S. C. 918). 
69. Compare Rama Suddarayalu v. Rengammal, A.I.R. 1962 Mad. 450 (court should follow that 
which seems more correct) with Januma Rai v. Chandradip Rai, A. I. R. 1961 Pat. 178 (earlier 
prevails) with Agurchand v. Deochand, A. I. R. 1960 A.P. 101 (more recent and more constant 
with general principles should prevail). 
70. Compare Shri Prithvi Cotton Mills v. Broach Municipality, A. I. R. 1968 Guj. 124 (decision 
reversed on point still remains good authority on another point) with Daya Krishnan v. Assessing 
Officer, A.I.R. 1966 Punj 490, (no longer good law). 
71. In the first few years of its existence, there were some instances in which justices of the Supreme 
Court delivered individual opinions. By 1953 it had become the invariable practice to have one 
justice driver a single opinion for the court. 
72. Doubt about less than majority decisions arises from Art. 145(5). See Shri Prithvi Cotton Mills v. 
Broach Municipality, A. I. R. 1968 Guj. 124 (minority judgment not binding); Jayantilal v. Rana, 
A. I. R. 1963 Guj. 80 (concurring opinion of a High Court judge not binding). 
­492­ 
justices which are uncontradicted by the majority may have some weak value as precedents.  73.   
The body of precedent consists of the judgments of the Supreme Court and the High Courts ­­ 
formally only of those officially reported, but in practice those reported anywhere.  74.   There is no 
reporting of the decisions of the subordinate courts. Nor are all of the judgments of the higher 
courts reported. There is no reporting of proceedings at the admission stage, which are so crucial to 
writ petition litigation. Judges are empowered to determine which of their judgments should be 
reported. Many cases which turn on questions of fact or which are felt merely to reiterate established 
principles are not reported, although the practice differs in the various High Courts. Inquires to 
lawyers and judges turned up only a few unreported cases in the compensatory discrimination area, 
none of which involved any important departure from the reported body of material. Reporting is 
slow ­­ a year or sometimes more may elapse before the decision is widely available. 
Indian lawyers and judges have an intense respect for English decisions, and the latter are often 
cited. Courts do regard such decisions as illuminating and sometimes persuasive guides, but not as 
binding by virtue of the doctrine of stare decisis.  75.   In the fields of preferences, as in many other 
areas of Indian constitutional law, there is little English precedent to go on. In dealing with 
constitutional question, courts and lawyers do look to American (and to a lesser extent to 
Australian) precedent where they are relevant. Again, such precedents are not binding and are not 
always persuasive. The practice of looking to American cases reached its zenith in the mid­fifties 
and has gradually been declining since then as a body of Indian precedent was built up. Although 
"equal protection" is an area in which American learning is available,  76.   there is little from America 
­­ or at least there wasn't until 
____________________ 
73. Venkatasubramanian ( 1965) says courts are bound by the views of dissenting judge where the 
majority did not express any view of the matter. Raji v. Deputy Tahsildar, A. I. R. 1963 Ker. 155. 
Cf. Ranjit Kumar v. State of West Bengal, A. I. R. 1968 Cal. 551. 
74. Although the Indian Law Reports Act, 1875 ( Act XVIII of 1875) makes only the official report a 
source of binding authority, in practice those unofficially unreported are freely cited­­and even 
unreported judgments are treated as good authority. Cf. Baij Nath Prasad v. State of Bhopal, A. I. 
R. 1957 S. C. 494. 
75. E.g., Income Tax Commissioner v. Vazir Sultan, A.I.R. 1959 S. C. 814. In some fields at least, 
English law is the residual source of law for Indian courts ( Phulchand v. Hukumchand, A. I. R. 
1960 Bom. 438). 
76. Art. 14 (equal protection) is an area in which Indian courts have been receptive to American case 
law. Cf. the observation of the Supreme Court that: "It may reasonably be assumed that our 
Constituent Assembly when it enshrined the guarantee of equal protection in our Constitution, 
was aware of its content delimited by judicial interpretation in the United States of America. In 
considering the authorities of the superior courts in the United States, we would not therefore be 
incorporating principles foreign 
­493­ 
recently ­­ or elsewhere that offers any guidance to Indian courts in the field of preferential 
treatment.  77.   This is a purely Indian area of the law and one erected from the bare constitutional 
texts in the last twenty­five years. In expounding the meaning of these constitutional provisions, the 
courts have been reluctant to go back to the Constituent Assembly Debates to determine the 
meaning of the text ­­ though this reluctance may be waning.  78.   (Indian courts do not consult 
legislative history to ascertain the meaning of particular enactments.)  79.   
One may wonder how the relatively few cases in this field could have generated so much law. Indian 
courts, unlike their more parsimonious American and British counterparts, tend to decide all ­­ or at 
least a great many ­­ of the issues of fact and law which arise in the cases before them. For example, 
where one of several alternative defences is upheld, an Indian court will often proceed to consider 
the remaining defenses on their merits, even though they are strictly speaking, not needed to reach a 
decision. The prevailing view is that all of these "alternative trains of reasoning" in a particular 
judgment have equal binding authority, even though the case could have been disposed of on only 
one ground.  80.   For example, in Balaji's case, the Supreme Court found three 
____________________ 
to our Constitution or proceeding upon the slippery ground of apparent similarity of expressions 
or concepts in an alien jurisprudence developed by a society whose approach to similar problems 
on account of historical or other reasons differ from our own" ( State of Uttar Pradesh v. 
Deoman, A.I.R. 1960 S.C. 1125 at 1131­32). An eminent commentator observes that Art. 14 "has 
been given content by practically taking over the theory of classification developed by the U.S. 
Supreme Court" ( Seervai 1968: 34). 
77. American cases are cited to bolster the "classification" argument in State of Kerala v. N. M. 
Thomas, A.I.R. 1976 S.C. 490. See discussion in chap. 11, § D, above. Earlier, purported 
American analogies had not proved persuasive when put forth in preference cases. See, e.g., 
Sukhdev v. Government of Andhra Pradesh, 1 Andh. W.R. 294. In Jagan Nath v. State of Jammu 
and Kashmir, A. I. R. 1958 J. & K. 1, American cases were relied upon ­­ but the holding there 
was directly under Art. 14 because at the time Art. 16(4) did not extend to Jammu and Kashmir. 
Ironically, one early preference case cited Plessy v. Ferguson, 167 U.S. 537 (1896), for the 
proposition that the burden lay with the petitioner to show that the classification is unreasonable 
( Om Prakesh v. State of Punjab, A. I. R. 1951 Punj. 93.). 
78. See Alexandrowicz 1957, chap. 1; Merillat 1967. 
79. State of Mysore v. Bidap, A. I. R. 1973 S. C. 2555, 2558 ("There is a strong case for whittling 
down the Rule of Exclusion followed in the British courts .... Where it is plain, the language 
prevails, but where there is obscurity or . . . other special circumstances it may be legitimate to 
take external assistance"); Anandji Haridas & Co. v. Engineering Mazdoor Sangh, (1975) 2 S. C. 
J. 415. 
80. Even where the case could have been disposed of on one point. See, e.g., Jaiwant Rao v. State of 
Rajastan, A. I. R. 1961 Raj. 250; Reddi v. Savitramma, A. I. R. 1961A.P. 305; Janardhana Rao v. 
Deputy Transport Commissioner, A.I.R. 1965 A.P. 115. But it is not unknown to read the 
previous case so that one train is dictum and not binding ( Ram Narain v. Director, A. I. R. 1965 
All. 172). 
­494­ 
independent grounds upon which the Mysore provisions violated the Constitution: first, the 
exclusive reliance on caste in measuring social backwardness; second, the inclusion among the 
Backward Classes of groups that were only relatively backward; third, the magnitude of the 
 reservation. The courts that have cited  Balaji
    have considered each of these pronouncements of
    
binding authority. 
The corpus of authoritative doctrine is enlarged by another feature of precedent in India: the 
assignment of greater weight to the dicta of the Supreme Court than to the holdings of the High 
Courts. The Constitution commands that all courts apply the law "as declared by the Supreme 
Court".  81.   This is usually taken to include obiter dicta, a view that has not completely restricted the 
flexibility of High Courts in handling Supreme Court dicta.  82.   The more assertive High Court 
judges find little difficulty in sidestepping Supreme Court pronouncements; the less bold, however, 
are likely to take the matter as settled by a Supreme Court dictum.
This orientation to the peak of the system is reflected in lawyers' perceptions and responses. They 
tend to be relatively unobservant about the doings of other High Courts. For example, the pre­Balaji 
litigation in Mysore clearly invalidated a number of arrangements widely employed elsewhere, 
especially in southern India. But there was no initiative taken elsewhere until the Supreme Court 
intervention, which was quickly followed by litigation in several states. 
The Indian judicial structure, as we have seen, is more unified, more pyramidal, and more heavily 
weighted at the top than, for example, its American counterpart. First, unlike the American, the 
unified Indian judiciary deals with all questions of state as well as central law.  83.   There is no 
exclusion of issues of local and state competence as disputes move up in the judicial system. 
Generally there is less narrowing of the issues as the case proceeds to higher levels in the system. 
The tendency for many issues in a case to remain open, and the generosity with which 
____________________ 
81. Art. 141. A view taken by a majority on which several judges reserved opinion is law declared by 
the Supreme Court and overrides earlier decisions of the court ( Bhavan v. Commercial Tax 
Officer, A. I. R. 1961 Mys. 3). 
82. Compare, e.g., Kalinath v. Nagendra Nath, A. I. R, 1959 Cal. 81, 83, where the Supreme Court 
dictum was followed even though the decision there was on other grounds, since "the Supreme 
Court Justices took a definite view on this point . . . it is binding on us," with Pardhasaradhi Rao 
v. Srinavasa, A. I.R. 1959 A.P. 512, 516, where the court carefully distinguished the Supreme 
Court case as "only authority for what it actually decides and the generality of expressions in it 
are not intended to be expositions of the whole law but are governed and qualified by the 
particular facts of the case. . . ." See Popkin 1962: 247 ff. 
83. On the Forty­second Amendment's attempt to constrict Supreme Court jurisdiction over state 
law, see n. 4 above. 
­495­ 
Indian judges decide points, make for doctrinal profusion at the same time that the system is 
institutionally unified. 
The broad appellate jurisdiction of the Supreme Court and the multiplicity of appeals make that 
Court a potent instrument of unification and integration.  84.   The centralizing tendencies are 
augmented by the relatively closely circumscribed recruitment practices and the lack of localizing 
devices (like juries, locally elected judges and prosecutors, local financial responsibility for courts 
and police).  85.   The top­heaviness is emphasized by the great social and professional distance 
between peak and lower agencies and by promotion on grounds of seniority and conformity. Like 
other formal structures in India, the lower echelons are severely constrained in their capacity to 
innovate and to accommodate new situations.  86.   (This top­heaviness is expressed in the reluctance 
of many lower court judges to hear elaborate legal arguments on the ground that any innovation is a 
monopoly of the High Courts.) All questions can be reopened at higher levels;  87.   there is little 
apportioning of power to make final decisions in different matters to subordinates at various levels. 
The competence of the higher authorities is viewed, not as another specialized skill, supervisory or 
managerial, but as a generalized potency which encompasses all of the skills of the subordinate 
staff. It is as if the lower courts were mere conduits through which the legal potency of the higher 
courts flows. 
In spite of their broad jurisdictional scope and the multiplicity of appeals, the role of the Supreme 
Court and the High Courts in making law uniform and predictable is undermined by a number of 
features: sitting in benches, high judicial turnover, overwork, and slow reporting. Discrepancies in 
the case law are attributable not only to ordinary judicial vagaries, but to the constant turnover of 
personnel who decide similar cases, sometimes unaware that a similar issue has been decided 
differently elsewhere. Judges are overworked, having little time for 
____________________ 
84. The Supreme Court's jurisdiction extends to cases where the High Court certifies that the case 
involves substantial constitutional questions; if refused, the Supreme Court may grant special 
leave to appeal (Art. 136). 
85. Cf. another centralizing factor, the right of the State to appeal acquittals in criminal cases, 
discussed in Law Commission 1958: II, 810 ff. The austere absence of overt channels through 
which local concerns and attitudes can gain expression may not be unrelated to the prevalence of 
covert devices by which the law is bent to local views. 
86. A similar feeling about the relation of subordinates and superiors informs the Indian notion of 
"due process" or reasonableness which includes having a sufficiently high official make the 
decision. See, e.g., Dwarka Prasad v. State of U. P. (1954) S. C. R. 803. 
87. Cf. Gledhill ( 1964: 13) observation that English procedure has been so adapted in India that "it 
seems based on the assumption that the lower courts are incompetent to do anything more than 
prepare a record of evidence and that the real contest is in the court of first appeal." 
­496­ 
reading and reflection (and, as we shall see, little critical stimulation). A distinguished leader of the 
bar observed: 
Those practicing before the [Supreme] Court know that on an average probably each of 
the Judges has to prepare in the course of a week about four judgments apart from his 
sitting in Court hearing cases and reading a number of miscellaneous matters during the 
five working days in a week. In the result the Court is often unaware of its own 
decisions and is apt to make inconsistent pronouncements.  88.   
This was epitomized in an incident in 1959 when on the same day two three­judge benches of the 
Supreme Court encountered the same question (of construction of a section of the Industrial 
Disputes Act) and decided it in opposite ways.  89.   
Notwithstanding their hierarchic organization and the formal doctrine of precedent and authority, 
the Indian courts diverge in practice from the model of tight pyramidal hierarchy in which legal 
rules are declared or fashioned precisely and unambiguously at the higher levels and applied 
automatically and inexorably at subordinate levels. It is a useful counter to this hierarchic image to 
imagine the courts as a collegial body whose members, by virtue of selection, socialization, habits 
of deference, and common work pressures, share affinities of outlook and style. 
The collegial image should be meaningful to anyone who has ever served on a committee made up 
of like­minded (if individualistic) peers which has to make a lot of decisions in a short time ­­ e.g., 
an admissions committee. Decisions tend to get made on the basis of a limited set of factors, often 
those conveniently found in the file, rather than by individualized weighing of all the particularities 
of the case after a fresh and exhaustive investigation. With insufficient time for everyone to brief 
himself thoroughly about each issue, there is an inclination to go along with what seems sensible 
and workable and to accept some variability among different sub­groups. Similarly, both formalism 
and tolerance for inharmonious strands of doctrine can be seen as a reflection of the work situation 
of the judges, who are asked to make quick and imponderable decisions on the basis of the limited 
materials supplied by the legal culture. 
____________________ 
88. Setalvad 1969: 1. 
89. The appeals turned on widely divergent readings of the term "gratuity." In Brachmachari 
Research Institute v. Their Workmen, A.I.R. 1960 S. C. 257, the workers received only gratuity, 
since "gratuity in the context of the award was held to mean nothing more than retrenchment 
compensation." But in Indian Hume Pipe Co. v. Their Workmen, A. I. R. 1959 S. C. 1081, the 
workers received both gratuity and retrenchment compensation on the ground that these were 
distinct benefits meant to redress different burdens on retrenched workers. 
­497­ 
In spite of all the unifying tendencies, there are many factors tending to allow the doctrinal system 
to be overloaded with conflicting principles, offering great leeway for subsequent courts and great 
opportunity for selective reading of the corpus of authority.  90.   In the compensatory discrimination 
area, where we start from conflicting principles, the institutional setting permits the co­existence of 
competing themes without overwhelming pressure for definitive resolution. This doctrinal 
indeterminacy accentuates the passive or reactive position of the courts.  91.   What they do depends 
very much on the cases that parties bring to the courts and the arguments that lawyers make to 
them. 
C. THE PATTERN OF LITIGATION 
With the exception of election appeals, almost all of the cases in the field of preferences have 
originated in the High Courts or in the Supreme Court under their special jurisdiction to issue writs, 
orders, and directions for the protection of Fundamental Rights.  92.   The writ jurisdiction is the most 
striking legal innovation in independent India.  93.   Parties are permitted to approach the highest 
courts directly with a petition that the court either invalidate legislation or set aside an 
administrative order that contravenes the petitioner's Fundamental Rights. The writ jurisdiction has 
proved extremely popular with litigants, and lawyers exercise all of their ingenuity to put their 
client's claims in the form of a writ. The number of writ petitions in India is immense: by 1956 the 
total in all of India was about 15,000.  94.   In 1973 nearly 4,000 were instituted in Karnataka alone. 95. 
The term "writ" has entered popular speech; the client who has no inkling of the substantive 
provision under which he seeks redress will insist to his lawyer that he requires a "writ." Its 
popularity is not difficult to understand: writ petitions are expeditious and they are relatively cheap. 
The filing fee is almost nominal, compared with the high ad valorem court fees which 
____________________ 
90. In the High Courts and Supreme Court, cases are argued, decided, and reported in English. 
Varying standards of English probably increase the distortion in communication, blunting 
precision and screening out subtleties. However, the effects of this are more pronounced below 
the level of the High Court. 
91. See Black 1973. 
92. Arts. 32, 226. On this jurisdiction and its use, see Moothan 1963; Law Commission 1958: II, 657 
ff. 
93. Alexandrowicz 1956: 35­45: Moothan 1963. 
94. Law Commission 1958: II, 659­60. Moothan ( 1966­67:261) observes that in all of England the 
annual number of applications to the High Court for prerogative writs is about thirty. Gledhill 
( 1964:12) gives an even lower estimate of twenty. 
95. I. L. R. 1973 Kar. (Annex) viii. 
­498­ 
must be paid in advance in an ordinary civil suit or appeal.  96.   Lawyers' fees (the costliest item in 
such litigation) and other costs are lower because the expensive process of presenting evidence is 
eliminated  97.   ­the facts being established by affidavit. Starting at the top, there is less likelihood of 
being embroiled in subsequent appeals. Perhaps the most attractive feature of the writ is the direct 
access that it affords to a high and respected authority. Although the writ jurisdiction has put an 
enormous burden on the higher courts, most observers credit this device with inspiring a widespread 
feeling that redress is available against governmental improprieties. (On the other hand, not only are 
the courts inundated, but the widespread use of writs may undermine the authority of other 
institutions ­­ e.g., universities.) 
The procedure for hearing writ petitions differs from court to court. Some High Courts hear all 
writs by a Division Bench; others hear them by a single judge, with a possibility of appeal to a 
larger bench. In all courts there is some kind of screening process ­­ an admissions stage in which 
writ petitions are examined to see whether they should be heard on the merits. A large majority ­­ 
one knowledgeable lawyer estimated 80% ­­ are eliminated at this stage. Those eliminated include 
cases involving disputed questions of fact. In writ proceedings no evidence is taken. In theory the 
facts are determined by the affidavits of the parties. Generally, writs are not admitted if there are 
serious questions of fact. However, in our compensatory discrimination litigation, we observe that 
courts vary considerably in their willingness to undertake extended factual inquiries.  98   
____________________ 
96. The Indian judiciary, as Alan Gledhill ( 1964: 8) has quipped, "is the most successful of the 
nationalized industries." The court fees paid by civil litigants support the entire judicial system 
and in many states yield a surplus. Avoidance of civil remedies is by no means confined to writ 
petitions; there is massive use of the criminal law to settle what might be regarded as "civil 
disputes" ­­ thus avoiding the delay and high court fees. Cf. the observations of Walsh ( 1930: 26) 
that "the Indian much prefers, if he can seek its assistance, the criminal court for the redress of 
his civil wrongs." 
97. Under Indian practice, lawyers' fees may be awarded to the winning party as part of his costs. But 
fees are allowed according to a schedule and do not necessarily reflect actual fees charged. For 
example, a winning petitioner in a writ case in a High Court may be awarded Rs. 100, although 
the lawyer may charge much more than that. For some rough estimates of lawyers' fees in this 
kind of litigation, see § D, below. 
98 Thus, in Kajari Saha v. State, A. I. R. 1976 Cal. 359 at 369, the judge emphasizes that "this Court 
in its Constitution Writ Jurisdiction is not a Court of facts. This Court does not sit in appeal over 
the decision of an authority whose order is challenged in this proceeding." But cf. Sardool Singh 
v. Medical College, A. I. R. 1970 J. & K. 45, at 47: "In order to test the validity of [petitioner's] 
argument [about the proceedings of the selection committee] we called for all the relevant 
records of the Selection Committee showing the marks fetched by the candidates at the last 
qualifying examination and those awarded to them at the interview by the Selection Committee. 
After an examina­ 
­499­ 
In virtually every instance, it is the government or a public body that is the opposite party, since the 
Fundamental Rights, with few exceptions, run only against the government (in the broad sense, 
including public corporations, nationalized banks, universities, and other components of the 
growing public sector) and do not afford protection against the restriction of these rights by private 
action.  99.   
The use to which the writ jurisdiction is put varies in its emphasis from state to state. For example, 
in Mysore a high percentage of writ petitions were "service matters" concerning the respective 
rights of government servants; most of these are questions arising out of the 1956 reorganization of 
the state and the consequent amalgamation of services from previously separate areas. Service 
matters are less prominent in other High Courts. In Bombay, for example, it is estimated that about 
80% of the writ petitions are concerned with rent and tenancy of agricultural land; in effect, the writ 
petition has transformed the High Court into a court of appeals from a non­appealable land tribunal. 
The number of writ petitions concerned with the policy of preferences is insignificant in every 
Indian state, even in Mysore, which has had more writs concerned with this subject than any other 
state.  100.   
Courts in India are viewed with a curious ambivalence; they are simultaneously fountains of justice 
and cesspools of manipulation.  101.   Litigation is widely regarded as infested with dishonesty and 
corrupt manipulation. But courts, especially the High Courts where this compensatory 
discrimination litigation takes place, are among the most respected and trusted institutions. 102. 
Compensatory discrimination tends to display the benign face of the Indian legal process. It is 
relatively inexpensive and quick; it focusses on genuine disputes ­there are few distracting false 
issues. And with the partial exception of some election cases, there is little occasion for the 
fabrication of evidence. The other face of Indian litigation peeks through  103   in the 
____________________ 
tion of these documents, we find that by and large candidates have been selected for admission 
to the Medical College purely in order of merit. . . . A perusal of these documents would further 
reveal that the Selection Committee did not try to plump up marks of a particular candidate so as 
to tilt the scale. . . ." 
100. Since there are no records kept of writs filed but not admitted, it is impossible to say how many 
more attempts there have been. Interviews with lawyers suggest there were not many unadmitted 
cases in this field. 
101. Kidder 1973, 1974; Morrison 1974. Cf. Galanter 1972b. 
102. Gadbois 1977. 
103 E.g., in Narasimha Reddy v. Bhupathi, 31 Election L. R. 211, 219­20 (1967), the judge refers to 
evidence that was "artificial and improbable" offered by a witness who 
99. Shamdasani v. Central Bank, A. I. R. 1952 S. C. 59. It remains an open question whether the 
courts would fashion writ remedies to protect against violations of those Fundamental Rights 
which run against private individuals rather than against the State (Articles 15[2], 17, 23). 
­500­ 
controversies over group membership and is fully exposed in some of the election controversies. 
Virtually all of the litigation about compensatory discrimination has involved reservations, even 
though preferences in the form of provisions of facilities, resources, and protections directly affect a 
much larger number of recipients. (See table 25.) A rough indication of this concentration is that in 
113 reported litigations about compensatory discrimination which I have found,  104.   only 12 have 
concerned matters other than reservations­­chiefly land acquisition and debt relief. 
This litigation has, by and large, been initiated, not by the beneficiaries of protective discrimination, 
but by others complaining of schemes which affect their interest. Thus of the 113 cases, if we put 
aside the 32 battles over reserved seats,  105.   we find that only 27 of the remaining 81 cases involved 
claims brought to the courts by would­be beneficiaries and that 54 were initiated by those 
challenging the existence or operation of the preference. 
As tables 25 and 26 suggest, litigation about compensatory discrimi­ 
TABLE 25 SUBJECT MATTER OF REPORTED LITIGATION ON COMPENSATORY 
DESCRIMINATION 1950­1977 
 Subject Matter of        
   Involving Reservation
              
   Not Involving
                 
              
   Total 
   
 Preference       
              
   Device 
              
              
              
   Reservation 
   
Education 41 1 42 Government Employment 26 1 27 Elective Office 31 1 32 Other 3 9 12 

TOTALS 101 12 113 

would "not lift his head and was looking down all the time while deposing as if he 
was ashamed of telling lies." Indeed, the court concluded that "his evidence is 
nothing but a tissue of lies." And in Rajgopal v. Arumugum, 1972(2) Mys. L.J. 240, 
253­54, the judge refers to one side's "belated inventions for the purposes of defence" 
and a bit of "tampering" with some documentary evidence by the prevailing party. 
104. On this 
universe, see 
n. 45 above. 
105. Electoral 
cases are 
omitted from 
this 
calculation 
because 
almost all of 
them are 
group 
membership 
problems in 
which rival 
claimants 
from the 
beneficiary 
group dispute 
as to the 
boundaries of 
that group. It 
is difficult to 
identify those 
claims in 
terms of the 
classifications 
used here. 
­501­ 
nation has not been spread evenly over all applications of that principle. It has concentrated heavily 
on the use of the reservation device. But it has not been spread over all uses of reservations. Instead, 
litigation has focussed on a few kinds of opportunities. There are a few persistent scenarios. 
The most common of all litigants in these cases is the prospective student seeking admission to a 
higher technical or professional college ­­ in most cases a medical college ­­ who is excluded 
because of what he regards as an unwarranted reservation. Of our 42 education cases, 41 concern 
reservations in medical and engineering colleges in 11 states. (These are in turn only a fraction of an 
even larger number of "admission writs" that attend the yearly selection of candidates to these 
coveted institutions. As one judge in Mysore, where litigation on this has been thickest, put it, the 
admission writs had become an "annual feature" in that High Court.)  106.   Although some are field 
by single petitioners, these petitions often come in large batches, so that these cases involve many 
hundreds of students. The largest in Mysore involved 473 petitioners. Although all members of the 
affected class might eventually benefit from a revision of the government's order, interim relief is 
available only to the individual litigant.  107.   Thus those in­ 
TABLE 26 INITIATING AND SUCCESSFUL PARTIES IN COMPENSATORY 
DISCRIMINATION LITIGATION 1950­1977 
Initiating Party Seeks To
Subject matter of Preferential Treatment  Overturn, Restrict,Curtail the Government's
Preference Enlarge, Restore, or
Amplify the Coverage
of the Government's
Preference 
Successful Unsucessful Successful Unsucessful 
Education 12 12 4 14 
Government 
Employment 10 10 0 7 
Other 5 5 0 2 
TOTALS 27 27 4 23 

106. Viswanath v. Government of Mysore, A. I. R. 1964 Mys. 132 at 134. 
107. Some High Courts have followed the English practice requiring separate writ applications from 
each party, while others have allowed joint or common applications in cases where the claimed 
right to relief arises from the same act or transaction and there is a common question of law or 
fact ( Annam Adinarayana v. State of A.P., A. I. R. 1958 A.P. 16). Such joint or common writs 
have not so far played a prominent part in litigation about compensatory discrimination. 
­502­ 
terested in obtaining admission in the current year would be inclined to file a writ petition of 
their own, and typically many coattail­riding petitions are filed whenever a college admission 
is under attack. 
All of these admissions cases are concerned with reservations by the state governments rather than 
the Centre, And virtually all of them are concerned with reservations for the Other Backward 
Classes, not for the Scheduled Castes and Scheduled Tribes. As table 26 suggests, those student 
petitioners seeking to overturn or curtail schemes of reservations have enjoyed a high degree of 
success.  108.   (Their 50% rate of success score is higher than that for any classes of parties in 
Gadbois' survey of Supreme Court litigation.)  109.   As we have seen, they have succeeded in 
establishing a variety of limits on the designation of Backward Classes, the extent of reservations, 
and the standards for the operation of preferences. 
Not all of the professional­college litigation is initiated by nonbeneficiaries. In 16 of our cases, it 
was initiated by putative beneficiaries who claimed that government policy wrongly confined the 
scheme or excluded them from it. Not only were they less frequently successful, winning only 3 of 
16 cases, but their successes were of a limited and correctional nature, holding the government to its 
announced policy rather than broadening that policy.  110.   
A less numerous class of litigants are competitors for government jobs who feel that they have been 
unfairly denied employment or promotion by unfounded or excessive reservations. Actually, there 
has been surprisingly little litigation in regard to initial recruitment into government service. In 24 
years there has been only a single reported case concerning initial recruitment into central 
government service! Of the 27 government employment cases that I have found, 19 were concerned 
with reservations in promotions (and a promotion element seems to be present in at least 2 others). 
Of the remaining 8 cases, 5 involved the post of munsif (minor civil judge); that is, the applicants 
were lawyers, presumably more inclined and better equipped to litigate. (At least one of the 
promotion cases also involved a legally trained petitioner.) Thus 
____________________ 
108. Gadbois ( 1970b: 48) calculated the rates of success for claimants under various Fundamental 
Rights provisions. Overall, the individual claimant won in 36 of 120 decisions involving the 
Constitution's equality provisions (Arts. 14­18). 
109. This was the same as the average for all Fundamental Rights claims. Gadbois also has rates of 
success for different classes of parties ( 1970b: 46). Only employees and unions with a 55% 
success rate did better in contests with the government than our medical applicants. For example, 
political candidates prevailed only 21% of the time in contests with the government. 
110. The three successful "affirmative" cases are identified in n. 126 below. The list there includes a 
fourth success (Mohan Rao), but that is a group membership case and hence excluded from the 
present calculation. 
­503­ 
the employment cases may be viewed less as communities fighting over entrance into 
government service than as government servants (and lawyers) fighting over opportunities for 
advancement. The battle focusses not at the top, where reservations have little impact on the 
chances of others, nor at the bottom, where they are hardly needed (and the others might not 
easily afford litigation), but in the middle ­particularly in the upper ranges of Class III and the 
jump from Class III to Class II.  111.   Naturally such litigation clusters where reservations are 
present and most effective ­­ e.g., within the central government in those offices which put into 
practice the provision for reservation in promotions. Eleven of our 13 central government 
cases concern the Railways, Posts and Telegraphs and the Central Secretariat. All of the 
litigation about central schemes concerned reservations for Scheduled Castes and Scheduled 
Tribes, but 9 of the 14 cases about state schemes concerned reservations for the Other 
Backward Classes. 
Not all of these were attempts by governments servants to fend off reservation. In 7 cases, putative 
beneficiaries took the initiative and attempted to get courts to extend (in 2 cases, reinstate) 
preferences. None were successful. The government servants, on the other hand, were successful in 
limiting the extent of reservations and eliminating practices unfavorable to them.  112.   In several 
states they succeeded in knocking out reservations for Other Backward Classes (at least 
temporarily), and in a number of states they brought about major revisions in state policy, including 
severe curtailment of the scope of preferential treatment. 
A third and much smaller and scattered group of litigants are those licensees, creditors, or property 
owners who object to particular government schemes for backward classes (debt relief, land 
acquisition) that interfere with their property or business. There were only 10 of these cases, all 
concerning various state schemes, of which 5 were successful. No pattern is evident here. 
Finally, there are the litigants in electoral cases ­­ usually losing candidates, but in a few cases 
would­be candidates whose nomination papers were rejected because of dispute about their 
membership in the 
____________________ 
111. Cf. our earlier estimate (chap. 4, § B, above) that it is precisely in Class III that reservations are 
most effective and it is only there that their effectiveness is declining. It is not suprising that this 
is where the serious resistance is to be found. 
112. Comparison with Gadbois' analysis of Supreme Court litigation provides a rough indication of 
just how successful they have been. He found that the overall success ratio of government 
employees litigating with the government was less than one in three 1970b: 53); claimants under 
the various equality provisions of the Constitution prevailed in only 30% of cases ( 1970b: 48). 
Claims of the type considered here are then a successful sub­group of both these classes. 
­504­ 
preferred group. Three cases involved the question of reservations as compartments or guaranteed 
minimums; in all, the courts decided in favor of the mimimum. All the other cases involved claims 
of group membership. Again, no pattern is evident in the gross outcomes. 
Except in the election disputes and a few cases involving protections (debt relief, etc.), the 
government is the opposite party. In many cases, where appointment or admission of others would 
be affected by the decree, it is necessary for the petitioner to join them as respondents. This is part 
of the reason that litigation is more popular when the number of others affected is small (e.g., in 
promotions rather than, for example, an annual recruitment of clerks by the state government, which 
might involve hundreds of potential parties). 113. 
Most of the litigant success has been at the expense of the state governments rather than the Centre. 
If we set aside the electoral cases, we find that in the remaining 81 cases, government loses (wholly 
or partly) in 30. But in only 3 of these is it the central government that loses. (Two of these 3 were 
High Court judgments on limited technical matters of administering preferences; 114. only the 
single Supreme Court case, Devadasan, was of prime importance ­­ and its major holding is now in 
doubt.) 115. By and large, the courts exercised little direct control over preferential treatment by the 
central government. On the other hand, the courts struck down 27 arrangements (20 for medical 
admissions) in eight states. Twenty of these cases concerned preferential treatment for Backward 
Classes. 
Litigation about preferences for Scheduled Castes, though less frequent, was spread fairly evenly 
about the country. But litigation concerning preferences for Backward Classes was, in the 1950s and 
early 1960s, largely a South Indian phenomenon. (See table 27.) But after 1966 the pattern changes, 
and Backward Classes issues are litigated in the North as well ­­ a trend which the recent surge of 
reservations in northern states promises to continue. 
The general pattern that we have detected in compensatory discrimination litigation emerges with 
even greater clarity if we consider only that litigation which reached the Supreme Court. 
As table 28 indicates, all the cases involving the Other Backward 
____________________ 
11 A further reason encouraging litigation about munsifs is that typically the number of successful 
3.  competitors is small, so that it is feasible to make all of them respondents (in contrast to clerks, 
where there may be hundreds of affected respondents who are difficult to track down because of 
transfers, etc.). 
11 Director General v. Natarajan, 1971 M. L. J. 79; Sudama Prasad v. Divisional Superintendent, A. 
4.  I. R. 1965 Raj. 109. 
11
Devadasan v. Union of India, A. I. R. 1964 S. C. 179. 
5. 
­505­ 
TABLE 27 REGIONAL DISTRIBUTION OF LITIGATION ABOUT (OTHER) 
BACKWARD CLASSES 
Period Four Southern States Rest of India 
1951­55 3 0 1956­60 3 1 1961­65 9 2 1966­70 11 10 1971­75 3 5 

Note: The units here are litigations, as defined in n. 45, above. Times are dates of the final
decision. 
Classes arose from state schemes; virtually all of the litigation about Scheduled Castes and Tribes 
arose from central government schemes. The cases decided by the Supreme Court deal almost 
exclusively with three kinds of disputes: (1) problems of reservation by state governments in South 
India for admission of Other Backward Classes to professional (especially medical) colleges (10 
cases); (2) problems of reservation of promotion posts in the central services (especially Railways 
TABLE 28 SUBJECT MATTER OF SUPREME COURT LITIGATION ABOUT 
COMPENSATORY DISCRIMINATION 1950­1977
EDUCATION GOVT. POSTS ELECTORAL OTHER TOTALS
GROUP Reserva­ tions in
profes­
sional
Colleges Other Initial
Recruit­
ment Promo­
tions Group
Member­
ship Other 
a
SC/ST 2(2) 0 0 6(2) 14(0) [1(0)    
 ] 1(1) 23(5) OBC 7(7) 0 0 3(3) 0 0 0 10(10) Both 3(3) 0 2(2) 0(0) 
b
0 0 0 5(5) TOTALS 12(12) 0 2(2) 9(5) 14(0) [1(0)] 1(1) 38    
 (20) 

Note: State schemes in ( ). 
aThis case ( Dora) is listed in group membership column ­­ hence adjustment of total. 

