STEVEN COLLINS

THE

DISCOURSE
(aa

ON
GAIr

WHAT

IS PRIMARY

S UTTA)

A n Annotated Translation

GENERAL INTRODUCTION (p. 30I ft.)

I

--

Some remarks on the context(s) of AS

(i) in world history: agrarian social order (p. 303)
(ii) in early Indian history: towns and small-scale polities (p. 306)
(iii) in early Indian cultural debate: competing hierarchical models (p. 119)

II -- Some remarks on the text of AS

(i) is the text as we have it a clumsy patchwork (p. 312)?
(ii) what is the nature of the 'satire' in AS (p. 313)?
(iii) AS in relation to other early texts (p. 316)
(iv) notes on the status of my i,~terpretation (p. 322)
III -- The
(i)
(ii)
(iii)

Story of origins, monastic Life and ideals, and the Vinaya

individual verbal reminiscences of the Monastic Code (p. 326)
the Five Impossible Things (p. 327)
'making a store'; the Fall of Mankind and Vinaya infractions (p. 328)

IV -- The structure of AS, and keywords (p. 331)

Notes to the Introduction (p. 334)
TRANSLATION (pp. 338ff.)
Introduction (p. 338)
The Discourse on What is Primary (p. 338)
NOTES TO THE TRANSLATION (pp. 349ff.)
Abbreviations (p. 378)
APPENDIX 1: On the word mahdsammata (p. 379)
APPENDIX 2: Is there a 'social contract' in AS? (p. 387)
BIBLIOGRAPHY (pp. 390ff.)

GENERAL

INTRODUCTION

A l t h o u g h the Aggafiiia Sutta ( h e r e a f t e r A S ) , N o . 27 of the Digha

Nikaya (I) III 8 0 - - 9 8 ) , is o n e o f t h e b e s t k n o w n e a r l y B u d d h i s t texts,

Journal of lndian Philosophy 21:301--393, 1993.

9 1993 KluwerAcademic Publishers. Printed in the Netherlands.

302

STEVEN COLLINS

cited and discussed by scholars in various disciplines, I believe there

are some central facts about it which have not been commented on
before. But I do not imagine for a moment that what I say here
exhausts the riches of this fascinating text; my intention is to encourage rather than foreclose discussion. I aim to do three things:
(i) to prove, by means of close linguistic analysis, that what I shall
call the 'parable of origins' in AS is permeated by references to the
Monastic Code, the Vinaya: this, I think, must condition our understanding of the parable, and our assessment of the spirit in which it
was offered;
(ii) to provide in this General Introduction a context for reading and
interpreting AS, on three levels: in relation to world history, to ancient
Indian society, and to other early Buddhist texts:
(iii) to illustrate by this one case what I think is a more general
desideratum in Buddhist Studies: that the familiar and standard translation of early Pali texts issued by the Pali Text Society, mostly at the
beginning of the century, should be treated not as definitive guides to
the original, as I suspect they are in practice, even by some Buddhologists not specialising in Pali, but as they were intended -- pioneering
attempts in need of constant revision as knowledge progresses. This is
the case, I hope to show, with the Rhys Davids's translation of AS
(1921; hereafter RhD), sub-titled 'A Book of Genesis', which has
become a cultural object in its own right, canonised and immortalised
on our library shelves in one volume of 'the Sacred Books of the
Buddhists'. While I have not catalogued every place where I have
disagreed with their version, I do refer to it often, since the majority
of those who have discussed AS have done so by reading it, rather
than the original. It was an admirable achievement in its day, but I
would like to think that the rendering given here represents an
improvement, in accuracy if not in elegance. (Other previous translations of AS are listed in the Introduction of my own below [p. 148]).
Part I of this General Introduction offers some remarks, of varying
levels of generality, on the socio-historical contexts in relation to
which I find it most helpful to read AS; part II discusses the text of
AS, and some other texts which I take to be relevant to it; part III
collects together the evidence on the basis of which I argue that the
story of the Fall/Evolution of Mankind in sections r 10ff. draws

DISCOURSE ON WHAT IS PRIMARY (AGGAlqlqA-SUTTA)

303

explicitly on themes from the life and ideals of the Monastic Order,
and on language from the Vinaya; part IV gives my view of the
structure of the sutta as a whole, citing occurences of what I see as the
keywords of the text. I refer to the translation notes, which are found
on pp. 159--88, by section and note number; notes of this Introduction follow it on pp. 144--7. The bibliography for both Introduction
and translation notes is at the end.
Readers unfamiliar with AS might read through the translation at
this point, without notes (pp. 148--58); some acquaintance with the
text is necessary if the following Introduction is to be comprehensible.

I. Some remarks on the context(s) of A S

(i) in world history: agrarian social order
The first context I will provide for AS is a very general one indeed. I
draw on Ernest Gellner's vision of the whole of human history thus
farJ For Gellner, mankind has passed through three main historical
stages, those of hunter-gatherers, agrarian society, and industrial
society. These stages are defined in relation to the means (or their
absence) of producing, accumulating and storing food and wealth, to
the forms of coercion and legitimation which accompany them, and
also, in the second and third stages, to the social distribution and
varieties of cognition they encourage. My summary version of his
argument will, of course, be unable to convey anything of its force and
persuasiveness; and it is true that, as a reviewer quoted on the back
cover of the 1990 paperback edition of Plough, Sword and Book
states, 'deductive history on this scale cannot be proved right or
wrong'. Nonetheless -- and leaving aside any problems which arise in
relation to the first and third of his stages -- some things he says
about agrarian society strike me as being plausible, and very revealing
as a background to the concerns and motifs of AS. The main advance
made in moving from hunter-gatherer to agrarian modes, he says, is
the greatly increased presence in the latter, and virtual absence in the
former, of the capacity to produce, accumulate and store food and
wealth. As a result of this, a small surplus is produced (small when
compared to industrial society), but one which is stable enough such

304

STEVEN COLLINS

that 'agrarian societies tend to develop complex social differentiation,

an elaborate division of labour. Two specialisms in particular become
of paramount importance: the emergence of a specialized ruling class,
and of a specialized clerisy (specialists in cognition, legitimation,
salvation, ritual)' (p. 17). These two groups he calls, variously, kingswarriors and clerisy/priests, or most simply thugs and legitimators.
Coercion can take two main forms, corresponding to the two specialisms (which can sometimes be combined): sheer physical force, or
the threat of it, and the imposition of social-ideological norms (the
latter significantly extended in the move to transcendentalist, universalist ideologies of 'salvation'). While in a direct contest, it is absurd
to think that legitimators could overpower the thugs, in the normal
course of settled life social order and control are maintained more
directly by the imposition and internalization of norms than by brute
force.
Although he does not say so directly, I think Gellner is aware of
the relevance of this analysis to two classic themes in the writing of
South Asian history. In the first place, given that the two specialisms
are only possible because of agriculture and the surplus it produces, a
tripartite structure of workers, warriors and priests is not so much a
special feature of Indo-European society, as Dumdzil and his followers
claim, but a structural feature of any agrarian economy producing a
reliable but small surplus. (Note that according to Gellner the enormously increased capacity to produce surplus wealth, and the correspondingly great increase in specialisation and the division of labour,
are among the main features of industrial society.) Secondly, thugs and
legitimators, where they are different, must, since they both exercise
related forms of coercion, come to some sort of mutual modus
vivendi: thus the complex and multivalent relations between kings and
priests, ksatriyas and brahmins 2 (which Louis Dumont saw in terms of
a difference between power and status, a view which has occasioned
much discussion) are again not specific traits of Indian society and
culture but general features of the agrarian order.
The legitimators, of course, are not always of only one party (and
this is one of the weaknesses of Dumont's view, which over-privileges
the Brahmanical vision of society: see below). In the next two sections
I shall apply Gellner's model to early South Asian history and cultural

DISCOURSE ON WHAT IS PRIMARY (AGGAIqNA-SUTTA)

305

debate, and include Buddhism, as a sub-division of what he calls ~the

wider clerisy' (see quote on p. 306). H e r e I want to remain a little
longer with one aspect of the analysis which has special relevance to
AS: the storage of food, and its relation to power. Although towns do
exist in agrarian society, as the seats of market, military and administrative activity, a
society with a small surplus cannot possibly become a generalized market society. By
contrast, in the developed modern world, endowed with an enormous surplus, the
individual hands over his labour and buys virtually all he needs with his wages. In the
physical sense, there generally isn't any thing to hand over: the individual simply takes
part in a very complex activity. With his remuneration, he draws from the market
what he needs for survival, when he needs it. An analogous procedure in agrarian
society would be absurd and disastrous. If the agricultural producer handed over his
entire output and then relied on purchasing what he needs, the first fluctuation in
prices, occasioned let us say by shortages in a neighbouring area, would leave him
starving. In consequence, a very large part of production is stored for safety. Agrarian
society is, in effect, a collection of protected storage units (p. 129).
Agricultural society is defined by the systematic production and storage of food,
and in a lesser measure of other goods. The existence of a stored surplus inevitably
commits the society to some enforcement of the division of that surplus, and to its
external defence. Hence violence, merely contingent amongst hunters, becomes
mandatory amongst agriculturalists (p. 275).
Two comments here: first, in general, this strikes me as a valuable
perspective from which to view South Asian (and other) religious
values of non-violence (ahim.sd), and their relation to the maintenance
of social order. To b o r r o w a phrase from Gunawardana's discussion of
Buddhist monasticism in medieval Sri Lanka (79: 344), there must
always be an 'antagonistic symbiosis' between legitimators who
expound non-violence, and the thugs whose (threat of) physical
coercion maintains the social order which allows the legitimators to
keep their economic and other resources, and to preserve an established role (which is, in ideological theory, outside the realm of
production and reproduction). Second, in relation to AS in particular,
not only is there an explicit connexion, in what I will call its parable of
origins, between the cultivation and storage of food and the origins of
violence and kingship (sections # 17--21), but also the motif of
'making a store' is the central figure around which I will group, on
linguistic and semantic grounds, what I hope to show are a series of
references to the Buddhist monastic code, the Vinaya. Monks, ideally,