bThe total here (38) is one less than the total of Supreme Court cases given earlier. In dealing

with judicial participation, the two Triloki Nath opinions were treated as two units, but in dealing
with litigations that case as a single unit. 
­506­ 
and the Central Secretariat) for Scheduled Castes and Tribes (5 cases); (3) problems of group 
membership raised by challenges to elections to legislative seats reserved for Scheduled Castes 
and Tribes (14 cases). Thus, these three strands of litigation account for 29 of the 38 cases in 
which the Supreme Court has encountered the compensatory discrimination policy. 
Considering the varied and diverse government programs of compensatory discrimination, how can 
we account for this clustering of litigation around these few problems? Both electoral and 
medicaladmission situations involve high stakes and claimants with status and resources (including, 
in the electoral cases, organizational backing). The government service promotion cases involve 
high stakes for those concerned ­­ government servants in the middle grades­­and often common 
stakes around which group support can be organized. In all of these cases there is at stake a 
substantial interest in some lifeopportunity for the petitioners. Few, if any, cases are deliberately 
brought as test cases ­­ that is, for the primary purpose of establishing a principle rather than 
advancing the interests of an individual litigant. 
But these cases do sometimes become a focus of group or communal struggle.  116.   A lawyer may 
develop a reputation as a champion of one particular interest, providing an informal focus for 
organized effort. In several of these cases we find low­grade government servants supporting 
expensive litigation that appears to be beyond their individual means,  117.   and in fact such litigants 
are often assisted either by organized associations or by ad hoc informal groups of supporters. 
Occasionally, organized communal associations have taken the initiative in instituting litigation, but 
more typically they have intervened in cases which have already been filed by interested litigants. 
For example, in Kerala, in the Jacob Matthew case, the Nair Service Society provided financial 
assistance to those challenging reservations, while the Sree Narayana Dharma Paripalana Yogam, 
an Ezhuva association, hired a leading lawyer to argue in favor of the reservations. A somewhat 
different pattern of intervention is found in Balaji's case, where several communal associations 
formally intervened, attacking the reservation on the grounds that they had been wrongfully 
excluded from the Backward Classes. Although such formal intervention is relatively rare in 
____________________ 
116. Sometimes there may be struggles within the government about reservation policy and litigation 
strategy. The extension of reservations at the promotion stage by the Railway Ministry and the 
conduct of the subsequent litigation (Rangachari) became a cabinet­level dispute between the 
Home and Railway Ministries. 
117. The petitioner in Rangachari v. General Manager, A. I. R. 1961 Mad. 35, rev'd A. I. R. 1962 S. 
C. 36, was supported by an informal association of his colleagues. The cost of this litigation was 
estimated by knowledgeable Indian lawyers as Rs. 3,000­4,000. 
­507­ 
Indian litigation, it seems to be slightly less rare in these compensatory discrimination cages. 
 118.   
None of this intervention has been by "universalistic" associations remote from the immediate 
parties or by disinterested amici curiae who seek the establishment of an abstract principle of law. 
There are, as we saw above, instances of participation by pre­existing and multi­purpose communal 
organizations, built around traditional ties. There are also instance of participation by organizations 
of a more "modern" type (e.g., the associations of government servants of a particular grade). An 
intermediate instance is provided by composite communal interest groups which are bound together 
specifically by an interest in the reservation or other schemes of preference ­­ for example, the 
Central Government Scheduled Castes and Scheduled Tribes Employees Welfare Association, which 
formally intervened in the C. A. Rajendran case; or the state­wide association of Backward Classes, 
which took the initiative in attempting to restore the Andhra Pradesh reservations by litigation; or 
the Kerala Harijan Samskarika Kshema Samity, at whose urging the Kerala government issued the 
extensions that led to the Thomas litigation.  119.   
Some associations of beneficiaries of preferences are little more than paper organizations or mere 
personal organs. But others, especially associations of Scheduled Castes and Tribes government 
employees, are enduring organizations with a wide membership base ­­ although the central 
government has adamantly refused to recognize them.  120.   
____________________ 
118. From figures supplied by Dhavan ( 1977: 34, 107) it appears that intervenors were allowed by the 
Supreme Court in only 301 reported cases from 1950 to 1970, only 14.3 per year and 7.4% of the 
4,098 reported cases during that period. In contrast, of the 39 reported Supreme Court cases 
about compensatory discrimination there were formal interventions in 6 (15.3%). 
119. A. I. R. 1976 S. C. 490 at 494. There are also organizations specifically for the purpose of 
opposing reservations. These include such ad hoc groups as the "antireservationist" 
organizations that were formed in Bihar in 1978, but also more enduring and specialized groups 
such as associations of government employees other than Scheduled Castes and Tribes, who 
have agitated against reservations in the services. 
120In 1975 the Commissioner for Scheduled Castes and Scheduled Tribes ( 197374: 129) noted that 
. the government refusal to recognize employees associations based on caste, tribe, or religion was 
misapplied to "service associations of Scheduled Castes and Scheduled Tribes associations as 
these are not actually based on any particular caste, tribe or religion." He held that the claims of 
these associations should be judged on their merits, which were considerable, since Scheduled 
Castes and Scheduled Tribes employees had distinctive interests in the working of the special 
safeguards and concessions in their behalf. These interests were not adequately protected by the 
general employees associations: "The Commissioner . . . has not come across even a single case 
where a general employees' union took up the case of implementation of the orders regarding 
reservations and concessions for Scheduled Castes and Scheduled Tribes." Nevertheless, a "High 
Power Committee" on reservations in the Central Services chaired by the 
­508­ 
Such associations have been active in making representations in individual grievances and 
pressing collective demands by lobbying and demonstrations, but have only infrequently 
resorted to litigation. For example, in early 1978 the All­India Scheduled Castes and 
Scheduled Tribes Railway Employees Association was supporting a case to disqualify a non­
Scheduled Castes person who had obtained a railway job with the help of a false caste 
certificate.  121.   Although they had a host of grievances about the situation of Scheduled Caste 
and Tribe employees ­­ ranging from administration of reservations, interviews, transfers, and 
quarters, to non­recognition of their organization to the excessive complexity and ambiguity of 
the orders ­­ limitations of financial resources and legal imagination left them unable to 
visualize the courts as a forum for securing redress. 
Although some of the participating associations have strong political links and some of the 
contesting groups are reputed to be linked with contending political parties, it is difficult to discern 
any general political lines in the fights over reservations. Reservations have been challenged by 
Communists and defended by Swatantra leaders and vice versa. Occasionally, the battles do take on 
a party tinge, but many more seem to be intra­Congress fights. 
The propensity to litigate these questions depends on a number of factors. One is clearly the feeling 
of being deprived of a scarce opportunity. Another is the question of expense: the cost of obtaining 
that opportunity by litigation must appear to be less than getting it by some other means. It is 
estimated that in Mysore, where the alternative to a state medical or engineering college is a private 
college with a high capitation fee, litigation was regarded as a much less expensive means of 
obtaining admission than payment of the fees in a private college. The tendency to have very large 
numbers of writ petitions whenever a reservation is challenged can be explained in part by 
decreasing marginal cost. Once a writ petition is filed, additional petitioners can usually arrange to 
be parties at much lower cost. Lawyers typically reduce their fees when there are a number of writ 
petitioners and arrange to process their petitions on a mass basis. Also, junior lawyers will file writ 
petitions along the line of one already filed by a senior, for a very much 
____________________ 
Prime Minister herself later that year decided against such recognition on the ground that 
recognition of separate associations would "give rise to internal friction and would not only 
prevent their absorption in the mainstream but would also lead to unhealthy repercussions 
amongst other employees who might also start clamouring for recognition of other similar 
groups or associations" ( RCSCST, 1974­75: 81). 
121. Interview, New Delhi, Feb. 1978. Several cases in which there were complaints about use of false 
certificates in connection with government posts were pressed by Scheduled Caste employees' 
associations ( RCSCST 1974­75: 261). 
­509­ 
lower fee. Again, the publicity attendant upon such litigation tends to draw more petitioners 
in, and the success of one petition encourages others to try this mode of obtaining resources. 
For the propensity to litigate depends, too, on the expectation of success. Let us take the negative 
case. One of the striking things about the nationwide pattern of litigation is that after the 1951 
amendment of the Constitution there was no reported litigation about reservations in Madras until 
1967. Sizable reservations for the Backward Classes in both employment and professional­college 
admissions existed in Madras; the Backward Class lists were formulated on the "caste basis" and 
thus were arguably open to attack on the basis of the Balaji case. Yet no litigants came forward to 
challenge the Madras scheme. In part this may be due to the fact that the Madras reservations are 
more modest than those that formerly obtained in Mysore, and thus the "advanced communities" do 
have substantial opportunities to enjoy the resources involved. Lack of contest in regard to 
government employment might be attributed in part to the shift in ambition among Madras 
Brahmins away from government service toward the private sector. But this would not explain the 
absence of contention about reservations in professional and technical colleges, which are the prime 
preparation for advancement in non­governmental employment.  122.   It appears, then, that the 
communities not on the backward list regarded the reservations as good as it was possible to get 
under the circumstances. There was a feeling that litigation would not be effective to prevent some 
kind of preference for the "Backward Classes." As one lawyer put it, "[i]f they get rid of it in one 
way, the majority communities would spare no efforts to introduce it in another way" ­­ a remark 
that must be placed in the context of the existence of selection committees  123.   as well as of 
reservations. That is, it was felt that the courts could not provide ultimately effective relief This 
might explain the lack of organized backing, since no widespread and permanent benefit could be 
anticipated, but it does not explain the lack of individual litigants, since these might gain some 
tangible advantage even though the excluded communities did not secure any lasting benefit.  124.   
Litigation, it seems, requires not only parties who would benefit but an example or model of success 
and an incident to serve as a catalyst. 
____________________ 
122. This may be due in part to the growth of alternative private medical colleges in which a 
capitation fee might be paid. There were more of these in Mysore than in Madras, and even with 
this safety valve, there was far more litigation in Mysore. 
123. Cf. the discussion of selection committees in chap. 13, §A, above. 
124. Apparently, there had been a few admission writs in the preceding years and the Backward 
Classes argument had been raised but never decided (letter from G. Ramaswamy of the Madras  
bar, 30 Nov. 1968). 
­510­ 
That is, present discomforts, possible gains, and acceptable costs and risks of litigation must 
be made vivid and palpable.  125.   In Madras the model was provided by Balaji and the spate of 
similar litigation in other southern states. The catalyst was proved in 1967 by the peculiar 
circumstance that the marks list in the medical school competition had been destroyed. This 
was widely publicized, and as one of the lawyers described it, a large number of disappointed 
applicants, hoping that the entire selection might be struck down, filed petitions. The paucity 
of "affirmative" litigation by the beneficiaries of preferences may be explained in part by lack 
of resources. But it also reflects the absence of such "models of success" ­­ there has been 
virtually no successful affirmative litigation  126.   ­­ and of "catalytic events" to trigger the 
attempt.  127.   
Summing up, we may surmise that the gross effect of litigation on the compensatory discrimination 
policy has been to curtail and confine it. Those who have attacked compensatory discrimination 
schemes in court have compiled a remarkable record of success,  128.   while those seeking to extend 
compensatory discrimination have been less successful. One way of expressing this gross effect is to 
restate table 26 by comparing the level of preferential treatment before and after judicial 
intervention. 
Thus, we see that after litigation the level of preference was about the same as that previously 
propounded or administered by the government in 49 of the 80 cases; it was less in 27 cases and 
more in only 4 cases. This scoring probably understates the curtailing effect of the judiciary. Courts, 
reluctant to strike down government schemes, will uphold arrangements while warning of the 
necessity of correcting unconstitu­ 
____________________ 
125. In Mysore, the catalytic event was provided in 1958 by the issuance of a new Backward Classes 
Order after the state's 1956 reorganization. The new Order impinged on groups outside old 
Mysore who were not previously covered. In Andhra, the event was the conjunction of Balaji's 
case with the state government's expansion of preferences. In Kerala and elsewhere it was the 
decision in Balaji's case with its promise of successful challenge of existing policy. 
126. Raghuramulu v. State of Andhra Pradesh, A. I. R. 1958 A.P. 129, established the important 
principle that reservations could not be used to limit the admission of backward classes and has 
been much cited. Guntur Medical College v. Mohan Rao, A. I. R. 1976 S. C. 1904, established a 
group membership principle of deep theoretical import, but one that few others would be in a 
position to invoke. The other two successes ( Ashok v. Dean, 69 Bom. L. Reporter 603, and 
Susila v. State of Madras, A.I.R. 1970 Mad. 399) secured technical corrections in administration 
which benefitted the backward class applicants but did not significantly broaden the scope of the 
preferences or establish any general principle. 
127. The two affirmative cases with a broader thrust were triggered by "catalytic events" ­­ the 
withdrawal (in Dasa Rayudu v. Andhra Pradesh Public Service Commission, A. I. R. 196 7 A. P. 
353) or curtailment (in C. A. Rajendran v. Union of India, A. I. R. 1968 S. C. 507) of previously 
existing reservations. 
128. Seen. 112, above. 
­511­ 
tional practices, thus coaching the government to introduce changes favorable to the 
petitioners. 
Looking back over the pattern of litigation, we see that the central focus of the judicial encounter 
with the compensatory discrimination policy has been, not the general outline and performance of 
that policy, but rather the way in which government's attempts at redistributing certain highly valued 
scarce opportunities have impinged on the chances of other competitors with the resources and 
incentive to raise such challenges. The initiative has lain largely with these competitors. Thus the 
beneficiaries of compensatory discrimination have been at a disadvantage in the development of the 
law; the competitors have occupied the strategic heights in the litigation battle. The working out of 
principles is inevitably shaped by the contingencies of the factual situations which formed the 
context in which they are formulated. It is the most expansive schemes and the most flagrant abuses 
which have come to court. The first round of litigation (the 1951 cases) came up from the old 
Madras system of communal quotas rather than from a scheme specifically designed to advance the 
backward classes. Subsequently, it was the Mysore situation, which was little removed from 
communal quotas, that produced litigation and Supreme Court law. Debate on backward classes 
policy has been strongly affected by the fact that the crucial and determinative litigation has come 
out of Mysore, where reservations for Backward Classes were the most flagrant and expansive. Had 
some state with a more modest scheme of preferences been the source of the important litigation, 
the law might have taken a somewhat different shape. 
More generally, it should be noted that litigation (and public debate) has tended to center almost 
exclusively around schemes for Other Backward Classes rather than the more extensive schemes for 
Scheduled Castes and Scheduled Tribes. In only two areas has there been any significant litigation 
about the latter: problems of group membership and problems of extent, particularly in regard to 
promotions. That is, the litigation in regard to Scheduled Castes and Scheduled Tribes has 
concerned the boundaries of the preference policy. All of the other litigation has concerned 
provisions for the Backward Classes, and here the thrust of the litigation has been to question the 
basic policy of compensatory discrimination as well as its borderline applications. Thus the law of 
Articles 15(4) and 16(4) is being made largely in reference to situations involving the Backward 
Classes. The courts have erected restrictions on these articles to counter excessive and unwarranted 
preferences for the Backward Classes. But these redound to restrict what the government may do for 
the Scheduled Castes and Scheduled Tribes. Similarly, in public debate and scholarly com­ 
­512­ 
mentary, the well­publicized abuses of the Backward Class category easily attach to the 
resentment against the preferences for the Scheduled Castes ­­ a confusion compounded by the 
widespread inability of the public to distinguish these groups (i.e., to distinguish the inclusive 
and narrower senses of "[b/B]ackward [c/C]lasses". 
Thus Scheduled Castes and Tribes may be indirectly "misrepresented" by spokesmen for the Other 
Backward Classes. More generally, the configuration of parties in this litigation may fail to make it 
a forum in which all the relevant interests can be heard. Chitralekha provides an example. In the 
High Court proceeding, judge Hegde upheld the state scheme, while finding serious deficiencies in 
its failure to do enough for those who should have been beneficiaries. The appeal was taken by the 
petitioners, who wanted the state order struck down ­­ i.e., their criticisms were precisely in the 
opposite direction from Judge Hegde's. The state on other hand wanted to maintain its Order. So 
both parties were hostile to his observations about the deficiencies of the state's scheme: there was 
no one to press the interests of those he felt wrongly excluded.  129.   The absence of adversary control 
may be responsible for the conceptual muddle that emerged.  130.   Although this is an extreme case, it 
is not difficult to believe that there may be a similar tendency in cases where the contending parties 
are the state and non­backward objectors to the state's schemes for the backward. 
We have seen how the normal leeways of a case­law system are amplified and structural factors limit 
the ability of the courts to achieve coherence. In such a setting, what courts do depends in large 
measure on the kinds of claims that are brought to them and on the kind of arguments that lawyers 
make. 
D. LAWYERS 
Between courts and litigants there stand the lawyers; it is the lawyers who translate the grievances 
of litigants into the claims presented to courts. Notwithstanding recurring complaints that India's 
legal system is unsuitable to her conditions, courts and lawyers are a familiar and much­utilized part 
of Indian life from great public issues to village disputes.  131.   India has a numerous legal profession 
­­ both absolutely 
____________________ 
129
Havanur 1965: 119. 

130See chap. 7, §A, above. Similarly, delicate problems of group membership with subtle 
. implications for the integrity of beneficiary groups have arisen in election cases where there may 
be no effective participation by these groups and where argument may be dictated by 
considerations of expediency for the candidate. Meshram is a striking example. See chap. 9, §E, 
above. 
131. See Galanter 1972a. 

­513­ 
and relatively.  132.   A recent estimate places the number of lawyers at 228,000­­approximately 
336 for each million people, a ratio far larger than in other Third World countries.  133.   In spite 
of differences of region, language, caste, and religion, these lawyers share a common legal 
culture, which they put into the service of a wide variety of interests. As middlemen between 
official policy and local interests, they disseminate official norms, rephrase local concerns in 
acceptable legal garb, and devise new organizational forms for forwarding these interests. In 
modern India, where local concerns and interests increasingly find expression by 
representation at centers of power rather than by enjoying a sphere of relative autonomy, the 
lawyers are crucial agents for the expression of diverse local and parochial interests at the 
same time that they rephrase (and perhaps re­form) these in terms of official norms. Thus in 
cases involving preferences, it is the lawyers who rephrase the claims of caste and stratum into 
constitutional categories. 
Almost without exception, all of the lawyers involved in these cases, like most Indian lawyers, have 
been trained in India. None that I encountered had joined the bar in England, but some had 
postgraduate legal training or teaching experience, including a few who had some exposure to 
American legal education. The basic educational background is that of the Indian law college. 134. 
Except for some desultory apprenticeship requirements, formal legal training in India takes place 
within universities as a two­year course (recently become three­year) after the bachelor's degree. 
Admission to law colleges is more or less unrestricted. The law colleges are crowded with students 
who have neither a professional nor a scholarly interest in the subject; only a small proportion go on 
to practice. There are some imaginative and innovative teachers, but for the most part teaching 
proceeds by following a rigid syllabus and concentrates on the rote learning of rules. The syllabus is 
not geared to current problems, nor to legal subjects appropriate to an expanding welfare state or a 
developing economy, but is with rare exceptions confined to the standard subjects of half a century 
ago. 
Most law colleges, especially in the South, are part­time operations, staffed by practicing lawyers 
who serve as part­time teachers. There are some full­time law colleges, some of better quality. But 
except for some very recent innovations, these too are characterized by overcrowding, 
____________________ 
132The rough portrait of the legal profession here is drawn from personal observation combined 
. with Rowe, Bastedo, Morrison, and Schmitthener and the Conference proceedings reported in 
Galanter and Kidder 1968­69, and the papers by Morrison ( 1972), Khare ( 1972), Kidder 
( 1974). 
133
Personal communication from Dr. N.R. Madhava Menon, Bar Council of India Trust, March, 
. 1982. 
 
134
.
On legal education, see Bastedo 1968­69; Sunshine and Berney 1970. 
 
­514­ 
obsolete curriculum, rote learning, inadequate facilities, heavy teaching loads, and an 
atmosphere which is not congenial to creative research or scholarship. The full­time teachers 
in these colleges have little contact with the practicing bar. They are, with few exceptions, not 
held in high esteem by the profession, nor regarded as repositories of special expertise in their 
fields, nor, with the rarest exceptions, used as consultants or specialists in professional work. 
The lawyer learns his trade, not at the law college, but by associating with the "court community" at 
the courts. The profession is overcrowded and competition is stiff. Work is unevenly distributed: 
"leading lawyers" are extremely busy and hurried while a large number of aspirants have very little 
to occupy them. Building a successful practice usually requires sufficient resources to wait out the 
early, lean years. Many with legal training take to other tasks after an unsuccessful fling at the 
profession. The risk and insecurity of a career as practicing lawyer have combined with increasing 
opportunities elsewhere to make law less attractive than it once was to the ambitious. There has 
been a general decline in the prestige of the legal profession in recent years and a drop in the quality 
of talent that is drawn into the profession.  135.   The leading role of lawyers in resisting Emergency 
Rule during 1975­77 may have boosted their position in public regard, at least temporarily. 
The profession is a unified one, with a single set of qualifications for practicing in any court or 
engaging in any aspect of legal practice.  136.   Yet both the public and the profession concur in 
visualizing the lawyer primarily in the role of courtroom advocate, rather than business advisor or 
negotiator, much less social planner. Although there is no formal solicitor­barrister distinction 137. 
Most lawyers see themselves as barristers whose principal function is arguing in court rather than 
advising their clients. Typically, clients appear at the lawyer's door at a relatively late stage in the 
dispute, when they are already intent upon litigation. Ties with clients tend to be episodic and ad  
hoc, not enduring. Rarely do lawyers keep extensive records that would be helpful in later cases. 
The lawyers tend to be oriented to the courts, rather than to other legal settings. This orientation to 
courts is vividly displayed in spatial 
____________________ 
135On the decline in the prestige of the bar and the quality of entrants, see Schmitthener 1968­69 
. and Bastedo 1968­69; on the ambivalence of lawyers to their profession, see Kidder 1974. 
136

On the older distinctions, see All­India Bar Committee ( 1953); Schmitthener ( 1968­69). 
137. With the limited exception of the High Courts of Bombay and Calcutta where the solicitor­
barrister distinction is observed on the "Original Side" jurisdiction over major cases arising in 
those cities. 
­515­ 
terms ­­ lawyers are literally to be found at the court when it is in session. The office, rarely 
separate from the home, does not serve as a staging area for operating in various arenas. The 
lawyer's business is usually at a court. He will not move with his client's case from court to 
court, but will more often pass the matter on to a lawyer attached to the appropriate court. 
Typically, the lawyer spends his working life at a particular level of the system. But in spite of 
the stratification that this entails, the profession is relatively undifferentiated in character. The 
profusion of appeals, the tendency to have little narrowing­down of the issues in higher courts, 
and the latter's broad original jurisdiction leads to courts at all levels dealing with the same 
kind of problems. Although those at the higher levels enjoy more prestige, lawyers at each level 
do much the same sort of thing. Within each level lawyers are stratified by skill, influence, 
prestige, and wealth, but there is little division of labor by specialization (beyond civil­
criminal) and little coordination in the form of partnerships and firms (though established 
and successful lawyers may have one or more juniors, as well as clerical help). Basically, all 
lawyers offer the same kind of service under conditions of chronic oversupply. 
The emphasis on litigation and the courtroom role reinforces lawyers' rule­mindedness. Where the 
lawyer's task is to win disconnected battles, rather than to pattern relationships, there is little to 
induce the practicing lawyer to go beyond the kind of conceptualism that is characteristic of most 
Indian legal scholarship and that pervades legal education. In his court­centered world, the lawyer 
works in a realm of rules with little attention to either the social policy underlying them or the 
problems of implementing them. 
In preparing his cases, the lawyer has available a plethora of treatises, mostly mechanical 
compilations by professional hacks in the pay of legal publishers. There is a smaller body of work 
prepared by law teachers and practitioners, some of it of excellent quality, but almost invariably 
confined to scholarship of a traditional type, with heavy emphasis on close textual analysis and 
rather less consideration of either underlying policy or the actual working of legal regulations. 
Although expressed in an idiom of deference, and in spite of a widespread feeling that judges are 
sensitive to criticism, outspoken critical assessment of the courts is not uncommon. 
The law regarding compensatory discrimination is not seen to constitute a field or a subject in its 
own right. In addition to passages in the standard works on the Constitution, there is a small body of 
literature produced by academic lawyers (and a little by practitioners who have been involved in 
preference cases) but much of it is in publications seen by few members of the bar. Occasionally, 
lawyers in preference cases 
­516­ 
have ransacked government reports and standard ethnographic works for evidentiary points, 
but there has been no attempt to marshall the empirical data that might be relevant. For 
practitioners, and critics as well, the attempt to grapple with the basic principles and practical 
effects has been hampered by the emphasis on textual criticism and by the absence of any 
tradition of assessing law in empirical terms. 
All of the lawyers involved in cases concerning preferences are lawyers whose practices center 
around the High Courts. There are no specialists in litigation about preferences as such. Although 
most High Court lawyers do some writ petition work, there are some specialists in writ petitions, 
whose practice may be made up almost completely of cases of this type. Some of these writ 
specialists have done much of the litigation in the preference field. There are a few instances in 
which a lawyer has become identified as a champion of one or the other side of the preference issue 
in a series of cases, most notably one writ specialist who has become famous as a persistent and 
redoubtable opponent of reservations in Mysore. On the other hand, a somewhat larger number of 
eminent lawyers have appeared on both sides of these issues. Thus the lawyer who successfully 
obtained a writ striking down the Andhra reservations subsequently argued a number of writs 
petitioning for their restoration. 
Where, as commonly happens, there are a number of writ petitions considered together, leading 
lawyers may handle several or even a large number of such petitions. There is at least one instance 
of a lawyer handling so many admissions cases that he had cyclostyled copies of the petition into 
which he filled the names of students. When a petition challenging some arrangement which affects 
a large number of persons, such as a college admission, is filed by a senior lawyer, other lawyers 
often file similar or identical petitions, and these are consolidated for a hearing. At the hearing, the 
most "senior" lawyer ­­ that is, the one most esteemed at the local bar ­­ will ordinarily give the 
principal presentation of the arguments, and the other lawyers will add any additional points if they 
wish. 
Lawyer's fees are steeply graded according to their eminence at the bar. A leading lawyer at the 
Supreme Court may charge as much as Rs. 4,000 for handling a writ petition; his counterpart at the 
High Court in Karnataka may charge Rs. 1,500 to 2,000 (some High Courts are reputed to be more 
expensive, some less). The average medical­admission writ in Karnataka might involve lawyers' fees 
of Rs. 250. Where there are a number of identical writ petitions, fees are ordinarily reduced for the 
multiple petitioners. Junior lawyers, riding on the coattails of a senior lawyer's petition, may charge 
a little as Rs. 100 to their clients. 
In these cases, Government is ordinarily represented by its Advocate 
­517­ 
General or one of his assistants. These government lawyers, it should be noted, are not 
specialists in this kind of litigation. The administrative departments concerned do not have 
their own legal staffs; nor do the general legal offices of the government have any specialists 
with expertise in the administration or enforcement of these compensatory discrimination 
policies.  138.   In important cases, the government sometimes retains eminent members of the 
bar, especially for appeals. In some cases, too, private parties are joined to the government as 
respondents and appear with their own counsel; some arrangement is then made between the 
Advocate General and the private lawyers as to the division of labor in the argument. 
Most lawyers come from the higher castes ­­ as do most members of the educated classes. Therefore 
it is not surprising that most of the lawyers in these preference cases are of high caste origin. Many 
come from families with a long legal tradition; but several prominent in these cases are first­
generation lawyers. In a number of cases the caste origin of the lawyers corresponds with a personal 
or a communal interest of their clients. But in a larger number of cases there is no such 
correspondence. There are occasional complaints that untouchables sometimes have difficulty in 
engaging leading lawyers, but the untouchable side has been represented by leading members of the 
bar in several cases. It is often the case that lawyers on both side are Brahmins. 
In these preference cases, some lawyers have become engaged through communal or political or 
kinship networks. Sometimes such considerations play a part in selecting a lawyer who is not so 
connected ­­ a state may prefer an out­of­state lawyer in the Supreme Court in a touchy case or may 
engage an eminent Brahmin lawyer from the state to defend its preferences which are detrimental to 
Brahmins. On the whole, professional standing seems to play a larger role than any other factor in 
the selection of counsel. Only a few lawyers have made a "cause" of this kind of case. Some are 
quite partisan on these questions, but many are generous­minded, understanding, and capable of 
considerable detachment on these issues. On the whole there is much less vehemence and bitterness 
among the lawyers than there is among other participants. 
The lawyers' method of work places a great deal of emphasis on oral skills. There is little reliance 
on written arguments. No briefs are submitted; there is only a skeletal memorandum. Oral 
argument, however, is elaborate and often extended. Oral argument in Balaji's case 
____________________ 
138. This follows the general pattern of lack of specialization and initiative in enforcement tasks in 
government legal work. E.g., the Law Ministry of the central government has a drafting 
department and a legal advice department, but no staff specialising in the enforcement of laws in 
specific policy areas. 
­518­ 
took up two full days. The course of the development of the law on preferences has been 
determined in part by lawyers' considerations of strategy. This strategy has been in the main 
dictated by the considerations of advantage to the litigants rather than any attempt to 
establish abstract principles. Thus in Meshram's case, arguments were advanced which would 
have compromised the whole Buddhist movement, for the purpose of securing a client his seat 
in the legislative assembly.  139.   On the other hand, responsiveness to clients may sometimes 
encourage counsel to omit a stronger line of argument. For example, in Balaji, the state took a 
hard line and attempted to defend the total reservation although Delhi counsel had strongly 
advised conceding that the state's expanded Backward Class list (including Lingavats) was 
invalid in the hope that they could save the bulk of the list and the principle of reservation on 
a caste basis. 
Questions of reservations in services or admissions to educational institutions present some special 
tactical difficulties. Basically, the writ jurisdiction permits two kinds of challenges: first, one may 
assert that the individual's rights are violated and seek a remedy for that violation; second, one may 
attempt to strike down the whole government order. In some cases only the first kind of relief has 
been requested and not the second. Where the appointments or admissions of others would be 
adversely affected by the decree, it is necessary to join all of the parties who have benefitted by the 
order that is being challenged. This makes government employment cases even more difficult than 
admissions cases, and it is what accounts for the concentration of government employment litigation 
around posts like munsifs where a small number are involved. Such a case is much easier to handle 
than a case in which there is a very large number of persons who are affected. The alternative is to 
forego immediate relief for the petitioners, securing instead a favorable ruling which will apply to 
subsequent government action.  140.   Thus in P. Rajendran, the Supreme Court allowed the 
proceeding without joinder of the 1,100 admitted medical students, since the petitioners disclaimed 
any desire to disturb the current year's competition.  141.   
____________________ 
139. In another respect this may have been one of the rare instances in which the preoccupations of 
counsel shunted even the interests of the client into the background. See chap. 9, § E, above. 
140. Cf. Ramesh Chander Garg v. State of Punjab, A. I. R. 1966 Punj. 476, where petitioner was 
ordered to be admitted in the subsequent year. 
141. P. Rajendran, with its prospective relief, and Guntur Vedical College v. Mohan Rao, where 
respondent was assured of his place no matter what the outcome of the state's appeal, come close 
to being "test cases" in the American sense. But in each instance the case was initiated and 
pursued by an individual seeking (and achieving) a tangible and individual benefit. 
­519­ 
Timing may be a crucial element. Especially in the case of college admissions, there is a danger 
that if an appeal is necessary, the whole matter may be moot before there is a decision. 
Another tactical consideration is whether to fight the case in the High Court or to go directly 
to the Supreme Court. The latter course requires a much greater financial outlay and, usually, 
a greater degree of organizational support to provide it. Lawyers have gone to the Supreme 
Court directly when they anticipate a less favorable response from their High Court (as in 
Madras?) or find the High Court is unable to control the state government (as in Jammu and 
Kashmir). In Mysore, the High Court had given several decisions that would have led one to 
expect a favorable reponse to an attack on the government's reservation policy. However, 
counsel was convinced that a change in the bench that was admitting writ petitions lessened 
his chances of success, so he took Balaji directly to the Supreme Court. 
Lawyers have succeeded in translating into contestable claims the individual and group grievances 
of those discomfited by compensatory discrimination and in utilizing the writ proceeding to curtail 
and limit it. They have shown less evidence of creativity in bringing into court the inchoate claims 
of the lowly for enlargement of preferences or prompt, vigorous, and sympathetic implementation. 
The profession has provided an agency for addressing the sins of commission ­­ the "wrong thing" 
and "too much" criticisms ­­ but has had less to offer to those who are aggrieved by the sins of 
omission. 
Legal services in India are supplied on a fee­for­services basis, usually by single practitioners to 
individual clients already embroiled in a dispute. It is a system that emphasizes litigation rather than 
counselling, negotiating, or preventive lawyering. And even this limited range of services is 
available only to those who can pay for it. There has been virtually no "legal aid" in practice. The 
long­standing but unfilled commitment  142.   to provide legal aid to the poor was brought to center 
stage during the 1975­77 Emergency Rule: indeed the Forty­second Amendment introduced a new 
Directive Principle that the State shall 
____________________ 
142On the pre­history of legal aid, see Koppell 1966; Murphyet al. 1973: 281 ff., and the 1974 report 
. of the Expert Committee on Legal Aid ( Justice Krishna Iyer, chairman), devastatingly assessed 
by Baxi ( 1975). The most extensive governmentsponsored legal aid until now has been various 
state schemes for Scheduled Castes and Scheduled Tribes. The desultory if not illusory quality of 
this aid is suggested by the fact that eight states for which information is available provided Rs. 
136,700 for legal aid for Scheduled Castes (including Scheduled Tribes in one of the states) for 
1970­71. Since these states had a combined total Scheduled Caste population of 24.7 million, 
this works out to a provision of about one two­hundredth of a rupee per head. Nevertheless, only 
58% of the total provided was actually expended ( Expert Committee on Legal Aid 1974: 110). 
­520­ 
"provide free legal aid . . . to ensure that opportunities for securing justice are not denied to 
any citizen. . . ."  143.   A National Committee on Juridicare reported in 1978 and recommended 
a massive program of legal aid, including "public interest litigation and class actions" and in 
particular "activist," "militant legal aid programmes" for Scheduled Castes and Tribes. 144. 
Such activist programs, were they to come about, under government or other auspices, might 
stimulate lawyers to prosecute a whole range of claims on behalf of the ostensible beneficiaries 
of compensatory discrimination and put the courts in a very different posture toward these 
policies than they have stood previously. 
____________________ 
143. Art. 39a. 
144. Ministry of Law, Justice and Company Affairs 1978, § 12.01, § 11.06. 

­521­ 
15 The Uses of Judicial Action 
A. THE IMPACT ON POLICY AND ADMINISTRATION 

THE DECISIONS of courts do not automatically bring about the state of affairs that they prescribe. 
A growing body of research has indicated that many factors may influence the effect that a judicial 
1.
decision has on its environment.      Or perhaps we should say environments, since there are many 
environments and many effects­­on the parties in the case, on others in more or less similar 
situations, on interest groups of various kinds, on government administrators and policy­makers, on 
other courts, and finally on various sections of the public. The impact of a decision on any of these 
may vary independently of its impact on the others. 
There is usually not a great deal of difficulty in assuring relief to the individual petitioner in 
litigation about compensatory discrimination. Sometimes merely filing a writ petition suffices to 
obtain relief for him­­a petitioner challenging a government order may be given his admission or job 
and the petition withdrawn.    2. Such settlements between the government and the petitioner do not 

become part of the reported case law. However, they sometimes induce the government to revamp 
its practice in the future in order to avoid similar petitions. Or they may simply provide a way for the 
government to buy out a complainant and avoid the necessity of shifting its policy­­a tactic that can 
work only so long as the complaint is one that cannot readily be emulated by others. 
If the case goes to judgment, the court ordinarily enters an order in regard to the admission, 
promotion, etc., of the petitioners involved. Since the higher court enters its own order rather than 
remanding to lower courts, there is no problem of evasion or distortion of higher court 
____________________ 
1. The literature on the impact of the decisions of the United States Supreme Court is reviewed in 
Wasby 1970 ; Becker and Feeley 1973. 
2. Thus in Meera Bai (unreported, High Court of Madras, 20 Sept. 1967), a series of writ petitions 
was withdrawn upon the government's discovering that some additional seats were available and 
agreeing to reassess the qualifications of the petitioners. 
­522­ 
3.
decrees by lower courts.    
 The form of the remedy ordered varies: sometimes it is just to 
4.
reconsider the petitioner's case, at other times it is to admit him.    
 Sometimes the court goes 
beyond this and not only orders relief for the parties before it but strikes down the whole 
regulation. Since courts are reluctant to upset an entire selection when the successful 
competitors are not before it as parties, they have often resorted to "prospective overruling"­­
5.
the decree affects future competitions but the present selection is left undisturbed.    
 This has 
been extended to protect petitioners who were admitted consequent to securing the striking 
down of a state order by the High Court. When the Supreme Court reversed, restoring the 
6.
state order, the admissions pursuant to the High Court decision were left intact.    
 
In several instances the courts have gone beyond simply striking down an unconstitutional scheme, 
by decreeing remedies that actively guided the government toward the construction of a scheme that 
passes constitutional muster.     7. In Periakaruppan v. State of Tamil Nadu, the Supreme Court ordered 
that state to constitute a new selection committee of specified composition to complete the year's 
selection.  8.   In an elaboration of "prospective overruling," the Kerala High Court in Hariharan 
9. decreed that the 
Pillai, finding the state's scheme defective and in need of a fresh investigation,    
present scheme would not be 
____________________ 
3. Discussion of this phenomenon in the American setting is found in Murphy 1959 ; Harvard Law 
Review 1954. 
4. Thus the petitioner in Dilip Kumar v. Govt. of U.P., A. I. R. 1973 All. 592, 596, was ordered 
admitted to a medical college four years after he was wrongfully rejected. In Sukhvinder Kaur v. 
State, A. I. R. 1974 H. P. 35, 41, the court ordered the medical college to add a seat to 
accommodate the petitioner. 
5. P. Rajendran v. State of Madras, A. I. R. 1968 S. C. 1012. The prospective overruling aspects of 
Rajendran are discussed by Blackshield 1968: 200 ff The problem of devising relief that does not 
disturb admissions already made is surveyed in Surendrakumar v. State of Rajasthan, A. I. R. 
1969 Raj. 182. In the employment area, cf. Venkatramana v. State of Madras, A. I. R. 1951 S. C. 
229. And note the Court's effort to avoid striking down the selection in B. N. Tewari v. Union of 
India, A. I. R. 1965 S. C. 1430. 
6. State of Andhra Pradesh v. Balaram, A. I. R. 1972 S. C. 1375, 1400. Cf. Guntur Medical College 
v. Mohan Rao, A. I. R. 1976 S. C. 1904, where the state, appealing from an adverse decision 
below, guaranteed the candidate a seat whatever the outcome. 
7. Cf the Constitution's provision that the Supreme Court may "pass such decree or order as is 
 
necessary for doing complete justice in any cause or matter pending before it" (Art 142 [1]). 
8. Periakaruppan v. State of Tamil Nadu A. I. R. 1971 S. C. 2303 at 2311. See also Meera Bai v. 
 
Director of Medical Education (High Court of Judicature at Madras, 20 Sept. 1967), where as 
part of a settlement agreement the state agreed to constitute a "High Power Committee" to assess 
the claims of the petitioners and others on the waiting list in the light of principles embodied in 
Supreme Court decisions. 
9. Hariharan Pillai v. State of Kerala, A. I. R. 1968 Ker. 42 at 51. 
­523­ 
valid after a certain date. And in Triloki Nath, the Supreme Court directed the High Court to 
carry out an investigation and report back.  10.   
The government, which is the respondent in a writ petition proceeding, is ordinarily responsive to 
the order of the court. Almost invariably, it removes those specific features of its scheme which the 
court has found objectionable or carries out specific directions of the court. Thus in Kerala the 
government established an investigatory commission as ordered by the court and implemented the 
committee's recommendations. In Mysore, the government successively modified its Backward 
Classes order to conform to a succession of High Court judgments and then to the judgment of the 
Supreme Court in Balaji. However, compliance does not necessarily imply a shift in underlying 
policy. In Mysore relatively prosperous and politically dominant groups had been using the 
provisions for Backward Classes for their own advantage. Although the state in 1963 dropped its 
explicit provisions for these groups, it adopted two new practices which produced much the same 
result without violating the commands of the courts. First, they adopted the income/occupation test 
of backwardness, which gave considerable advantage to these groups. Second, they adopted a 
system of interview marks, which permitted a tincture of communal considerations to be introduced 
while ostensibly adhering to secular and non­ascriptive criteria. Thus policies favoring these groups 
became less visible, but it is not clear that they were less effective, or that the state adhered more 
closely to the Court's admonition that the state direct its aid to the "really backward."  11.   
There have been instances of sustained resistance, the clearest in Jammu and Kashmir.  12.   The 
Constitution's Fundamental Rights were made applicable to that state in 1954. In 1958 the High 
Court held "void and illegal" a cabinet order authorizing appointments "to remove the communal 
disparity," on the ground that this did not fall within the "reservations of posts" permitted by Article 
16.  13.   
Nevertheless, the state maintained its system of communal quotas in distributing government posts. 
 14   In 1966 a challenge of this arrange­ 
____________________ 
10. Triloki Nath Tiku v. State of J. & K., A.I.R. 1967 S.C. 1283. 
11. On the effects of the income test, see chap. 8, §G, above. Of course, the requirement that 
preferential treatment be confined to the "really backward" was relaxed if not abandoned in 
Chanchala v. State of Mysore, A. I. R. 1971 S. C. 1762. See discussion in chap. 12, §D, above. 
12. A less dramatic instance of state inaction is found in the wake of Nanda Kishore Sharma v. State 
of Bihar, A. I. R. 1965 Pat. 372. A commission to revise the state's list of backward classes was 
established seven years after the decision holding the list unconstitutional ( Blair 1980 ). 
13. State of J. & K. v. Jagar Nath, A. I. R. 1958 J. & K. 14. 
14 The conflict over medical admissions in Jammu and Kashmir took a somewhat different course. 
In early 1966 the High Court struck down a scheme of communal 
­524­ 
ment was taken directly to the Supreme Court by two teachers seeking to quash promotions to 
posts of headmasters. After laying down the requirements for reservations under Article 16(4), 
the Supreme Court directed the High Court to gather necessary material on the population 
and backwardness of the several communities, to ascertain the criteria being used by the state, 
and to submit a report within two months.  15.   Informed by the material gathered by the High 
Court, the Supreme Court in April 1968 struck down the state scheme as outside the scope of 
Article 16(4): 
In effect the State policy . . . was a policy not of reservations of some appointments or 
posts; it was a scheme of distribution of all the posts community­wise.  16.   
The state was invited to devise a scheme for reservations for Backward Classes as allowed by 
Article 16(4). But, as the Court later recounted, "no such scheme as was envisaged was ever 
devised."  17.   Instead, the state officials "thought of an ingenious device of giving ostensible effect to 
the decision of this court. Those teachers . . . whose promotions became illegal in view of the 
pronouncement of . . . [the Supreme C]ourt . . . were 'allowed to work against the posts they were 
holding prior to their reversion' on temporary basis. . . . A new nomenclature was evolved for the 
post of Head Masters. They were called Teachersin­charge. . . . [Each of the reverted teachers thus 
retained] his emoluments and his posting as Head of an institution. . . ."  18.   
Other teachers who had not been parties to the earlier petition "were all promoted to the gazetted 
cadre . . .subsequent to the previous decision in complete defiance of the law laid down by this 
Court."  19.   The Supreme Court dismissed as "wholly untenable and misconceived" the contention 
that those who were not parties to the earlier decision were not governed by that decision. 
The Judgment which was delivered did not merely declare the promotions granted to the 
respondents in the writ petition filed at the previous stage as unconstitutional but also 
laid down in clear and unequivocal terms that the 
____________________ 
quotas ( Lalita Shuri Tikku v. State of J. & K., A. I. R. 1966 J. & K. 101). The state instituted a 
system of selections by a selection committee with broad discretion to give marks for the 
interview, but the selection was set aside because the committee's failure to take academic merit 
into consideration was "patently unreasonable, arbitrary and subjective" ( Jalali v. Principal, A. 
I. R. 1967 J. & K. 106). The operation of the selection committee was scrutinized by the High 
Court again in 1969 and this time it passed muster ( Sardool Singh v. Principal, A. I. R. 1970. J. 
& K. 45). 
15. Trilok Nath Tiku v. State of J. & K. [I], A.I.R. 1967 S.C. 1283. 
16. Triloki Nath Tiku v. State of Jammu and Kashmir [II], A. I. R. 1969 S. C. 1, at 4. 
17. Makhan Lal Waza v. State of J. & K., A. I. R. 1971 S. C. 2206, at 2208. 
18. Id., at 2208­9. 
19. Id. at 2209. 
­525­ 
distribution of appointments, posts or promotions made in implementation of the communal 
policy was contrary to the constitutional guarantee of Article 16. The law so declared by this 
court was binding on the respondent State and its officers and they were bound to follow it 
whether a majority of the present respondents were parties or not to the previous petition.  20.   
Consequently, all of the promotions made pursuant to the state's communal policy were "illegal and 
unconstitutional" and "would have to be revised and reconsidered and appropriate orders must be 
passed by" the state.21
Almost a year after the first Triloki Nath decision, the state set up a one­man Commission of 
Inquiry. composed of P. B. Gajendragadkar, the author of the Balaji judgment, who had retired as 
Chief Justice of India in 1966. The second Triloki Nath decision was handed down while the 
Commission was sitting. In its report, submitted in November 1969, the Commission recommended 
employment of multiple criteria of backwardness and urged that the state appoint a high­powered 
committee to draw up a list of Backward Classes. Accordingly, a Backward Classes Committee 
chaired by the retired chief justice of the High Court was appointed in February 1969 and reported 
in November of that year. While the committee was sitting, the state had essayed the evasive 
maneuver in the case of the headmasters that came before the Supreme Court in Waza. By the time 
this was struck down in February 1971 the state had issued new reservations rules and had used an 
elaborate interview procedure to fill the now vacated headmaster's posts­arriving at appointments 
bearing marked resemblance to the previous selection.  22.   These were challenged in Janki Prasad, 
which upheld the designation of Backward Classes, with minor modifications, but found the 
interview procedure entirely unsatisfactory, "a travesty of selection," and set it aside as 
unconstitutional.  23.   The state was ordered to undertake another selection in which work records as 
well as interviews would be taken into account. 
It was seven years since the teachers had first filed their petition and fifteen years after the 
"communal quota" was struck down in Jagar Nath. But it is clear that Jammu and Kashmir had 
moved, slowly and grudgingly, toward alignment with the national norms on reservations. 
____________________ 
20. Id., at 2209. 
21. Id., at 2210. In June 1971 the Prime Minister encountered a deputation protesting "the reversion 
of 350 odd teachers and others following a Supreme Court judgment" ( Hindu, 30 June 1971, p. 
8). 