306

STEVEN COLLINS

like the beings in the parable when in the 'paradisial' state before their
'Fall' into agriculture and ordered society, neither produce nor store
their food.
I shall say m o r e below on the place of ascetic ideologies as hierarchical models of society in South Asia. F o r the m o m e n t I will end this
section by quoting a little m o r e f r o m Gellner on the issue. Although,
as mentioned,
agrarian society is doomed to violence.., it does not always place violence at the
summit of excellence, though the Western equation of nobility with military vocation
does so. Sometimes it places the scribes/legitmators above the swordsmen, though
we must remember that it is the scribes who write the record and formulate the
principles...
Agraria does on occasion invert values. They [i.e. values] may conspicuously defy,
rather than mirror, the social hierarchy. It may commend ascesis or humility rather
than display, conspicuous consumption and assertiveness. These inversions of values,
of the utmost importance in the history of mankind, can be seen in part as devices
employed by rival elements within the wider clerisy. One way the legitimators gain
influence and power is by being outside the formal system, by opting out, and ascesis
or humility constitutes a kind of conspicuous self-exile. The logic of the agrarian
world, however, does not allow such values to be implemented consistently and
universally. (pp. 154, 155--6; cp. 225) 3
(ii) in early Indian history: towns and small-scale polities
An account such as Gellner's, of course, operates at a very great level
of generality, and describes socio-economic structures very much of
the historical longue durde. It is thus reassuring when a specialist
scholar of early Indian archaeology and history writes that by the
6th--4th. centuries B.C. 'the technological base of the economy in this
period [had] already reached a level not to be significantly exceeded
until the 20th century' - - that is, until the coming of industrialism. 4 In
this section I give a brief summary of what I believe to be a scholarly
consensus on some aspects of the history of North India before the
Mauryan empire, which began in the 4th. century B.C. and reached its
apogee under Agoka in the 3rd. The study of early Indian history
continues to struggle with the problem of assessing the relative weight
of textual and archaeological evidence, but a reasonably clear picture
can be drawn, to the best of our available knowledge. 5 I am not
concerned with precise dating: so m u c h depends on the date of the
Buddha, at present under m u c h discussion. 6 The most likely time for

D I S C O U R S E ON W H A T IS P R I M A R Y ( A G G A N l q A - S U T T A )

307

the Buddha and early Buddhism, it now seems, is the 5th--4th

centuries B.C. There are three main points in what follows: first,
during this period Brahmanism was more strongly established in the
countryside than in the rising urban centres, where a competing
purality of ideologies was emerging; second, these urban centres,
which arose from and encouraged a food surplus, were the market,
military and administrative centres of small-scale polities, not metropolitan capitals of large empires; third, these polities were ruled, in the
earlier part of the period, by oligarchies, and only gradually turned to
monarchy, at the time of the Buddha himself and immediately thereafter. The society apparently presupposed by AS fits just this picture.
At this time, then, as far as the evidence allows us to know, the
countryside of north India was permeated by Brahmanical ideology;
more so in the west than the east, since it had been established there
longer. This was suited to a rural society, where it makes sense to
suggest that Brahmanical social hierarchy could be more easily
stabilised and social order more easily enforceable on the basis of
ritual alone. Urban centres and state formations had begun to arise,
especially strikingly in the north-east, along the Ganges~ The later
Vedic texts produced by Brahmins, the dharma-s~tra-s, show both an
uneasiness about urban life and a concern with 'the laying down of
explicit codes of c o n d u c t . . . That rulers are now explicitly enjoined to
enforce correct behaviour, signals a change in the orientation of
government away from rituals and in the direction of secular administration based o n force'. 7 The connexion between early Buddhism and
urban government and trade, as suggested in its texts, has long been
known. 8 Recently, moreover, Olivelle has argued that the appearance
within Brahmanism of ascetic thought and practice at this time, as
evident from the Upanisad-s, may be the result of urbanized Brahmins
accomodating a trend towards asceticism within their own tradition, as
against the continuing opposition of their (culturally-speaking) 'country
cousins'? Although there is extensive evidence of urban centres at this
9 time, and of a more complex social differentiation than earlier, there
is no evidence of the larger kind of imperial metropolis which arose
from the time of the Mauryan empire. There were a number of
regional divisions, called janapada-s or rnahdjanapada-s, whose names
are given variously in different texts but which are usually said to