22. "[W]hen previously the appointments were made on the communal basis, 178 posts had gone to 
Muslims and 134 to Jammu Hindus. Now after selection, 177 posts go to Muslims and 134 to 
Hindus" ( Janki Prasad v. State of J. & K., A.I.R. 1973 S.C. 930 at 934). 
23. Id., at 936. 

­526­ 
Communal quotas couldn't be arranged openly, and it was increasingly difficult to arrive 
there by tacit means. 
But the courts do not always have the last word. The Supreme Court's decision in Dora, permitting 
both seats in a double­member constituency to be captured by contenders for the reserved seats, 
"administered the major shock" leading to Parliament's abolition of the double­member 
constituencies.  24.   And there was, of course, the early and dramatic instance of a constitutional 
amendment to reverse a judicial decision: Article 15(4) was added to counter the Dorairajan 
decision and legitimate preferential treatment outside the government employment area.  25.   But even 
there the status quo ante was not restored; the post­amendment State policies were more constricted 
than those that prevailed earlier. 
On the other hand, there have been instances in which a state used a court decision as the occasion 
for a striking change in policy that went far beyond the order of the court. Thus in Andhra when the 
High Court struck down the list of Backward Classes for educational reservations the state 
government used the occasion to abolish its reservations in admissions to educational institutions 
and in government employment as well.  26.   
In many cases, litigation transforms a settled policy into an open issue. Thus in Mysore, Andhra, 
and Kerala, the most striking result of litigation was the formation of commissions to formulate a 
new state policy in regard to reservations and a period of intense political concern with this issue. It 
should be noted that litigation had this effect in Kerala, in spite of judicial upholding of the existing 
reservations. 
The results of litigation may be felt in other states as well as in the place of its origin. Although 
government officials who deal with these matters are often uninformed or misinformed about what 
the courts have done, there are some who are more aware and who scrutinize judgments for 
guidance­­usually with an eye to maintaining existing policy and avoiding unconstitutional features. 
The effects seem more pronounced on the side of the prospective petitioners: in several instances a 
judgment concerning another state served to catalyze local discontent into litigation against existing 
preferences. Thus, for example, Balaji stimulated petitions in Andhra and in Kerala. But the delay 
____________________ 
24. Dushkin 1972: 192. 
25. See the debates over this amendment, discussed in chap. 6, §B, and chap. 11, §A, above. To some 
extent the Dora case contributed to the abolition of double­member constituencies (see chap. 3, 
§A, and chap. 13, §B, above), although this is an instance of avoiding the result, rather than over­
turning it. 
26. Discussed in chap. 7, §A, above. Note that the state later reinstated reservations along the lines of 
those abandoned. 
­527­ 
in Madras suggests that such models may be insufficient to overcome features of the local 
scene that militate against litigation.  27.   
Indirectly, litigation may alter the constellation of groups which support or oppose a preference 
policy. The striking down of a scheme may disarm groups that enjoyed a reservation. Since these 
groups sometimes have little to unify them other than the reservation itself, the loss of the 
reservation may dissolve them into a series of disparate elements with conflicting views on what the 
new scheme ought to be. On the other hand, the expectation of success in litigation may bring 
together groups to oppose the existing scheme. A new policy of reservations may reflect (or 
generate) a new convergence of interests. Consider, for example, the situation in Mysore where, 
before Balaji, the thrust for preference for Other Backward Classes came from the populous and 
politically powerful middle castes, acting in conjunction with (or partially on behalf of) the great 
multitude of less articulate and less politically forceful communities, mostly quite poor. It was 
mainly the Brahmin community which was the target, and it was mainly Brahmins who opposed the 
policy through litigation. The post­Balaji shift to the economic test tacitly shifted the alignment of 
beneficiaries.  28.   Not only the politically dominant middle castes but Brahmins were now major 
beneficiaries of this policy; the former had managed in effect to co­opt their most determined and 
resourceful opponents and shift the burden (here, of exclusion from medical colleges, etc.) to those 
least likely to mount a campaign against it. And, of course, the new standard had the added virtue of 
satisfying courts, publicists, and the nationwide educated audience that it was secular and non­
communal. (Here, as in other instances, the flow of tangible benefits shifts from the have­nots to the 
haves, but under the banner of secularism and progress.)  29.   Eventually, political mobilization of 
lower groups and the resentment of some advanced groups against the extensive benefits enjoyed by 
Lingayats combined with a campaign of intellectual criticism to undermine this system. 
We have been speaking in terms of judicial initiative and government response, but Balaji should 
remind us that the relation of courts and government may be more complex. There the Supreme 
Court not only acted as critic of the state government's policy; it also acted as an agent 
disseminating the central government's policy. The state governments had continued to ignore the 
Centre's recommendations as to the makeup of the Backward Classes. Private litigants in effect used 
the Court to 
____________________ 
27. See discussion in chap. 14, §C, above. 
28. See chap. 8, §G, above, especially table 21. 
29. Generally, on the relation between political symbolism and the distribution of benefits, see 
Edelman 1967. 
­528­ 
force the state to adopt a policy along the lines of that approved by the central government. 
The Supreme Court served as a conduit for central policy. 
The flow of influence in this case is somewhat obscured. Later, central government spokesmen 
attributed the government's position to the influence of the Supreme Court: 
following the judgment of the Supreme Court [in Balaji] the Central Government had 
accepted economic factors and not caste as the basis for determining the backwardness 
of people.  30.   
Partisans of the new policy concurred in this view: thus the Times of India editorialized that "the 
Government has not been slow in acting on this admirable judgment."  31.   There are two significant 
transformations here. First, the central government had taken a firm and explicit position against the 
use of caste units in enumerating the backward classes at least as early as a year before the case was 
decided.  32.   Second, the Balaji judgment, as we have seen, did not rule out the use of caste and 
communal units; instead it invalidated caste status as an exclusive test of their backwardness while 
indicating a strong preference for purely economic tests. It was not the technical holding of the 
judgment that served to reinforce central policy, but the conflation of the holding with the Court's 
policy preferences, flowing from the Court's failure to disentangle the dual meanings of caste. The 
government was thus provided with rhetorical ammunition in its previously launched campaign 
against the use of caste units. It was relieved of having to debate on its merits the question of using 
caste units, for it could now claim that it was constrained by the Supreme Court judgment (and the 
Constitution itself). This result was perhaps not unwelcomed or uninvited by the Court in view of its 
warm commendation of economic tests applied directly to individuals. 
The Court provided ammunition and reinforcement not only to government but also to other 
opponents of caste units: intellectuals and newspaper editorialists joined some government officials 
in embracing an undifferentiated and oversimplified version of Balaji as interdicting the use of caste 
units in conferring preferences. In part, this is due to imperfect transmission, but it also involves an 
element of perceptual distortion to fit the decision into the pre­existing preferences of these groups. 
 33.   
____________________ 
30. R. M. Hajarnavis, LSD (3rd Series), Vol. 48, No. 16, cols. 3973­76 ( 25 Nov. 1965). 
31. Times of India, 30 Nov. 1965. 
 
32. See chap. 6, §C above. 
33. See chap. 7, above. Cf. the discussion of the perceptual distortion attending the United States 
Supreme Court's school prayer decision in Dolbeare and Hammond 
­529­ 
Thus the Supreme Court not only reinforced central government policy and served as a 
conduit for disseminating and legitimating it: it also served as an instrument for bringing the 
states into conformity with it. Earlier the Centre had urged its powerlessness to change the 
state lists of Backward Classes. Now it claimed to have adopted "the path of persuasion" and 
in 1965 it claimed that eight states complied with the centrally­favored economic criteria. But 
inspection of the government's list of compliant states reveals that the impact was less than it 
claimed.  34.   The list included two states ( Gujarat and Maharashtra) that had retained caste 
units outside the scholarship area; one state ( Punjab) that not only retained the caste test 
outside the scholarship area, but employed a list of communities in the scholarship area as 
well, albeit in conjunction with an income test; one state ( Orissa) that had no schemes of its 
own for Other Backward Classes, and two states ( Assam and West Bengal) that had none 
other than scholarships. Only two states with a high level of benefits to OBC adopted non­
communal tests: Mysore and Andhra Pradesh. The latter reverted to caste units less than a 
year later­­as Mysore did more than a decade later. 
States with low levels of benefits for Other Backward Classes tended to go along with the Centre's 
shift to economic tests. These tended to be the states where there had been no litigation. In the states 
with higher levels of benefits for OBC­­and these were the states where there had been litigation­­
the effect was not to stimulate a changeover to an economic basis but to stimulate the pruning and 
rationalization of the community­wise list. For the most part, central pronouncements against 
communal units were given effect where the evils deplored were least present and the quieter 
qualifications permitting communal units with safeguards were given effect where extensive 
benefits to communal groups were at issue. (The exception of course was Mysore, where the 
economic test remained for more than a decade.) 
The impact of court decisions, as of other governmental policies, is limited by decentralization. 
Even where, as in Balaji, the Supreme Court supports central policy, there is little the Centre can do 
administratively to bring around recalcitrant states­­especially in areas where it has neither 
legislative power to exercise  35.   nor financial inducements 
____________________ 
( 1971), where the decision was perceived so as to strain out unwelcome elements and make it 
compatible with prevailing practices. (There, as here, there was also misperception of prevailing 
practices.) In Balaji, the process works to magnify favored aspects of the decision. Actually, the 
newspaper version closely anticipated Chitralekha, but no notice was taken when that case 
effectively dropped out of sight. 
34. See chap. 6 at n. 124, above. 
35. "Education, including technical education, medical education and universities. . . ." was a state 
subject until 1976 when it was moved from the state list to the 
­530­ 
to offer.  36.   (Of course, if the state ministry is politically dependent on the central government, 
there may be political inducements ­­ but these may run in the other direction as well.) There 
are no central enforcement agencies for carrying out such policy. The function of 
implementing and enforcing central policies is left to the states. So while the system is very 
centralized in declarations of policy, including judicial policy, it is very decentralized in 
implementing these declarations. The problem of implementation is compounded by the lack 
of reliable and systematic data about what is going on. The Centre has no 
informationgathering apparatus of its own but obtains all of its information about the 
implementation of policy from the states, usually from those very officials in charge of the 
implementation. Provisions for the Backward Classes is an area in which the states have not 
been very forthcoming with information. 
Thus the effect of a court decision is not necessarily as widespread or as automatic as we might tend 
to assume from a formal chart of Indian government. The effect of a decision may be as much 
through cultural and political channels as through structures of command and administration, and 
there may be a dissociation of symbolic changes from administrative ones.  37.   Even within the 
hierarchy of courts, the effect of a decision on the decisions in subsequent cases  38.   is not so 
automatic as our discussion of the doctrine of precedent in India might suggest. 
____________________ 
concurrent list by the 42nd Amendment. It was subsequently restored to the state list by the 
Constitution ( 43rd Amendment) Act, 1977. 
36. The Centre had applied its new views in the administration of the postmatriculation scholarship 
scheme. This is not an area in which the Centre felt strongly enough to provide other 
inducements or sanctions. For a criticism of this, see Ghurye 1969: 428H. 
37. A consideration of the layers of symbolic transformation and distortion in this instance will 
surely induce modesty in attempting any estimate of effects. The Balaji court 
misperceived/misrepresented the policy of the Maharashtra government; the central government 
misperceived/misrepresented the holding of Balaji; the press and commentators 
misperceived/misrepresented the relation between the judgment and the central government's 
position; the central government misperceived/misrepresented the state government response to 
central urging, etc., etc. We have here a symbolic reality, a tissue of illusions built on illusions. 
This kind of dissociation of symbolic from administrative reality is not necessarily typical of 
policy­making in India, but I suspect it is not unique, especially in carrying out policies which 
are redistributive to low­status, inarticulate, and unorganized beneficiaries. 
38. The interpretation of the constitutional provisions regarding compensatory discrimination forms, 
for the most part, a rather isolated, self­contained area of Indian constitutional law. Decisions 
here do not have wide ramifications in other areas. A notable exception is the Rangachari case, 
which in the course of bringing promotions under Article 16 (4) firmly established that the 
equality guaranteed by Article 16(1) and (2) applied to promotions, a point which gave rise to 
much litigation by government servants. 
­531­ 
Outright evasion of prior judgments is not so formidable a problem in India as it has 
sometimes been in the United States. The system is more centralized, and the checks, formal 
and informal, against innovation and initiative by lower courts are stronger. (And this is 
especially true in compensatory discrimination litigation, where the matter is entirely within 
the higher judiciary.)  39.   Yet the Indian judiciary should not be viewed as a machine for the 
automatic and relentless application of previously established rules. 
The body of doctrinal learning propounded by the judiciary is constantly exposed to two opposing 
kinds of erosion. On the one hand, a subsequent court may screen out subtleties and minor themes 
and vulgarize a delicate calculus into a rigid and mechanical black­letter rule. On the other hand, 
receptivity to the subtleties and minor themes of a complex judgment may succeed in dissolving 
away the thrust of the judgment, a process often aided by the normative overload of the original 
decision. Balaji again provides an example of a judgment that has undergone both kinds of erosion: 
formalization (rigidification) and deformalization (diffusion). Thus Balaji's holding about the 
permissible extent of reservations (which we might paraphrase as "reasonable under the 
circumstances, but never more than 50%") was transformed by subsequent courts into a flat 50% 
rule.  40.   On the other hand, Balaji's teaching about caste was so overloaded ­­ (1) disapproving caste 
units while (2) permitting them but (3) striking down caste standing as an exclusive test ­­ that 
subsequent courts, administrators, and commentators could (and did) find support there for almost 
any position. 
The dangers of distortion are greater if we consider that judicial doctrines in this area have to be 
applied not only by courts, but by a host of administrators and clerks with less education and less 
legal sophistication than judges. The more subtle and pragmatic (in the sense of taking into account 
all the complexities of the given case) the rules that the courts lay down, the greater the distortion 
that will attend their application ­­ as routine administration by harried functionaries generates 
pressures toward either formalism (rigid rules) or broad dis­ 
____________________ 
39. Transmission of rules within the system is attended by friction and slippage. Thus High Court 
judges may misunderstand the Supreme Court's pronouncements; or they may believe that with 
their better information about local conditions, they can make the Supreme Court's rule more 
suitable to those conditions. See, e.g., Viswanath v. Government of Mysore, A. I. R. 1964 Mys. 
132, as an emendation on Balaji. 
40. This may illustrate Friedman observation ( 1967: 815) that in recurrent crisis situations courts are 
likely "to develop a rule that can be delegated to other authorities for administration." The courts 
have gravitated toward selecting those elements in the Balaji rule that can be quantitatively 
expressed, relieving themselves of imponderable judgments. 
­532­ 
cretion.  41.   Thus the courts are in something of a dilemma. The limits of their own energy and 
resources require that they make rules that can be administered by others. Yet the more 
sophisticated, subtle, and pragmatic the rules they propound, the less translatable into routine 
administration and the more likely to invite further recourse to the courts they become. 42. 
One solution would be to have more sophisticated and sensitive administrators and clerks, 
with greater resources for investigation and more time for deliberation.  43.   But this entails 
costs of its own. And the diversion of resources to such a reform lies outside the scope of 
judicial remedies. 
B. JUDICIAL REVIEW: THE DOCTRINAL ACCOMPLISHMENT 
The Constitution confronts both government and courts with the problem of reconciling the 
conflicting principles of equal treatment and compensatory discrimination. The sweeping language 
of Articles 15(4) and 16(4) suggests that the framers of the Constitution relied primarily on the 
discretion of the politicians and administrators of the future, rather than on the courts, to effect such 
a reconciliation. But while these provisions give the executive and the legislatures broad discretion 
in their application, judicial review is not entirely excluded. The initial leverage for judicial 
intervention was the placement of these provisions as exceptions to the judicially enforceable 
Fundamental Rights. These rights can only be vindicated to the extent that the courts undertake to 
see that the government has used its powers within the limit of the constitutional authorization. 
Thus the courts may scrutinize the government's designation of backward classes to see that the 
beneficiaries are indeed the backward classes. Similarly, the courts may examine the government's 
schemes to see that they work in favor of the intended beneficiaries and not to their detriment and 
that the extent or method of operation of the schemes does not unduly impair the rights of others. 
Such review would seem essential not only to vindicate the rights of the non­backward but to 
effectuate the policy of these provisions. 
____________________ 
41. Even though discretion implies increased power, those who exercise it may find it too costly and 
may prefer to divest themselves of it. Thus it was reported that in Mysore, the practice of scoring 
the interview was dropped by the interview committee. See chap. 13, n. 52, above. 
42. Friedman 1967: 818. 
43. This dilemma may characterize all judicial control over re­distributive schemes that have to be 
administered on a day­to­day basis by an army of clerks. On the limits of judicial capacity to 
control the administration of complex welfare programs in the United States, see Handler 1966. 
­533­ 
The main lines of policy­formation ­­ the use, extent, method, recipients, duration, and 
administration of preferences ­­ lie with the government. But courts may still serve to keep 
their use consistent with constitutional boundaries. If the most important determinants of the 
success of compensatory discrimination lie beyond the power of the judiciary, there is still 
much that the judiciary can contribute to their working. In our survey of the law as developed 
in the courts, we encountered a number of substantial accomplishments: 
1.  The courts have decisively prevented the use of these provisions as justification for a regime of 
communal quotas. 
2.  The courts have inhibited runaway expansion of the category of Other Backward Classes. (In 
restraining expansion they have protected not only the non­backward but also the intended 
beneficiaries of these policies, preventing diversion and dilution of benefits.) 
3.  The courts have firmly established that gross over­inclusion will not be tolerated; the Backward 
Classes must be confined to those who are demonstratively backward by some tangible criteria; 
that the application of these criteria must be supported by adequate data, recent enough to be 
relied upon as reflective of current conditions. These requirements have not only 
institutionalized periodic reassessment and encouraged collection of information which may 
facilitate better administration, but they have clearly contributed to making discourse on the 
subject of Backward Classes more precise and cogent.  44.   
4.  The courts have encouraged flexibility and scope in compensatory discrimination policy by 
helping to free the government from the notion that the beneficiaries must be selected by some 
single pre­ordained criterion; by pointing out a variety of ways in which the backward might be 
selected; by dispelling the notion that the same data and standards could be used indefinitely. 
5.  By refusing to accept reservations as compartments or maximum quotas, and by refusing to 
confine special treatment to quantitative representation, the courts have insured that special 
provisions are not used to curtail the opportunities of the beneficiaries. 
6.  The courts have succeeded in confining the quantum of preferences, thus assuring that other 
important interests will not be ignored. 
In developing these responses, the courts have proceeded sometimes by narrow technical 
construction of the constitutional provisions, sometimes by undertaking to construe these provisions 
with an eye to the 
____________________ 
44. As is evident from a comparison of the more recent commission reports­Government of Gujarat 
1976; Government of Karnataka 1975; Andhra Pradesh Backward Classes Commission 1970; 
Kerala ( Kumara Pillai Commission) 1966­­with e.g., the reports of the Mysore Backward 
Classes Commission 1961 or the [Central] Backward Classes Commission 1955. 
­534­ 
various broader social interests affected by preference policies. They have, in effect, read the 
words "reasonable" into Articles 15(4) and 16 (4) ­­ an approach that permits them to weigh 
the issues in terms of the broad lines of policy set forth in the Constitution. 
The courts have served policy­makers and the nation well by keeping open options in pursuing its 
commitment to compensatory discrimination. They have rejected both the "modernist" view that 
communal units are impermissible and the "historical" view that only communal units may be used 
to designate Backward Classes, in favor of the "elastic" view that enables the government to utilize a 
variety of ways of selecting Backward Classs.  45.   Again, the courts have refused to reduce 
compensatory discrimination to compassionate adjustments appended to a regime of formal 
equality. And they have for the most part held at a distance the view that the Constitution provides a 
blueprint for a regime of substantive equality and mandates the general use of compensatory 
classification for equalizing purposes.  46.   By refusing to accept a single view of either the goals or 
the means of compensatory discrimination, the courts have been loyal to the original ambivalence 
and have kept open options for the future. 
Having provided both critical standards and flexibility, the judicial achievement is marked by a 
number of conceptual problems which could well impede the effectuation of the compensatory 
discrimination policy. 
First, there is the abiding confusion surrounding the use of (1) castes as the units or classes that are 
deemed backward, and (2) the use of caste rank or standing as a measure of backwardness. 47. 
Notwithstanding their general disapproval of caste, the courts have given little guidance to the states 
as to the limitations on the use of communal units in selecting Backward Classes. They have failed 
to discuss the kind of units that can be used or the kind of tests that can be used to select them. 
While they have put government on notice that caste standing is not by itself an acceptable test of 
backwardness, they have left unclear just how it may be measured and just what role it may play in 
the determination of backwardness. The failure of the courts to make clear the distinction (between 
castes as units and castes as ranks) has made it difficult to solve either of these problems. It has also 
led to considerable confusion and, in particular, it has obscured and diffused the original 
commitment to overcome the heritage of caste distinctions. 
Second, while encouraging movement away from exclusive reliance on communal factors, the 
courts have portrayed Indian society as 
____________________ 
45. See chap. 7, §B, above. 
46. Cf. chap. 11, §D, and chap. 12, §D, above. 
47. See chap. 7, §A, above. 

­535­ 
comprising closed and mutually exclusive compartments, arranged in an unambiguous rank 
order: for example, by formal construction of the schedules, by permitting religious tests for 
membership in Scheduled Castes, and by intruding notions of caste rank into the 
determination of tribal status.  48.   The courts not only have allowed religious discrimination 
and community ranking to be obtruded into the law of preferences, but have given currency to 
views of Indian society that are neither empirically persuasive nor congenial to the principles 
of the Constitution. 
Third, while promoting non­communal tests of backwardness, the courts have not indicated how 
these can be assessed by the high standards they have used to criticize communal tests.  49.   They 
have failed to reconcile economic tests with the doctrines set up for judging the acceptability of 
communal units. Nor have they shown how economic tests may be directed to the purpose of 
overcoming accumulated inequalities. 
Fourth, although imposing some quantitative limits on preferences, the courts have not penetrated 
beyond the superficial question of the specified amount of preference to address the real quantitative 
question of the net effects of preferences ­­ and the way in which this is entwined with the question 
of allowable methods of administering reservations.  50.   
Fifth, in their concern with the abuses and excesses of some schemes of preference, the courts have 
come close to propagating the notion that the government must designate some single group of 
"backward" classes who must be treated uniformly. This tendency has appeared in the judicial 
insistence that backward classes be both socially and educationally backward.  51.   It has appeared 
again in the judicial hostility to compartmental reservations and "layers" of preference.  52.   In spite 
of their recognition that backwardness comes in kinds and degrees, the courts appear to be imposing 
on the government limitations which not only have little constitutional warrant, but could hamper 
the flexibility of governmental schemes for the backward. The State should not be confined to 
addressing its efforts to an undifferentiated group called "backward classes" but should be free to 
select backward classes by criteria appropriate to the particular scheme. Such judicial restrictions 
would make difficult a transition to more differentiated criteria of backwardness. Again, the courts 
should not foreclose efforts to see that the smaller and most underprivileged groups are not 
overlooked while 
____________________ 
48. See chap. 9, §§ A, D, E, above. 
49. See chap. 8, §G, above. 
50. See chap. 12, §B, and chap. 13, §B, above. 
51. See chap. 8, §C, above. 
52. See chap. 13, §C, above. 

­536­ 
the lion's share of the benefits go to the most advanced and assertive sections of the backward. 
Sixth, while indicating some of the interests that have to be balanced against the thrust of 
compensatory preference, the courts have not developed any standards for assessing the weightiness 
of such interests.  53.   They have, at times, slipped into a facile equation of the national interest with 
that of the advanced classes or of the non­backward applicants and students before them.  54.   Of 
course, such assessments are not susceptible of exact quantitative statement but rest upon judgments 
that proceed from value commitments as well as from estimates of fact. But the coherence and 
consistency of such judgments may be increased, not only by more incisive analysis, but by the 
presence of data that is relevant, comprehensive, and of high quality. Courts have required that in 
designating Backward Classes, government act on the basis of adequate data rather than mere 
assertion or casual opinion.  55.   No comparable standards have yet developed for the kind of factual 
basis that must serve as the foundation for assertions about efficiency, morale, effective use of 
talent, etc. 
Seventh, while developing justifications for compensatory treatment as consistent with the 
Constitution's equality provision, the courts have left unclear the relation between compensatory 
discrimination for historically disadvantaged sections of society and the measures to remedy various 
kinds of personal and circumstantial deprivation.  56.   By diffusing the special authorization, they 
have obscured the commitment to overcome the heritage of historical differences in a hierarchic 
society. Nor have they convincingly assessed the difference between the policy of a favorable or 
equalizing tilt in favor of the Other Backward Classes and the commitment to intensive forced draft 
measures to promote the inclusion of the Scheduled Castes and Scheduled Tribes. 
Finally and perhaps foremost, there is a problem not of questionable doctrine but of the absence of 
doctrine. The courts have not equipped themselves with any doctrine by which they might reach the 
"affirmative" problems of compensatory discrimination policy ­­ that is, by which they might assure 
that the deserving are included among the beneficiaries, that preferences are of sufficient scope and 
amount, that they are implemented in a timely and effective fashion, etc. In part this flows from the 
basic constitutional posture ­­ Fundamental Rights are justiciable, Directive Principles are not ­­ 
which casts the courts as a 
____________________ 
53. Thus, e.g., in C. A. Rajendran v. Union of India A. I. R. 1968 S. C. 507, 514, the Court simply 
accepts the government's satisfaction as sufficient evidence that efficiency would be adversely 
affected by extending promotion reservations to Classes I and II. 
54. See, e.g., the discussion of Devadasan in chap. 12, §B, above. 
55. See chap. 8, §F, above. 
56. See chap. 11, §D, and chap. 12, above. 
­537­ 
brake and a baffle, rather than as stimulant and energizer of the compensatory discrimination 
policy. 
The decline of the older notion that Directive Principles were subordinate to Fundamental Rights 
has not really changed that posture. Although the Directive Principles have been elevated in 
constitutional dignity,  57.   and the powers of the legislature have thereby been amplified, the 
Directive Principles still lack self­executing legal force. They do not create any justiciable rights in 
favor of any party,  58.   nor in current view are courts competent to compel the government to carry 
out any of these directives.  59.   Thus the Directive Principles do not supply the courts with a lever for 
increased responsiveness to the beneficiaries of compensatory discrimination. 
But judicial unresponsiveness to beneficiary claims is as much the cause as the result of the lack of 
doctrine to justify an affirmative judicial role. Judicial unresponsiveness in turn reflects basic 
ideological presuppositions about the role of courts, but also the limited capacities of the courts as 
institutions (e.g., to gather information or monitor implementation). But even if courts were not 
resistant, such an affirmative role is not something that can be produced by judges. It is the joint 
product of courts and of competent and resourceful parties and lawyers who can identify grievances, 
devise workable remedies, erect legal theories to justify those remedies, and press their demands in 
a sustained and persuasive manner in courts and other fora. The limits of the court are compounded 
by the weaknesses of the other elements of legal activism. In this area, as we saw, group 
participation has been feeble and intermittent: there are few recurrent organizational actors with the 
resources or legal imagination to mount a persistent campaign to use the courts affirmatively. Nor is 
any significant continuity supplied by the lawyers. Although a few writ specialists have handled a 
number of such claims, cases involving compensatory discrimination do not form the mainstay of 
the practice of a single one of India's 228,000 lawyers. There is no occasion to develop any 
professional expertise regarding the administration of preferences. Nor has such expertise developed 
in the realm of legal scholarship Judges, lawyers, academics, and potential 
____________________ 
57. The Constitution (25th Amendment) Act, 1971, added Art. 31C, which provided that no law 
 
giving effect to the directives in Art. 39(b) and (c) should be deemed void as inconsistent with 
Arts. 14, 19, or 31. The Constitution (42nd Amendment) Act, 1976, §4, enlarged this to cover all 
of the Directive Principles. The pre­1976 law regarding the relation of fundamental rights to 
Arts. 13, 16, and 20 is left intact. 
58. Kesavananda v. State of Kerala, A. I. R. 1973 S. C. 1467 (para. 134, 139, 1714); In re Thomas, A. 
I. R. 1952 Mad. 21. 
59. Kesavananda, op cit., above; In re Kerala Education Bill, A. I. R. 1958 S. C. 956. 

­538­ 
parties share a view of courts as critics and restraints on paternalistic government, but not as 
prods or energizers to make the government live up to its commitments. 
The setting and task of the courts expose them to opposing dangers of, on the one hand, incoherence 
(ambivalence, inconsistency, overload) and, on the other hand, of formalism (conceptualism, false 
precision). Not surprisingly, both are found in our examination of the work of the courts. 
Courts have to respond to obvious abuses and to address intractable problems and imponderable fact 
situations. It is not surpirsing that they often erect a general rule in response to the formal 
characteristics of the situation rather than a rule calibrated to distinctions of fact and purpose. (Thus 
the "both socially and educationally" rule and the "no compartments" doctrine.) Or they substitute 
rule­manipulation for complex factual determinations (as in reading the lists or dealing with 
converts). As a result, the courts promulgate rules that might well undermine the effectuation of the 
policies they are enunciating. 
In an area of law founded on the constitutional embrace of conflicting principles, it should not be 
expected that courts would provide an enduring synthesis that transcends and encompasses them 
and settles disputed issues with finality. Rather, we would expect ­­ if the courts are at all 
representative of the larger society ­­ some ambiguity and vacillation. And the courts have 
vacillated, sometimes emphasizing compensatory discrimination; other times, formal equality; 
sometimes, rectification of communal disparities; other times, non­recognition of communal units. 
We have seen that a number of factors may operate to relax the pressure for consistency and 
economy and to facilitate the admission into the corpus of authority of a variety of doctrines, not all 
of them reconcilable, which reflect the tensions between antagonistic principles. 
There are powerful forces operating to deflect the thrust toward consistency into a tolerance for ad 
hoc elaboration: heavy work pressures, high turnover on courts, a multiplicity of small benches of 
changing composition. Judges address the exigencies of different factual situations and have to 
apply doctrine which they have encountered infrequently and without the help of expertise on the 
part of counsel. Rajeev Dhavan attributes judicial tolerance for conflicting strands of precedent to 
"ineptness of counsel and the lack of interest in the judges in trying to evolve a comprehensive and 
complete picture of the disputed points of law on a particular subject."  60.   He goes on to cite the 
decisions about the caste basis for designating Backward Classes as 
____________________ 
60. R. Dhavan 1977: 451. 
­539­ 
exemplifying the observation that in India "new ideas are added without old ideas being 
discarded."  61.   Thus the courts generate a kind of doctrinal overload. In a setting of weak 
hierarchic structure, the presence of inconsistent doctrinal strands leads to the ambivalence 
and vacillation we noted earlier. As Dhavan remarks of the Supreme Court, "law does not 
evolve in the Supreme Court, it meanders. . . ."  62.   
Thus both formalism and incoherence are built into the structure of the situation. A conceptualistic 
intellectual tradition combines with imponderable questions and limited fact­finding capacities to 
produce overgeneral rules. Heavy work loads, sitting in small and shifting benches, lack of expertise 
on the part of advocates, and a lack of a drive for synthesis lead to the elaboration of various 
inconsistent strands of authority. The two tendencies are mutually reinforcing: the plurality of rules 
adds to overload and ambiguity, which is temporarily resolved by further rule elaboration. 
In the area of compensatory discrimination, both incoherence and formalism have yet another 
source: they flow from, and are accentuated by, the failure of the courts to come to grips with the 
immense variation and complexity of the group structure of Indian society. In examining the judicial 
handling of compensatory discrimination we can discern not only divergent proposals for combining 
and implementing the principles of the Indian Constitution, but also competing (if inchoate) views 
of what Indian society is like. The judiciary are inevitably engaged in the delicate task of mediating 
between social actualities and the farreaching reconstruction of society envisaged by the 
Constitution. They are not only authoritative interpreters of these goals but assessors of the 
changing actuality in which these are to be realized. These are, of course, not wholly separate 
undertakings, for the perception of what is tends to be informed by one's goals, and the 
interpretation of goals and values takes on content and color from one's picture of what is. What one 
imagines to be desirable and attainable and what one imagines to be real are interdependent. 
The contribution of the judiciary to social reconstruction, then, is affected by the clarity, scope, and 
penetration of its view of Indian society. By virtue of their training, experience, and positions, 
judges hold views of society that are to some extent partial and distorted.  63   
____________________ 
61. Dhavan 1977: 452. 
62. Dhavan ( 1977: 452) submits that this is "India's version of the 'happy mean'. a broad margin of 
tolerance, which is a variety within the common­law pattern and not to be judged by the 
standards of highly integrated and consolidated commonwealth countries." 
63 See chap. 9 generally. A recent example from a judge extremely sympathetic to the lowly, 
suggests a kind of curious foreshortening in the judicial view of lower strata. 
­540­ 
This is of course not peculiar to Indian but is common to at least all large and diverse 
societies. And it is certainly not peculiar to judges. Each participant in society can see it only 
from his own vantage point and can apprehend it only in terms of available concepts and 
symbols. Along with other men of affairs, judges may have more elaborate and intricate 
images of social reality, but this makes them even more dependent on the quality of the 
concepts and data available to them. And along with other men of words, they come to cherish 
the symbols not only as instruments but as an independent source of grievance or satisfaction. 
A society is a symbolic system as well as a structure of social relations. For those charged with 
designing and effectuating policies to change fundamental aspects of social structure, 
deficiencies in the quality of these images can be a special disability. 
We have suggested earlier that the learning available to the courts may be an insufficient guide to 
the complexities of social structure and social process in India. Compared with other decision­
makers, judges have the decided advantage of dealing with concrete factual situations, but they do 
so from the perspective of a relatively closed system of normative learning about that society, one 
whose fixed categories, formal classifications, and emphasis on rules may make it difficult for them 
to develop the kind of understanding of society appropriate to guide this transformation. There is 
little in their milieu to assist them in overcoming the deficiencies of their training and the limits of 
their experience. It remains to be seen whether Indian scholarship can generate and provide the 
information (e.g., about the actual effects of judicial decisions) and concepts that will enable the 
constitutional goals to serve as guides and monitors for (rather than as symbolic substitutes for) 
social transformation. 
C. THE LIMITS OF JUDICIAL ACTION 
We have seen several reasons that contribute to giving the courts a central role in the working out of 
the compensatory discrimination policy. There has been an absence of other agencies with specific 
responsibilities to oversee the implementation of these policies (like, e.g., equal opportunity 
commissions, attorneys general with enforcement staffs);  64   there are no actors in the legal arena 
with detailed 
____________________ 
Thus Justice Krishna Iyer thinks that the Kerala provision for clerks in the Thomas case may 
contribute to "an equalitarian break through" because "lower division clerks are likely to be 
drawn from the lowest levels of harijan humanity" (A. I. R. 1976 S. C. 490 at 531­32). 
64 The Commissioner for Scheduled Castes and Scheduled Tribes is an information­ 
­541­ 
policies, expert staffs, and institutional momentum. Indeed, the government in implementing 
these policies avoids using the legal arena­the courts are not used affirmatively as an 
instrument for implementing government policy.  65.   They are brought in almost invariably by 
someone objecting to government action­­government is always the defendant in 
compensatory discrimination cases. The courts encounter the policy in its incidence on 
aggrieved individuals. The higher judiciary enter directly through writ petitions, not by the 
indirect filtering route of appeals from lower courts or from an agency with expertise in 
implementing policy in this area. 
The higher judiciary, then, gets directly involved, not on its own initiative or that of the government 
as part of a plan to implement these policies or as part of a process of monitoring their 
implementation, but in reaction to private grievances arising out of their administration. Thus the 
judicial encounter with compensatory discrimination is selective in several ways: 
First, those who invoke the courts are those who not only have something substantial at stake, but 
also have financial or organizational resources to defend or pursue it­­e.g., excluded medical 
students, unpromoted middle­level government employees, defeated candidates.  66.   The student who 
fails to get his stipend on time or the disappointed aspirant for a reserved government post are not 
likely to litigate­­much less the hutment dweller with no decent water supply. 
Second, as we have seen, preferential treatment tends to be concentrated in practice and tends to 
impinge heavily on relatively few individuals. It is from these areas of heavy concentration and 
these targets of heaviest impingement that the cases come and the courts get their picture of 
compensatory discrimination policy. The doctrinal 
____________________ 
gathering agency with no enforcement powers. (On the limitations of this office, see chap. 3, 
§C). Since 1969 it has been augmented by a Joint Parliamentary Committee on the Welfare of 
Scheduled Castes and Scheduled Tribes which has investigative powers but no enforcement or 
policy responsibilities. In July 1978 the five­member Scheduled Castes and Scheduled Tribes 
Commission, was established. See chap. 3, §C, at n. 126, above. 
65. This is not suprising, for the targets of the compensatory discrimination policy are almost 
entirely within the government itself. Government does not use courts to compel its agencies to 
carry out its policies. Presumably, they are reachable through the chain of command (which may, 
however, be somewhat illusory). 
66. That is, these are parties who lost the first round. Of course, those who are most amply provided 
with wealth and influence may find it possible to secure admissions (or other prizes) through 
other channels and thus have no need to pursue these goals through the courts. Cf. Dolbeare 
observation ( 1967) that plaintiffs in an American trial court tended to be drawn from the 
prosperous, but those who had failed to prevail in more political forums. 
­542­ 
overload (i.e., the plurality of irreconcilable principles) in this area inclines courts to adopt a 
"balancing approach." Naturally, in these cases they see their job as one of balancing the 
claims of the backward (as asserted by the government) against the claims of those 
immediately affected (e.g., government servants in posts subject to heavy reservations). 
Balancing or deciding what is "reasonable" becomes a weighing of these limited interests, not 
all of those involved. The courts encounter only fragments of the total picture. They have no 
information about (and no power to act in regard to) those who escape entirely from the 
incidence of compensatory discrimination. Nowhere in the balance do we find the government 
officer in a post effectively immune from reservations or persons in private employment 
(where there are no reservations at all). Courts thus never get a chance to look at the whole 
but see only the fragment before them.  67.   
Third, the setting and the configuration of parties make it difficult for courts to give great weight to 
the long­term interests embodied in the compensatory discrimination principle. The claims that 
come to court focus on the distribution of scarce opportunities in the short run. The battle before 
them is about allocation this year, with its vivid and palpable effects. The courts naturally incline to 
visualize the preference policy as an attempt to be fair among competing claimants in the short run 
rather than in the light of long­term historical recompense. It is difficult for the courts to assess the 
long­term interests that are invoked on both sides. Although courts talk in terms of interests like 
efficiency, morale, utilization of talent, etc., as well as fairness to the parties, they typically have 
little to go on in judging these wider effects. Little data on such effects have been generated by 
Indian social science; lawyers are not accustomed to locating such data or arguing from them. In 
dealing with such interests the courts tend to fall back on an admixture of commonsense 
impressions and aprioristic surmise­­which may be more or less acute and current.  68.   The long­run 
interests of the nation in equalization are less tangible and immediate than interests in morale, etc., 
and even more difficult to assess. And the configuration of ad­ 
____________________ 
67. On the distortion built into the forum's reliance on individual complaints, compare Mayhew 
observation ( 1968: 159) that the pattern of complaints to an anti­discrimination agency "did not 
correspond to the structure of discrimination. By and large, the complaints did not attack the 
major bastions of discriminations. Rather, the complaints reflected the current structure of Negro 
employment; they tended to be directed toward areas where racial barriers had already, fallen." 
On the deficiencies of reliance on individual complaints, see Lockard 1968: 98. 
68. See. e.g., the remarks in Balaji (A.I.R. 1963 S.C. 659 at 662) about the inevitable fall in 
standards. This apriorism is aggravated by the writ petition proceeding with its absence of 
opportunity to adduce and challenge evidence. 
­543­ 
versaries may not bring before the courts those parties who could best assert these long­term 
and imponderable interests. The interests of the disappointed competitor do not necessarily 
correspond with those of the public; the interests of the government, eager to defend its 
existing schemes, do not necessarily correspond with those of the backward classes. 
While providing a forum responsive to what I have called the "too much" and "wrong thing" 
criticisms of compensatory discrimination policy, the courts have not been responsive to the "not 
enough" or "not the right thing" criticisms. They have done little directly to offer remedies for the 
deficiencies of implementation of existing schemes­­to see that reservations are filled, scholarships 
are distributed on time, programs are well suited to circumstances and pursued with sufficient 
energy, resources, and attention. In part, this is due to the posture of the Constitution, which 
provides no explicit authorization for affirmative judicial action. Claims under Directive Principles 
are not justiciable­so there is no justiciable claim that preferences be awarded. But courts, abetted by 
lack of inventiveness at the bar, have extended this passivity to claims that those preferences which 
are awarded be implemented effectively. 
If courts could provide more in the way of judicial remedies for abuses, maladministration, and 
omissions which hurt beneficiaries of preference, they are not necessarily well suited to addressing 
the larger problems of pursuing the long­term goal of equalization. The limitations of the courts can 
be appreciated if we visualize compensatory discrimination as a special irrigation scheme, designed 
to bring lifegiving waters to lands which have long been parched and sterile as a result of neglect 
and oppression by the owners of the more prosperous lands among which they are situated. The 
system is instituted on the assumption that without it the available water would inevitably flow to 
the fields that are already fertile. The capabilities of the courts as keepers of the sluice gates are 
very different for the different groups involved. There is little that they can do for the parched 
lands­­they can close down illicit diversions and stop the state from placing illegitimate limits on the 
flow in existing channels. But they cannot dig new channels or deepen the existing ones. They 
cannot insure that the channels will be maintained in good repair and that all steps will be taken to 
effectuate delivery. They cannot insure that the special water is in addition to the trickle that might 
reach them through ordinary channels. And, they can do nothing to increase the flow of water into 
the system. But for the prosperous neighbors, they can stem the flow if it is so excessive that it 
depletes the supply of water available for the rich land; they can see that water doesn't flow to the 
wrong fields or through 
­544­ 
channels that damage adjacent interests. The sluice gates that the courts use to protect these 
neighboring interests may impede the legitimate flow. 
Of course, the courts do not tend all the sluice gates themselves­­only those major ones which have 
been the subject of complaint by the owners of the richer lands. The vast network of channels is 
tended by keepers with less education and sophistication, fewer resources for fact­finding, less time 
for deliberation. The more subtle and complex the rules that the courts establish at their gates, the 
more difficult it will be for the attendants along the channels to apply them. But the courts cannot 
deploy resources to upgrade and equip the attendants. They can act only in those cases which are 
brought to them. 
Thus the courts are a check on the system, not an expediter of it except in some marginal cases. 
They cannot put water into the system; they cannot supply any moving power; they cannot improve 
the channels or their attendants. These things can be done by politicians and administrators­­
including the occupants of reserved seats (and reserved posts). It would be wrong to visualize the 
courts as enemies or even as inadvertent wreckers of compensatory discrimination policy, somehow 
responsible for its deficiencies and shortcomings. These are not due mainly to courts. If courts have 
in a few instances played a restrictive role, this has been far outweighed by their positive 
contributions. Probably the greatest of these is to give compensatory discrimination legitimacy. By 
giving remedies in cases where others are hurt by it, the courts have made crucial publics feel, if not 
enthusiasm, that at least it is not out of control and is somehow compatible with the welfare of all. 
By containing and curtailing the system, the courts have helped to maintain and preserve it, though 
at a level of performance below that of paper commitments. But the shortcomings are not in what 
the courts did, but in what other agencies of government (and outside government) did not do. The 
question is to what extent might the courts be instrumental in stimulating and energizing a higher 
level of performance by these agencies. 
In part, the inability of courts to institutionalize compensatory discrimination at a higher level of 
performance flows from the basic architecture of the common­law court.  69.   Courts as we know 
them lack independent fact­finding apparatus; they lack capacity for systematic and continuing 
monitoring and surveillance; they take up "cases" rather than address whole problems; they are 
limited in the nature of the remedies they can deploy (e.g., they have no scope to address problems 
by devising new administrative machinery). These structural 
____________________ 
69. See Galanter 1974. 
­545­ 
constraints are overlaid and accentuated by a number of features of the Indian legal milieu­­
narrow recruitment and high turnover of judges, an intellectual tradition of conceptualism 
and positivism, weak connection to social scientific or other sources of data. These 
institutional limitations of the courts are articulated to other features of the legal process. On 
the level of rules, there is the absence of doctrinal grounding for justiciability and for class 
actions. The bar has no specialists with enduring relationships to this kind of party, and these 
parties themselves are poorly organized, lack expertise, have no facilities for systematic 
gathering of information, etc. The absence of pressure from the courts for a higher level of 
performance of compensatory discrimination is the result of a complex of interlocking factors. 
But there is no reason to think that the legal system is incapable of being mobilized to give 
greater support to these policies. One cannot say in advance what mix of organization of 
parties, supply of legal services, doctrinal inventiveness, and changes in institutional 
incumbents it would take to get new affirmative input from the legal system. But once such an 
affirmative role is undertaken, it may very well be self­sustaining: court success might 
stimulate organization of parties, the development of expertise, generation of data, the 
stimulation of further doctrinal innovation, etc. There is no reason inherent in the legal 
process that its elements (doctrine, courts, lawyers, and parties) could not be joined in a 
pattern supporting increased effectiveness, just as they are at present joined in a 
lowperformance equilibrium. 
­546­ 
CONCLUSION 
16 "The Little Done, the Vast Undone" 
A. A COSTLY SUCCESS 
IF EXAMINING the response of the courts to compensatory discrimination policies teaches us 
something about Indian courts, what does it teach us about the policies themselves? Have they 
"worked"? What results have they produced? and at what cost? Our earlier tabulation of alleged 
costs and benefits suggested the complexity hidden in these apparently simple questions. 1. 
Performance is difficult to measure: effects ramify in complex interaction with other factors. 
Compensatory policies are designed to pursue a multiplicity of incommensurable goals in 
unspecified mixtures that vary from program to program, from time to time, and from proponent to 
proponent. Evaluation of a specific scheme for a specific group during a specific period is itself a 
daunting undertaking. 
Does our inquiry tell us anything about the thrust and meaning of this protean cluster of measures 
and models? What has the commitment to compensatory discrimination done to the shape of Indian 
society and of lives lived within it? Employing some of the rubrics from our earlier list of alleged 
costs and benefits, I shall draw a very crude sketch of the policy's effects in their largest outline. The 
limited clarity of such a sketch is dimmed by the necessity to distinguish between compensatory 
discrimination for the Scheduled Castes and Tribes on the one hand and for the Other Backward 
Classes on the other. The following summary focusses on programs for Scheduled Castes (SC) and 
Scheduled Tribes (ST) and adds some qualifications in the light of experience with schemes for the 
Other Backward Classes (OBC). 
Undeniably, compensatory discrimination policies have produced substantial redistributive effects. 
Reserved seats provide a substantial 
____________________ 
1. See the table of alleged costs and benefits in chap. 3, § D, above, and the specification of that list 
to the job reservation area in chap. 4, above. 
­547­ 
legislative presence and swell the flow of patronage, attention, and favorable policy to SC and 
ST. The reservation of jobs has given to a sizable portion of the beneficiary groups earnings, 
and the security, information, patronage, and prestige that goes with government employment. 
At the cost of enormous wastage, there has been a major redistribution of educational 
opportunities to these groups. (Of course not all of this redistribution can be credited to 
preferential policies, for some fraction would presumably have occurred without them.) 
Such redistribution is not spread evenly throughout the beneficiary groups. There is evidence for 
substantial clustering in the utilization of these opportunities. The clustering appears to reflect 
structural factors. (e.g., the greater urbanization of some groups) more than deliberate group 
aggrandizement, as is often charged.     2. The better situated among the beneficiaries enjoy a 

disproportionate share of program benefits.     3. This tendency, inherent in all government programs ­­ 
quite independently of compensatory discrimination ­­ is aggravated here by passive administration 
and by the concentration on higher­echelon benefits. Where the list of beneficiaries spans groups of 
very disparate condition ­­ as with the most expansive lists of OBC ­­ the "creaming" effect is 
probably even more pronounced. 
Reserved seats afford a measure of representation in legislative settings, though the use of joint 
electorates deliberately muffles the assertiveness and single­mindedness of that representation. The 
presence of SC and ST in legislative settings locks in place the other programs for their benefit and 
assures that their concerns are not dismissed or ignored. Job reservations promote their presence in 
other influential roles, and educational preferences provide the basis for such participation. Of 
course, these positions are used to promote narrower interests ­­ although we should not assume 
automatically that those they displace would bestow the benefits of their influence more broadly. If, 
for example, reserved­seat legislators are disproportionately attentive to the concerns of their fellows 
who already have something, it is not clear that this is more the case with them than legislators in 
general seats. 
Preference programs are integrative in several ways. Reserved legislative seats are occupied by 
members of national political parties. They must aggregate broad multi­group support in order to 
get elected and, once elected, must participate in multi­group coalitions in order to be effective. In 
the office setting, too, there are relations of reciprocity and interdependence. The broad 
participation afforded by reserved seats and reserved jobs is for many others a source of pride and a 
warrant of security. 
____________________ 
2. Shah and Patel 1977: 149 ff. 
3. Malik 1979: 158. 
­548­ 
If the separate and special treatment entailed by preferential programs wounds and alienates 
the members of beneficiary groups, this is amplified by the hostility experienced on being 
identified as a recipient. As sources of alienation, these experiences must be placed against the 
background of more devastating manifestations of hostility, such as the much publicized 
assaults and atrocities perpetrated on Scheduled Castes. 
At the policy­making level, reserved seats have secured the acceptance of SC and ST as groups 
whose interests and views must be taken into account. In every legislative setting they are present in 
sufficient numbers so that issues affecting these groups remain on the agenda. Anything less than 
respectful attention to their problems, even if only lip service, is virtually unknown. Overt hostility 
to these groups is taboo in legislative and many other public forums. But there is evidence that SC 
and ST are not accepted politically. Very few members of these groups are nominated for non­
reserved seats, and only a tiny number are elected. There is massive withdrawal by voters from 
participation in election for reserved seats in the legislative assemblies. Apparently, large numbers 
of people do not feel represented by these legislators and do not care to participate in choosing 
them.  4.   
In the long run, education and jobs help weaken the stigmatizing association of SC and ST with 
ignorance and incompetence, but in the short run they experience rejection in the offices, hostels, 
5. Resentment of 
and other settings into which they are introduced by preferential treatment.    
preferences may magnify hostility to these groups, but rejection of them obviously exists 
independently of compensatory programs. 
Compensatory programs provide the basis for personal achievement and enlarge the beneficiaries' 
capacity to shape their own lives. But in other ways the programs curtail their autonomy. The design 
of the legislative reservations ­­ the dependence on outside parties for funds and organizations and 
the need to appeal to constituencies made up overwhelmingly of others ­­ tends to produce 
compliant and accommodating leaders rather than forceful articulators of the interests of these 
groups. The promise of good positions offers a powerful incentive for individual effort. But 
reservations in government service ­­ and educational programs designed to provide the requisite 
qualifications ­deflect the most able into paths of individual mobility that remove them from 
leadership roles in the community. Constraints are intruded into central issues of personal identity 
by eligibility requirements that 
____________________ 
4. This evidence is presented in Galanter 1979. 
5. Cf. Malik finding ( 1979: 50) that middle­class Scheduled Castes experience more exclusion than 
do their less­edticated fellows. 
­549­ 
penalize those who would solve the problem of degraded identity by conversion to a non­
Hindu religion. 
Although preferential treatment has kept the beneficiary groups and their problems visible to the 
educated public, it has not stimulated widespread concern to provide for their inclusion apart from 
what is mandated by government policy. (This lack of concern is manifest in the record of private­
sector employment­as it was in publicundertaking employment before the introduction of 
reservations.) Against a long history of such lack of concern, it is difficult to attribute its current 
absence to compensatory discrimination policy. But this policy has encouraged a tendency to 
absolve others of any responsibility for their betterment on the ground that it is a responsibility of 
6. The pervasive overestimation of the amount and effectiveness of preferential 
the government.    
treatment reinforces the notion that enough (or too much) is already being done and nothing more is 
called for. 
Compensatory preference involves a delicate combination of selfliquidating and self­perpetuating 
features. Reservations of upperechelon positions should become redundant as preferential treatment 
at earlier stages enables more beneficiaries to compete successfully, thus decreasing the net effect of 
the reservations. A similar reduction of net effect is produced by the extension to others of benefits 
previously enjoyed on a preferential basis (e.g., free schooling). Judicial requirements of more 
refined and relevant selection of beneficiaries (and of periodic reassessment) and growing use of 
income cut­offs provide opportunities to restrict the number of beneficiaries. 
Reserved seats in legislatures are self­perpetuating in the literal sense that their holders can help to 
produce their extension. But extension requires support from others. The periodic necessity of 
renewal provides an occasion for assessment and curtailment. Programs for SC and ST are for a 
delimited minoriy and pose no danger that the compensatory principle will expand into a 
comprehensive and self­perpetuating system of communal quotas. Although restrained by the 
courts, the provisions for OBC are open­ended: a majority may be beneficiaries, and the dangers of 
self­perpetuation cannot be dismissed. 
The diversion of resources by compensatory discrimination programs entails costs in the failure to 
develop and utilize other talents. The exact extent of this is unclear. It seems mistaken, for example, 
to consider compensatory discrimination a major factor in the lowering of standards that has 
accompanied the vast expansion of educational facilities since Independence. The pattern in 
education has been less 
____________________ 
6. Cf. Lelah Dushkin observation ( 1979:666): "In the course of my visits to India over two decades 
I have noticed an erosion and virtual disappearance of a liberal­minded public opinion 
supporting private efforts to improve opportunities for the S. C." 
­550­ 
one of excluding others than one of diluting educational services while extending them 
nominally to all. Similarly, the effect of SC and ST on the effectiveness of a much enlarged 
government bureaucracy is overshadowed by a general lowering of standards combined with 
the assumption of a wide array of new and more complex tasks. 
The most disturbing costs of preferential programs may flow not from their exclusion of others but 
from their impact on the beneficiaries. What do the programs do to the morale and initiative of 
those they purport to help? The numbers who fall by the educational wayside are legion. How 
rewarding is the educational experience of those who survive? Compensatory discrimination 
policies are not the source of the deficiencies of Indian education that impinge with special force on 
the beneficiaries. 
As a forced draft program of inclusion of SC and ST within national life, compensatory 
discrimination has been a partial and costly success. Although few direct benefits have reached the 
vast mass of landless laborers in the villages, compensatory discrimination has undeniably 
succeeded in accelerating the growth of a middle class within these groups ­­ urban, educated, 
largely in government service. Members of these groups have been brought into central roles in the 
society to an extent unimaginable a few decades ago. There has been a significant redistribution of 
educational and employment opportunities to them; there is a sizable section of these groups who 
can utilize these opportunities and confer advantages on their children; their concerns are firmly 
placed on the political agenda and cannot readily be dislodged. But if compensatory discrimination 
can be credited with producing this self­sustaining dynamic of inclusion, there is at the same time a 
lesser counter­dynamic of resentment, rejection, manipulation, and low selfesteem. And these gains 
are an island of hope in a vast sea of neglect and oppression. This mixed pattern of inclusion and 
rejection, characteristic of urban India and of the "organized" sector, is echoed in the villages by a 
pattern of increasing assertion and increasing repression. 
Since Independence, India had undergone what might crudely be summarized as development at the 
upper end and stagnation at the bottom. With the boost given by compensatory discrimination, a 
section of the SC and ST has secured entry into the modern class manning the organized sector. 
What does this portend for the bulk of untouchables and tribals who remain excluded and 
oppressed? Are they better or worse off by virtue of the fact that some members of their descent 
groups have a share in the benefits of modern India? The meaning of these achievements ultimately 
depends on how one visualizes the emergent Indian society and the role of descent groups in it. 
­551­ 
Even this kind of crude characterization of the overall impact of policies is not possible in 
dealing with measures for Other Backward Classes. Policies diverge from state to state, and 
very different groups of people are involved. In some states the OBC category is used to 
address the problems of a stratum of lowly groups who are roughly comparable in 
circumstance to the SC and ST. In other places this category has been used to tilt the 
distribution of government benefits in favor of a major section of the politically dominant 
middle castes. The latter doubtless produce substantial redistributive effects, if less in the way 
of including the most deprived. But these expansive preferences for OBC are of immense 
consequence for the SC and ST. They borrow legitimacy from the national commitment to 
ameliorate the condition of the lowest. At the same time they undermine that commitment by 
broadcasting a picture of unrestrained preference for those who are not distinctly worse off 
than non­beneficiaries, which attaches indiscriminately to all preferential treatment. And 
because the OBC categories are less bounded and are determined at the state rather than at 
the Centre, they carry the threat of expanding into a general regime of communal allotments. 
B. FAIRNESS AND HISTORY 
Arguments about the utility of compensatory preference entwine with arguments about its fairness. 
Apart from the other costs, is compensatory preference so tainted by unfairness that it is illegitimate 
to promote the general welfare by this means? We have encountered several arguments about the 
fairness and unfairness of these programs. We shall examine four here: (1) Is it unfair to depart from 
judgments on "individual merit" to favor the beneficiaries over other contenders for valued 
resources? (2) Is it unfair to compensate members of some groups for injustice perpetrated on their 
ancestors? (3) Is it unfair to compensate some victims and not others? (4) Is it unfair that some 
should bear more of the burden of compensation than others? Before taking up these fairness 
arguments it may be helpful to recall the justifications that may be advanced to the compensatory 
discrimination policy. Although they are often entwined in practice, we can separate out three sorts 
of justifications for these measures, which I label the non­discrimination, the general welfare, and 
reparations themes. 
The non­discrimination theme 
Compensatory discrimination may be viewed as an extension of the norms of equal treatment, an 
extension invited by our awareness that even when invidious discriminatory standards are 
abandoned there 
­552­ 
remain subtle and tenacious forms of discrimination and structural factors which limit the 
application of new norms of equality. Aspiring members of previously victimized groups 
encounter biased expectations, misperceptions of their performance, and cultural bias in 
selection devices; they suffer from the absence of informal networks to guide them to 
opportunities, entrenched systems of seniority crystallize and perpetuate the results of earlier 
discriminatory selections. Thus norms of non­discrimination in present distributions are 
insufficient to erase or dislodge the cumulative effects of past discrimination. Compensatory 
preference operates to counter the residues of discrimination and to overcome structural 
arrangements which perpetuate the effects of past selections in which invidious discrimination 
was a major determinant. In this view compensatory preferences serves to assure personal 
fairness to each individual applicant. Group membership is taken into account to identify 
those individuals who require special protection in order to vindicate their claim for selection 
on "merit" grounds. (The justification for much American affirmative action is often cast in 
these terms ­as an extension of classical individualistic non­discrimination principles.) 
The general welfare theme 
One the other hand, compensatory discrimination may be advocated, not as a device to assure 
fairness to individuals, but as a means to produce desired social outcomes ­­ e.g., to reduce group 
disparities, afford representation, encourage the development of talent, and so forth. Arrangements 
for reservations in British India were justified on such "functional" grounds as are the various 
7. Americans are familiar with the "balanced ticket" 
preference for "Oriental" Jews in Israel today.    
and other arrangements by which shares are apportioned among various constituencies in the 
expectation that abrasive disparities are kept in bounds, participation is spread out, representation is 
secured, and responsiveness is assured. The units in such functional "welfare" calculations are 
groups rather than individuals. The chances of individuals are affected by the rearrangement of the 
chances of groups. But the purpose is not to rectify discriminatory selection among individuals, but 
to introduce a standard quite apart from personal desert. 
The contrast of non­discrimination and welfare themes is displayed in imaginary alternative 
proposals for admitting more members of Group X to medical colleges. The non­discrimination 
proposal might argue that selection procedures be revised to eliminate subtle bias 
____________________ 
7. On these programs, see Smooha 1978 ; Alder et al. 1975. 

­553­ 
which impinged on individual Xs ­­ e.g., culturally biased tests or differences in networks for 
acquiring recommendations. The general welfare proposal might argue that more Xs should 
be admitted in order to equalize the distribution of medical services or for the represenation 
of X views in making health policy or to afford non­X doctors the experience of fellowship 
with Xs. The goal is not a non­discriminatory selection of Xs among individual applicants, but 
a selection that optimizes social goals. Such a selection might diverge from that which would 
be dictated merely by the prospects of individual performance on the job, because it defines 
the job to include the symbolic, representational, and educational aspects that may not be 
included in the job description. 
The reparations theme 
In some cases, compensatory policies have another root ­­ that a history of invidious treatment has 
resulted in accumulated disabilities which are carried by certain groups. No matter how fair and 
unbiased the measures now employed for distributing benefits, the victims of past injustice will not 
fare well in terms of current performance. To distribute benefits by neutral standards will perpetuate 
and amplify unjust exactions and exclusions in the past. Fairness, then, demands that present 
8. 
distributions be arranged to undo and offset old biases, not to perpetuate them.    
Like the non­discrimination theme this is a fairness argument, rather than a welfare argument. But it 
emphasizes groups as the carriers of historic rights rather than as indicators of individual 
victimization. And it looks to a very different time frame. Welfare arguments are prospective; non­
discrimination looks at the present situation and seeks to refine out lingering inequalities. The 
reparations theme sees the present as an occasion to reckon accounts for past injustice. 
Do preferential programs unfairly confer benefits on grounds that depart from evenhandedness, 
merit, etc., that should govern the distribution of opportunities and resources? Along the lines 
referred to 
____________________ 
8. The notion of restitution for collective misdeeds is familiar in the indemnity and reparation 
payments exacted from defeated nations (see Angell 1934 ). Something closer to the 
compensatory discrimination situation is found in the reparations payments paid by Germanv to 
Israel as the representative of Jews victimized by the Nazis (see K. Grossman 1954). In these 
instances the lapse of time between misdeed and reparation is relatively short. But cf. the 
occasional claims for "three hundred years of back pay" raised in the American setting ( Lecky 
and Wright 1969; Schuchter 1970 ; Bittker 1973 ). But a scheme of preferences may lack this 
theme of justification entirely, as in the Israeli case (n. 7, above), where the privileged and 
deprived groups had little previous contact. 
­554­ 
above as the non­discrimination theme, proponents might respond that some of the preference 
accorded is not departure from evenhandedness but its extension in substance rather than 
form to individual members from the beneficiary groups. In this view, compensatory 
discrimination arrangements counter subtle discriminations and overcome the structural 
arrangements which entrench the results of past selections from which the beneficiaries were 
excluded. 
In the Indian setting, few would argue that compensatory discrimination seeks only to protect merit 
against subtle or structural bias. Preferential treatment is accepted as a departure from merit 
selection in order to promote such goals as redistribution, integration, and representation. Is it unfair 
to combine these with merit as a basis for distributing benefits? 
Let us take merit to mean performance on tests (examinations, interviews, character references, or 
whatever) thought to be related to performance relevant to the position (or other opportunity) in 
question and commonly used as a measure of qualification for that position. (In every case it is an 
empirical question whether the test performance is actually a good predictor of performance in the 
position, much less of subsequent positions for which it is a preparation.) Performance on these tests 
is presumably a composite of native ability, situational advantages (stimulation in the family setting, 
good schools, sufficient wealth to avoid malnutrition or exhausting work, etc.), and individual effort. 
The latter may be regarded as evidence of moral desert, but neither native ability nor situational 
advantages would seem to be. The common forms of selection by merit do not purport to measure 
the moral desert dimension of performance. Unless one is willing to assume that such virtue is 
directly proportionate to the total performance, the argument for merit selection cannot rest on the 
moral deservingness of individual candidates. Instead, it rests upon the supposed consequences: 
those with more merit will be more efficient or productive; awarding them society's scarce resources 
will produce more indirect benefits for their fellows. A regime of rewarding merit will maximize 
incentive to cultivate talents; the demoralizing effects of departing from merit outweigh supposed 
advantages, based on calculations of imponderables. The argument for merit is an argument for 
production of more social well­being. 
Many sorts of effects flow from any allocation of resources. Benefits are multiple; they include not 
only tangible production, but symbolic affirmations and the creation of competences. The allocation 
of education, government jobs, or medical careers arguably has consequences for the distribution of 
incentives, levels of participation, disparities in the delivery of services. Which dimensions of 
benefit are to be taken into 
­555­ 
account in designing a given selection? In settings where there has been a broad consensus 
that a single­minded test of performance is appropriate, what is the argument for a shifting to 
a broader, promotional basis of selection? Compensatory discrimination schemes involve 
enlargement of the basis of selection to include other criteria along with the productivity 
presumably measured by "merit" ­­ representation, integration, stimulation, and so forth. 
This enlargement is justified on the ground that without it society would be deprived of the 
various benefits thought to flow from the enhanced participation of specified groups in key 
sectors of social life. The argument is that the combination maximizes the production of good 
results. Of course, there is always the empirical question of whether the promised results are 
indeed produced, but the supplementation of merit with other instrumental bases of selection 
hardly seems unfair in principle. Pursuit of other worthy results can be balanced against 
merit, as one resultoriented justification for unequal allocations against another. 
Compensatory discrimination is both more and less than a reformulation of selection criteria. It is 
less because typically merit (in narrow performance terms) is left intact for the main part of the 
selection. The criteria are modified to require the inclusion of certain groups, an inclusion thought 
to produce a wide spectrum of beneficial results. But the new mixed standards are not applied 
across the board to the whole selection. So compensatory discrimination involves something more: 
the demarcation of those groups in whose behalf the broader promotional standards should be 
employed. 
To prefer one individual over another on grounds of caste, religion, or other ascriptive criteria is 
specifically branded as unfair by the antidiscrimination provisions of the Indian Constitution. The 
ban on the use of these criteria is, as we have seen, qualified to allow preferential treatment of a 
certain range of groups, whose history and condition seemed distinctive. There was agreement that 
some groups were burdened by a heritage of invidious discrimination and exclusion (and/ or 
isolation), which made their condition distinct from that of their fellow citizens; the deprivations of 
their past and present members were thought to justify a special effort for their improvement and 
inclusion. 
Spokesman far backward classes sometimes call for measures specifically to remedy the wrongs of 
the past. If one thinks of the blighted lives, the thwarted hopes, the dwarfing of the human spirit 
inflicted on generations of untouchables, or of the oppression and exploitation of tribal peoples, the 
argument for measured vindication of these historic wrongs has an initial appeal. But there are many 
kinds and grades of victimization; deprivations are incommensurable. Perpetrators and 
­556­ 
victims sometimes stand out in stark clarity, but infirm and incomplete data often leave 
unclear precisely who were brutually exploitative, who willing or reluctant collaborators, who 
inadvertent beneficiaries of what we now see as systems of oppression. These arrangements 
interact with many other factors ­­ climate, invasions, technology ­­ in their influence on the 
present distribution of advantages and disadvantages. The web of responsibility is tangled, 
and as we try to trace it across generations, only the boldest outlines are visible. Without 
minimizing its horrors, we may concede that the past provides a shaky and indistinct guide for 
policy. It is beyond the capacity of present policy to remedy these wrongs: in the literal sense 
these injustices remain irremediable. 
But if our perception of past injustice does not provide a usable map for distributing reparative 
entitlements, it can inform our vision of the present, sensitizing us to the traces and ramifications of 
historic wrongs. The current scene includes groups which are closely linked to past victims and 
which seem to suffer today from the accumulated results of that victimization. In a world in which 
only some needs can be met, the inevitable assignment of priorities may take some guidance from 
our sense of past injustice ­­ thus providing the basis for a metaphoric restitution. 
All remedies involve new distinctions and thus bring in their train new (and it is hoped lesser) forms 
of unfairness. Singling out these historically deprived groups for remedial attention introduces a 
distinction among all of the undeserved inflictions and unfairnesses of the world. One batch of 
troubles, but not others, are picked out for comprehensive remedy using extraordinary means. Those 
afflicted by other handicaps and misfortunes are left to the succor and aid that future policy­makers 
find feasible and appropriate within the framework of competing commitments (including 
commitments to equal treatment). But drastic and otherwise outlawed remedies were authorized for 
victims of what was seen as a fundamental flaw in the social structure. The special quality of the 
commitment to correct this flaw is dramatized by the Constitution's simultaneous rejection of group 
criteria for any other purpose. 
The line of distinct history and condition that justifies compensatory discrimination is of course less 
sharp in practice than theory. There are border lines, grey areas, gradual transitions. There is 
disagreement about just where the line should be drawn. And once it is drawn, the categories 
established are rough and imperfect summations of need and desert; there are inevitable "errors" of 
under­inclusion and overinclusion. 
­557­ 
We arrive, then, at an ironic tension that lies at the heart of the compensatory discrimination 
policy. Since the conditions that invite compensatory treatment are matters of degree, special 
treatment generates plausible claims to extend coverage to more groups. The range of 
variation among beneficiaries invites gradation to make benefits proportionate to need. These 
preferential policies create new discontinuities, and it is inviting to smooth them out by a 
continuous modulated system of preferences articulated to the entire range of need and/or 
desert. But to do so is to establish a general system of group allotments. 
Compensatory discrimination replaces the arbitrariness of formal equality with the arbitrariness of a 
line between formal equality and compensatory treatment. The principles that justify the preference 
policy counsel flexibility and modulation. We may shave away the arbitrary features of the policy in 
many ways. But we may dissolve the arbitrary line separating formal equality and preferential 
treatment only at the risk of abandoning the preference policy for something very different. 
If there is to be preferential treatment for a distinct set of historically victimized groups, who is to 
bear the cost? Whose resources and life chances should be diminished to increase those of the 
beneficiaries of this policy? In some cases, the costs are spread widely among the taxpayers, for 
example, or among consumers of a "diluted" public service. But in some cases major costs impinge 
on specific individuals, like the applicant who is bumped to fill a reservation. Differences in public 
acceptance may reflect this distinction. Indians have been broadly supportive of preferential 
programs ­­ e.g., the granting of educational facilities and the sharing of political power ­­ where the 
"cost" of inclusion is diffused broadly. Resentment has been focussed on settings where the life 
chances of specific others are diminished in a palpable way, as in reservations of jobs and medical 
college places. 
There is no reason to suppose that those who are bumped from valued opportunities are more 
responsible for past invidious deprivations than are those whose well­being is undisturbed ­­ or that 
they were disproportionately benefitted by invidious discrimination in the past. Reserved seats or 
posts may thus be seen as the conscription of an arbitrarily selected group of citizens to discharge an 
obligation from which equally culpable debtors are excused. The incidence of reservations and the 
effectiveness with which they are implemented tends to vary from one setting to another. 
Reservations impinge heavily on some careers and leave others virtually untouched. The 
administration of compensatory discrimination measures seems to involve considerable unfairness 
of this kind. If some concentration of benefits is required by 
­558­ 
9.
the aims of the preference policy,    
 it seems clear that more could be done to distribute the 
burden among non­beneficiaries more widely and more evenly. 
C. SECULARISM AND CONTINUITY 
Fairness apart, to many Indian intellectuals compensatory discrimination policies seems to 
undermine progress toward the crucial national goal of a secular society. Secularism in this setting 
implies more than the separation of religion and State (religious freedom, the autonomy of religious 
groups, withdrawal of State sanction for religious norms, and so forth). It refers to the elimination 
(or minimization) of caste and religious groups as categories of public policy and as actors in public 
life.  10.   In the 1950s and 1960s this was frequently expressed as pursuit of a "casteless" society. 
Proponents of such a transformation were not always clear whether they meant the disestablishment 
of social hierarchy or the actual dissolution of caste units. But at the minimum, what was referred to 
was a severe reduction in the salience of caste in all spheres of life. 
The Constitution envisages a new order as to the place of caste in Indian life. There is a clear 
commitment to eliminate inequality of status and invidious treatment and to have a sociey in which 
government takes minimal account of ascriptive ties. But beyond this the posture of the legal system 
toward caste is not as single­minded as the notion of a casteless society might imply. If the law 
discourages some assertions of caste precedence and caste solidarity, in other respects the 
prerogatives previously enjoyed by the caste group remain unimpaired. The law befriends castes by 
giving recognition and protection to the new social forms through which caste concerns can be 
expressed (caste associations, educational societies, political parties, religious sects).  11.   
If the legal order's posture toward caste is ambivalent, public denuciation of caste has universal 
appeal. For lower castes it provides an opportunity to attack claims of superiority by those above 
them; for the highest castes it is a way to deplore the increasing influence of previously subordinate 
groups, either the populous middle castes that have risen to power with adult suffrage or the lowest 
castes whose inclusion is 
____________________ 
10. On the aspiration to secularism, see the literature cited in chap. 9, n. 85, above. 
11. See Galanter 1968. 
9. Cost spreading may itself entail costs in terms of the benefits delivered to the beneficiaries. For 
example, rotation of reserved legislative seats among constituencies might cure unfair clustering 
of costs, but it would vitiate the value of the reserved seats, rendering impossible the 
accumulation of experience, seniority, and political strength. 
­559­ 
mandated by compensatory discrimination programs. Looking up, the call for castelessness is 
an attack on the advantages retained by those who rank high in traditional terms; looking 
down, it denies legitimacy to the redistributive claims of inferiors and insists on evenhanded 
application of individual merit standards. 
The use of caste groups to identify the beneficiaries of compensatory discrimination has been 
blamed for perpetuating the caste system, accentuating caste consciousness, injecting caste into 
politics, and generally impeding the development of a secular society in which communal affiliation 
is ignored in public life.  12.   This indictment should be regarded with some skepticism. Caste ties and 
caste­based political mobilization are not exclusive to the backward classes. The political life within 
these groups is not necessarily more intensely communal in orientation;  13.   nor are the caste politics 
of greatest political impact found among these groups. Communal considerations are not confined 
to settings that are subject to compensatory discrimination policies, but flourish even where they are 
eschewed. Although it has to some extent legitimated and encouraged caste politics, it is not clear 
that the use of caste to designate beneficiaries has played a preponderant role in the marriage of 
caste and politics. Surely it is greatly overshadowed by the franchise itself, with its invitation to 
mobilize support by appeal to existing loyalties. But the avowed and official recognition of caste in 
compensatory discrimination policy combines with the overestimation of its effects to provide a 
convenient target for those offended and dismayed by the continuing salience of caste in Indian life. 
The amount of preference afforded the Scheduled Castes and Tribes is widely overestimated. The 
widespread perception of ubiquitous and unrestrained preferment for these groups derives from 
several sources. First, there is the chronic overstatement of the effects of reservation: large portions 
of reservations (especially for cherished higher positions) are not filled; of those that are filled, 
some would have been gained on merit; diversion of benefits to a few may be perceived as a 
deprivation by a much larger number. The net effect is often considerably less than is popularly 
perceived. Second, ambiguous nomenclature and public inattention combine to blur the distinction 
between measures for Scheduled Castes and Tribes and those for Other Backward Classes. The 
____________________ 
12. See chap. 3, § D, above. 
13. Consider, for example, the fascinating finding of Eldersveld and Ahmed ( 1978: 205), studying 
political participation, that "the overwhelming majority of . . . activists are politically conscious 
of caste. They know how the caste leader voted. Among upper castes the ratio is 6 to 1 that 
activists are aware of caste, in the middle castes it is better than 8 to 1, and for the lower castes 
and Harijans it is 3 to 1 that activists are caste conscious." 
­560­ 
resentment and dismay engendered by use of the OBC category to stake out massive claims on 
behalf of peasant­middle groups (particularly in some southern states) are readily transferred 
to discredit the more modest measures for SC and ST. 
If caste has displayed unforeseen durability, it has not remained unchanged. Relations between 
castes are increasingly independent and competitive, less interdependent and cooperative. 
"Horizontal" solidarity and organization within caste groups have grown at the expense of "vertical" 
integration among the castes of a region. The concerns of the local endogamous units are 
transformed as they are linked in wider networks and expressed through other forms of organization 
­­ caste associations, educational societies, unions, political parties, religious societies. 
If secularism is defined in terms of the elimination of India's compartmental group structure in favor 
of a compact and unitary society, then the compensatory discrimination policy may indeed have 
impeded secularism. But one may instead visualize not the disappearance of communal groups but 
their transformation into components of a pluralistic society in which invidious hierarchy is 
discarded while diversity is accommodated. In this view compensatory discrimination policy 
contributes to secularism by reducing group disparities and blunting hierarchic distinctions. 
The development of a secular society in which the hierarchic ordering of groups is not recognized 
and confirmed in the public realm is a departure from older Indian patterns. The compensatory 
discrimination policy is a major component in the disestablishment of a central part of the 
traditional way of ordering the society. But this break with the past is itself conducted in a familiar 
cultural and institutional style. The administration of preference programs reflects older patterns in 
the fecund proliferation of overlapping schemes, the fragmentation of responsibility, and the broad 
decentralization of authority under the aegis of unifying symbols. When these policies encounter 
the judiciary, what purports to be a pyramidal hierarchy establishing fixed doctrine turns out to be a 
loose collegium presiding over an open textured body of learning within which conflicting 
tendencies can be accommodated and elaborated. 
The compensatory principle of substantive equality is added to the constitutional scheme of formal 
equality, but it does not displace it. This juxtaposition of conflicting principles is an instance of 
what Glanville Austin admiringly describes as one of 
India's original contributions to constitution­making, [that is] accommodation . . . the 
ability to reconcile, to harmonize, and to make work without 
­561­ 
changing their content, apparently incompatible concepts­­at least concepts that appear 
conflicting to the non­Indian, and especially to the European or American observer. Indians 
can accommodate such apparently conflicting principles by seeing them at different levels of 
value, or, if you will, in compartments not watertight, but sufficiently separate so that a 
concept can operate freely within its own sphere and not conflict with another operating in a 
separate sphere . . . . With accommodation, concepts and viewpoints, although seemingly 
incompatible, stand intact. They are not whittled away by compromise but are worked 
simultaneously.  14.   
The expectation that these principles could co­exist has been fulfilled. The compensatory principle 
has been implemented, but it has not been allowed to overshadow or swallow up opposing 
commitments to merit and to formal equality. 
The compensatory discrimination policy is not to be judged only for its instrumental qualities. It is 
also expressive: through it Indians tell themselves what kind of people they are and what kind of 
nation. These policies express a sense of connection and shared destiny. The groups that occupy the 
stage today are the repositories and transmitters of older patterns. Advantaged and disadvantaged 
are indissolubly bound to one another. There is a continuity between past and future that allows past 
injustices to be rectified. Independence and nationhood are an epochal event in Indian civilization 
which makes possible a controlled transformation of central social and cultural arrangements. 
Compensatory discrimination embodies the brave hopes of India reborn that animated the freedom 
movement and was crystallized in the Constitution. If the reality has disappointed many fond hopes, 
the turn away from the older hierarchic model to a pluralistic participatory society has proved 
vigorous and enduring. 
D. EX ORIENT LUX? 
The Indian experience has obvious resonances with the contemporary American encounter with 
"affirmative action" and/or "reverse discrimination." India and the United States are two of a 
number of democratic countries which employ policies of formal preference for members of 
disadvantaged groups for the purpose of erasing group disparities.  15   I have tried to portray the 
Indian policies in the setting of Indian society and in its distinct political and legal culture. It would 
be 
____________________ 
14. Austin 1966: 317­18. 
15 Others are programs for Oriental Jews in Israel (see n. 7 above) and programs for the Burakumin 
in Japan (see Hah and Lapp 1978 ; DeVos and Wagatsuma 1966.) Cf. 
­562­ 
inappropriate to assume that patterns found there would accompany cognate measures in 
very different settings. But Indian policies are more than expressions of unique and peculiar 
Indian conditions. The government that conducts them shares basic architectural features that 
are found elsewhere (constitutionalism, the rule of law, a "neutral" civil service, an 
independent judiciary, etc.,); these policies are oriented to values that are widely shared. 
India and the United States are both large, complex, democratic societies with heterogeneous 
populations in which sizable descent groups were in the past systematically suppressed, deliberately 
deprived of resources, skills, and opportunities, and branded as socially inferior. Each is embarked 
on an attempt to offset this heritage of invidious distinction by policies of preferential treatment for 
these historically disadvantaged groups. In both of these countries preferential policies are framed 
by a commitment to formally equal treatment of individuals. In India as in the United States courts 
occupy an extraordinarily prominent position as forums for individual grievances, monitors of 
governmental policy, and producers of political symbols. In both societies courts are called upon to 
play a key role in resolving the tensions between commitments to formal equality and to 
compensatory preference. Notwithstanding the colossal differences in history, culture, and wealth 
that separate the two countries, there are enough parallels and affinities to suggest that India's 
encounter with the perplexities of these policies may illuminate our own. 
At the least the Indian experience should make us wary of hasty and tendentious inferences about 
the necessary characteristics of compensatory discrimination policies. It displays a range of options 
that are open. If it alerts us to some general tendencies in such policies, it suggests ways of 
tempering or counteracting these features. It cautions us about the difficulties of assessing such 
programs. 
Perhaps the most important lesson is that there is no single big lesson. That is, there is no large 
general consequence that flows inexorably from embrace of a principle of compensatory preference. 
Compensatory discrimination does not necessarily extinguish commitments to merit and 
evenhandedness, and it does not necessarily metastasize into a comprehensive system of communal 
quotas. On the other hand, it does not automatically produce the sought­after redistribution and it is 
not costless. 
The Indian experience demonstrates that the compensatory commit­ 
____________________ 
Malaysia's extension of similar preferential devices into a comprehensive scheme of communal 
quotas favoring its politically dominant but economically backward Malay majority ( Suffian 
1972 ; MacDougall 1968 ; Ratnam 1965 ). 
­563­ 
ment can co­exist with commitments to formal equality. Tensions at the level of principle can 
be contained; conflicting principles can be accommodated in a relatively stable pattern. The 
effects of the compensatory policy cannot be predicted from analysis of the principles that 
animate it. Nor can the performance of the policy be judged by the legal doctrine or public 
discourse that accompanies it. The symbolic dimensions of the policy are important in their 
own right, but they do not necessarily reflect crucial successes and failures in the design and 
implementation of programs. 
The Indian experience indicates that recognition of group identity in public life does not inevitably 
erode evenhandedness, generate invidious exclusions, or justify oppression. This instance of a 
relatively benign jurisprudence of group identity should remind us that South Africa and Nazi 
Germany are not the most relevant examples of the use of group membership as a category for 
policy. But it should also be clear that the development of authoritative official learning on the 
question of group identity is an unavoidable concomitant of compensatory preference policies. 
However commonsensical the categories may appear at first blush, programs which turn on descent 
or ethnic categories inevitably encounter questions of who is included within these categories. 
Where descent, affiliation, avowal, and public regard coincide, such questions are easy to answer; 
where they diverge, we are drawn into a small forest of less tractable puzzles. These puzzles reflect 
the overlap and indeterminacy of social identities; their malleability is stimulated by the incentive 
offered by such programs to align identities with official categories. These tendencies appear in 
India, where these policies could address bounded corporate groups; surely they are no less present 
in the United States, where identities such as Black or Hispanic are less bounded and more diffuse. 
The development of such a jurisprudence reflects in miniature the irony of employing group identity 
in the service of reducing disparities among groups. To the extent that the policy aims to dissolve or 
diffuse the boundaries between groups, application of rigid categories based on "objective" criteria 
should be eschewed in favor of determinations that recognize movement, overlap, and voluntarism. 
It requires more resources to administer flexible standards than mechanical ones; and the 
administration of flexible standards broadcasts the problematic quality of group identity and points 
to the relative character of the selection of beneficiaries. 
Questions about which individuals are within the beneficiary groups flow into the problem of 
selection ­­ why these groups and not others? ­­ and thus into the justification of the programs. If 
justification is logically prior to group selection, in practice it is likely to be the other 
­564­ 
way around. Some group is widely acknowledged to require or deserve special treatment. The 
emergent response to that paradigm case ­"untouchables" in India or Blacks in the United 
States ­­ is summarized and codified in a set of justifications which present both barriers and 
precedents for other claimants. Once some group histories or conditions are accepted as 
justifying preferential treatment, there are inevitable questions of comparison. Are the 
deprivations of forced migration or peonage or religious persecution comparable to those of 
slavery? How should incidents of personal misfortune ­­ family handicap, disease, accident, 
disaster, victimization­­compare with those inflicted on whole groups? Should preference be 
triggered by the need of the victims? or by the accountability of others for their deprivations? 
or by the potential for promotion of general well­being? Is there to be a single line separating 
those deserving of preferential treatment? Or should it be modulated by degrees? 
Once deserving beneficiaries are selected and levels of benefits set, the system is attended by other 
ironies that can be displayed with a simple numerical example. Let us take as a paradigm case a 
selection for some coveted benefit in limited supply ­­ e.g., medical college admission. Let us 
suppose that there is an uncontroverted way of measuring ability to perform as a medical student 
and that there is general agreement that it is good to choose medical students on the basis of this 
test. Suppose that there are twenty available places and that two of them are reserved for members 
of "backward class" X. Hundreds of students take the test, including dozens of Xs. Suppose that the 
highest ranking X is number 6 and the next highest is number 46. In accordance with the 
reservation, number 20 is deprived of the twentieth seat and it is given to number 46 instead. 
(This example presents the basic preferential mechanism in the starkest form. The edges may be 
softened in various ways: there may be less than full agreement about the suitability of the test; the 
selection may involve a summing of incommensurables; and student performance may not be 
unambiguously ranked; who the Xs are may be open­ended; the reservation may be merely a 
guideline; and number 46 might be included without excluding number 20, presumably diluting ever 
so slightly the training given to each student.) 
This example points to some of the salient features of many compensatory discrimination schemes. 
First, there is the tendency of the program to cast a symbolic shadow much longer than the program 
itself, producing demoralization and resentment far disproportionate to the benefits delivered. In our 
example, not only is number 20 displaced, but numbers 21 to 45 are "passed over" and someone 
with a lower score is chosen before them. Those below 46 may come away feeling that their 
­565­ 
failure to gain the prize was due to unfair preference for the Xs. Even the successful non­Xs 
may resent the additional jeopardy to which their ambitions were exposed. 
Second, it is likely that all of the Xs are stigmatized as the recipients of preferential treatment, even 
though in our example 50% of the Xs won their place on "merit." Third, in contrast to the non­Xs 
who see ubiquitous and unrestrained preference, to the dozens of Xs who didn't make it the amount 
of preference may seem paltry. Fourth, anyone who seeks admission to a medical college may be 
deprived in the select company of his fellow applicants, but ranks high in education and advantages 
compared with the whole society and especially with the Xs. 
Ironically, benefits justified by the need and deprivation of the Xs tend to flow to those Xs who are 
least needful and deprived. Number 46 is elevated because his selection will benefit the Xs ­­ by a 
percolation of tangible benefits, by representation, by their vicarious participation in his success, or 
in some other way. In India, where programs deal with endogamous corporate groups, the diffusion 
of benefits throughout the group is problematic; in the United States, where beneficiary classes are 
less bounded and cohesive, the diffusion of benefits is even more problematic. 
Similarly, there is the converse of this problem. In our example, the direct cost of the preference was 
borne by number 20, who lost his opportunity to attend medical college. To the extent that the 
compensatory discrimination policy is based upon promotion of social benefit or recompense for 
historic wrongs, the transfer from number 20 to number 46 is a payment for a collective good. These 
costs should be spread to the whole society (or some appropriate segment) rather than being borne 
by those individuals who happen to be located in the path of the policy. 
Like all remedial redistributions, compensatory discrimination imposes its own arbitrariness and 
unfairness. This is amplified to the extent that benefits are large and non­decomposable higher­
echelon benefits (i.e., open only to those who have acquired demanding preliminary qualifications). 
They can be diminished by concentrating on benefits that can be spread more generally (like early 
educational opportunities) and the cost of which can be monetized and spread widely. Such 
widespread costs and benefits impinge less dramatically on individual expectations and are far less 
provocative of resentment and resistance. 
While benefits and costs should be spread, they must also be bounded. Compensatory 
discrimination not only creates resentment of particular redistributions, but arouses fears that, by 
broadening the 
­566­ 
categories and modulating benefits, it will gravitate into a comprehensive system of group 
allotments. To prevent such a slide it is important that boundaries be clearly established ­­ for 
what groups and for what amount of time ­­ with self­liquidating devices built into preference 
programs. 
Insight into the cumulative structural character of inequality has taught us that a regime of formal 
equality tends to perpetuate disparities we find intolerable. But if we find formal equality an 
insufficient remedy, we shrink from the uncharted plunge into comprehensive individualized 
substantive justice. Compensatory discrimination offers a way to leaven our formalism without 
entirely abandoning its comforts. The Indian example is instructive: India has managed to pursue a 
commitment to substantive justice without allowing that commitment to dissolve competing 
commitments to formal equality that make law viable in a diverse society with limited consensus. 
The Indian experience displays a principled eclecticism that avoids suppressing the altruistic 
fraternal impulse that animates compensatory policies, but that also avoids being enslaved by it. 
From afar it reflects to us a tempered legalism ­­ one which we find more congenial in practice than 
in theory. 
­567­ 
Appendix 
RELEVANT PROVISIONS OF THE INDIAN CONSTITUTION [AS OF 1975 EXCEPT AS 
NOTED] 
PREAMBLE 
WE, THE PEOPLE OF INDIA, having solemnly resolved to constitute India into a SOVEREIGN 
DEMOCRATIC REPUBLIC and to secure to all its citizens: 