308

STEVEN COLLINS

number sixteen. These were not at first equivalent to political units,

but were areas which contained powerful clans, ruled by local chiefs
grouped in tribal oligarchies. But the janapada-s gradually came to be
political units, particularly after the Kosalans conquered the Buddha's
own clan, the Sfikyans (see ~ 8 and # 8.2.); and the transition to
monarchical rule was in process during the Buddha's life. After the
Buddha's death, the 16 janapada-s were reduced to four main rivals;
and eventually that of Magadha became dominant, thus laying the
foundation for the Mauryan empire, centred on Magadha. 1~
After such imperial formations were known, it became possible to
imagine imperial cities to have existed at any period. Erdosi (88:11-2) cities such a description of Ayodhyfi from the Rdmdyana, but adds
'that no such city existed in Ayodhya in the 7th century B.C. (the date
assigned to events described in the Ramayana) is clear from the
archaeological record; so is the fact that the author of the [passage
cited] used the impressive cities of the post-Maurya p e r i o d . . , as his
models' (sic, without diacritics). To say this, of course, is not ipso facto
to criticise the Rdrndyana, since texts are not obliged to provide us
with historical data, nor avoid anachronism. But to say this is to refute
the suggestion that this passage of that text pre-dates the Mauryan
empire. In this light, therefore, the absence of any depiction of
imperial cities and larger-scale political formations in those Pali texts
usually accepted to be early 11 renders it more plausible to trace them
back to the pre-Mauryan period (though we will never be able to
prove this). The view of kingship in AS is not that of a 'universal
emperor', the cakkavatti, found in some canonical texts (although
elements of the AS story were attached to that ideology by the later
tradition). 12 What is perhaps the best-known episode in AS does
concern a single figure, the 'Great Appointee' (mahdsammata -- I
argue in # 21.1 and Appendix 1 that in AS this was a title, not a
proper name as it became later); but this is no more than a narrative
device, as with the first individual who eats the 'earth-essence' in # 12,
the first couple to have sex in # 16, the series of individuals in # 17
who store food, the being in ~ 19 who originates theft, and the series
of individuals from each class who give birth to the 'ascetic-group' in
# 26. In each of these cases the narrative either states explicitly or
implies by ellipsis that the practice spreads from the first individual to

DI:SCOURSE ON WHAT IS PRIMARY (AGGAIqNA-SUTTA)

309

others. Likewide, after the first king is described in ~r 20--1, section

~r 22 refers to rdjadhdniyo, 'royal cities', in the plural; in the text itself
there is no hint of a single territory ruled by one 'universal' monarch,
encompassing (ideally) the whole world; rather, power is held by a
plurality of local chieftains (see # 22.3). As Sharma says (68: 69), 'the
closing passage [i.e. # 21] . . . relates the origin of the khattiyamandala, namely, the ruling oligarchy'.
The rise of urban centers in this period and the production of a
food surplus are obviously connected processes; it would seem that
the two were mutually encouraging. 13 Although no doubt many factors
lay behind the increase in food production, one plausible candidate is
the increased growing of rice (a higher-yield crop than the earlier
dominant barley), and in particular the technique of transplanting rice
in wet-land cultivation. According to Sharma 'although the later Vedic
people grew rice, vrihi [the word then used] was a rainy season crop
whose yield was limited on acccount of its being sown in the field.
Obviously the people did not know the art of paddy transplantation,
or wet paddy production, which appeared later as a winter crop (83:
161--2). Rice grown by the latter method was known as ~dli (ibid.
96). A number of motifs in AS are also found in later Brahmanical
asceticism, which like Buddhism reverses the earlier Vedic celebration
of food as a cosmogonic force; TM they cannot therefore be linked to
any specific period of history. But if we locate AS in the pre-Mauryan
period, it is then not coincidental that its 'origin myth' deals with the
aetiology of the cultivation of gdli (Pall sdli), a surplus of which was
used to create the institution of kings.
Although none of this definitively proves that AS is pre-Mauryan,
in the absence of positive reasons to doubt it, I think it reasonable to
assume so, since the society presupposed in AS's parable fits so well
with what else we know about that period. It is, of course, possible to
remain sceptical; but in the present state of knowledge, if we are to
locate AS in a historical context, the pre-Mauryan period seems the
best candidate.
(iii) in early Indian cultural debate: competing hierarchical models
The main point of the historical specification attempted in the previous section is to give a richer sense of the cultural and ideological