JUSTICE, social, economic and political;
LIBERTY of thought, belief, faith and worship;
EQUALITY of status and of opportunity; and to promote among them all
FRATERNITY assuring the dignity of the individual and the unity of the Nation;
IN OUR CONSTITUENT ASSEMBLY this twenty­sixth day of November, 1949, do 
HEREBY ADOPT, ENACT AND GIVE TO OURSELVES THIS CONSTITUTION. 
Part III Fundamental Rights 
Art. 14. The State shall not deny to any person equality before the law or the equal protection of the 
laws within the territory of India.Art. 15. (1) The State shall not discriminate against any citizen on 
grounds only of religion, race, caste, sex, place or birth or any of them. 
2.  No citizen shall, on grounds only of religion, race, caste, sex, place of birth or any of them, be 
subject to any disability, liability, restriction or condition with regard to­­ 
a.  access to shops, public restaurants, hotels and places of public entertainment; or 
b.  the use of wells, tanks, bathing ghats, roads and places of public resort maintained wholly 
or partly out of State funds or dedicated to the use of the general public. 
3.  Nothing in this article shall prevent the State from making any special provision for women and 
children. 
4.  Nothing in this article or in clause (2) of article 29 shall prevent the State from making any 
special provision for the advancement of any socially and educationally backward classes of 
citizens or for the Scheduled Castes and the Scheduled Tribes.    1. 

____________________ 
1. Added by the Constitution (First Amendment) Act, 1951, s.2. 
­569­ 
Art. 16. (1) There shall be equality of opportunity for all citizens in matters relating to 
employment or appointment to any office under the State. 
2.  No citizen shall, on grounds only of religion, race, caste, sex, descent, place of birth, residence 
or any of them, be ineligible for, or discriminated against in respect of, any employment or 
office under the State. 
3.  Nothing in this article shall prevent Parliament from making any law prescribing, in regard to a 
class or classes of employment or appointment to an office under the Government of, or any 
local or other authority within, a State or Union territory, any requirement as to residence 
within that State or Union territory prior to such employment or appointment. 
4.  Nothing in this article shall prevent the State from making any provision for the reservation of 
appointments or posts in favour of any backward class of citizens which, in the opinion of the 
State, is not adequately represented in the services under the State. 
5.  Nothing in this article shall affect the operation of any law which provides that the incumbent 
of an office in connection with the affairs of any religious or denominational institution or any 
member of the governing body thereof shall be a person professing a particular religion or 
belonging to a particular denomination. 
Art. 17. "Untouchability" is abolished and its practice in any form is forbidden. The enforcement of 
any disability arising out of "Untouchability" shall be an offence punishable in accordance with 
law.Art. 29. (2) No citizen shall be denied admission into any educational institution maintained by 
the State or receiving aid out of State funds on grounds only of religion, race, caste, language or any 
of them.Art. 32. (1) The right to move the Supreme Court by appropriate proceedings for the 
enforcement of the rights conferred by this Part is guaranteed. 
2.  The Supreme Court shall have power to issue directions or orders or writs, including writs in 
the nature of habeas corpus, mandamus, prohibition, quo warranto and certiorari, whichever 
may be appropriate, for the enforcement of any of the rights conferred by this Part. 
3.  Without prejudice to the powers conferred on the Supreme Court by clauses (1) and (2), 
Parliament may by law empower any other court to exercise within the local limits of its 
jurisdiction all or any of the powers exercisable by the Supreme Court under clause (2). 
4.  The right guaranteed by this article shall not be suspended except as otherwise provided for by 
this Constitution. 
­570­ 
Part IV Directive Principles of State Policy 
Art. 37. The provisions contained in this Part shall not be enforceable by any court, but the 
principles therein laid down are nevertheless fundamental in the governance of the country and it 
shall be the duty of the State to apply these principles in making laws.Art. 46. The State shall 
promote with special care the educational and economic interests of the weaker sections of the 
people, and, in particular, of the Scheduled Castes and the Scheduled Tribes, and shall protect them 
from social injustice and all forms of exploitation. 
Part VI The States 
Art. 226. (1) Notwithstanding anything in article 32, every High Court shall have power, throughout 
the territories in relation to which it exercises jurisdiction, to issue to any person or authority, 
including in appropriate cases any Government, within those territories directions, orders or writs, 
including writs in the nature of habeas corpus, mandamus, prohibition, quo warranto and 
certiorari, or any of them, for the enforcement of any of the rights conferred by Part III and for any 
other purpose. 
Part XV Election 
Art. 325. There shall be one general electoral roll for every territorial constitutency for election to 
either House of Parliament or to the House or either House of the Legislature of a State and no 
person shall be ineligible for inclusion in any such roll or claim to be included in any special 
electoral roll for any such constitutency on grounds only of religion, race, caste, sex or any of them. 
Part XVI Special Provisions Relating to Certain Clauses 
Art. 330. (1) Seats shall be reserved in the House of the People for­­ 
a.  the Scheduled Castes; 
b.  the Scheduled Tribes except the Scheduled Tribes­­ 
i.  in the tribal areas of Assam; 
ii.  in Nagaland; 
iii.  in Meghalaya; 
iv.  in Arunachal Pradesh; and 
v.  in Mizoram; and 
c.  the Scheduled Tribes in the autonomous districts of Assam. 
2.  The number of seats reserved in any State or Union territory for the Scheduled Castes or the 
Scheduled Tribes under clause (1) shall bear, as nearly as may be, the same proportion to the 
total number of seats allotted to that State or Union territory in the House of the People as the 
population of the Scheduled Castes in the State or Union territory or of the Scheduled 
­571­ 

Tribes in the State or Union territory or part of the State or Union territory, as the case may be, 
in respect of which seats are so reserved, bears to the total population of the State or Union 
territory. 
3.  Notwithstanding anything contained in clause (2), the number of seats reserved in the House of 
the People for the Scheduled Tribes in the autonomous districts of Assam shall bear to the total 
number of seats allotted to that State a proportion not less than the population of the Scheduled 
Tribes in the said autononmous districts bears to the total population of the State. 
Art. 332. (1) Seats shall be reserved for the Scheduled Castes and the Scheduled Tribes, except 
the Scheduled Tribes in the tribal areas of Assam, in Nagaland and in Meghalaya, in the 
Legislative Assembly of every State. 
2.  Seats shall be reserved also for the autonomous districts in the Legislative Assembly of the 
State of Assam. 
3.  The number of seats reserved for the Scheduled Castes or the Scheduled Tribes in the 
Legislative Assembly of any State under clause (1) shall bear, as nearly as may be, the same 
proportion to the total number of seats in the Assembly as the population of the Scheduled 
Castes in the State or of the Scheduled Tribes in the State or part of the State, as the case may 
be, in respect of which seats are reserved, bears to the total population of the State. 
Art. 334. Notwithstanding anything in the foregoing provisions of the Part, the provisions of this 
Constitution relating to­­ 
a.  the reservation of seats for the Scheduled Castes and the Scheduled Tribes in the House of the 
People and in the Legislative Assemblies of the States; and 
b.  the representation of the Anglo­Indian community in the House of the People and in the 
Legislative Assemblies of the States by nomination, 
2.
shall cease to have effect on the expiration of a period of forty years    
 from the commencement of 
this Constitution: 
Provided that nothing in this article shall affect any representation in the House of the 
People or in the Legislative Assembly of a State until the dissolution of the then existing 
House or Assembly, as the case may be. 
Art.335. The claims of the members of the Scheduled Castes and the Scheduled Tribes shall be 
taken into consideration, consistently with the maintenance of efficiency of administration, in the 
making of appointments to services and posts in connection with the affairs of the Union or of a 
State. 
____________________ 
2. Originally ten years, but extended for successive ten­year periods by the Constitution (Eighth 
Amendment) Act, 1959; the Constitution (Twenty­third Amendment) Act, 1969; and the 
Constitution (Forty­fifth Amendment) Act, 1980. 
­572­ 
Art. 338. (1) There shall be a Special Officer for the Scheduled Castes and Scheduled Tribes to 
be appointed by the President. 
2.  It shall be the duty of the Special Officer to investigate all matters relating to the safeguards 
provided for the Scheduled Castes and Scheduled Tribes under this Constitution and report to 
the President upon the working of those safeguards at such intervals as the President may 
direct, and the President shall cause all such reports to be laid before each House of Parliament. 
3.  In this article references to the Scheduled Castes and Scheduled Tribes shall be construed as 
including references to such other backward classes as the President may, on receipt of the 
report of a Commission appointed under clause (1) of article 340, by order specify and also to 
the Anglo­Indian community. 
Art. 339. (1) The President may at any time and shall, at the expiration of ten years from the 
commencement of this Constitution by order appoint a Commission to report on the administration 
of the Scheduled Areas and the welfare of the Scheduled Tribes in the States.The order may define 
the composition, powers and procedure of the Commission and may contain such incidental or 
ancillary provisions as the President may consider necessary or desirable. 
2.  The executive power of the Union shall extend to the giving of directions to a State as to the 
drawing up and execution of schemes specified in the direction to be essential for the welfare 
of the Scheduled Tribes in the State. 
Art. 340. (1) The President may by order appoint a Commission consisting of such persons as he 
thinks fit to investigate the conditions of socially and educationally backward classes within the 
territory of India and the difficulties under which they labour and to make recommendations as to 
the steps that should be taken by the Union or any State to remove such difficulties and to improve 
their condition and as to the grants that should be made for the purpose by the Union or any State 
and the conditions subject to which such grants should be made, and the order appointing such 
Commission shall define the procedure to be followed by the Commission. 
2.  A Commission so appointed shall investigate the matters referred to them and present to the 
President a report setting out the facts as found by them and making such recommendations as 
they think proper. 
3.  The President shall cause a copy of the report so presented together with a memorandum 
explaining the action taken thereon to be laid before each House of Parliament. 
­573­ 
Art. 341. (1) The President may respect to any State or Union territory, and where it is a State 
after consultation with the Governor thereof, by public notification, specify the castes, races or 
tribes or parts of or groups within castes, races or tribes which shall for the purposes of this 
Constitution be deemed to be Scheduled Castes in relation to that State or Union territory, as 
the case may be. 
2.  Parliament may by law include in or exclude from the list of Scheduled Castes specified in a 
notification issued under clause (1) any caste, race or tribe or part of or group within any caste, 
race or tribe, but save as aforesaid a notification issued under the said clause shall not be varied 
by any subsequent notification. 
Art. 342. (1) The President may with respect to any State or Union territory, and where it is a State, 
after consultation with the Governor thereof, by public notification, specify the tribes or tribal. 
communities or parts of or groups within tribes or tribal communities which shall for the purposes 
of this Constitution be deemed to be Scheduled Tribes in relation to that State or Union territory, as 
the case may be. 
2.  Parliament may by law include in or exclude from the list of Scheduled Tribes specified in a 
notification issued under clause (1) any tribe or tribal community or part of or group within 
any tribe or tribal community, but save as aforesaid a notification issued under the said clause 
shall not be varied by any subsequent notification. 
Part XIX Miscellaneous 
Art. 366. [Definitions] 
24.  "Scheduled Castes" means such castes, races or tribes or parts of or groups within such castes, 
races or tribes as are deemed under article 341 to be Scheduled Castes for the purposes of this 
Constitution. 
25.  "Scheduled Tribes" means such tribes or tribal communities or parts of or groups within such 
tribes or tribal communities as are deemed under article 342 to be Scheduled Tribes for the 
purposes of this Constitution. 
­574­ 
Bibliography 
BOOKS, MONOGRAPHS, ARTICLES 

Abraham, Amrita
1978 "A Report from Marathwada." Economic and Political Weekly 13 : 1536­40. 
Adam, William
1840 The Law and Custom of Slavery in British India in a Series of Letters to Thomas Powell 
Buxton, Esq. Boston: Weeks, Jordan & Co. 
Adler, Chaim, Reuven Kahane, and Amy Avgar
1975 The Education of the Disadvantaged in Israel: Comparisons, Analysis and Proposed Research. 
Jerusalem: The Hebrew University of Jerusalem School of Education. 
Agarwal, C. B.
1934 The Harijans in Rebellion. Bombay: Taraporevala Sons & Co. 
Agarwal, R. N.
1964 Inter­Caste Tensions Among Children. Agra: Lakshmi Narain Agarwal. 
Aggarwal, Partap C.
1980 Halfway to Equality. Forthcoming. 
Aggarwal, Partap C., and Mohammed Siddiq Ashraf
1976 Equality through Privilege: A Study of Special Privileges of Scheduled Castes in Haryana. 
New Delhi: Shri Ram Centre for Industrial Relations and Human Resources. 
A "Harijan" (pseud.)
c. 1946 Gandhi or Ambedkar. Madras: Gandhi Era Publications. 46 pp. 
Ahir, D. C.
1962 Babasaheb's Message. Delhi: Institute of Buddhist Thought. 16 pp. 
Ahmed, Karuna
1978 "Towards Equality: Consequences of Protective Discrimination." Economic and Political 
Weekly 13:69­72. 
Aiyappan, A.
1965 Social Revolution in a Kerala Village: A Study in Culture Change. New York: Asia Publishing 
House. 
Alexandrowicz, Charles Henry
1957 Constitutional Developments in India. Bombay: Oxford University Press. 
­575­ 
All­India Backward Classes Federation
1954 Memorandum to the Backward Classes Commission. Amravati: Shri Shivaji Printing 
Press. 
1955 Report of Working of the All­India Backward Classes Federation from the Date of Its 
Formation up to 30th September, 1955. 10 pp. 
1955 Report. Amravati: Shri Shivaji Printing Press. 
1956­59 Reports. New Delhi: New India Press. 
1961 Review. New Delhi: New India Press. 
All­India Bar Committee
1953 Report. Delhi: Manager of Publications. 
All­India Congress Committee
n.d. Indian National Congress 1930­34: Being the Resolutions Passed by the Congress . . . During 
the Period Between January 1930 to September 1934. . . . Allahabad. 
All­India Federation of Scheduled Castes, Scheduled Tribes and Other Backward Classes
1964 Memorandum for the Implemenation of the Report of the Backward Classes Commission, 
Government of India. New Delhi: S. M. Kaushik. 7 pp. 
All­India Scheduled Castes Pratinidhi Sabha (held at Nanded)
1965 The Four Immediate Needs of Twelve Crore Supressed Human Beings in India (Resolutions 
Passed Unanimously). Hyderabad: B. Sham Sunder. 
Ambedkar, Bhimrao Ramji
1943 Mr. Gandhi and the Emancipation of the Untouchables. Bombay: Thacker & Co. 
1945 Annihilation of Castes with a Reply to Mahatma Gandhi. 3rd ed Bombay: Bharat Bhushan 
Printing Press. 
1946 What Congress and Gandhi have Done to the Untouchables. 2d ed Bombay: Thacker & Co. 
1948 The Untouchables: Who Were They and Why They Became Untouchables? New Delhi: Amrit 
Book Co. 
1947 States and Minorities: What Are Their Rights and How to Secure Them in the Constitution of 
Free India. Bombay: Thacker & Co. 
1964 Ranade, Gandhi and Jinnah (Address Delivered on the 101st Birthday Celebration of M. G. 
Ranade). Jullundur City: Bheem Patrika Publications. 
1968The Annihilation of Castes with a Reply to Mahatma Gandhi [1945] and Castes in India: Their  
Mechanism, Genesis and Development [1916]. [Reprint with an introduction by Bhagwan Das] 
Jullundur City: Bheem Patrika Publications. 
Ambedkar, Bhimrao Ramji, and Rao Bahadur R. Srinivasan
1932 "Supplementary Memorandum on the Claims of the Depressed Classes for Social 
Representation." Appendix 7 in Indian Round Table Conference (Second Session) 7th September, 
1931­1st December 1931. Proceedings of Federal Structure Committee and Minorities Committee 
(Vol. 3). Calcutta: Government of India. 
Anandan, P. I.
1911 "The Coronation and the Depressed Classes," Hindustan Review 24:529­32. 
­576­ 
Anant, Santokh Singh
1972 The Changing Concept of Caste in India. Delhi: Vikas Publications. 
1978 "Changing Caste Hindu Attitudes Toward Harijans­­A Followup After Four Years." In G. R. 
Gupta, ed., Cohesion and Conflict in Modern India. Durham, N. C.: Carolina Academic Press. 
Angell, James W.
1934 "Reparations." In Encyclopaedia of the Social Sciences 13:300­8. New York: MacMillan. 
Anonymous
c. 1913 The Depressed Classes: An Enquiry into Their Conditions and Suggestions for Their Uplift. 
Madras: G. A. Natesan & Co. 
1935­37 The Depressed Classes: A Chronological Documentation. Part I, Ranchi: Rev. Fr. J.Jans, 
S.J., Catholic Press. Parts II­­VII, Kurseong: St. Mary's College. 
Aryamane, R. V.
1964­66 "Backward Classes in the Indian Constitution," Political Scientist (Ranchi) 1­2:27­36. 
Association of Indian Universities
1977 Handbook of Medical Education. New Delhi: Association of Indian Universities. 
Austin, Granville
1966 The Indian Constitution: Cornerstone of a Nation. Oxford: Clarendon Press. 
Ayyar, R. S. Vaidyanatha.
1939 After the Reforms, Communal Versus Democratic Psychology. Karur: Maheswari Publishing 
House. 118pp. 
Ayyar, S. V.
1965 A Study of Opinion Regarding the Various Ameliorative Measures to Improve the Conditions 
of Backward and Scheduled Castes. Annamalainagar: Annamalai University. 
Bailey, F. G.
1960 Tribe, Caste and Nation: A Study of Political Activity and Political Change in Highland 
Orissa. Manchester: Manchester University Press. 
1961 '"Tribe' and 'Caste' in India. In Contributions to Indian Sociology, No. 5, edited by Louis 
Dumont and D. Pocock. Paris: Mouton & Co. 
Baines, Athelstan
1912 "Ethnography of India." Grundriss der Indo­Arischen Philologie und Altertumskunde. Vol. 2 
pt. 5. Strassburg: Karl J. Trubner Verlag. 
Balasundaram, S. N.
1972 "Higher Education and Mass Politics." In The Crisis in Higher Education; Proceedings, 
edited by G. D. Parikh. Bombay: Leslie Sawhney Programme of Training for Democracy. 
FriedrichNaumann Stiftung. 
Banaji, D. A.
c. 1933 Slavery in British India. Bombav: D. B. Taraporewala Sons & Co. 2d ed., n. d. 
­577­ 
Banerjee, Anil Chandra
1948 The Making of the Indian Constitution 1939­1947. Vol. 1, Documents. Calcutta: A. 
Mukherjee & Co. 
Banerjee, D. N.
1960 Our Fundamental Rights: Their Nature and Extent (as Judicially Determined). Calcutta: 
World Press. 
Barnabas, A. P., and Subhash C. Mehta.
1965 Caste in Changing India. New Delhi: The Indian Institute of Public Administration. 
Barnett, Marguerite Ross
1976 The Politics of Cultural Nationalism in South India. Princeton: Princetion University Press. 
Barrier, Norman Gerald
1968 "The Punjab Government and Communal Politics, 1870­1908." Journal of Asian Studies  
27:523­39. 
Bastedo, T. G.
1968­69 "Law Colleges and Law Students in Bihar." Law & Society Review 3:269­94. 
Basu, Durga Das
1977 Constitutional Law of India. New Delhi: Prentice­Hall of India. 
Baxi, Upendra
1967 "'The Little Done, The Vast Undone'­­Some Reflections on Reading Granville Austin's The 
Indian Constitution." Journal of the Indian Law Institute 9:323­430. 
1975 "Legal Assistance to the Poor: A Critique of the Expert Committee Report." Economic and 
Political Weekly 10 (July 5):1005­13. 
Beals, Alan R.
1962 Gopalpur: A South Indian Village. New York: Holt, Rinehart & Winston. 
Becker, Theodore L., and Malcolm M. Feeley, eds.
1973 The Impact of Supreme Court Decisions: Empirical Studies. 2d ed. New York: Oxford 
University Press. 
Beidelman, Thomas O.
1959 A Comparative Analysis of the Jajmani System. New York: J. J. Augustin. 
Berreman, Gerald D.
1963 Hindus of the Himalayas. Berkeley: University of California Press. 
Béteille, André
1965a "The Future of the Backward Classes: The Competing Demands of Status and Power." In 
Perspectives [Supplement to the Indian Journal of Public Administration, Vol. XI, no. 1], pp. 1­39. 
Reprinted in India and Ceylon: Unity and Diversity. Edited by Philip Mason. New York: Oxford 
University Press. 1967. 
1965b Caste, Class and Power: Changing Patterns of Stratification on a Tanjore Village. Berkeley: 
University of California Press. 
1974 Studies in Agrarian Social Reform. London: Oxford University Press. 
­578­ 
Bhandare, R. D.
1965 The New Converts to Buddhism and Their Place in Indian Society: A Note Presented to the  
Honorable Prime Minister of India. Bombay. Bound typescript. 55 pp. 
Bhatt, Anil
1975 Caste, Class and Politics: An Empirical Profile of Social Stratification in Modern India. Delhi: 
Manohar Book Service. 
Bittker, Boris
1973 The Case for Black Reparations. New York: Random House. 
Black, Donald
1973 "The Mobilization of Law," Journal of Legal Studies 2:125­49. 
Blackshield, A. R.
1968 "Fundamental Rights and the Economic Viability of the Indian Nation. Part, III: Prospective 
Overruling." Journal of the Indian Law Institute 10:183­240. 
Blair, Harry W.
1980 "Rising Kulaks and Backward Classes in Bihar: Social Change in the Late 1970s," Economic 
and Political Weekly 15 (Jan. 12):64­74. 
Blalock, Hubert M., Jr.
1967 Towards a Theory of Minority­Group Relations. New York: John Wiley & Sons. 
Blaug, Mark, Richard Layard, and Maureen Woodhall
1969 The Causes of Graduate Unemployment in India. London: Allen Lane. 
Blunt, E. A. H.
1931 The Caste System of Northern India: With Special Reference to the United Provinces of Agra 
and Oudh. London: Oxford University Press. 
Blunt, E. A. H., ed.
1938 Social Service in India: An Introduction to Some Social and Economic Problems of the Indian 
People. London: His Majesty's Stationery Office. 
Bombay Presidency Social Reform Association
1939 Social Reform Annual, 1939. Bombay: Bombay Vaibhav Press. 
Bonarji, N. B.
1956 "Untouchability and the Law." Times of India, July 8. 
Bopegamage, Albert, and P. V. Veeraraghavan
1967 Status Images in Changing India. Bombay: Manaktalas. 
Borale, P. T.
1968 Segregation and Desegregation in India: A Socio­Legal Study. Bombay: Manaktalas. 
Bose, Nirmal Kumar
1954 "Who are the Backward Classes?" Man in India 34:89­98. Reprinted in Culture and Society in 
India. Bombay: Asia Publishing House, 1967. 
Bose, Nirmal Kumar, ed.
1960 Data on Caste: Orissa. Memoir No. 7. Calcutta: Anthropological Survey of India. 
­579­ 
Brandt, Richard B., ed.
1962 Social Justice. Englewood Cliffs, N.J.: Prentice­Hall. 
Brass, Paul R.
1966 Factional Politics in an Indian State: The Congress Party in Uttar Pradesh. Berkeley: 
University of California Press. 
Brown, F. H.
1914 "The Social Dawn in India." The Nineteenth Century and After (Twentieth Century) 75:538­48. 
Calabresi, Guido, and Phillip Bobbit.
1978 Tragic Choices. New York: W.W. Norton & Co.. 
Carter, Robert, Dorothy Kenyon, Peter Marcuse, and Loren Miller
1965 Equally. New York: Pantheon Books. 
Central Government Scheduled Castes and Scheduled Tribes Employees Welfare Association
1966 Memorandum on Reservation of Posts in Government Services for the Scheduled Castes and 
the Scheduled Tribes. New Delhi, Feb. 1, 1966 
Chakravarti, Satis Chandra, and Sarojendra Nath Ray, eds.
1933 Brahmo Samaj, the Depressed Classes and Untouchability. Calcutta: Sadharan Brahmo Samaj. 
87 pp. 
Chandidas, R.
1969 "How Close to Equality Are Scheduled Castes?" Economic and Political Weekly 4 (June 
14):975­79. 
Chandidas, R., W. Morehouse, L. Clark, and R. Fontera
1968 India Votes (A Sourcebook on Indian Elections). New York: Humanities Press. 
Chatterjee, B. B., S. S. Singh, and D. R. Yadav
1971 Impact of Social Legislation on Social Change. Calcutta: Minerva Associates. 
Chatterjee, N. C.
1965 Judicial Review and Fundamental Rights under the Constitution of India. Bangalore: 
Bharadwaja Publications. 
Chaudhary, N. D.
1975 "Educational Development among the Scheduled Tribes in Rajasthan." Education Quarterly 
27 ( 3):26­30. 
Chauhan, Brij Raj
1969 "Scheduled Castes and Scheduled Tribes." Economic and Political Weekly 4 (Jan. 25):257­63. 
Chauhan, Brij Raj, G. Narayana, and T. N. Singh
1975 Scheduled Castes and Education. Meerut: Anu Publications. 
Chauhan, D. N. S., and Chopra, S. L.
1969 "A Comparative Study of the Socio­Economic Background of the Members of the Lok Sabha 
Elected from the Reserved and the General Constituencies." Journal of Constitutional and 
Parliamentary Studies 3 ( 2):94­100. 
Chayes, Abram
1976 "The Role of the Judge in Public Law Litigation." Harvard Law Review 89:128­1316. 
­580­ 
Chidambaram P. Pillai
1933 Right of Temple Entry. Nagercoil: Chidambaram Pillai. 
Chintamani, C. Yajnesvara
1901 Indian Social Reform: In Four Parts, Being a Collection of Essays, Addresses, Speeches, etc., 
with an Appendix. Madras: Thompson and Co. Chitnis, Suma 
1974 Literacy and Educational Enrolment Among the Scheduled Castes of Maharashtra. Bombay: 
Tata Institute of Social Sciences. 
1977 A Long Way to Go . . . (Report on a Survey of Scheduled Caste High School and College 
Students in Fifteen States of India). Surat: Centre for Social Studies. Mimeographed. 
Chowdhury, S. R.
1961 "Equality Before the Law in India." Cambridge Law Journal 19: 223­38. 
Cohen, Stephen P.
1969 "The Untouchable Soldier: Caste, Politics and the Indian Army." Journal of Asian Studies  
28:453­68. 
Cohn, Bernard S.
1965 "Anthropological Notes on Disputes and Law in India." American Anthropologist 67 (6), Part 
2:82­122. 
Coupland, R.
1944 The Constitutional Problem in India. ( 3 volumes in 1: Part I. The Indian Problem, 1833­1935; 
Part II. Indian Politics, 1936­1942; Part III. The Future of India.) London: Oxford University Press. 
Dahl, Robert
1961 Who Governs? Democracy and Power in an American City. New Haven: Yale University 
Press. 
Dalal, M. N.
1940 Whither Minorities? Bombay: D. B. Taraporevala Sons and Co. 
Dalton, Dennis
1967 "The Gandhian View of Caste, and Caste After Gandhi." In India and Ceylon: Unity and  
Diversity, edited by Philip Mason. New York: Oxford University Press. 
Damaska, Mirjan
1975 "Structures of Authority and Comparative Criminal Procedure," Yale Law Journal 84:480­544. 
Das, Amal Kumar, and M. K. Raha
1962 West Bengal Tribes: Through Photographs (Special Series No. 6, Bulletin of the Cultural 
Research Institute). Calcutta: Temple Press. 
Das, Amal Kumar, B. K. R. Choudhury, and M. K. Roha
1966 Handbook on the Scheduled Castes and Scheduled Tribes of West Bengal. Special Series No. 
8, Bulletin of the Cultural Research Institue, Tribal Welfare Dept. Calcutta: Temple Press. 
Das, Bhagwan, comp.
1963 Thus Spoke Ambedkar (Selected Speeches). Vol. I. Jullundur: Bheem Patrika Publications. 
1969 Thus Spoke Ambedkar: Selected Speeches. Vol. II. Jullunder: Bheem Patrika Publications . 
­581­ 
Dave, P. K.
1966 "Promotion and Incentives in Public Services." Indian Journal of Public Administration  
12:533­43. 
Davis, David Brion
1966 The Problem of Slavery in Western Culture. Ithaca: Cornell University Press. 
Davis, Kingsley
1951 The Population of India and Pakistan. Princetion: Princeton University Press. 
Derrett, J. Duncan M.
1958 "Statutory Amendments of the Personal Law of the Hindus since Indian Independence." 
American Journal of Comparative Law 7:38093. 
1963 Introduction to Modern Hindu Law. Bombay: Oxford University Press. 
1968 Religion, Law and the State. London: Faber & Faber. 
De Anthony Reuck, and Julie Knight, eds.
1967 Caste and Race: Comparative Approaches. London: J. & A. Churchill. 
Desai, C. C.
1969 "Reservation of Seats in the House of the People and State Assemblies for Scheduled Castes 
and Scheduled Tribes and Nominations of Anglo­Indians thereto: Case for Extension of the Period." 
Journal of Constitutional and Parliamentary Studies 3 ( 2):8693. 
Deshpande, V. V.
1967 "Educational Planning: Some Aspects of Protective Discrimination." In Educational 
Planning: Its Legal and Constitutional Implications in India, edited by G. S. Sharma. Bombay: N. 
M. Tripathi. 
Deva, Shanthi, and C. M. Wagh
1965 Dr. Ambedkar and Conversion. Hyderabad: Dr. Ambedkar Publications Society. 46 pp. 
De George Vos, and Hiroshi Wagatsuma
1966 Japan's Invisible Race: Caste in Culture and Personally. Berkeley: University of California 
Press. 
Dharmalingam, A. M. (Dharmu)
1951 Blasted Hopes: or Democracy in India; Bangalore: The Kural Nilayam. 
Dhavan, Rajeev
1977 The Supreme Court of India: A Socio­Legal Critique of its Juristic Techniques. Bombay: N. M. 
Tripathi. 
Dhaven, S. S., ed.
1964 Aspects of Indian Constitution. Allahabad: Central Book Depot. 
Dolbeare, Kenneth M.
1967 Trial Courts in Urban Politics. New York: John Wiley & Sons. 
Dolbeare, Kenneth M., and Phillip Hammond
1971 The School Prayers Decisions: From Court Policy to Local Practice. Chicago: University of 
Chicago Press. 
Driver, Edwin D.
1963 Differential Fertility in Central India. Princeton: Princeton University Press. 
­582­ 
D'Souza, Victor S.
1969 "Education, Social Structure and Democracy in India." In S. P. Ruhela , ed., Social  
Determinants of Educability in India. New Delhi: Jain Brothers. 
Dua, Veena
1971 "Elections in a Scheduled Caste Neighbourhood: A Punjab Town." Economic and Political 
Weekly 6:2323­27. 
Dubey, Harihar Prasad
1968 A Short History of the Judicial Systems of India and Some Foreign Countries. Bombay: N. M. 
Tripathi. 
Dubey, S. N.
1973 Administration of Social Welfare Programmes in India. Bombay: Somaiya Publications. 
Dumont, Louis
1970 Homo Hierarchicus: The Caste System and Its Implications. London: Weidenfeld and 
Nicholson. 
Dushkin, Lelah
1957 "The Policy of the Indian National Congress Toward the Depressed Classes: An Historical 
Study." Master's thesis, University of Pennsylvania. 
1961 "The Backward Classes: Special Treatment Policy." Economic Weekly 13:1665­68, 1695­1705, 
1729­38. 
1972 "Scheduled Caste Politics." In The Untouchables in Contemporary India, edited by J. Michael 
Mahar. Tucson: University of Arizona Press. 
1974 "The Nonbrahman Movement in Princely Mysore." Ph. D. dissertation, University of 
Pennsylvania. 
1979 "Backward Class Benefits and Social Class in India, 1920­1970." Economic and Political 
Weekly 14:661­67. 
Dutta, Ratna
1969 "The Party Representative in Fourth Lok Sabha." Economic and Politial Weekly 4:179­89. 
Edelman, Murray
1967 The Symbolic Uses of Politics. Urbana: University of Illinois Press. 
Eldersveld, Samuel J., and Bashiruddin Ahmed
1978 Citizens and Politics: Mass Political Behavior in India. Chicago: University of Chicago Press. 
Eldersveld, Samuel J., V. Jagannadham, and A. P. Barnabas
1968 The Citizen and the Administrator in a Developing Democracy: An Empirical Study in Delhi 
State. New Delhi: Indian Institute of Public Administration. 
Fisher, Fred
1919 India's Silent Revolution. New York: Macmillan. 
Fisher, Marguerite J.
1970 "Problems of Implementation in India's Office of the Commissioner for Scheduled Castes." 
Western Political Quarterly 23:715­32. 
­583­ 
Fox, Richard D.
1969 "Varna Schemes and Ideological Integration in Indian Society." Comparative Studies in  
Society and History 11:27­45. 
Friedman, Lawrence M.
1966 "On Legalistic Reasoning­­A Footnote to Weber." Wisconsin Law Review 1966:148­71. 
1967 "Legal Rules and the Process of Social Change." Stanford Law Review 19:786­840. 
Gadbois, George H., Jr.
1968­69 "Indian Supreme Court Judges: A Portrait." Law & Society Review 3:317­36. 
1970a "Indian Judicial Behavior." Economic and Political Weekly 5:1­11. 
1970b "The Supreme Court of India: A Preliminary Report of an Empirical Study." Journal of 
Constitutional and Parliamentary Studies 4 ( 1):33­54. 
1977 "The Emergency, Mrs. Gandhi, the Judiciary and the Legal Culture." Mimeographed. 
Gaikwad, B. K., M. P.
1966 Speech delivered on the occasion of the 5th Session of the Republican Party of India held at 
Delhi on 1 May 1966. 14 pp. 
Gajendragadkar, P. B.
1951 The Hindu Code Bill. Two lectures delivered at the Karnatak University at Dharwar on 12­13 
April 1951. Dharwar: Karnatak University. 52 pp. 
Galanter, Marc
1965 "Secularism East and West." Comparative Studies in Society and History 7: 133­59. 
1968 "Changing Legal Conceptions of Caste." In Structure and Change in Indian Society, edited by 
Milton Singer and Bernard S. Cohn. Chicago: Aldine. 
1971 "Hinduism, Secularism and the Indian Judiciary." Philosophy East and West 21:467­87. 
1972a "The Abolition of Disabilities: Untouchability and the Law." In The Untouchables in 
Contemporary India, edited by J. Michael Mahar. Tucson: University of Arizona Press. An earlier 
version appeared in 1969, Economic and Political Weekly 4:131­70. 
1972b "The Aborted Restoration of 'Indigenous' Law in India." Comparative Studies in Society and 
History 14:53­70. 
1974 "Why the 'Haves' Come Out Ahead: Speculations on the Limits of Legal Change." Law & 
Society Review 9:95­160. 
1975 "Afterword: Explaining Litigation." Law & Society Review 9:347­68 
1976 "Delivering Legality: Some Proposals for the Direction of Research." Law & Society Review 
11:226­46. 
1978 "Who are the Other Backward Classes? An Introduction to a Constitutional Puzzle." Economic 
and Political Weekly 13:1812­28. 
1979 "Compensatory Discrimination in Political Representation: A Preliminary Assessment of 
India's Thirty­Year Experience withReserved Seats in Legislatures." 
­584­ 
Reserved Seats in Legislatures." Economic and Political Weekly 14:437­54. 
1983 "Emerging Patterns of Legal Services in India," Journal of the Indian Law Institute. 
forthcoming, 1983. Forthcoming. "Adjudication, Litigation and Related Phenomena". To appear in a 
handbook of law and social science, to be published under the auspices of the Social Science 
Research Council. 
Galanter, Marc, ed.
1965 Legal Materials for the Study of Modern India. Chicago: Syllabus Division, University of 
Chicago Press. 
Galanter, Marc, and Robert Kidder, eds.
1968­69 "Lawyers in Developing Countries with Special Reference to India". Law & Society Review 
3, Special Issue. 
Galanter, Marc, Frank Palen, and John M. Thomas
1979 "The Crusading Judge: Judicial Activism in Urban Trial Courts." Southern California Law 
Review 52:699­741. 
Gandhi, M. K.
1954a For Workers Against Untouchability. Ahmedabad: Navajivan Publishing House. 34 pp. 
1954b The Removal of Untouchability. Ahmedabad: Navajivan Publishing House. 
1965 My Varnashrama Dharma. Edited by Anand T. Hingorani. Bombay: Bharatiya Vidya Bhavan. 
n.d. Untouchability. Edited by Dewan Ram Parkash. Lahore: Gandhi Publications League. 
Gauba, K. L.
1973 Passive Voices: A Penetrating Study of Muslims in India. New Delhi: Sterling Publishers. 
Ghouse, Mohammad
1969 "Judicial Control of Protective Discrimination." Journal of the Indian Law Institute 11:371­81. 
1973 Secularism, Society and Law in India. Delhi: Vikas Publishing House. 
Ghurye, G. S.
1932 Caste and Race in India. London: Routledge and Kegan Paul. 
1950 Caste and Class in India. 2d ed. Bombay: Popular Book Depot. 
1957 Caste and Class in India. 3rd. ed. Bombay: Popular Book Depot. 
1961 Caste, Class and Occupation. 4th ed. Bombay: Popular Book Depot. 
1963 The Scheduled Tribes. 3d. ed. Bombay: Popular Prakashan. 
1969 Caste, Class and Occupation. 5th ed. Bombay: Popular Prakashan. 
Glazer, Nathan
1975 Affirmative Discrimination: Ethnic Inequality and Public Policy. New York: Basic Books. 
Gledhill, Alan
1964 "The Expansion of the Judicial Process in Republican India." In Some Aspects of Indian Law 
Today: International and Comparative Law Quarterly Supplementary Publication No. 8. 
­585­ 
Gough, E. Kathleen
1960 "Caste in a Tanjore Village." In Aspects of Caste in South India, Ceylon and North­west 
Pakistan, edited by E. R. Leach. Cambridge: Cambridge University Press. 
Gould, Harold
1963 "The Adaptive Functions of Caste in Contemporary Indian Society." Asian Survey 3: 427­38. 
Government of Malaysia
1964 Citizenship of Malaysia. Kuala Lumpur: Government Press. 136 pp. 
Grimshaw, Allen D.
1959 "The Anglo­Indian Community: The Integration of a Marginal Group." Journal of Asian 
Studies 18:227­40. 
Grossman, Joel, and Richard Wells
1972 Constitutional Law and Judicial Policy Making. New York: John Wiley & Sons. 
Grossman, Kurt Richard
1954 Germany's Moral Debt: The German­Israel Agreement. Washington: Public Affairs Press. 
Groves, Harry E.
1963 "Equal Protection of the Laws in Malaysia and India." American Journal of Comparative Law 
12:385­96. 
Gupta, Khadija A.
1971 "General Elections of 1967 in a Small U.P. Town." Economic and Political Weekly 6:1881­86. 
Gupta, Madan Gopal
1964 Aspects of Indian Constitution. 2d ed. Allahabad: Central Book Depot. 
Gupta, Raj Kumar
1969 "Justice: Unequal but inseparate." Journal of the Indian Law Institute 11:57­86. 
Gurdon, P. R. T.
1914 The Khasis. 2d. ed. London: Macmillan and Co. 
Hah, Chong­Do and Christopher C. Lapp
1978 "Japanese Politics of Equality in Transition. The Case of the Burakumin." Asian Survey 
18:487­504. 
Handler, Joel F.
1966 "Controlling Official Behavior in Welfare Administration." In The Law of the Poor, edited by 
Jacobus tenBroek and the Editors of the California Law Review. San Francisco: Chandler 
Publishing Co. 
Hardgrave, Robert L., Jr.
1965 The Dravidian Movement. Bombay: Popular Prakashan. 
1969 The Nadars of Tamilnad: The Political Culture of a Community in Change. Berkeley: 
University of California Press. 
Harijan Sevak Sangh
1953­1970/71 Annual Report. Delhi: Udyogshala Press. 
Harrison, Selig S.
1960 India: The Most Dangerous Decades. Princeton: Princeton University Press. 
­586­ 
Hart, Henry C.
1971 "Parliament and Nation­Building: England and India." In Modern Parliament: Change  
or Decline? edited by Gerhard Loewenberg. Chicago: Aldine­Atherton. 
Harvard Law Review 1954 Note, "Evasion of Supreme Court Mandates in Cases Remanded to State 
Courts Since 1941." Harvard Law Review 67:1251­59. 
Hasan, Saiyid Zafar
1963 Federal Grants and Public Assistance: A Comparative Study of Policies and Programmes in 
U.S.A. and India. Allahabad: Kitab Mahal. 
Havanur, L. G.
1963 "Backward Classes." Mysore Law Journal 1963 ( 2):15­19. 
1965 Specifying the Backward Classes Without the "Caste" Basis. Bangalore: Mysore State 
Backward Classes Welfare Association. 
c 1970 Judicial Blunders: Governmental Sins Against Backward Classes. Speech delivered by L. G. 
Havanur, Advocate, at the Preliminary Meeting of the SC, ST, and BC held Nov. 8, 1970, in 
Bangalore. N.p. 8 pp. 
Hazari (pseud.) 1951 An Indian Outcaste: The Autobiography of an Untouchable. London: 
Bannisdale Press. 
Heimsath, Charles H.
1964 Indian Nationalism and Hindu Social Reform. Princeton: Princeton University Press. 
Heller, Celia S.
1969 Structured Social Inequality: A Reader in Comparative Social Stratification. New York: 
Macmillan. 
Holderness, T. W.
1911 Peoples and Problems of India. London: Williams and Norgate. 
Horowitz, Donald L.
1977 The Courts and Social Policy. Washington, D. C.: The Brookings Institution. 
Huang­Thio, S. M.
1963 "Equal Protection and Rational Classification." Public Law 1963:412­40. 
Hutton, J. H.
1961 Caste in India: Its Nature, Function, and Origins. 3d ed. Bombay: Oxford University Press. 
Imam, Mohammed
1966 "Reservation of Seats for Backward Classes in Public Services and Educational Institutions." 
Journal of the Indian Law Institute 8:44149. 
Indian Institute of Public Administration 1959 Morale in the Public Services: Report of a 
Conference, January 3­4, 1959. New Delhi: New India Press. 
Indian Institute of Public Opinion 1964 Monthly Public Opinion Survey. No. 109­10. 
1966 Monthly Public Opinion Survey. No. 124. 
­587­ 
Irschick, Eugene F.
1969 Politics and Social Conflict in South India: The Non­Brahman Movement and Tamil  
Separatism, 1916­1929. Berkeley: University of California Press. 
1977 Social Movements and the Indian Nationalist Movement: Tamilnad in the 1930's. (Chapter I: 
"The Theory and Practice of Backwardness.") Mimeographed Draft of Chapter I. (forthcoming.) 
Issacs, Harold R.
1965 India's Ex­Untouchables. New York: John Day. 
Jackson, A. M. T.
1907 "A Note on the History of the Caste System." Journal of the Asiatic Society of Bengal (New 
Series) 3:509­15. 
Jagadisan, T. N., and Shyamlal, eds.
1949 Thakkar Bapa: Eightieth Birthday Commemoration Volume. Madras [no publisher given] 
Jain, M. P.
1961 "Reservation of Posts for Scheduled Castes and Article 16(4)­General Manager, Southern 
Railway v. K. Rangachari." Journal of the Indian Law Institute 3:366­71. 
Jennings, Sir Ivor
1953 Some Characteristic of the Indian Constitution. London: Oxford University Press. 
Jha, M. L.
1973 Untouchability and Education. Meerut: Namita Publications. 
Jones, Kenneth W.
1968 "Communalism in the Punjab: The Arya Samaj Contribution." Journal of Asian Studies  
28:39­54. 
Joshi, Barbara
1978 "Scheduled Caste Political Reservations: Who Benefits?" Paper presented to the Annual 
Meeting of the Association for Asian Studies, Chicago. 
Kannan, C. T.
1963 Intercaste and Inter­community Marriages in India. Bombay: Allied Publishers. 
Karve, Irawati
1961 Hindu Society: An Interpretation. Poona: Deccan College. 
1968 Maharashtra: Land and Its People. Maharashtra State Gazetteers, General Series. Bombay: 
Directorate of Government Printing. 
Karve, Irawati, and Y. B. Damle
1963 Group Relations in Village Community. Deccan College Monograph Series, no. 24. Poona: 
Deccan College Post­Graduate and Research Institute. 
Katzenstein, Mary F.
1979 Ethnicity and Equality: the Shiv Sena Party and Preferential Policies in Bombay. Ithaca: 
Cornell University Press. 
Keer, Dhananjay
1954 Dr. Ambedkar: Life and Mission. Bombay: A. V. Keer 
­588­ 
1962 Dr. Ambedkar: Life and Mission. 2d ed. Bombay: Popular Prakashan. 
1964 Mahatma Jotirao Phooley: Father of Our Social Revolution. Bombay: Popular Prakashan. 
Kennedy, Duncan
1973 "Legal Formality." Journal of Legal Studies 2:351­98, 
1976 "Form and Substance in Private Law Adjudication" Harvard Law Review 89:1685­1778. 
Ketkar, Shridhar V.
1909 The History of Caste in India. Ithaca: Taylor and Carpenter. 
1911 An Essay on Hinduism: Its Formation and Future. London: Luzac and Co. 
Khare, R. S.
1972 "Indigenous Culture and Lawyer's Law in India." Comparative Studies in Society and History 
14:71­96. 
Khobrogade, B. D.
1966 Report of the General Secretary, Republican Party of India. New Delhi. 7 pp. 
Khosla, G. D.
1949 Our Judicial System. Allahabad: University Book Agency. 
Kidder, Robert L.
1973 "Courts and Conflict in an Indian City: A Study in Legal Impact." Journal of Commonwealth 
Political Studies 11:121­39. 
1974 "Formal Litigation and Professional Insecurity: Legal Entrepreneurship in South India." Law 
& Society Review 9:11­37. 
Kirpal, Viney
1978 "Higher Education for the Scheduled Castes and Scheduled Tribes." Economic and Political 
Weekly 13:165­69. 
Kirpal, Viney, and Govind S. Kelkar
1976 "Scheduled Castes and Tribes: Are They 'Government Brahmins'?" Illustrated Weekly of India 
97 ( 16):4­9. 
Kochanek, Stanley A.
1968a The Congress Party of India: The Dynamics of One­party Democracy. Princeton: Princeton 
University Press. 
1968b "The Relation between Social Background and Attitudes of Indian Legislators." Journal of 
Commonwealth Political Studies 6:34­53. 
Koppell, Oliver
1966 Note, "Legal Aid in India." Journal of the Indian Law Institute 8:22451. 
Krishna, Gopal
1966 "One Party Dominance­­Development and Trend." Perspectives. Supplement to the Indian 
Journal of Public Administration 12. 
1967 "Electoral Participation and Political Integration." Economic and Political Weekly 2:179­90. 
Krishna K. P. Shetty
1969 Fundamental Rights and Socio­Economic Justice in the Indian Constitution. Allahabad: 
Chaitanya Publishing House. 
­589­ 
Kumar, Dharma
1965 Land and Caste in South India: Agricultural Labour in Madras Presidency in the  
Nineteenth Century. Cambridge: Cambridge University Press. 
Kuppuswamy, B.
1978 Backward Class Movement in Karnataka. Bangalore: Bangalore University. 
Lajpat Lala Rai
1914The Arya Samaj. London; n.p. (1967 rev. ed. by Sri Ram Sharma, A History of the Arya Samaj. 
Bombay: Orient Longman). 
Lakoff, Sanford A.
1964 Equality in Political Philosophy. Cambridge: Harvard University Press. 
Lal, Sheo Kumar, and Umed Raj Nahar
1978 Higher Education: Scheduled Castes and Scheduled Tribes. Jodhpur: Jainsons. 
Lander, Stephen
1965 "'Backward Classes' in the Indian Constitution." Mimeographed. Chicago: University of 
Chicago. 
Latthe, A. B.
1924 Memoirs of His Highness Shri Shahu Chhatrapati Maharaja of Kolhapur. 2 vols. Bombay: 
Times Press. 
Leach, E. R., ed.
1960 Aspects of Caste in South India, Ceylon and North­west Pakistan. Cambridge Papers in Social 
Anthropology, no. 2. Cambridge for the Dept. of Archaeology and Anthropology: University Press. 
Lecky, Robert S., and H. Elliott Wright, eds.
1968 Black Manifesto­­Religion, Racism and Reparations. New York: Sheed and Ward. 
Lewis, Oscar
1958 Village Life in Northern India: Studies in a Delhi Village. New York: Vintage Books. 
Lockard, Duane
1968 Toward Equal Opportunity: A Study of State and Local Antidiscrimination Laws. New York: 
Macmillan. 
Lohia, Ram Manohar
1964 The Caste System. Hyderabad: Navahind. 
Luthera, Ved Prakash
1964 The Concept of the Secular State and India. London: Oxford University Press. 
Lynch, Owen M.
1969 The Politics of Untouchability: Social Mobility and Social Change in a City of India. New 
York: Columbia University Press. 
McCormack, William
1966 Caste and the British Administration of Hindu law. Journal of Asian and African Studies  
1:25­32. 
MacDougall, John Arthur
1968 Shared Burdens: A Study of Communal Discrimination by the PoliticalParties of Malaysia and 
Singapore
­590­ 
Parties of Malaysia and Singapore. Cambridge, Mass.:Harvard University Press. 
Maddison, Angus
1971 Class Structure and Economic Growth: India and Pakistan Since the Moghuls. New York: W. 
W. Norton. 
Madras, Henry
1909 "The Out­Castes of India." The Nineteenth Century 66:1022­31. 
Majumdar, Nabendu Dutta
1955 "The Backward Classes Commission and Its Work." In Social Welfare in India (issued on 
behalf of the Planning Commission). New Delhi: Government of India. 
Malik, Hafeez
1963 Moslem Nationalism in India and Pakistan. Washington. D. C.: Public Affairs Press. 
Malik, Suneila
1979 Social Integration of Scheduled Castes. New Delhi: Adhinav Publications. 
Mandelbaum, David G.
1970 Society in India. 2 Vols. Berkeley: University of California Press. 
1974 Human Fertility in India: Social Components and Policy Perspectives. Berkeley: University of 
California Press. 
Marriott, McKim
1959 "Interactional and Attributional Theories of Caste Ranking." Man in India 39:92­104. 
1960 Caste Ranking and Community Structure in Five Regions of India and Pakistan. Deccan 
College Monograph Series, no. 23. Poona: Deccan College Postgraduate and Research Institute. 
1976 "Hindu Transactions: Diversity Without Dualism." In Transaction and Meaning: Directions in 
the Anthropology of Exchange and Symbolic Behavior, edited by B. Kapferer. Philadelphia: Institute 
for the Study of Human Issues. 
Marriott, McKim, ed.
1955 Village India: Studies in the Little Community. Chicago: University of Chicago Press. 
Marriott, McKim, and Ronald Inden
1974 "Caste Systems." Encyclopaedia Britannica. Macropaedia 3:982­91. 15th ed. Chicago:William 
Benton. 
Mathur, P. C.
1966 "Politics and Anti­Casteism in India." Political Science Review 5 ( 1): 82­93. 
Mayakrishnan, V., and N. Radhakrishnan
1965 "A Critical Note on State of Kerala v. R. Jacob" Madras Law Journal 1965 ( 1):23­28. 
Mayer, Adrian C.
1960 Caste and Kinship in Central India: A Village and Its Region. Berkeley: University of 
California Press. 
­591­ 
Mayhew, Leon H.
1968 Law and Equal Opportunity: A Study of the Massachusetts Commission Against  
Discrimination. Cambridge: Harvard University Press. 
1975 "Institutions of Representation: Civil Justice and the Public." Law & Society Review 9:401­29. 
Mehta, Ashoka
1965a The Politics of Poverty. Feroze Memorial Lecture. New Delhi: Planning Commission. 
1965b Freedom and Justice in a Changing Society. New Delhi: Government of India. 9 pp. 
Mehta, Subhash Chandra
1963 "Persistence of the Caste System: Vested Interest in Backwardness." Quest no. 36, pp. 20­27. 
Merillat, H. C. L.
1967 "The Soundproof Room: A Matter of Interpretation." Journal of the Indian Law Institute 
9:521­46. 
Miller, Robert J.
1966 "Button, Button . . . Great Tradition, Little Tradition, Whose Tradition." Anthropology 
Quarterly 36:26­42. 
Mittal, J. K.
1964 "Special Representation in Legislative Bodies in India." Indian Advocate 4:23­48. 
1965a "Educational Equality and the Supreme Court of India." Indian Advocate 5( 1­2):31­39. 
1965b "Equality of Opportunity in Matters of Public Employment and the Indian Supreme Court." 
Malaya Law Review 7:113­26. 
1965c "Right to Equality and the Indian Supreme Court." American Journal of Comparative Law 
14:422­58. 
Mohapatra, Manindra Kumar
1973 "The Scheduled Caste and the Tribal Political Elites: An Empirical Exploration of the 
Attributes, Ascent and Role Perceptions of 'Token Elites' in an Indian State." International 
Behavioural Scientist 5 ( 4):1­23. 
1974 "College Youth and Legislators' Role in a Modernizing Legislature: Preliminary Report of an 
Empirical Study." Paper prepared for the American Political Science Association's 70th meeting, 
Chicago. 
1978 "The Scheduled Caste Members of the Indian Lok Sabha in 1977: An Empirical Study of 
Developmental Change in Legislative Recruitment." Paper presented to the Seventh Annual 
Conference on South Asia, University of Wisconsin, Madison, November 3­4, 1978. 
Mootham, Sir Orby
1963 "Constitutional Writs in India." In Changing Law in Developing Countries, edited by J. N. D. 
Anderson. London: Allen and Unwin. 
1966­67 "The Indian High Courts Before and After Independence." In South Asia Seminar, 
Contributions to the Study of Indian Law andSociety
­592­ 
Society. Philadelphia: South Asia Regional Studies, University of Pennsylvania. 
Morkhandikar, R. S.
1978 "Marathwada Riots: The Background." Economic and Political Weekly 13:1461­2. 
Morris­Jones, W. H.
1957 Parliament in India. Philadelphia: University of Pennsylvania Press. 
Morrison, Charles.
1972 "Kinship in Professional Relations: A Study of North Indian District Lawyers." Comparative 
Studies in Society and History 14:10025. 
1974 "Clerks and Clients: Paraprofessional Roles and Cultural Identities in Indian Litigation." Law 
& Society Review 9:39­61. 
Mukherjee, B. K.
1967 "Parliamentary and Constitutional Developments: Government's Snap Defeat." Journal of 
Constitutinal and Parliamentary Studies 1 ( 4):55­56. 
Mukherji, Panchanandas
1920 The Indian Constitution . . . and all Relevant Documents Relating to the Indian Constitutional 
Reforms of 1919. Calcutta: Thacker, Spink & Co. 
Murdoch, J., ed.
1896 Caste: Its Supposed Origin; Its History; Its Effects; the Duty of Government, Hindus and 
Christians with Respect to It; and Its Prospects. Compiled from Muir, Max Weber, Sherring, 
Wilson, Monier Williams, Cornish and Others. Madras: The Christian Literature Society for India. 
62 pp. 
Murphy, Arthur, B. Errabi, N. R. M. Menon, Narasimhaswamy Mysore, and B. Sivaramayya
1973 Law and Poverty: Cases and Materials. Bombay: N. M. Tripathi. 
Murphy, Walter
1959 "Lower­Court Checks on Supreme Court Power." American Political Science Review 
53:1017­31. 
Murthy, B. S.
c. 1972 Depressed and Oppressed (Forever in Agony). New Delhi: S. Chand & Co. 
Nair, Kusum
1962 Blossoms in the Dust: The Human Factor in Indian Development. New York: Frederick A. 
Praeger. 
Namboodiripad, E. M. S.
1979 "Caste Conflicts vs. Growing Unity of Popular Democratic Forces." Economic and Political 
Weekly 14:329­36. 
Naoroji, Dadabhai
n. d. Speeches and Writings of Dadabhai Naoroji. 2d ed. Madras: G. A. Natesan & Co. 
Narayana, G.
1974 "The Political Elite among the Scheduled Castes." Journal of Constitutional and 
Parliamentary Studies 8:214­19. 
1978 "Social Background of Scheduled Caste Lok Sabha Members, 1962­71." Economic and  
Political Weekly 13:1603­8. 
­593­ 
1980 "Rule Making for Scheduled Castes: Analysis of Lok Sabha Debates, 1962­71" Economic  
and Political Weekly 15:433­40. 
Narayana K. V. Rao
1970 State and the Harijan in Andhra Pradesh. Ministry of Home Affairs, Research and Policy 
Division. New Delhi. 
Natarajan, D.
1972 Extracts from the All­India Census Reports on Literacy (Census Centenary Monograph No. 9). 
New Delhi: Office of the Registrar General, Ministry of Home Affairs. 
Natarajan, Swaminath
1959 A Century of Social Reform in India. London: Asia Publishing House. 
National Convention of Christian Leaders on the Plight of Christians of Scheduled Caste Origin
1978 The Story of 30 Years Struggle for Justice: A Consolidated Account of 30 Years' Effort in the 
Cause of Christians of Scheduled Caste Origin. Bangalore: National Centre (NBCLC) 
National Council of Applied Economic Research
1970 Survey of Backward Districts of Andhra Pradesh. New Delhi: National Council of Applied 
Economic Research. 
National Council of Educational Research and Training
1971 Educational and Economic Condition, and Employment Position of Eighteen Tribes. New 
Delhi: Delhi Press. 
Nehru, S. S.
1932 Caste and Credit in the Rural Area, A Survey. Calcutta: Longmans, Green & Co. 
Newell, William H.
1961 "The Comparative Study of Caste in India and Japan." Asian Survey 1:3­10. 
Nigam, S. S.
1960 "Equality and the Representation of the Scheduled Classes in Parliament." Journal of the 
Indian Law Institute 2:297­320. 
O'Malley, L. S. S.
1932 Indian Caste Customs. Cambridge: University Press. 
O'Malley, L. S. S., ed.
1941 Modern India and the West: A Study of the Interactions of their Civilizations. London: Oxford 
University Press. 
Omvedt, Gail
1976 Cultural Revolt in a Colonial Society: The Non­Brahman Movement in Western India. 1873 to 
1930. Bombay: Scientific Socialist Education Trust. 
O'Neil, Robert M.
1975 Discriminating Against Discrimination: Preferential Admissions and the DeFunis Case. 
Bloomington: University of Indiana Press. 
Oommen, T. K.
1968 "Strategy for Special Change: A Study of Untouchability." Economic and Political Weekly 
3:933­36. 
1972 Charisma, Stability and Change: An Analysis of the Bhoodan­ Gramdan Movement. New 
Delhi: Thomson Press. 
­594­ 
Ouwerkerk, Louise
1945 The Untouchables of India. London: Oxford University Press. 47 pp. 
Pai V. A. Panandiker
1964 "Classification of Services in the Central Government." Indian Journal of Public 
Administration 10:647­61. 
Palkhivala, N. A. ed.
1973 A Judiciary Made to Measure, Bombay: M. R. Pai. 
Pandya, B. V.
1959 Striving for Economic Equally. Bombay: Popular Book Depot. 
Panikkar, K. M.
1933 Caste and Democracy. London: Hogarth Press. 39 pp. 
1961 Hindu Society at Cross Roads. 3d ed. New York: Asia Publishing House. 
Parvathamma, C.
1971 Politics and Religion: A Study of Historical Interaction between Sociopolitical Relationships  
in a Mysore Village. New Delhi: Sterling Publishers. 
Parvathamma, C., and R. T. Jangam
1969 "India's Scheduled Caste MPs: A Socio­Economic Profile." Journal of Karnatak University­
Social Sciences 5:153­66. 
Patankar, Bharat, and Gail Omvedt
1979 "The Dalit Liberation Movement in Colonial Period." Economic and Political Weekly 
14:409­24. 
Patawardhan, Sunanda
1973 Change Among India's Harijans: Maharashtra­­A Case Study. New Delhi: Orient Longman. 
Pennock, J Roland, and John W. Chapman
1967 Equality. New York: Atherton Press. 
Phadnis, Urmila
1969 "Neo­Buddhists in India and Ceylon." Economic and Political Weekly 4:1897­98. 
Pickett, J. Waskon
1933 Christian Mass Movements in India: A Study with Recommendations. New York: Abingdon 
Press. 
Ponnarangam, V. A.
1958 The Discovery of India's Plan. Madras: n. p. 
Popkin, William D.
1962 "Prematurity and Obiter Dictum in Indian Judicial Thought." Journal of the Indian Law 
Institute 4:231­60. 
Pradhan, M. C.
1966 The Political System of the Jats of Northern India. Bombay: Oxford University Press. 
Prakash, Kailash
1967 "The Annual Assessment and Promotion Systems in Indian Bureaucracy." Indian Journal of 
Public Administration 13:268­82. 
Pyarelal (Nair)
1932 The Epic Fast. Ahmedabad: Navajivan Publishing House. 
­595­ 
Rachiah, N.
1962 Rachiah's Speeches. Bangalore: Government Press. 
Radhakrishnan, N.
1964 "Reservation to the Backward Classes." Indian Yearbook of International Affairs 13:293­345. 
1965 "Units of Social, Economic and Educational Backwardness: Caste and Individual." Journal of 
the Indian Law Institute 7:262­72. 
1966 "Chitralekha v. State of Mysore:Art. 15(4)­­Meaning of Socially and Educationally Backward 
Classes." Jaipur Law journal 6:11321. 
Rajagopala N. Aiyangar
1937 The Government of India Act, 1935, with a Commentary, Critical and Explanatory. Madras: 
Madras Law Journal Office. 
Rajagopalachari, C.
1933 Plighted Word. Being an Account of the History of the Untouchability Abolition and Temple 
Entry Bills. Delhi: Servants of Untouchables Society. 34 pp. 
1946 Ambedkar Refuted. Bombay: Hind Kitabs. 39 pp. 
Rajah, M. C.
1925 The Oppressed Hindus. Madras: Huxley Press. 
Rama T. S. Rao
1967 "Protective Discrimination and Educational Planning." In Educational Planning: Its Legal and  
Constitutional Implications in­India, edited by G. S. Sharma. Bombay: N. M. Tripathi. 
Ramanathan, G.
1965 Indian Babu (A Study in Social Psychology). New Delhi: Sudha Publications. 
Ramaswamy, K. V.
1934 India's Untouchable Saints. Madras: G. A. Natesan & Co. 47 pp. 
Ramaswamy, Uma
1978 "The Belief System of the Non­Brahmin Movement in India: The Andhra Case." Asian Survey 
18:290­300. 
Rao, D. M.
1963 "Disparities of Representation Among Direct Recruits to the I.A.S." Indian Journal of Public 
Administration 9:88­94. 
Rao, M. S. A.
1967 "Education, Social Stratification and Mobility." In Papers in the Sociology of Education in 
India, edited by M. S. Gore, I. P. Desai, and Suma Chitnis. New Delhi: National Council of 
Educational Research and Training. 
1968 "Political Elite and Caste Association: A Report of a Caste Conference." Economic and  
Political Weekly 3:779­82. 
Rao, V. K. R. V.
1972 Keynote Address on "Social Change." Bangalore: Polygraph Press. 15pp. 
Ratnam, K. J.
1965 Communalism and the Political Process in Malaya. Kuala Lumpur: University of Malaya 
Press. 
­596­ 
Real Indian National Congress Party
1945 Destroy This Brahmin Raj. Bezwada: Sath Brahmana Ashram. 72 pp. 
Republican Party of India (B. D. Khobrogade, General Secretary)
1964 The Charter of Demands: Submitted to Shri Lal Bahadur Shastri, the Honorable Prime 
Minister of India. New Delhi. 32 pp. 
Retzlaff, Ralph H.
1963 "The Problem of Communal Minorities in the Drafting of the Indian Constitution." In 
Constitutionism in Asia, edited by R. N. Spann . Bombay: Asia Publishing House. 
Revankar, Ratna G.
1971 The Indian Constitution: A Case Study of Backward Classes. Rutherford, N.J.: Faireleigh 
Dickinson University Press. 
Risley, Sir Herbert
1908 Peoples of India. London: W. Thacker Co. ( 2d. ed. 1915). 
Robbin, Jeanette.
1964 Dr. Ambedkar and His Movement. Hyderabad: Dr. Ambedkar Publication Society. 
Roy, Sarat Chandra
1928 Oraon Religion and Customs. Ranchi: "Man in India" Office. 
Rudolph, Lloyd I., and Susanne Hoeber Rudolph
1967 The Modernity of Tradition: Political Development in India. Chicago: University of Chicago 
Press. 
Rudolph, Susanne Hoeber, and Lloyd I. Rudolph, eds.
1972 Education and Politics in India: Studies in Organization, Society and Politics. Delhi: Oxford 
University Press. 
Saberwal, Satish
1972 "The Reserved Constituency: Candidates and Consequences." Economic and Political Weekly 
7:71­80 
Sachchidanada, assisted by A. K. Lal and S. N. Tiwary.
1977 The Harijan Elite: A Study of Their Status, Networks, Mobility and Role in Social 
Transformation. Faridabad: Thomson Press. 
Saksena, R. N.
1965 "Factor in Social Tension." Seminar 70 (June):22­25. 
Sanjana, J. E.
1946 Caste and Outcaste. Bombay: Thacker & Co. 
Santhanam, K.
1946 The Fight Against Untouchability: A Brief Review of the Phenomenal Progress Made During 
the Last Three Decades in the Eradication of an Ancient Evil. New Delhi: Hindustan Times. 
Santhanam, K.
1964 Transition in India and Other Essays. Bombay: Asia Publishing House. 
Sapru, Tej Bahadur
1945 Constitutional Proposals of the Sapru Committee. Moradabad: Sapru Committee. 
Saraswathi, S.
1974 Minorities in Madras State: Group Interests in Modern Politics. Delhi: Impex India. 
­597­ 
Sastri, D. Raghavachandrayya Sath.
1946 The Brahmin Non­Brahmin Hindu Problem. Bezwada: Sath Brahmana Ashram. 30 pp. 
Sathi, Chedi Lal
n.d. Neliru Sarkar Dvara Pichara Vargo ke Adhikaro ki Hatya? [Crushing Down of Rights of 
Backward Classes in Nehru Government?], Lucknow. [c. 1960] 
1967 Educational Planning: Its Legal and Constitutional Implications in India. Bombay: N. M. 
Tripathi. 
Sharma, R. C, and C. L. Sapra
1969 Wastage and Stagnation in Primary and Middle Schools in India. New Delhi: National 
Council of Educational Research and Training. 
Sharma, R. S.
1958 Sudras in Ancient India: A Survey of the Position of the Lower Orders down to circa A.D. 500. 
Delhi: Motilal Banarsidass. 
Sheridan, L. A.
1961 The Federation of Malaya Constitution: Text, Annotations and Commentary. Singapore: 
University of Malaya Law Review. 
1962 "Equal Opportunity of Public Employment (Non­discrimination in Public Service in India)." 
International and Comparative Law Quarterly 11:782­805. 
Shiva B. Rao (Chairman of Project Committee), and others
1968 The Framing of India's Constitution: A Study. New Delhi: Indian Institute of Public 
Administration. 
Shrivastava, B. N.
1968 Exploitation in Tribal Areas. New Delhi: Bharatiya Adimjati Sevak Sangh. 
Shukla, V. N.
1967 "Equal Protection Clause, Selective Techniques and Educational Planning." In Educational 
Planning: Its Legal and Constituional Implications in India, edited by G. S. Sharma. Bombay: N. 
M. Tripathi. 
Silverberg, James, ed.
1967 Social Mobility in the Caste System in India, an Interdisciplinary Symposium, The Hague: 
Mouton Publishers. Supplement III to Comparative Studies in Society and History. 
Sinder, Leon
1963 Caste Instability in Moghul India. Seoul, Korea: International Culture Research Center, 
Chung­ang University. 
Singer, Milton, and Bernard S. Cohn, eds.
1968 Structure and Change in Indian Society. Chicago: Aldine Publishing Co. 
Singh, Harjinder
1967 "Social Grading of Castes and Occupations in an Indian Village." Indian Journal of Social 
Work 27:381­96. 
Singh, Mohinder
1947 The Depressed Classes: Their Economic and Social Condition. Bombay: Hind Kitabs. 
­598­ 
Singh, Parmanand
1976 "'Equal Opportunity' and 'Compensatory Discrimination': Constitutional Policy and 
Judicial Control," Journal of the Indian Law Institute 18:300­19 
Singh, Saint Nihal
1913 "India's 'Untouchables'". Contemporary Review 103:376­85. 
Saxena, Ishwar Chandra
1963 "The Doctrine of Precedent in India: A Study of Some of Its Aspects." Jaipur Law Journal 
3:188­214. 
Sayaji Rao, III, Maharaja of Baroda
c. 1912 "The Depressed Classes." In The Depressed Classes: An Enquiry into Their Condition and 
Suggestions for Their Uplift. Madras: G. A. Natesan & Co. [Reprinted from the Indian Review of 
Dec. 1909]. 
1928 Speeches and Addresses. London: Macmillan Co. 
Schermerhorn, R. A.
1969 "Scheduled Caste Welfare: Public Priorities in the States." Economic and Political Weekly 
4:397­401 
1970 Comparative Ethnic Relations. New York: Random House. 
Schmitthener, Samuel
1968­69 "A Sketch of the Development of the Legal Profession in India." Law & Society Review 
3:337­82. 
Schuchter, Arnold
1970 Reparations: The Black Manifesto and Its Challenge to White America. Philadelphia: 
Lippincott. 
Schwartzberg, Joseph E., ed. 1978 A Historical Atlas of South Asia. Chicago: University of Chicago 
Press. 
Scott, Roland W.
1953 Social Ethics in Modern Hinduism: The Religious Quest of India. Calcutta: n.p. 
Seervai, H. M. 1968 Constitutional Law of India. Bombay: N. M. Tripathi. 
Seminar
1965 "Secularism." No. 67. New Delhi. 
Setalvad, M. C.
1969 "President's Page." Indian Advocate 9 (9):1. 
Shah, Vimal P., and Tara Patel
1977 Who Goes to College? Scheduled Caste/Tribe Post­Matric Scholars in Gujarat. Ahmedbad: 
Rachana Prakashan. 
Shapiro, Martin
1975 "Courts," In A Handbook of Political Science, edited by N. Polsby and F. Greenstein. 
Massachusetts: Addison­Wesley Publishing Co., Inc. 
Sharma, B. A. V.
1968 "Secular State and Civil Service." In Secularism in India, edited by V. K. Sinha. Bombay: 
Lalvani. 
Sharma, G. S.
1965 "Concept of Leadership Implicit in the Directive Principles of StatePolicy in the Indian 
Constitution." 
­599­ 
Policy in the Indian Constitution." Journal of the Indian Law Institute 7:174­88. 
Sharma, G. S., ed. 1966 Secularism: Its Implications for Law and Life in India. Bombay: N. M. 
Tripathi. 
1980 "Some Reflections on the Indian Experience with the Policy of Reservation". Unpublished ms. 
Sinha, D. N.
1967 "Educational Planning: Discrimination and the Equal Protection of the Laws." In Educational 
Planning: Its Legal and Constitutional Implications in India, edited by G. S. Sharma. Bombay: N. 
M. Tripathi. 
Sinha, V. K., ed. 1968 Secularism in India. Bombay: Lalvani. 
Sivertsen, Dagfinn
1963 Men Caste Barriers Fall: A Study of Social and Economic Change in an Indian Village. New 
York: Humanities Press. 
Smith, Donald E.
1963 India as a Secular State. Princeton: Princeton University Press. 
Smith, Donald E., ed 1965 South Asian Politics and Religion. Princeton: Princeton University Press. 
Smith, William Roy
1938 Nationalism and Reform in India. New Haven: Yale University Press. 
Smooha, Sammy
1978 Israel: Pluralism and Conflict. Berkeley: University of California Press. 
Srinivas, M. N.
1956 "A Note on Sanskritization and Westernization." Far Eastern Quarterly 15:81­96. 
1962 Caste in Modern India and Other Essays. London: Asia Publishing House. 
1966 Social Change in Modern India. Berkeley: University of California Press. 
Srinivas, M. N., Y. B. Damle, S. Shahani, and André Bétéille
1959 "Caste: A Trend Report and Bibliography." Current Sociology/La Sociologie Contemporaine 
7:135­83. 
Srivastava, L., and others
1971 Educational and Economic Condition and Employment Position of Eighteen Tribes. New 
Delhi: National Council of Educational Research and Training. 
Subbiah, A. 1963 "The Caste System: Vested Interests of the High Castes." Quest no. 37, pp. 85­7. 
Subramaniam. V.
1971 Social Background of India's Administrators. New Delhi: Publications Division, Ministry of 
Information and Broadcasting. 
­600­ 
Suffian, Tan Sri Mohamed
1972 An Introduction to the Constitution of Malaysia. Kuala Lumpur: Dichetak di­Jabatan 
Chetak Kerajaan. 
Sunshine, Russell B., and Berney, Arthur L.
1970 "Basic Legal Education in India: An Empirical Study of the Student Perspective at Three Law 
Colleges." Journal of the Indian Law Institute 12:39­118. 
Taub, Richard P.
1969 Bureaucrats under Stress: Administrators and Administration in an Indian State. Berkeley: 
University of California Press. 
Thakkar, A. V.
1949 "The Problem of Aborigines in India." (The R. R. Kale Memorial Lecture, Gokhale Institute of 
Politics and Economics, 1941). In Thakkar Bapa: Eightieth Birthday Commemoration Volume, 
edited by T. N. Jagadish and Shyamlal. Madras, n.p. 
Tinker, Hugh
1962 India and Pakistan: A Political Analysis. New York: Praeger. 
Titus, Murray T.
1959 Islam in India and Pakistan. Calcutta: YMCA Publishing House. 
Tope, T. K. 1967 "Protective Discrimination and Educational Planning." In Educational Planning:  
Its Legal and Constitutional Implications in India, edited by G. S Sharma. Bombay: N. M. Tripathi. 
Tripathi, P. K.
1966 "Secularism: Constitutional Provision and Judicial Review." Journal of the Indian Law 
Institute 8:1­29. 
1972 Some Insights into Fundamental Rights. Bombay: University of Bombay. 
Vagiswari, Alladi
1972 Income Earning Trends and Social Status of the Harijan Community in Tamil Nadu. Madras: 
Sangam Publishers. 
Varadarajalu T. Naidu, comp. 1948 The National Dharma: Life, Speeches and Writings of Dr. P. 
Varadarajulu Naidu. Salem: The Tamil Nadu Co. 
Varadarajulu T. Naidu
1931 The Justice Movement. Madras: Justice Printing Works. 
Venkatarangiengar, S. K.
1961 "Article 16 of the Constitution and the Application of the Equality Clause to Civil Servants." 
Indian Advocate 1 (3):29­34. 
Venkatasubramanian, C. S.
1965 "Separate Concurring judgments in the Supreme Court," A.I.R. 1965 Journal 3. 
Verba, Sidney, Bashiruddin Ahmed, and Anil Bhatt
1971 Caste, Race and Politics: A Comparative Study of India and the United States. Beverly Hills: 
Sage Publications. 
Verma, Brind Bihari
1956 Agriculture and Land Ownership System Among the Primitive People of Assam. Delhi: 
Bharatiya Adimjati Sevak Sangh. 
­601­ 
Vivekananda, Swami
1955 Caste, Culture and Socialism. Calcutta: Advaita Ashrama. 
Von Christoph Furer­Haimendorf
1967 "The Position of the Tribal Populations in Modern India." In India and Ceylon: Unity and  
Diversity, edited by Philip Mason. London: Oxford University Press. 
Von Arthur T. Mehren
1963a "The judicial Process with particular reference to the United States and to India." Journal of 
the Indian Law Institute 5:271­80 
1936b "Some Observations on the Role of Law in Indian Society and on Indian Legal Education." 
Jaipur Law Journal 3:13­18. 
Wadhwa, Kamlesh Kumar
1975 Minority Safeguards in India (Constitutional Provisions and Their Implementation). Delhi: 
Thomson Press. 
Wagley, Charles, and Marvin Harris
1958 Minorities in the New World: Six Case Studies. New York: Columbia University Press. 
Walsh, Sir Cecil
1930 Crime in India, with an Introduction on Forensic Difficulties and Peculiarities. London. 
Ernest Benn, Ltd. 
Wasby, Steven
1970 The Impact of the United States Supreme Court: Some Perspectives. Homewood, Ill.: Dorsey 
Press. 
Weiner, Myron, and John Osgood Field
1975 "How Tribal Constituencies in India Vote." In Electoral Politics in the Indian States: Three 
Disadvantaged Sectors, edited by Jagdish Bhagwati et al. Vol. II. Studies in Electoral Politics in the 
Indian States. Delhi: Manohar Book Service. 
Weitbrecht­Stanton, H. U.
1920 "The Untouchables of India and Their Enumeration." Asiatic Review 16:171­84. 
Whitcombe, Elizabeth
1972 Agrarian Conditions in Northern India. Vol. 1, The United Provinces Under British Rule, 
1860­1900. Berkeley: University of California Press. 
Wilson, John
1877 Indian Caste. 2 vols. Edinburgh: William Blackwood 1976, New Delhi: Deep Publications. 
Wiser, William Henricks
1958 The Hindu Jajmani System. A Socio­economic System Interrelating Members of a Hindu 
Village Community in Services. Lucknow: Lucknow Publishing House. 
Wiser, William H., and Charlotte Viall Wiser
1963 Behind Mud Walls. 1930­1960. Berkeley: University of California Press. 
World Bank
1975 Health: Sector Policy Paper. Washington D.C: World Bank. 
­602­ 
Yadava, Jai Narayan Singh
1955 The Problems of India's Backward Classes. Delhi: Pahari Dhiraj, Ahiram. 
Zelliott, Eleanor
1966 "Buddhism and Politics in Maharashtra." In South Asian Politics and Religion, edited by 
Donald E. Smith. Princeton University Press. 
1969 "Dr. Ambedkar and the Mahar Movement." Ph. D. dissertation, University of Pennsylvania. 
Zinkin, Taya
1962 Caste Today. Issued under the auspices of the Institute of Race Relations. London: Oxford 
University Press. 
ZoBell, Karl M.
1959 "Division of Opinion in the Supreme Court: A History of Judicial Disintegration." Cornell 
Law Quarterly 44:186­214. 
GOVERNMENT DOCUMENTS 
Pre­Independence 
Bombay (Presidency). Depressed Classes and Aboriginal Tribes Committee [O.H.B. Starte, 
Chairman]
1930 Report, March 1930. Bombay: Printed at the Government Central Press. 
Great Britain. India Office. Indian Constitutional Reforms 1919 Reports of the Franchise Committee 
and the Committee on Division of Functions [Lord Southborough, Chairman]. Calcutta: 
Superintendent of Government Printing. 
Great Britain. Indian Franchise Committee [Marquess of Lothian, Chairman]
1932 Report. Vol. 1 Calcutta: Government of India, Central Publication Branch. 
Great Britain. Indian Round Table Conference [Second Session, 7 Sep. 1931­1 Dec. 1931]
1932 Sub­committees' Reports; and Prime Minister's Statement. Calcutta: Government of India, 
Central Publications Branch. 
Great Britain. Indian Statutory Commission. 1930a Memoranda Submitted by the Government of 
India and the India Office. Vol. 5. Calcutta: Government of India, Central Publications Branch. 
Great Britain, Indian Statutory Commission
1930b Report, Vol. I (Survey). Calcutta: Government of India, Central Publications Branch. 
Great Britain. Indian Statutory Commission. Auxilliary Committee on Growth of Education [P. 
Hartog, Chairman]
1929 Interim Report of the Indian Statutory Commission. London: His Majesty's Stationery Office. 
­603­ 
India. Indian Central Committee [Sankaran Nair, Chairman]
1929 Report. Calcutta: Government of India, Central Publication Branch. 
India. Reforms Enquiry Committee [A.P. Muddiman, Chairman]
1925 Report of the Reforms Enquiry Committee, 1924 . . . and Connected Papers. London: His 
Majesty's Stationery Office. 
Post­Independence: Central Government (India, Dominion) 
Constituent Assembly
Debates: Official Report. v. 1­12; Dec. 9, 1946­Jan. 24, 1950. Delhi: Manager of Publications. 
Post­Independence: Central Government (India, Republic) 
Administrative Reforms Commission
1969 Report of the Study Team on Recruitment, Selection, U.P.S.C./State P.S.Cs. and Training. 
Delhi: Manager of Publications. 
All­India Agricultural Labour Enquiry
1954 Agricultural Labour: How They Work and Live; Essential Statistics. Balakrishna Ramamurti 
for the All­India Agricultural Labour Enquiry. Forward by V. V. Giri. Delhi: Manager of 
Publications. 
1955 Report on Intensive Survey of Agricultural Labour, Employment, Underemployment, Wages 
and Levels of Living. Delhi: Manager of Publications. Vol. 1: All India; Vol. 2: North India; Vol. 3: 
East India; Vol. 4: South India; Vol. 5: West India; Vol. 6: Central India; Vol. 7: North West India. 
Backward Classes Commission, (K. Kalelkar, Chairman)
1955 Report. 3 Vols. Delhi: Manager of Publications. 
Backward Classes Commission. (B.P. Mandal, Chairman)
1981 Report: 7 Vols. in 2. New Delhi: Controller of Publications 
Central Social Welfare Board
1966­68, 1972­76 Reports. Mimeo. New Delhi. 
Commission on the Hill Areas of Assam
1966 Report, 1965­66. [H. V. Pataskar, Chairman]. New Delhi: Government of India. 
Commissioner for Scheduled Castes and Scheduled Tribes
1951­56, 1957­58 to 1978­79 Reports. New Delhi: Manager of Publications. 
1968 Handbook on Scheduled Castes and Scheduled Tribes. Delhi: Manager of Publications. 
Department of Community Development
1966 Panchayati Raj at a Glance as on 31 March 1966. New Delhi: Ministry of Food, Agriculture, 
Community Development and Cooperation. 
Department of Personnel
1972 Brochure on Reservation for Scheduled Castes and Scheduled Tribes in Services. ( 3d ed.) 
Delhi: Manager of Publications. 
Department of Social Security
1965 Report of the Advisory Committee on the Revision of the Lists of Scheduled Castes and 
Scheduled Tribes [Lokur Committee]. Delhi. 
­604­ 
Department of Social Welfare
1965­66 Reports. New Delhi: Manager, Government of India Press.
to
1975­76 
1966 Report of the Committee on Customary Rights to Scavenging [N. R. Malkani, Chairman]. 
Faridabad: Government of India Press. 
1968 A Statistical Handbook of Tribal Welfare and Development. New Delhi: Manager, Government 
of India Press. 
1969 Report of the Committee on Untouchability, Economic and Educational Development of the 
Scheduled Castes [ Elayaperumal Committee] Delhi. 
Directorate of Backward Classes Welfare
1973­74 The Newsletter. New Delhi: Ministry of Home Affairs.
Election Commission 
1955 Report on the First General Elections in India, 1951­52. 2Vols. Delhi: Manager of 
Publications. 
1959 Report on the Second General Elections in India, 1957. 2 Vols. Delhi: Manager of 
Publications. 
1966 Report on the Third General Elections in India, 1962­66. 2 Vols. Delhi: Manager of 
Publications. 
1968 Report on the Fourth General Elections in India, 1967. 2 Vols. Delhi: Manager of 
Publications. 
1973 Report on the Fifth General Elections to the House of the People in India, 1971. Vol. 2 
(Statistical.) New Delhi: Manager of Publications. 
1976a Delimitation of Parliamentary and Assembly Constituencies Order, 1976. New Delhi: 
Government Press. 
1976b Report on the General Elections to the Legislative Assemblies in India, 1970­72. Vol. 2­A 
(Statistical) except U.P., Nagaland, and Pondicherry. Delhi: Government Press. 
Expert Committee on Legal Aid [V. K. Krishna Iyer, Chairman]
1974 Processual justice to the People. Report of the Expert Committee on Legal Aid. May, 1973. 
New Delhi: Department of Legal Affairs, Ministry of Law, justice and Company Affairs. 
Law Commission
1958 Fourteenth Report (Reform of Judicial Administration), Vol. I. Delhi: Manager of Publications. 
Ministry of Community Development and Cooperation, Department of Cooperation
1964 Conference of Registrars of Cooperative Societies, and Directorsin­Charge of Backward 
Classes and the State Ministers­in­Charge of Welfare of Backward Classes and State Minister­in­
Charge of Cooperation; May, 1964. Proceedings and Agenda Notes. Simla: Manager, Government 
of India Press. 
Ministry of Education
1962 Report of the Committee on Emotional Integration [Sampurnand, Chairman] Delhi. 
Ministry of Education. Education Commission
1966 Report of the Education Commission 1964­66. Delhi: Manager of Publications. 
­605­ 
1970 Report of the Education Commission, 1964­66, Vol. 2 (Supplement). Delhi: Ministry of 
Education and Youth Services. 
Ministry of Education and Scientific Research
1957 Progress of Scholarships for Scheduled Castes, Tribes and Other Backward Classes. 4th ed. 
Delhi: Manager, Government of India Press. 
Ministry of Education and Social Welfare
1955 Education in India, 1951­52. Delhi: Manager of Publications. 
1968 Education in India, 1961­62. Delhi: Manager of Publications. 
1973 Report, 1972­73. New Delhi: Manager of Publications. 
Ministry of Finance
1959 Report of the Commission of Enquiry on Emoluments and Conditions of Service of Central 
Government Employees, 1957­1959. Delhi. 
Ministry of Home Affairs.
1956 Memorandum on the Report of the Backward Classes Commission. Delhi: Government of 
India Press. 
1962a Conference of State Ministers in Charge of Welfare of Backward Classes. Agenda and Notes; 
26­27 July 1962. Mimeographed. 71 pp. 
1962b Report of the Special Working Group on Cooperation for Backward Classes [M.P. Bhargava, 
Chairman]. 2 vols. New Delhi: Manager, Government of India Press. 
Ministry of Information and Broadcasting
1952 Harijans Today. Delhi: Director, Publications Division. 
1953 India: A Reference Annual. New Delhi. 
1963 Welfare of the Backward Classes. Delhi: Director, Publications Division. 
Ministry of Law, justice and Company Affairs
1977 judges of the Supreme Court and the High Courts, as on 1­4­77. Delhi: Ministry of Law, 
Justice and Company Affairs. 
1978 Report on National Juridicare: Equal Justice­­Social Justice [P.N. Bhagwati, Chairman]. New 
Delhi: Ministry of Law, Justice and Company Affairs. 
Parliament
1950­52 Parliamentary Debates. Official report. 
Parliament. House of the People (Fourth Lok Sabha). Committee on the Welfare of Scheduled 
Castes and Scheduled Tribes
1969­70 Reports1­19. New Delhi: Lok Sabha Secretariat. 
Parliament. House of the People (Fifth Lok Sabha). Committee on the Welfare of Scheduled Castes 
and Scheduled Tribes. 1971­76 Reports1­14, 16­38, 40­51. New Delhi: Lok Sabha Secretariat. 
Parliament. Lok Sabha [House of the People] 1952­ Debates. Delhi: Lok Sabha Secretariat. 
Parliament. Rajya Sabha
1952­ Debates. Delhi: Rajya Sabha Secretariat. 
Planning Commission
1951 The First Five Year Plan: A Draft Outline. New Delhi: Manager of Publications. 
1953 First Five Year Plan: Questions and Answers. Delhi: Publications Division, Ministry of 
Information and Broadcasting. 
­606­ 
1960 Social Welfare in India. Forward by Jawaharlal Nehru. Revised and abridged ed. Delhi: 
Publications Division, Ministry of Information and Broadcasting. 
1961 Third Five Year Plan. Delhi: Manager of Publications. 
1965 Report of the Seminar on Employment of Scheduled Castes and Scheduled Tribes. Delhi: 
Manager of Publications. 
1970 Fourth Five Year Plan, 1969­74. Delhi: Government of India. 
1976 Fifth Five Year Plan, 1974­79. Delhi: Controller of Publications. 
Planning Commission, Programme Evaluation Organization
1956 Study of Village Artisans. Delhi: Government Press. 
Scavenging Conditions Enquiry Committee [N. R. Malkani, Chairman]
1960 Report. Delhi: Manager of Publications. 
Scheduled Areas and Scheduled Tribes Commission [U. N. Dhebar, Chairman]
1961 Report, 1960­61. 2 vols. Delhi: Manager of Publications, Civil Lines. 
Study Group on the Welfare of the Weaker Sections of the Village Community [J. P. Narayan, 
Chairman]
1961, 1963 Report, 2 vols. Delhi: Ministry of Community Development, Panchayati Raj and 
Cooperation. 
Study Team on Social Welfare and Welfare of Backward Classes [Renuka Ray, Chairman]
1959 Report. 1 vol. New Delhi: Committee on Plan Projects. 
Study Team on Tribal Development Programmes
1970 Report. New Delhi: Committee on Plan Projects, Planning Commission. 
University Education Commission
1950­51 Report. Dec. 1948­Aug. 1949. 2 vols. Delhi: Ministry of Education. 
Post­Independence: States 
Andhra Pradesh. Backward Classes Commission [Manohar Pershad, Chairman]
1970 Report. Hyderabad. 
Andhra Pradesh. Department of Social Welfare
1975 Performance Budget, 1975­76. Hyderabad: Government Central Press. 
Bihar
1968 A Report on the Situtation in the Tribal Areas of Bihar. Patna: Printed by the Superintendent, 
Secretariat Press. 
Bihar. Welfare Department
1966a Annual Administration Report. 1964­65. Patna: Superintendent, Secretariat Press. 
1966b The Bihar Welfare Manual: Compendium of Important Circulars and Letters Relating to the 
Programme for the Welfare of Backward Classes. Patna: Superintendent, Secretariat Press. 
Bombay. Backward Classes Department
1957, 1958 Administration Report on the Working of the Backward Classes Department 1954­55, 
1955­56. Poona: Printed at the Yervada Prison Press. 
­607­ 
Bombay. Director of Backward Class Welfare
1955 Privileges Provided by Bombay State Government for Backward Classes. Poona: Yervada 
Prison Press. 
Bombay. Ex­Criminal Tribes Rehabilitation Committee [K. B. Antrolikar, Chairman]
1951 Report, 1950. Bombay: Government Central Press. 
Gujarat. Directorate of Social Welfare
1963­64 Annual Administration Report on the Working of the Social Welfare Department. 
Ahmedabad: Government Printing, Publications and Stationery. 
Gujarat. Education Department
1959­60 Administration Report. Ahmedabad: Government Central Press. 
Gujarat. Socially and Educationally Backward Classes Commission [A. R.. Bakshi, Chairman]
1976 Report of the Socially and Educationally Backward Classes Commission. 2 vols. Ahmedabad: 
The Commission. 
Haryana. Legislative Assembly. Committee on the Welfare of Scheduled Castes and Scheduled 
Tribes
1975 Report. 1st. January. Chandigarh: Vidhan Sabha Secretariat. 
Jammu and Kashmir. Backward Classes Committee [J. N. Wazir, Chairman]
1969 Report. Srinagar: Printed at the Government Press. 
Karnataka. Karnataka Backward Classes Commission [L. G. Havanur, Chairman]
1975 Report, Karnataka Backward Classes Commission, Constituted under the Commission of 
Inquiry Act, 1952 (Central Act 60 of 1952). 4 vols. in 5 parts. Bangalore: Government of Karnataka. 
Kerala
1960 Kerala Services Manual. Vol. 1. Trivandrum: Superintendent of Government Printing. 
Kerala. Administrative Reforms Committee [E. M. S. Nambudiripad, Chairman]
1958 Report. Vol. 1, Part 1, 1958. Trivandrum: Printed at the Government Press. 
Kerala. Backward Classes Reservation Commission [Nettur P. Damodaran, Chairman]
1971 Report, 1970. 2 vols. Trivandrum: Printed at the Government Press. 
Kerala. Commission for Reservation of Seats in Educational Institutions
[G. Kumara Pillai, Chairman]
1966 Report. Trivandrum: Printed at the Government Press. 
Kerala. Harijan Welfare Department.
"Administration Report. Annual. 1956­57 to 1963­64; 1967­68 to 1971­72". Ernakulam: Printed at 
the Government Press. 
1967 First Supplement to the Hand Book of the Department of Harijan Welfare: Department Rules 
and Standing Orders., as on 30th September 1965. Shoranur: Government Press. 
­608­ 
Kerala. Legislative Assembly. Committee on the Welfare of Scheduled Castes and Scheduled 
Tribes
1974a Report on the Alienation of Tribal Lands, 1st, 1973­74. Trivandrum: Secretariat of the 
Kerala Legislature. 
1974b Second Report on the Representation of Scheduled Castes and Scheduled Tribes in the 
Services of Various Departments under the Government of Kerala. Trivandrum: Secretariat of the 
Kerala Legislature. 
Kerala. Revenue Department
1964 A Handbook of the Department of Harijan Welfare (as on 31 August 1962). Trivandrum: The 
Government Press. 
Madras
n.d. Harijan Welfare in Madras State. Madras: Director of Information and Publicity, Government 
of Madras. 
Madras. Harijan Welfare Department
Administration Report, Annual. 1955­56; 1958­59 to 1960­61; 197172. Madras: Government Press. 
Maharashtra
n.d. Schemes for the Welfare of Backward Classes and Provisions Made Therefore in the Budget 
Estimates for 1973­74. Bombay: Government Central Press. 
Maharashtra. Backward Class Hostels Committee
1962 Report. Poona: Printed at the Yervada Prison Press. 
Maharashtra. Backward Classes Welfare Wing
1960­61 Annual Administration Report on the Welfare of Backward Classes. to Bombay: Director. 
Government Prinitng. Publications and 1966­67 Stationary. 
Maharashtra. Committee on Reservation for Backward Classes in the Service [B. D. Deshmukh, 
Chairman]
1964 Report. Bombay: Printed at the Government Central Press. 
Maharashtra. Legislature. Committee on the Welfare of Scheduled Castes and Scheduled Tribes
1973 Fifth Report: Education. Bombay: Maharashtra Legislative Secretariat. 
Maharashtra. Public Service Commission
1963 Report, 1961­62. Bombay: Director, Government Printing, Publications and Stationery. 
Maharashtra. Social Welfare Department
1960­61 Annual Administration Report of Scheduled Areas in Maharashtra State. to Bombay: 
Director, Government Printing, Publications and 1971­72 Stationery. 
Mysore
1965 Administration Report on the Welfare of Backward Classes for the Year 1963­64. Bangalore: 
Cyclostyled. 84 pp. 
Mysore. Backward Classes Committee [R. Nagan Gowda, Chairman]
1960 Interim Report. Bangalore: The Government Press. 
1961 Final Report. Bangalore: The Government Press. 
­609­ 
Mysore. Bureau of Economics and Statistics
1972 Bidar District: A Survey of a Backward Area. Bangalore: n.p. 
Mysore. Commissioner for Depressed Classes
1957 Summary Record and Main Recommendations of the Mysore State Depressed Class Seminar 
Held at Kengeri, February 11th to 14th, 1956 Mysore: Deputy Director, Govenment Branch Press. 
Mysore. Education Department
1969 Handbook of Scholarships and Educational Concessions for the Students of Mysore State (as 
on 2nd January 1969). 2 vols. Bangalore: The Government Press. 
Mysore. General Administrative Department
1962 Scheduled Castes and Scheduled Tribes Appointments Committee Report. Coorg: Manager, 
Government Branch Press. 
Mysore. Legislature. Committee on the Welfare of Scheduled Castes and Scheduled Tribes
1973 First Report: Education of Scheduled Castes and Scheduled Tribes. Bangalore: Mysore 
Legislature Secretariat. 
Mysore. Planning, Housing and Social Welfare Department
1962 The Welfare of Scheduled Castes in Mysore State. Bangalore: The Government Press. 
Mysore. Social Welfare Department
1961 Detailed Administration Report, 1958­59. Bangalore: The Government Press. 
1968­69 Annual Administration Report (draft). Mimeographed. 
1970 Welfare of the Weaker Sections and the Handicapped, 1969­70. (Report of the Social Welfare 
Department for]969­70.) Bangalore: The Government Press. 
Orissa. Tribal and Rural Welfare Department
Administration Report. Annual. 1965­66; 1968­69; 1970­71; 197172. Cuttack: Orissa Government 
Press. 
1958 Welfare of Backward Classes in Orissa. Bhubaneswar: Orissa Government Press. 
1972 Report on the Annual Administration of Scheduled Areas in Orissa for the Year 1970­71. 
Cuttack: Orissa Government Press. 
Punjab. Welfare Department
1966 Report of the Evaluation Committee on Welfare Regarding the Welfare of Scheduled Castes, 
Backward Classes and Denotified Tribes in Punjab State for the Period Commencing from 15 
August 1947 (December 1965­August 1966). Chandigarh: Controller of Printing and Stationery, 
Punjab. 
Tamil Nadu. Backward Classes Commission [A. N. Sattanathan, Chairman]
1971 Report. 3 vols. Madras: Director of Stationery and Printing. 
Tamil Nadu. Directorate of Backward Classes 1971­72 Performance Budget. Madras: Director of 
Stationery and Printing. 
Tamil Nadu. Harijan Welfare Department
1971­72 Performance Budget. Madras: Government Press. to 1973­74 
­610­ 
Travancore­Cochin. Committee Enquiring into the Condition of Scheduled Hill Tribes in 
Travancore­Cochin [W. S. S. Mackay, Chairman.]
1956 Preliminary Report of the Committee Enquiring into the Condition of Scheduled Hill Tribes  
in Travancore­Cochin, 1955­56. Munnar: Imperial Press. 
West Bengal. Bureau of Applied Economics and Statistics.
1971 Backwardness of the Districts of West Bengal­­A Comparative Study. Calcutta: Reliance 
Printing Works. 
West Bengal. Development Department
n.d. India's Villages. A collection of articles originally published in the Economic Weekly of 
Bombay. Calcutta: n.p. 
West Bengal. Legislature. Legislative Assembly. Committee on the Welfare of Scheduled Castes and 
Scheduled Tribes.
1976 Third Report, 1974­76. Calcutta: West Bengal Legislative Assembly Secretariat. 
West Bengal. Tribal Welfare Department. Progress, Statistics and Coordination Cell
1961­62., Administration Report. Alipore: Government Press
1962­63,
1963­66 
West Bengal. Tribal Welfare Department. Cultural Research Institute
1962­63, Bulletin. Annual. Calcutta
1963­64 
1962 Impact of Industrialization on the Life of the Tribals of West Bengal. Calcutta: Government of 
West Bengal. 
West Bengal. Tribal Welfare Department
1958 Report on Social Welfare and Welfare of Backward Classes in West Bengal. Alipore: 
Government Press. 
­611­ 
Index 
Abraham, Amrita, 40 
Adam, William, 19 
Adi Dravida, 315, 327 ­32 
Adler, Chaim, 553 
Adoption, 337 ­40, 341 
Agarwal, C. B., 34 
Aggarwal, Partap C., 66, 67, 68, 78, 108, 109, 110, 356, 357 
Ahmad, Nazirrudin, 146 
Ahmed, Karuna, 560 
Ahmed, S. A., 429 
Aiyappan, A., 16 
Alberuni, 13 
Alexandrowicz, C. H., 2, 366, 494, 498 
All­India Backward Classes Federation, 162, 174, 178, 179 
Ambedkar, Bhimrao Ramji, 18, 29 ­40 passim, 45, 127, 129, 307, 319, 326 ; in Constitutent 
Assembly, 39, 222, 368, 399 ­400; in First Amendment debate, 160, 166, 167 
Anandan, P. I., 20 
Anant, Santokh Singh, 75, 76, 77, 78 
Anantanarayanan, M., 315 
Andhra Pradesh Backward Classes Commission ( 1970), 101, 227, 236, 237, 244, 248, 253, 
254, 260, 466 
Angell, James W., 554 
Anglo­Indians, 42, 294 
Anti­Brahminism, 209, 447 
Anti­disabilities legislation, 28, 36. see also Untouchability (Offences) Act, 1955 
Arya Samaj, 23, 29, 292 ­93, 303 
Ashraf, Mohammed Siddiq, 66, 67, 68, 78, 108, 356, 357 
Austin, Granville, 44, 85, 363, 561 ­62 
Ayyangar, M. A., 166 
Ayyar, S. V., 79 
Azad, Maulana, 162 
Backward Class: definition of, 3, 154 ­59, 185 ­86; education, 62, 64, 162, 177 ; estimated 
population, 43, 158, 161, 168 ­69, 195 ; discussion in Constituent Assembly, 159 ­61; Centre's 
policy regarding, 167 ­79; varied state policies regarding, 161 ­63, 179 ­87; judicial control on 
permissible "classes" 188 221 passim; judicial determination of backwardness, 222 ­81 passim, 
341 ­48; compartments within and layers of, 463 ­72; most backward classes, 468 70; litigation 
concerning, 501 ­13 
Backward Classes Commission ( 195355), 165, 166, 167 ­74, 176, 179, 180, 217, 241, 245, 260, 
262 
Backward Classes Commission ( 197880), 187 
Baines, Athelstan, 123 
Bajpai, G. S., 123 
Bakhale, R. R., 125, 130 
Balaji, M. R., 428 
Balakrishnan, B. G., 75 
Banaji, D. A., 19 
Banerjee, A. C., 159 
Banias, 223 
Barnabas, A. P., 73, 84, 175 
Baroda, Maharaja of [Gaikwad] (Sayaji Rao III), 25, 30 
Barrier, N. G., 25 
Basi, P. Reddy, 446 
Bastedo, T. G., 514, 515 
Baxi, Upendra, 416, 520 
Becker, Theodore L., 522 
Bedars, 288 
Beg, M. H., 220, 385, 389, 424, 481 
Behara, Baidhar, 69 
Beidelman, Thomas O., 9 
Berney, Arthur L., 514 
­613­ 