310

STEVEN COLLINS

milieu in which AS was produced. The fact that it seems to presuppose a society like that we conjecture to have existed in northeast
India in the 5th--4th. centuries does not mean that it merely passively
'reflects' its environment; rather, to place it there is to see it as an
interlocutor in an ongoing cultural debate, conducted by various
groups within the ruling strata of the time. In section (i) above, I used
Gellner's basic dichotomy of rulers and legitimators, albeit quoting
from him on 'rival elements within the wider clerisy'; I wish now to
specify more precisely what those rival elements were in ancient India,
and how they, and kings, produced different hierarchical models of
society. I take this phrase from Richard Burghart's 'Hierarchical
Models of the Hindu Social System') 5 Restricting himself to the
'Hindu' world, Burghart claims that kings, brahmins and ascetics each
produced ideologies of the social world which hierarchised it, and,
naturally, placed themselves at the top of whatever value-scale the
model embodied. 'Brahmans, ascetics and the king each claimed their
superiority in the particular world in which they lived', and 'each
person based his claim in terms of a particular hierarchy which was
the exhaustive and exclusive order of social relations'. This Brahmanical hierarchy is expressed in terms of ritual purity and 'the
sacrificial body of Brahma' (sic). Ascetic hierarchy is expressed in
terms of 'the cycle of confused wanderings'; that is, rebirth. (I would
add in the Buddhist case a universal morality which over-rides social
hierarchy of all kinds, at the same time as it upholds the Buddhist
ascetic one: see pp. 130--1 below.) Kingly hierarchy, which is found
more in 'panegyrical and epigraphic sources' than in the kind of text
usually studied by historians of religion, is expressed in terms of a
'tenurial hierarchy which was derived from [the king's] lordship over
the land', a lordship construed as a divine marriage between god-king
and the earth (78: 520--1). We may add to the list of sources for
kings' perspectives texts like the Artha-ddstra, redacted in its final form
not before the 3rd. century A.D. but nonetheless usable for the earlier
period; 16 and, from a later period still the whole tradition of sophisticated court poetry (especially its erotic forms) drama (especially
comedy), and the like. 17 Of course kings would also use themes from
'religious' hierarchical models, particularly in their public pronouncements (as did Agoka): but it is now clear that, for example, Dumont's

DISCOURSE ON WHAT IS PRIMARY (AGGAIqNA-SUTTA)

311

vision of a monolithic India, where everyone was agreed on the status

inferiority of the (power-ful) king compared to the Brahmin, is a
reproduction of one Brahmanical hierarchical model rather than a
comprehensive historical account re cognising a plurality of voices
within India.
The ascetics to whom Burghart refers are, in the 'Hindu) context,
often though not always brahmins. The co-existence of ascetic and
non-ascetic ideologies within Brahmanism has been called by Olivelle
an 'inner conflict of [that] tradition', consciously taking up Heesterman's phrase referring to the relationship between kings and brahmins. 18 In the pre-Mauryan and Mauryan periods one of the most
important facts of society and culture in India as a whole was the
existence and success of what used to be called 'heterodox' (= antiBrahmanical) groups: Buddhists, Jains, and others. As we now know,
in so far as we can speak of a 'Hindu orthodoxy', such a thing was
developed, as a tension-filled amalgam, precisely as a response to
them. Speaking more generally for a moment, the most extensive
evidence we have of non-Brahmanical traditions throughout Indian
history is of Buddhists and Jains: one of the most difficult but most
pressing tasks of Indology, it seems to me, is to explore how far these
traditions (which were obviously not monolithic and without subdivisions of their own) and Brahmanism shared the same language,
both on occasion in the literal sense (Sanskrit) and in the wider sense
of a shared cultural vocabulary, repertoire of stories, etc. To share a
language is not to say the same things in it; and equally difficult and
pressing is the task of assessing how far these various groups used
similar concepts, narrative motifs, and the like, to say quite different
things. Students of 'myth' have, of course, seen this: O'Flaherty (76:
33; cp. 25) introduced a summary of the AS story of origins by saying
'most cosmogonic myths in Buddhism are probably intended as satires
on Hindu myths', both traditions drawing on a common fund of
stories. Recently Obeyesekere (90: 128; cp. 130ft.) has written of 'the
idea of debate [as] the hidden discourse that underlies myth variations'. AS has much in common with other origin stories in India;
Gombrich (92a) has shown that there are specific and pointed references to Brahmanical motifs -- in the form they are available to us, to
specific Vedic texts, which he identifies. The target of the satire in AS

312

STEVEN COLLINS

is, simultaneously, Vedic cosmogony and the social claims of the

Brahmin class: as Smith (89), (92) has shown, the Veda as authoritative text and the social hierarchy were inter-related themes in preBuddhist Brahrnanical cosmogony. Here, as in so many early texts, the
Buddha is represented as knowing very well indeed the Brahmanism
he rejects.
But it is not enough to position the authorial/redactorial voice of
AS in a pluralist, contested milieu of debate, and to speak of it as
presenting a Buddhist-ascetic hierarchical model of society; nor is it
enough to say that Buddhists and Brahmins (to keep to the two
groups relevant to AS) speak the same language but say different
things in it. For there may be differences in the tone of voice in which
things are said; differences not just of content and meaning but also of
style. The scholarly tradition of juxtaposing 'motifs' or 'themes',
familiar in folklore, structuralist (and other) studies of myth, and
elsewhere, may serve to hide from our view the particular qualities of
a text which derive from its tone, its style. But to discuss this is to
move from the context(s) of AS to the text itself; and for this I must
stop and backtrack a little.
II. Some remarks on the text of A S
(i) is the text as we have it a clumsy patchwork?
Some previous scholarship on AS has taken the form of a textual
analysis which sees inconsistencies and illogicalities in it, and then
attempts to separate out its 'earlier' f o r m . 19 I find this approach
unattractive for a priori reasons; I hope to show that it is inappropriate in relation to AS. It is true that the transition from # 9 to # 10
is sudden; and it is true that the 'origin myth', from # 10 up to various
points in # 21, is found separately in later Buddhist texts, usually as a
genealogy of the g~kyan family. But since early Buddhist texts were
composed and transmitted orally, it is no more than common-sense to
assume that different tellings of a tale, in different discursive contexts,
would be different, use elements from a repertoire differently, and so
on. Given this, the mania -- which is what I think it is -- for an 'Urtext' is entirely misplaced. Regardless of its origins in oral composition

DISCOURSE ON WHAT IS PRIMARY (AGGANlqA-SUTTA)

313

and transmission, the tradition has preserved AS in a particular

(written) form; we must, I think, in the first instance seek for meanings
in it as it has been redacted to us. 2~ (I shall return to these issues in [iv]
below.) We should approach the text as we have it respectfully,
looking not to make hasty and superficial judgements about its
disunity, but to seek out principles of structure and sequence which
can give us a sense of why this particular crystallization 2~ of meanings
took the form it did. I think such principles can be found: they are
given in Parts III and IV below, and in the translation notes.
(ii) what is the nature o f the ~satire' in A S ?
Let me return here to the question of the tone of voice in AS. Earlier
I cited O'Flaherty's description of it as satire. That there are humorous
elements in the text has been accepted by all its modern readers.
T. W. Rhys Davids wrote of AS, before any translation of it had
appeared (1899: 107): ~we may not accept the historical accuracy of
this legend. Indeed, a continual note of good-humoured irony runs
through the whole story, with its fanciful etymologies of the names of
the four vann& and the aroma of it would be lost on the hearer who
took it au grand s&ieux'. Tambiah (89) quotes these words of Rhys
Davids with approval, and argues, as he had done in (76) that 'behind
the mockery directed at Brahmanical beliefs, there is a positive
countervailing Buddhist account of the origins and evolution of the
world, kingship and social differentiation'. 22 He follows here, as
earlier, Reynolds (72: 18), who wrote of a 'positive interpretation of
the nature and function of royal authority' in AS. Gombrich (88: 85),
seemingly to the contrary, writes that AS is 'an extended satire on
brahminical ideas, full of parody and p u n s . . . As a debunking job I
think the sermon is serious: its main aim is to show that the caste
system is nothing but a human invention'. He adds there that 'I cannot
go here into all the reasons why I think that the positive statements in
the myth are satirical and not meant to be taken literally'. In his (92a)
article he does so.
If there is disagreement here, what exactly is at issue? When Rhys
Davids wrote of the text's 'historical accuracy', he was of course
speaking in 19th. century historiographical terms: the story in AS, for
us, cannot be taken to be, in von Ranke's phrase, wie es eigentlich

314

STEVEN COLLINS

gewesen ist. No-one, I take it, would wish to disagree. (Even so, the
quotation from Rhys Davids just given continues 'but it reveals a
sound and healthy insight, and is much nearer to the actual facts than
the Brahman legend it was intended to replace'.) There would seem to
be an equivocation over the word 'positive'. Gombrich is using it
partly in a historiographical sense, much as did Rhys Davids. Tambiah
and Reynolds, it seems to me, are using the word in much the same
sense that Gombrich intends when he says that the sermon is 'serious';
but they see the text, in addition to satirising Brahmanism, as intended
to provide a non-satirical Buddhist charter for social arrangements.
The issue then becomes one of what, if anything, we can say about the
original sense and motivation of AS. Tambiah explicitly eschews the
question: 'I must confess I was not there when the Buddha gave this
discourse, nor was I able to ask him what he actually meant'; one
should not, he thinks, 'take an absolutist stand, and insist on a single,
unambiguous formulation of authorial intent' (89: 120; 102). (It is thus
unclear, to me at least, exactly what exegetical status Tambiah accords
to his own assertion that 'behind the m o c k e r y . . , there is a positive
. . . account'.) Gombrich, on the contrary, wants to 'discover the
original meaning of the Buddha's sermons' (92a: 160). I do not accept
that we have only two options, either finding an 'original meaning' or
abandoning ourselves to a free-for-all relativism, in which a text 'has
no objective or inherent meaning' (ibid. 159). Varying readings of any
text are always possible: but I think we have a responsibility to argue
for different readings, some of which must be judged better than
others. In this article I argue that AS was intended by its earliest
composer(s) and redactors to be a humorous parable: its serious intent
was as moral commentary rather than as a 'myth of origins -- charter
for society' or an account intended to be 'factually' or 'historically'
accurate.
But what, again, is at issue here? One of the most discussed aspects
of AS is its apparent proposal of a Social Contract theory of kingship
(see Appendix 2). In western political thought it is still an uncertain,
and in some ways now unimportant question how far social contract
theorists believed their accounts of 'the Original Contract' to be
historically factual descriptions of an event or allegories giving the
justification for legal sanctions. 23 If western political thinkers in the