Berreman, Gerald D., 12 
Béteille, André, 329 
Bhandare, R. D., 267, 321 
Bhangis, 35 
Bhargava, Thakur Das, 166 
Bhovis, 285 
Bittker, Boris, 554 
Black, Donald, 498 
Blaug, Mark, 85 
Bobbitt, Phillip, 437 
Bonarji, N. B., 85 
Bopegamage, Albert, 235 
Bose, Nirmal Kumar, 168, 169 
Brahmins, 10, 222, 223, 235 
Brown, F. H., 123 
Buddhists, 16, 39 ­40, 144, 307 ­10, 311, 319 ­23, 337 
Calabresi, Guido, 437 
Caste: terminology, 7 ; system described, 7 ­13; legal position under British rule, 19 ­21; 
reform of caste as public issue, 23 ­25; as unit and as measure of backwardnessness, 139, 189 
­90, 229 ­30, 535 ; lost on conversion, 315 ­22, 332 ; judicial views of, 318, 331 ­32; 
excommunication, 333 ­35; constitutional position of, 350 ; role in Indian life, 559 ­61 
­­­­­relation to backwardness: correlational and causal, 209 ­13; absence of data, 213 ­15 
Census: definition of Untouchables, 127 29; eschews caste data, 164 ; use of old data, 259 ­60; 
collection of caste data, 260 ­61; impact on identities, 354 
Chamars, 14, 137, 286, 288, 289 
Chandrachud, Y.V., 445, 481 
Channiah, T., 160 
Chauhan, D. N. S., 51 
Chaurasia; S. D. Singh, 158, 161, 168, 172, 234 
Chavan, Y. B., 321 
Chayes, Abram, 395 
Chidambaram Pillai, P., 21 
Chintamani, C. Y., 125, 130 
Chitnis, Suma, 77, 79 
Chopra, S. L., 51 
Christians, 16, 223, 291, 293, 294, 312 ­13, 330, 335 ­36 
"Classes": communities as classes, 188 204; historic, modernist and elastic views, 188 ­89, 204 
­08, 535 
Cohen, Stephen P., 349 
Cohn, Bernard S., 20 
Compensatory Discrimination Policies: terminology, 2 ­3; expenditures under five year plans, 
56 ­57; waiver of age and other requirements, 87, 372, 389, 420 ­21; priority of commitment to 
SC and ST diluted, 391, 393, 435 ­37, 537 ; lessons from India's experience, 563 ­67 
­­­­­administration: shortfalls in expenditures, 65 ; benefits separate rather than additional, 65 
­67; absence of goals, 67 ­68; low prestige and priority, 68 ­69; of economic tests of 
backwardness, 342 ­45, 347 ­48; implementation of judicial rules, 351, 359, 532 ­33 
­­­­­perceptions of: by beneficiaries, 66, 77 ­79; by intellectuals, 73 ­75, 77 ; by public, 75 ­77, 
560 ­61 
­­­­­measurement of performance: 54 55, 72 ­75, 79 ­83; government employment, 93 ­117, 213 
­15, 356 ; problems of, 417 ­18; limitation of courts in effectuating, 541 ­46; effects 
summarized, 547 ­52: 
­­­­­justifications: fairness, 114 ­15, 552 ­59; nondiscrimination theme, 552 ­53; general 
welfare theme, 553 54; reparations theme, 554 ­555, 556 ­59 
­­­­­maldistribution: "lion's share" problem, 137, 270, 276 ; "creaming", 468 ­69 
Commission for the Scheduled Castes and Scheduled Tribes, 71 
Commissioner for Scheduled Castes and Tribes, 70 ­71, 161, 541 
Communal Award, 31 ­33 
Community, Indian usage of term, 16 ­17 
Constituent Assembly, 37, 39, 159 ­61, 363, 368, 371, 399 ­400 
Constitution of India: basic format, 37 39; Directive Principles, 41, 365, 366, 397, 443, 538 ; 
Article 338, 70 ; debate over Article 15(4), 165 ­67; forbidden classifications, 215 ­16; burden 
of proof as to backwardness, 218 ­20; tests for allowable classifications, 218, 314 ; judicial 
requirement of "reasonableness" in compensatory discrimination provi­ 
­614­ 

sions, 220 ­21; Article 12, 313, 129 ; antidiscrimination provisions, 364 ; Article 335, 368, 398 
; its relation to Article 16(4), 369 ; Article 320(4), 369 ; Article 46, 397 ; Forty­Second 
Amendment, 478, 520, 530, 538 
Conversion: among Scheduled Tribes, 138, 316 ­17; to Christianity, 312 ­13, 324 ­25; loss of 
caste theory, 315, 327 31, 332 ; reconversion. 327 ­28; to obtain benefits, 337 
Coupland, R., 375 
Dahl, Robert, 85 
Dalal, M. N., 375 
Dalton, Dennis, 18, 37 
Damaska, Mirjan, 359 
Das, S. R., 429 
Das, Seth Govind, 166 
Daschowdhury, B. K., 317 
Datar, B. N., 320, 325 
Dave, P. K., 100 
Davis, Kingsley, 259 
De, Arunangshu, 172 
Deb, Dasaratha, 151 
Denotified Tribes, 43 
Depressed Classes, see Untouchability, Scheduled Castes 
Derrett, J. Duncan M., 9, 19, 300, 301, 305, 306, 325 
Desai, K. K., 445 
Desai, Morarji, 341 
Deshmukh, Punjabrao, 45, 162, 165, 166, 369 
De Vos, George, 562 
Dharmalingam, A. M., 209, 213, 447 
Dhavan, Rajeev, 479, 481, 485, 489, 490, 508, 539, 540 
Dhavan, S. S., 479, 485 
Dhebar Commission, 66 
Dhobis, 137 ­42 
Directive Principles, see Constitution of India 
Dohars, 286 
Dolbeare, Kenneth M., 529, 542 
Double­Member Constituencies, see Legislative Bodies: Reserved seats 
Dravida Munnetra Kazagham, 236 
Driver, Edwin, 259 
D'Souza, Jerome, 165 
D'Souza, Victor S., 61 
Dubey, H. P., 478, 479 
Dumont, Louis, 8 
Dusadhs, 141 
Dushkin, Lelah, 18, 27, 32, 33, 43,, 46, 53, 54, 63, 64, 66, 85, 128, 129, 130, 137, 149, 209, 
222, 322, 325, 456, 527, 550 
Dutta, Ratna, 51 
Edelman, Murray, 528 
Education: increase and deficiencies in education of SC and ST, 58 ­62; backward classes, 62, 
64, 162, 177 ; postmatriculation scholarships, 62, 162, 177, 267 
Ekbote, Gopal Rao, 446 
Elayaperumal Committee, 53, 60, 70, 102 
Eldersveld, Samuel J., 84, 560 
Emergency Rule ( 1975­77), 139, 483, 520 
Equality: formal vs. substantive, 378, 391 ­93; "horizontal and vertical" views, 379 ­80; 
classification to promote substantive equality, 384 ­89; equal opportunity, varied meanings, 
404 ­06; accommodation of substantive equality and formal equality, 561 ­62 
Ezhuvas, 16, 23, 195, 235 ­36 
Fazl Ali, S. Murtaza, 382, 388, 419, 420, 424, 439 
Feeley, Malcolm, 522 
Fisher, Fred, 123 
Fisher, Marguerite J., 70 
Fox, Richard D., 11, 295, 349 
Franchise (Southborough) Committee 123, 125 
Friedman, Lawrence M., 359, 483, 533 
Gadbois, George H., 479, 480, 481, 482 
   484
   , 489, 490, 500, 503, 504 
Gaikwad, B. K., 320 
Gajendragadkar, P. B., 232, 302, 381, 410 
   486
   , 488, 526 
Galanter, Marc, 18, 19, 20, 21, 52, 54, 103 
   215
   , 305, 318, 358, 359, 364, 395 
Gandhi, Indira, 75, 477, 480, 481, 509 
Gandhi, M. K., 18, 29 ­40 passim, 375 
Gauba, K. L., 51 
Ghurye, G. S., 8, 13, 18, 21, 22, 23, 28 
   111
   , 147, 153, 213 
Gledhill, Alan, 496, 498, 499 
Gopalan Nambiar, V. P., 218, 242, 248 
Gough, E. Kathleen, 329 
­615­ 