DISCOURSE ON WHAT IS PRIMARY (AGGAIqICqA-SUTTA)

315

last few centuries can be equivocal on the issue or unconcerned with

it, why should we assume that the redactors of an ancient Indian
narrative of the kind found in AS either did, or would have wanted to
distinguish in the m o d e r n manner between 'empirical history' and
'legitimatory allegory'? I am certainly not saying that such a distinction
would have been in all circumstances impossible: I am saying that it is
not important in the interpretation of AS. Some words of T. S. Eliot
seem apposite here: 'this alliance of levity and seriousness (by which
the seriousness is intensified) is a characteristic of the wit we are
trying to identify'. 24 One might adduce another analogy from the
history of ideas in Europe. The utopian fables of Lucian seem to
modern scholars to have been intended as satires; but later utopian
writers regularly used themes from them non-satirically. 25 But here
again, the distinction between 'literal' and 'allegorical' intent is often
difficult to draw, and often unnecessary. The founder of the modern
genre, Thomas More, seems to have intended the title of his Utopia
specifically as an ambiguous (Greek) pun on ou-topia, 'No-place', and
eu-topia, 'Good Place'. The ambiguity, both intended and not, between
utopian texts as descriptions of real (actual or possible) societies and
as fictions whose purpose is to criticize the writer's actual society runs
throughout the utopian tradition. Many texts may be counted as much
part of the tradition of (serious) political thought as of 'utopianism',
simpliciter. 26
Indeed -- to bring this discussion to a close -- it seems to me that
this issue can be seen to arise, in its most general form, from the
nature of AS as a text. It is a commonplace nowadays to say that a
text by definition presents us with a world, which cannot be the
world. (I leave aside the difficult issue of how we can have cognitive
access to the world outside textual representations of it.) ff this is so,
if the world of any text is necessarily ou-topia, a 'No-place', in relation
to the real places of the material-historical world, then any composer/
redactor/recounter of a text must present a world whose reference to
the world is always up for discussion. It seems reasonable to assume
not only that the ambiguities of that reference are not necessarily a
matter of explicit concern to a text and its users, but also that they
might offer -- as in the case of western Social Contract and Utopian
traditions -- an opportunity for creativity. 27 I would suggest that AS

316

STEVEN COLLINS

remains a lively and interesting text just because of this question

(amongst other reasons), not in spite of it.
One last point needs to be dealt with here: that of what Rhys
Davids called the 'fanciful etymologies' in AS. He was referring to the
eight terms for social classes given in # 21--5, and also by implication
to the other things (surprisingly -- see notes # 13.4., 16.6, and
Appendix 1) called akkhara-s; that is, the phrases 'oh the taste!'
( # 13) and 'we've had it, it's given out on us!' ( # 15), and the custom
of throwing dirt, etc., at weddings ( # 16). In the case of the aetiologies
given for the two phrases and the non-linguistic custom, I think there
is clearly deliberate humor; and likewise with the derivations given for
two of the terms for the brahmin class in # 22--3 (see notes ad loc.). I
argue ( # 21.1. and Appendix 1) that there are associations of the term
sammata which render the choice of the title mahdsammata witty. But
with the other five terms there is no humor obvious on the surface -indeed the etymology given for rgtja in # 21 also occurs in the
Sanskrit MahSbhSrata, without overt humor; and the last two are now
rather opaque. Similar word-derivations are found in Pali canonical
and commentarial texts where there is no humorous intent. 28 Such
word-derivations were a long and established tradition in both
Sanskrit and Pali; they are called nirukti-s in Sanskrit (the Pali subcommentary to the first use of akkhara in AS # 21 explains it as
nirutti, D A T III 59--60). 29 For modern historical linguistics, these are
indeed 'fanciful etymologies'; but the tradition of nirukti existed
alongside that of vySkarana, which we can recognise as very unfanciful, indeed as 'scientific' grammar. Both of these traditions,
however, two of the traditional six kinds of Vedic study, are equally
a-historical: 3~ one might say that, for us, vydkarana grammar is
a-historical but scientific, nirukti etymology neither historical nor
scientific. My conclusion in relation to AS is that while such wordderivations are not intrinsically humorous or 'fanciful', they are offered
in this text, along with the three aetiologies, in such a way as to add to
the tone of ironic and polemical wit. But the wit arises from the
content of the aetiologies and 'etymologies', not their form.
(iii) A S in relation to other early texts
I want in this section to discuss some other texts, which like AS we