Government employment, attraction of, 84 ­86 
Government employment, reservation in: origins, 86 ; age, fee and other concessions, 87, 372, 
389 ; measurement of effects, 93 ­117; pre­examination coaching centers, 99 ; promotions, 99 
101, 370 ­71, 446 ; public sector undertakings, 102 ­03 
Government of India Act 1919, 27 
Govinda Menon, P., 255 
Grossman, Joel, 489 
Grossman, Kurt R., 554 
Grover, A. N., 481 
Gujarat Backward Classes Commission ( 1976), 258 
Gupta, A. C., 388 
Gupta, M. G., 480, 482 
Gurdon, P. R. T., 294 
Gupta, Raj Kumar, 378 
Gurung, Ari Bahadur, 159 
Hah, Chong­Do, 562 
Hajarnavis, R. M., 529 
Halbas, 143 
Hammond, Phillip, 529 
Handler, Joel F., 533 
Hanumanthaiya, K., 186, 317, 341 
Hardgrave, Robert L., 16, 23, 171, 262 
Harijan Sevak Sangh, 34 
Harijans, see Untouchability, Scheduled Castes 
Harrison, Selig S., 175 
Hart, Henry C., 47, 52 
Hartog Committee ( 1928), 123, 156 
Havanur Commission, 228, 239, 240, 245, 256 ­58, 260, 454 
Havanur, L. G., 190, 192, 194, 206, 210, 211, 222, 226, 231, 232, 239, 256, 262, 268, 513 
Hayden, Robert, 428 
Hazari, 324 
Hegde, K. S., 199, 206, 380, 481, 513 
Heimsath, Charles H., 18, 23 
Hidayatullah, M., 366, 384, 486 
Hindu: population, 33 ; definition of, 305 11 
Hindu Mahasabha, 35 
Holderness, T. W., 123 
Horowitz, Donald L., 395 
Hutton, J. H., 8, 9, 12, 28, 126, 127 ­28, 129, 130, 134, 143, 300 
Income Ceilings, 62, 243, 263 ­66, 273 
Income Tests, see Compensatory Discrimination Policies: Administration 
Inden, Ronald, 8, 10 
Indian Central (Nair) Committee ( 1929), 124, 156 
Indian Franchise (Lothian) Committee, 124 ­25, 127, 129 
Indian National Congress: excludes social reform questions, 24 ; adopts first antidisabilities 
resolution, 26 ; 1931 Program of Fundamental Rights, 34 ; pluralities dependent on reserved 
seats, 138 
Indian society: compartmental and hierarchic organization, 7 ­17; empirical and formal 
approaches to group membership, 348 ­51; "realist" and "relativist" views of I. S., 352 ­58 
Interviews, see Selection Committees 
Irschick, Eugene F., 18, 27, 154, 343, 447 
Isaacs, Harold R., 108, 110, 113, 319, 324 
Israel, preferences for oriental Jews, 553 
Jackson, A. M. R., 19 
Jaganmohan Reddy P., 485 
Jagannadham, V., 84 
Jajmani System, 9 
Jammu and Kashmir Backward Classes Committee ( 1969), 227 
Janata Party, policy on OBC's, 186 ­87 
Jȧngam, R. T., 51 
Jatavs, 137, 286 
Jennings, Ivor, 38 
Jha, A. N., 178 
Joint Parliamentary Committee on the Welfare of Scheduled Castes and Scheduled Tribes, 71 
Jones, Kenneth W., 25 
Judiciary: Supreme Court, 477 ­78; high courts, 477, 478 ­80; appointment, 479 ­80; turnover, 
481, 488 ; formalistic style, 483 ­85, 539 ; citation of authorities, 485 ­86, 493 ; dissenting and 
concurring opinions, 489 ­90; hierarchy, 490 ­92, 495 ­97, 531 ­32; writ jurisdiction, 489 ­99, 
500, 317 ; tendencies to incoherence and diffuseness, 539 ­40; effectuation of compensatory 
discrimination policies, 541 ­46 
­616­ 

Kalelkar, K. K., Chairman of B. C. 172, 260 
Kamble, N. M., 321, 323 
Kapur, J. L., 303 
Karve, Irawati, 12, 322, 349 
Katzenstein, Mary F., 427, 436 
Kayasthas, 223 
Keer, Dhananjay, 18, 23, 30, 36, 319, 324, 326 
Kelkar, G. S., 98, 113 
Kennedy, Duncan, 483 
Kerala Backward Classes Reservation Commission 1971, 227, 237, 244, 245, 256, 273 ­74, 
467, 468 
Ketkar, Sridhar V., 25, 123 
Khandekar, H.J., 160 
Khanna, H. R., 423, 435, 481 
Khare, R. S., 514 
Khasis, 294 
Khatiks, 141 
Khobrogade, B. D., 321 
Khosla, G. D., 479 
Kidder, Robert L., 500, 514, 515 
Kirpal, Viney, 98, 113, 422 
Kochanek, Stanley A., 53 
Kolhapur, Maharajah of (Shahu Chhatrapati), 30, 154 
Konda Kapus, 140 
Koppell, Oliver, 520 
Krishna, Gopal, 47, 48, 51, 54 
Krishna Iyer, V. R., 147, 249, 302, 386, 387 ­88, 393, 419, 420, 424, 436 ­37, 468, 486, 520, 
540 
Krishnamachari, T. T., 160 
Krishnan, G. Y., 285 
Kshatriyas, 10, 23, 295 ­302 
Kuchelar, G., 340 
Kumar, Dharma, 19 
Kumar, Santosh, 136 
Kumara Pillai Commission, 197, 227, 236, 255, 263 
Kunzru, Hirday Nath, 159 
Kuppuswamy, B., 462 
Lajpat Rai, Lala, 292 
Lapp, Christopher C., 562 
Latchanna, G., 72, 212 
Latin Catholics, 195 
Latthe, A. B., 154 
Lawyers: training, 514 ­15; orientation to courts, 515 ­16; emphasis on litigation, 515 ­16; low 
degree of specialization, 516, 517 ­18, 538 
Layard, Richard, 85 
Lecky, Robert S., 554 
Legal Aid, 520 ­21 
Legal Scholarship, 485, 514 ­15, 516 ­17, 540 ­41 
Legislative Bodies, reserved seats: separate electorates, 25, 31 ­32, 45, 375 ; extension of 
constitutional authorization. 46, 375 ­76; designation of constituencies, 46 ­; doublemember 
constituencies, 46, 456, 457 58; local bodies, 50 ; profile of occupants, 51 ­52, performance of 
occupants, 52 ­54; scope, 375 76 
Legislative Bodies, SC's & ST's in nonreserved seats, 49, 50 ­51 
Lingayats, 16, 225, 226, 231, 232 
Litigation: to enforce or challenge caste prerogatives, 20 ; regarding reservations in government 
service, 104 ­05; claims for inclusion among beneficiaries of compensatory discrimination, 
279 ­81; about group membership, 282 ­83; capability of litigants. 358 ­59; possible litigation 
for affirmative pursuit of substantive equality, 394 ­95; government success in Supreme Court, 
482, 487 ; about compensatory discrimination, 486 ­87, 500 ­07, 509 ­13; interest group 
participation, 507 ­09; tactical problems in reservations cases, 5, 19 ; remedies, 522 ­24; 
compliance and resistance to, 524 ­27; relation to government policy, 528 ­29, 530 ­31; relation 
to public opinion, 529 ; absence of litigation to implement compensatory discrimination policy, 
396 ­98, 537 ­39 
Lockard, Duane, 543 
Lohia, Ram Manohar, 234 
Lokur, B. N., 136 
Lokur Committee, 53, 136 ­38, 152, 286 
Luthera, Ved Prakash, 305 
Lynch, Owen M., 23, 108, 111 
McCormack, William, 19 
McDonald, Ramsay, 31 
MacDougall, John Arthur, 563 
Madhavan, K. K., 323 
Madigas, 335 ­36 
Madras, Henry, 123 
­617­ 

Mahanubhava Panth, 290 ­92 
Mahars, 14, 21, 29, 137, 290 ­92 
Majithia, Sunder Singh, 127 
Majumdar, N. D., 152, 162, 168 
Malas, 137 
Malik, Suneila, 548, 549 
Mallahs, 219 
Mandal, B. P., 187 
Mandelbaum, David G., 8, 9, 259, 349 
Mariappa, T., 172 
Marriage, as means of acquiring group membership, 138, 294 ­95, 340, 341 
Marriot, McKim, 8, 10, 11, 235, 302, 349 
Mathew, K. K., 382, 385, 387, 423, 424, 486 
Maurya, B. P., 138 
Mayakrishnan, V., 195 
Mathew, Leon H., 357, 543 
Mehta, S. C., 73, 175 
Merillat, H. C. L., 494 
Miller, Robert J., 291, 319 
Minto­Morley Reforms ( 1909), 25 ­26 
Misra, Lokanath, 323 
Misra, Sheokaran Nath, 158 
Mochis, 286, 288, 289 
Mohapatra, M. K., 51, 52, 53, 75 
Moka Doras, 295 ­302 
Montagu, Edwin, 154 
Mookerjee, S. P., 166 
Moonje, B. S., 33 
Moothan, Orby, 498 
Morkhandikar, R. S., 40 
Morrison, Charles, 500, 514 
Mudholkar, J. R., 194, 402 
Mukerji, P., 156 
Mundas, 289, 294 
Munshi, K. M., 160 
Murphy, Arthur, 520 
Murphy, Walter, 523 
Muslims, 16, 25, 26, 51, 195, 198, 201, 202, 222, 231, 247, 306 
Nadars, 16, 23, 171, 262, 304 
Nagan Gowda Committee, 224, 230, 232, 234, 262 
Nagappa, Shri, 212 
Nair, Kusum, 321 
Nalavade, R. M., 159 
Namasudras, 137 
Naoroji, Dadabhai, 24 
Narayan, Jayaprakash, 176 
Narayana, G., 51, 53, 71 
Narula, R. S., 272 
Natarajan, D., 58, 59 
Natarajan, K., 25 
Natarajan, S., 18, 23, 25, 27 
Nayakas, 287, 288 
Nehru, Jawaharlal, 159, 161, 162, 165, 166, 167, 169, 173, 241, 379, 482 
Nehru, S. S., 157 
Non­Brahmin Movements, 27, 36, 156 
Noronha, Dionysius, 329 
Occupational Test of Backwardness, see Compensatory Discrimination Policies: 
Administration 
O'Malley, L. S. S., 9, 12 
Omvedt, Gail, 14, 18, 23, 29, 31, 32, 156 
Oraon, Kartik, 317 
Oraons, 293 
Other Backward Classes, see Backward Classes 
Pal, G. S., 127 
Palekar, J. J., 445 
Palen, Frank, 395 
Pant, G. B., 325 
Paraiyans, 14, 312, 329 
Parsis, 223 
Parvathamma, C., 51 
Patankar, Bharat, 14, 29, 31, 32 
Patars, 289 
Patel, Tara, 61, 548 
Patel, Vallabhbhai J., 170 
Patil, M. S., 225, 232 
Phadnis, Urmila, 321 
Phule, Jyotirao, 23, 29 
Pickett, J. Waskon, 324 
Ponnarangam, V. A., 209 
Poona Pact, 32 ­33 
Prakash, Dharam, 159 
Precedent, 490 ­95 
Protection of Civil Rights Act, see Untouchability (Offences) Act 
Pyarelal, 18, 32 
Race, 217, 364 
Radhakrishnan, N., 190, 194, 195, 206, 485 
Rai, Kalp Nath, 323 
Rajagopala Ayyangar, N., 371 
­618­ 

Rajah, M. C., 33, 85, 157 
Rajbanshis, 137 
Rajshekar Shetty, V. T., 74 
Ram, Chandrika, 160 
Ram, Jagjivan, 86 
Rama Rao, T. S., 235, 447 
Ramaswami, V., 429 
Ramaswamy, G., 510 
Ramaswamy, Uma, 27, 36 
Ranade, M. G., 24 
Rao, M. S. A., 179 
Ratnam, K. J., 563 
Ray, A. N., 145, 382, 384, 388, 419, 481, 486 
Reddi, K. Brahmananda, 136, 323 
Reforms Enquiry (Muddiman) Committee ( 1925), 27, 123, 156 
Religion: Religious test for Scheduled Castes, 143 ­44; religious groups as Backward Classes, 
201 ­02. See also Conversion, Buddhists, Christians, Muslims, Sikhs 
Reservations: layers of preference, 273 74, 467 ­68; carry­forward of reservations from year to 
year, 403, 407 ­09, 410 ­11; for groups other than Backward Classes, 425 ­31; "over and above 
merit" vs. guaranteed minimums, 439, 455 ­63; compartments, 463 ­67, 536 ; "carry­over" of 
unfilled places to other groups, 474 76. See also Government employment, Legislative Bodies, 
Education 
Retzlaff, Ralph H., 363 
Risley, Herbert, 122 
Round Table Conference, 30 ­31, 33 
Rowe, Peter, 514 
Roy, Sarat Chandra, 293 
Rudolph, Lloyd I., 2, 18, 19, 74, 321, 349 
Rudolph, Susanne Hoeber, 2, 18, 19, 74, 321, 349 
Saberwal, Satish, 52 
Sachchidananda, 77 
Sahas, 286 ­87 
Sahib, Mhd. Ismail, 160 
Saksena, R. N., 73 
Santhanam, K., 75 
Sapra, C. L., 60 
Saraswathi, S., 36, 154, 159 
Saraswati, Dayananda, 23 
Sathi, Chedi Lal, 179, 206 
Saxena, I. C., 491 
Scheduled Areas, 147 ­51 passim 
Scheduled Castes: legislative reservations, 44 ­55 passim; population, 42, 133 ; education, 55 
­64 passim; government employment, 84 ­105 passim; designation of S.C., 122 ­31; inclusion 
of Sikhs, 144, 305, 323 ; determination of membership in, 283 ­93, 305 ­41. See also 
Untouchability 
­­­­­S.C., Lists: 134 ­35; 1936 Order, 190 ; 1951 Order, 132 ; 1976 amendments, 135, 139 ; area 
limitations, 139 ­42; religious test, 143 ­44; judicial interpretation, 144 ­45, 283 ­90 
Scheduled Tribes: legislative reservations, 44 ­55 passim; education, 55 ­64 passim; 
government employment, 84 105 passim; population, 148 ­50; determination of membership in, 
283 90, 293 ­305, 316 ­17 
­­­­­S. T. Lists: 1936 list, 148 ; 1950 Order, 149 ; 1976 amendments, 139 
Schermerhorn, R. A., 69 
Schmitthener, Samuel, 514, 515 
Schuchter, Arnold, 554 
Schwartzberg, Joseph E., 11 
Secularism, 305, 559, 561 
Seervai, H. M., 479, 494 
Selection Committees, 446 ­55 
Separate electorates, see Legislative Bodies: reserved seats 
Setalvad, M. C., 497 
Shah, J. C., 198, 199, 370 
Shah, K. T., 146, 164, 166, 368 
Shah, P. G., 172 
Shah, Vimal P., 61, 548 
Shankaraiya, Shri, 165 
Shapiro, Martin, 359 
Sharma, B. A. V., 73 
Sharma, G. S., 305 
Sharma, R. C., 60 
Sharma, R. S., 13, 14 
Sharp, Henry, 123 
Shelat, J. M., 481 
Shiva Rao, B., 322 
Sikhs, 16, 35, 144, 305, 312, 323 
Silverberg, James, 12, 234 
Simon Commission ( 1930), 30, 124, 125, 157 
Singer, Milton, 349 
Singh, Anup, 172 
­619­ 

Singh, Charan, 187 
Singh, Deva Charan, 161, 162 
Singh, Harjinder, 235 
Singh, Hukum, 160, 305 
Singh, K., 126 
Singh, M., 130 
Singh, Parmanand, 2 
Singh, Saint Nihal, 26, 123 
Singh, V. P., 178 
Sinha, V. K., 305 
Smith, Donald E., 305, 326 
Smooha, Sammy, 553 
Somnath Iyer, A. R., 372 
Sonars, 292 ­93 
Srinivas, M. N., 12, 16, 17, 18, 175, 295, 349 
Srinivasan, R., 34, 85 
Starte Committee ( 1930), 157 
Subba Rao, K., 196, 198, 205 ­06, 247, 377, 408, 481, 488 
Subbiah, A., 175 
Sudras, 10, 13, 35, 145, 158, 234 
Suffian, Tan Sri Mohamed, 563 
Sunris, 286 ­87 
Sunshine, Russell, 514 
Suryabashis, 141 
Tainbe, S. B., 125, 130 
Tambe, Y. S., 445 
Tamil Nadu Backward Classes Commission ( 1971), 234, 236 ­37, 460 
Taub, Richard P., 77 
Terminology, Regarding Compensatory Discrimination Policies, 2 ­3; Backward Classes, 3, 
154 ­59, 185 ­86; caste, 7 ; untouchability, 13, 14, 25, 34 ; community, 16 ­17 
Thakkar, A. V., 161 
Thomas, John M., 395 
Tinker, Hugh, 26, 36 
Titus, Murray T., 324 
Tripathi, P. K., 180, 193, 204, 481 
Tussman, Joseph, 486 
Untouchability: general account, 13 ­16; terminology, 13, 14, 25, 34 ; emergence as political 
issue, 25 ; evangelical and secular approaches to abolition, 28 ; dispute about population of, 
122 ­25, 129, 130, 132 ; dispute about meaning of, 125 ­130; among Muslims and Christians, 
143 ; definition of, 131 ­32, 134, 145 ­47 
Untouchability (Offences) Act, 1955, 103, 44, 54 
Untouchables, see Scheduled Castes: Untouchability 
Vaishyas, 10 
Varna, 10 ­11, 12, 23, 29, 35, 37, 145, 211, 234 ­35, 300 ­02, 324, 349 ­50 
Veeraswami, K. N., 315 
Venkata Ranga Iyengar, S. K., 428 
Venkataraman, S., 166, 167 
Venkataraman, S. R., 158 
Venkatasubramanian, C. S., 493 
Venkataswamy, Shri, 179 
Voddars, 285 
Vokkaligas, 225, 226, 231, 232 
Von Furer­Haimendorf, Christoph, 151 
Von Mehren, Arthur T., 484 
Wagatsuma, Hiroshi, 562 
Walsh, Cecil, 499 
Wanchoo, K. N., 37 ­410 
Wasby, Steven, 522 
Weitbrecht­Stanton, H. U., 25, 123 
Wells, Richard, 489 
Wiser, W. H., 9 
Women, deemed backward class, 168 ; reservations for, 217, 426 
Wood, Glynn, 455 
Woodhall, Maureen, 85 
Wright, H. Elliott, 554 
Yadav, O. K., 210 
Yadav, Shyam Lal, 323 
Zelliot, Eleanor, 18, 21, 22, 25, 26, 27, 28, 29, 30, 31, 32, 33, 36, 39, 40, 108, 291, 319, 320, 
326 
ZoBell, Karl M., 489 
­620­ 
Index of Cases 
Abdul Aziz v. State of Maharashtra, 445 
Abdul Kadir v. Dharma, 301, 318 
Abdul Latif v. State of Bihar, 414 ­17 
Abhoy Pada Saha v. Sudhir Kumar Mondal, 286, 487 
Abodha Kumar v. State of Orissa, 216, 445 
Abraham v. Abraham, 291 
A. R. V. Achar v. State of Madras, 201, 222, 376 
Agurchand v. Deochand, 492 
All­India Station Masters v. General Manager, 370 
Anandji Haridas & Co. v. Engineering Mazdoor Sangh, 494 
Anjali v. State of West Bengal, 426 
Annam Adinarayana v. State of Andhra Pradesh, 502 
K. Appa Rao v. Director of Posts and Telegraphs, 140 
C. M. Arumugam v. S. Rajgopal [II], 316, 329, 330, 334 ­35, 338 ­39, 501 
Ashok v. Dean, 337, 511 
Atmaram v. King­Emperor, 145 
Baij Nath Prasad v. State of Bhopal, 493 
Balaji v. State of Mysore, 178, 180, 191 ­97, 201 ­02, 204, 207, 220, 222, 226 ­29, 232 ­33, 235 
­36, 238, 241 ­42, 246 ­48, 250 ­51, 254 ­55, 258, 262, 267, 273, 275 ­77, 314, 371, 377, 381, 
389, 393, 397, 400, 402 ­04, 410 ­16, 420, 431 32, 438 ­39, 441, 466 ­68, 470, 472, 474, 475 
­76, 491, 495, 507, 510 ­11, 518 20, 524, 526, 528 ­32, 543 
Balaram v. State of Andhra Pradesh, 459 
Balchand v. Laxminarain Mateh, 286 
Basavalingappa v. Munichinnappa, 136, 285 
Basunta Kumar v. Chief Electrical Engineer, 491 
Baula v. Chief Executive Officer, 491 
Bhagwan Bakhsh Singh v. Drigbijai Singh, 306 
Bhagwan Koer v. Bose, 306 
Bhaiya Lal v. Harikishan Singh, 140, 285, 288 
Bhaiya Ram Munda v. Anirudh Patar, 283, 289 
Bhau Ram v. Bay Nath, 215 
Bhavan v. Commercial Tax Officer, 491, 495 
Bhopalsingh v. State, 215, 314, 376 
Bidi Supply Co. v. Union of India, 217, 314 
Biman Chandra v. Governor, 313 
Bisanchand Lalchand v. Union of India, 491 
Bombay Education Society v. State of Bombay, 42 
Brahmachari Research Institute v. Their Workmen, 497 
Budhan Chaudhry v. State of Bihar, 217, 314 
Cannabasaviah v. State of Mysore, 445 
Chait Ram v. Sikander, 197, 200, 219, 255, 373 
Chamaraja v. State of Mysore, 456, 473 
Champakam Dorairajan v. State of Madras [see State of Madras v. Champakam Dorairajan] 
Chanchala v. State of Mysore, 384, 393, 402, 426 ­27, 433, 435, 524 
Chandhlal v. Babulal, 492 
Chandra Sekhera v. State of Mysore, 372 
Chandrasekhara Mudaliar v. Kulandaivelu Mudaliar, 306, 311 
Chatturbhuj Vithaldas Jasani v. Moreshwar Parashram, 290, 292 ­93, 295 ­300, 304, 315 ­17, 
327, 333, 348, 484 
Chauhan v. State of Gujarat, 408, 420, 445 
Chhotey Lal v. State of Uttar Pradesh, 184, 432, 467 
Chinnaswammy v. Anthonyswamy, 306 
Chitra Ghosh v. Union of India, 428, 429, 434 
­621­ 

Chitralekha v. State of Mysore, 192, 194 ­98, 204, 207, 233, 247, 264, 314, 402, 448, 485, 488, 
491, 513, 530 
Chunku Manjhi v. Bhabani Majhan, 296, 306 
Daryao v. State of Uttar Pradesh, 479 
Dasa Rayudu v. Andhra Pradesh Public Service Commission, 196, 242, 397, 446, 511 
Dass Bank v. Kali Kumari, 491 
Dattatraya v. State of Bombay, 217, 369, 376, 416 
Daya Krishnan v. Assessing Officer, 492 
Daya Ram v. State of Haryana, 384, 403, 427, 451 
Devadasan v. Union of India, 93, 217, 267, 371, 372 ­73, 377, 379, 390, 398, 403 05, 408, 410 
­21, 431, 442, 445, 476, 488, 505, 537 
Devarajiah v. Padmanna, 134, 146, 309 
Devchand Totaram v. Ghaneshyam, 324 
Dhanki Mahajan v. Rana Chandubha, 492 
Didar Singh v. Sohan Singh, 289 
Digambar Rao Bindu v. Dev Rao Kamble, 406, 458 
Dilip Kumar v. Government of Uttar Pradesh, 228, 243, 258, 277, 421, 459, 492, 523 
Dippala Suri Dora v. V. V. Giri [See V. V. Girl, v. Dippala Suri Dora] 
Director General of Posts and Telegraphs v. Natarajan, 409, 505 
Dulichand v. Union of India, 376 
Durgaprasada Rao v. Sundarsanaswami, 319, 328 
Dwarka Prasad v. State of Uttar Pradesh, 496 
Ellappa v. Ellappa, 329 
Emperor v. Dholraram, 491 
Gadipalli Parayya v. Boyina Rajayya, 294, 298 
Ganpat v. Presiding Officer, 308, 325 
Garg v. State of Punjab, 207 
General Manager v. Rangachari, 100, 121, 217, 369 ­372, 409 ­412, 421, 487 ­88, 507, 531 
V. V. Giri v. Dippala Sun Dora, 47, 295, 298 ­99, 302, 348, 406, 456, 458, 488, 506, 527 
Gokul Prasad v. Sohani, 217 
Golak Nath v. State of Punjab, 367, 384 
Gopal v. Hanmant, 300 
Gudigar v. State of Mysore, 452 
Gullapalli Nageswara Rao v. Principal, 426 
Gupta v. Union of India, 421 ­22 
Guntur Medical College v. Mohan Rao, 335, 341, 503, 511, 519, 523 
Gurinder Pal Singh v. State of Punjab, 183, 265, 277, 432 
Gurmukh Singh v. Union of India, 310, 312 13, 319 
Gurunath v. Kamalabai, 491 
Gurusami Nadar v. Irulappa Konar, 319, 328 
Hadibandhu v. Banamali, 146, 333 
Hariharan Pillai v. State of Kerala, 196, 202, 218, 242, 248, 255 ­56, 261, 269, 378, 402, 416, 
459, 464, 476, 523 
Harpartap Singh v. Union of India, 244, 440, 456 
Horo v. Jahan Ara, 294 ­95, 335, 340, 348 49 
Hridaya Narain Singh v. Mhd. Sharif, 197, 219, 241 
I. ­ T. Officer Shillong v. Rymbai, 279 
Income Tax Commissioner v. Vazir Sultan, 493 
Income Tax Officer v. Nadar, 491 
Inder Singh v. Sadhan Singh, 301, 318 
Indian Hume Pipe Co., Ltd. v. Their Workmen, 497 
Jacob Mathew v. State of Kerala [See State of Kerala v. Jacob Mathew] 
Jagan Nath v. State of Jammu and Kashmir [See State of J & K v. Jagar Nath] 
Jagdish Rai v. State of Haryana, 392 ­93, 435 
Jagwant Kaur v. State of Bombay, 367, 442 
Jai Kaur v. Sher Singh, 492 
Jaisri v. Rajdewan, 492 
Jaiwant Rao v. State of Rajasthan, 494 
Jalali v. Principal, 525 
Janardhan Surraraya v. State of Mysore, 475 
Janardhana Rao v. Deputy Transport Commissioner, 494 
Jankilal v. Jabarsingh, 304 
Janki Prasad Parimoo v. State of Jammu and Kashmir, 92, 183, 191, 200 ­01, 207, 217, 228, 
238 ­39, 243, 248 ­49, 269, 274, 277, 377, 422 ­23, 438, 452, 453, 526. 
Januma Rai v. Chandradip Rai, 492 
Jayantilal v. Rana, 492 
Jayasree v. State of Kerala, 92, 183, 204, 238, 243, 249, 264, 280 
Jayaswal v. Principal, 275, 342 
Jiwan Khan v. Habib, 306 
Joseph Thomas v. State of Kerala, 216 
Joshi v. Madhya Bharat, 216 
Jyoti Bhushan v. Bodh Ram Muritram, 141 
­622­ 

Kajari Saha v. State of West Bengal, 283, 287, 348, 499 
Kalinath v. Nagendra Nath, 495 
Karkare v. Shevde, 313 
Kartik Oraon v. David Munzni, 293, 295, 298, 317, 348 
Karwadi v. Shambharkar, 307 
Kathi Raning Rawat v. State of Saurashtra, 218 
Kelkar v. Chief Controller of Imports and Exports, 408 
In re Kerela Education Bill, 1957, 42, 370, 478, 538 
Kesava Iyengar v. State of Mysore, 201, 245, 249, 369, 399 ­400, 438 
Keshavananda v. State of Kerala, 481, 538 
Kishan Chand v. Ram Babu, 491 
Kishori v. Board of Revenue, 216 
Kolandei v. Gnanavarum, 326 
Konkani v. Shankar Rao 285 
Kulkarni v. State of Mysore, 372 
Kushma Joshi v. Pro­Vice Chancellor, 384, 427, 445 
Laila Chacko v. State, 195, 263 
Lalita Shuri Tikku v. State of Jammu and Kashmir, 252, 426, 444 ­45, 525 
Lalwani v. Collector, 141 
Laxman Siddappa Naik v. Kattimani Chandappa Jampanna, 287, 329, 487 
Mahadeolal v. Administrator­General, 491 
Maharajah of Kolhapur v. Sundaram Iyer, 296 
Mahendra Nath Pathak v. State of Assam, 217, 282, 291, 364, 373, 446 
Makhan Lal Waza v. State of Jammu and Kashmir, 423, 453, 525 
Maktul v. Manbhari, 492 
Mangilal v. Mittilal­Radheylal Rastogi, 445 
Manickam v. Poongavanammal, 145 
Marthamma v. Munuswami, 328 
Mastanaiah v. Delimitation Commissioner, 48 
Meera Bal v. Director of Medical Education, 315, 522 ­23 
Michael v. Venkataswaran, 306, 309, 312 ­13, 318 
Michael Pillai v. Barthe, 318 
Mira Devi v. Aman Kumari, 306 
Mohd. Abdul Mabood Khan v. State of Maharashtra, 445 
Mohd. Hanif Qureshi v. State of Bihar, 218, 366 
Moosa v: State of Kerala, 373, 442 
Moti Das v. S. P. Sani, 215, 218, 310 
Mulai v. Lal Dan Bahadur Singh, 302 
Municipal Committee v. Hazra Singh, 491 
Muralidhar v. State of Andhra Pradesh, 216 
Muthusami Mudaliar v. Masilamani, 145, 300, 319, 349 
Nain Sukh Das v. State of U.P., 45, 215, 305, 375 ­76 
Nanda Kishore Sharma v. State of Bihar, 196, 524 
Narasimha Reddy v. Bhupathi, 327, 500 
Narayana Swami v. State of Mysore, 431 
Nataraja v. Selection Committee, 306, 339, 343 
Om Prakash v. State of Punjab, 366, 494 
S. G. Pandit v. State of Maharashtra, 439, 467, 475 
Panduram v. Biswambar, 333 
Pandurangarau v. Andhra Pradesh Public Service Commission, 218 
Parameswaran Moothathu v. Vasudeva Kurup, 146 
Pardhasaradhi Rao v. Srinavasa, 495 
Pars Ram v. Shiv Chand, 288 
S. A. Partha v. State of Mysore, 191, 218, 220, 222, 250, 279 ­80, 406, 456 ­57, 465, 471, 474 
­75 
Pasha v. State of Mysore, 345 
Periakaruppan v. State of Tamil Nadu, 199, 259, 380, 426, 438, 451, 460, 484, 523 
Periakaruppan Chettiar v. State of Tamil Nadu, 219, 452, 454 
Phulchand v. Hukumchand, 493 
Plessy v. Ferguson, 494 
Pradip Tandon v. State of Uttar Pradesh, 180, 200, 204, 278, 374, 421, 435, 492 
Pramanick v. Union of India, 446 
Pritam Kaur v. State of Pepsu, 218 
Punjabrao v. Meshram, 144, 307, 310 ­11, 315, 320, 327, 487, 513 
Puppala Sudarsan v. State of Andhra Pradesh, 456 
Raghuramulu v. State of Andhra Pradesh, 406, 455, 458, 464 ­65, 511 
Raj Narain v. Indira Nehru Gandhi, 477 
S. Rajagopal v. C. M. Arumugam [I], 315 16, 318, 327 ­28, 331 
S. Rajagopal v. C. M. Armugam [II] [See C. M. Arumugam v. S. Rajgopal] 
Rajasthan State Electricity Board v. Mohan Lal, 370 
C. A. Rajendran v. Union of India, 377, 379, 396 ­98, 408, 508, 511, 537 
­623­ 

P. Rajendran v. State of Madras, 197 ­98, 216, 233, 252, 262, 267, 426, 477, 488, 519, 523 
Raji v. Deputy Tahsildar, 493 
Raju v. Chief Electoral Officer, 373, 384 
Raju v. State of Gujarat, 420 
Ram Kishore v. Union of India, 491 
Ram Narain v. Director of Consolidation, 491, 494 
Rama Suddarayalu v. Rengammal 492 
Ramakrishna Singh v. State of Mysore, 175, 180 ­81, 190, 201, 222, 241, 246, 249, 314, 400, 
438, 464, 472, 475 
Ramalingam v. Boddu Abraham, 484 
Ramaswami v. Chandra Kotayya, 491 
Ramayya v. Josephine Elizabeth, 328 
Ramchandra Mahton v. State of Bihar, 416 
Ramchandra Vishnu v. State of Madhya Pradesh, 426 ­27 
Ramesh Chander Garg v. State of Punjab, 265, 431 ­32, 459, 474, 476, 519 
Rangachari v. General Manager [See General Manager v. Rangachari] 
Ranjit Kumar v. State of West Bengal, 493 
Ratansi D. Morarji v. Admr. General of Madras, 300, 306 
Rattan Singh v. Devinder Singh, 307, 310 
Reddi v. Savitramma, 494 
Registrar v. Shashi Pal Singh, 428 
Resham Bibi v. Khuda Bakhsh, 306 
Rex v. Ram Dayal, 491 
Rhagava Dass v. Sarju Bayamma, 292 
Rura Ram v. Gurbachna, 207 
Sagar v. State of Andhra Pradesh [See State of Andhra Pradesh v. Sagar] 
Sahdeonarain v. Kusumkumari, 296 
Saifuddin Saheb v. State of Bombay, 146, 334 
Sangannagonda v. Kallangonda, 301 
Sanghar Umar v. State, 215, 314 
Sankalinga Nadan v. Raja Rajeshwari Dorai, 145 
Sankaram Namboodri v. Madhavan, 303 
Sarda v. Mizo District Council, 487 
Sardool Singh v. Principal, Medical College, 277, 384, 427, 451, 499 
Sastri Yagnapurushadasji v. Muldas Bhundardas Vaishya, 146, 309 ­10, 327 
G. Satyanandhan and others v. State of Andhra Pradesh, 446 
Shamdasani v. Central Bank, 500 
Shameem v. Medical College, 263, 275, 344 
Shanshikalabai v. Election Officer, 457 
Shantha Kumar v. State of Mysore, 339, 347, 484 
Sharda v. Visveswariah Engineering College, 444 
Sheokaransingh v. Daulatram, 215 
Sherbert v. Verner, 325 
Shetty v. State of Mysore, 200, 268, 342, 345 47 
Shimoga Durga Bovi v. Returning Officer, 406 
Shri Prithvi Cotton Mills v. Broach Municipality, 492 
Shyam Behari Tewari v. Union of India, 369 
Shyamsundar v. Shankar Deo Udgir, 292, 295, 298, 316, 348 
Sidhrajbhai Sabbaj v. State of Gujarat, 370 
Sri Venkataramana Devaru v. State of Mysore, 146, 334 
Sridhara v. Director of Technical Education, 269, 345, 346 
Srinivasa Aiyar v. Saraswati Ammal, 215 
State of Andhra Pradesh v. Balaram, 184, 200, 227, 238, 248, 254, 377, 426, 438, 460, 466, 
487, 523 
State of Andhra Pradesh v. Sagar, 196, 198 201, 211, 216, 218 ­20, 233, 248, 250, 252 ­53, 377, 
389, 418, 426 ­28, 431, 485 
State of Bihar v. Abdul Majid, 491 
State of Bombay v. Narasu Appa, 215, 301 
State of Bombay v. Yagna Sastri Purushadasji [See Sastri Yagnapurushadasji v. Muldas 
Bhundardas Vaishya] 
State of Jammu and Kashmir v. Jagar Nath, 201, 222, 245, 305, 494, 524, 526 
State of Kerala v. Jacob Mathew, 180, 195, 202, 216, 235 ­36, 254, 258, 426 ­27, 440, 460, 464, 
476, 507 
State of Kerala v. Krishna Kumari, 264 ­65 
State of Kerala v. N. M. Thomas, 145, 147, 220, 244, 249, 302, 361, 379, 382 ­84, 389 ­95, 418 
­20, 423 ­25, 435 ­37, 439, 468, 484 ­85, 487, 494, 508, 541 
State of Madhya Pradesh v. Puranchand, 309 
State of Madras v. Champakam Dorairajan, 164 ­65, 217, 314, 365 ­67, 369, 406, 
State of Madras v. V. G. Row, 221 
State of Mysore v. Narasinga Rao, 217 
State of Mysore v. Bidap, 494 
State of Punjab v. Hira Lal, 92, 220, 425, 440, 487 
State of Rajasthan v. Pratap Singh, 215, 305 
­624­ 

State of Uttar Pradesh v. Deoman, 494 
State of Uttar Pradesh v. Pradip Tandon, 201, 216 ­18, 228, 239, 278, 421, 435, 487 
State of Uttar Pradesh v. Ram Chandra, 491 
Subhash Chandra v. State of Uttar Pradesh, 277 ­78, 492 
Subhash Mohan Jalali v. Principal, Medical College, 101, 450 
Subhashini v. State of Mysore, 426 ­28, 431 32, 450 
Subrao v. Radha, 300 ­301 
Sudama Prasad v. Divisional Superintendent, 445, 505 
Sudha v. Selection Committee, 345 
Sudharkar v. State of Orissa, 425 
Sudhir Kumar Ghose v. Ank Kameya, 121 
Sukhdev v. Government of Andhra Pradesh, 196, 218, 248, 250, 372, 421, 426 ­27, 474, 494 
Sukhnandan Thakur v. State of Bihar, 429, 430, 435 
Sukhvinder Kaur v. State of Himachal Pradesh, 183, 428, 523 
Sundarambal v. Suppiah, 325 
Sunder Devi v. Jheboo Lal, 300 
Superintendent v. Corporation of Calcutta, 491 
Superintendent v. Ram Ajodhva, 491 
Surendrakumar v. State of Rajasthan, 427, 523 
P. Susila v. State of Madras, 282, 304, 444, 511 
Swain v. Secretary, 219 
B. N. Tewari v. Union of India, 403, 476, 523 
In re Thomas, 312, 319, 538 
Thomas v. State of Kerala [See State of Kerala v. N. M. Thomas] 
Trilochan Choudhury v. Dayanidhi Patra, 491 
Triloki Nath Tiku v. State of Jammu and Kashmir, 181, 198, 202, 211, 220, 222, 242, 244, 245, 
252, 261, 379, 399, 423, 444, 453, 506, 524 ­26 
Umesh Chandra Sinha v. V. N. Singh, 197, 427 
University of Madras v. Shantha Bai, 369, 426 
Upoma Kachain v. Bholaram, 300 
Urmila Ginda v. Union of India, 340 ­41, 484 
Vinod Sagar Sood v. State of Madhya Pradesh, 217 
Varadiah v. Parthasarathy, 333 
Venkataramana v. State of Madras, 165, 167, 190, 201, 217, 220, 245, 367, 369, 399 400, 406, 
474, 523 
Vinayak Shamrao v. Moreshwar Ganesh, 491 
Vinod Sagar Sood v. State of Madhya Pradesh, 427 
Vishnu v. State of Madhya Pradesh, 217 
D. G. Vishwanath v. Government of Mysore, 192, 207, 233, 258, 271, 276, 346, 380, 402, 411, 
416, 440, 448 ­49, 451, 454, 502, 532 
Wilson Reade v. C. S. Booth, 294 ­95, 298, 348 
Wasudeo v. Union of India, 141, 143 
Yusuf Abdul Aziz v. State of Bombay, 217 
Yusuf Beg v. Maliq, 318 
­625­ 

Vous aimerez peut-être aussi