DISCOURSE ON WHAT IS PRIMARY (AGGAI~lqA-SUTTA)

317

have no reason to deny are early, and which seem to me to help clarify
its meaning and style. First let me summarise my approach to AS. I
take it to be a story whose raison d'etre is to present a Buddhistascetic hierarchical model of society, offered with satirical and ironic
wit in the manner of moral commentary, and with the discursive form
of an aetiology. (For this reason I prefer to call its story of origins a
parable rather than a myth.) Buddhist monasticism and morality order
the logic of values and social relations: Brahmanical values are
satirized and kingly values subordinated, albeit that neither the
Brahmanical hierarchy of discrete social classes nor kingship are
contested as 'social facts' (in the Durkheimian sense). This is the
overall theme which structures the whole of AS as we have it, and
which gives it unity and coherence. Brahmanical social classes are seen
not as a cosmogonic 'success story' but as a 'Fall': this Fall/Evolution
of Mankind is expressed in language and values derived from the
Monastic Order. The earliest Community of Beings, what one might
call the earliest Sam.gha, falls from 'glorious mind-made' celibacy to
the contemporary, embodied Brahmanical social order, by a series of
deeds which are described in the language used for the corresponding
contraventions of the Monastic Rule, the Vinaya. Thus social classes
and kings constitute a 'Fall' from an originally Buddhistic community
(aggaFlfia in the temporal sense of 'what is primary'); Buddhist moral
values, which are laid out systematically in AS twice, before and after
the story of origins, in # 5--7 and # 27--30, are 'what is primary'
(aggatMa in the evaluative sense: see p. 331 below). There are references to Brahmanicat texts and practices in all parts of the text, and
the story of king Pasenadi in # 8 is thematically continuous with the
understanding of kingship and the ks.atriya class given in the parable.
Leaving aside whatever general value these introductory remarks
might have, I hope that what I take to be my discovery of references
to the Vinaya in AS represents a genuine contribution to knowledge.
But if it is accepted that these references exist, one might then raise
the general issue of what such 'reference' means: how far, and in what
ways, can we use other texts to elucidate AS? Overall, there seems to
me no a priori solution to the question of inter-textuality. There can
be no way of proving that all, most or any audiences for individual
Buddhist texts would have interpreted them in the light of others: we

318

STEVEN COLLINS

must argue from internal evidence. I think that the references to

Brahmanical ideas and texts for whose presence Gombrich argues, and
those to the Vinaya I adduce here, suggest that AS was composed in
and for an educated milieu familiar with both styles of thought, one
which could smile at its wit, and appreciate its serious intention. If that
were the case, then perhaps we may suppose, not that other texts were
automatically in the minds of its audience in an academic-commentarial manner, but that certain lines of interpretation would be made
possible (or impossible) by what is found in other texts on comparable
themes, and with which an educated Buddhist audience might also be
expected to be familiar. That such texts are directly relevant to AS
must be shown not merely by thematic resemblance, but by specific
textual cues: similarity of phrasing, shared characters, etc. On this
assumption, then, I will describe briefly some texts from the Sutta
Nipdta and Majjhima Nikdya which I think help us to appreciate the
nature of the ascetic hierarchy presented in AS, and to get closer to
assessing the spirit in which its parable of origins should be read.
The Buddha's discussion in AS is said to have taken place with two
young brahmins, 'aspiring to become monks', who bear the names of
two famous old Brahmanical families, Vfisettha (Sanskrit Vfisistha) and
Bhfiradvfija. They appear in two other well-known texts in which
Brahmanism is criticised, the Tevijjd Sutta of the Digha Nikdya (no.
13), and the Vdset.thaSutta, redacted in both the Sutta Nipdta (Sn. pp.
115--123) and the Majjhima Nikdya (no. 98). (This very fact suggests
that the Vdset.thaSutta might be well-known to Buddhist audiences.)
At the start of the latter, before their conversion to Buddhism, they
declare themselves 'adept in the three v e d a s . . , philologists, grammarians, like our teachers in (vedic) recitation', 31 but in disagreement
as to whether a person becomes a brahmin by birth (jdti: cp. AS
# 3--4) or by action (karma, Pali kamma). The Buddha's reply first
states that although 'manifold indeed are [the] species of living beings'
(the word for 'species' here is also jdti, which is in addition the usual
word for 'caste'), there are no such sub-divisions among human beings:
'among men difference in spoken of as a matter of designation'
(samafi~d: cf. Appendix 1 for this as a commentarial gloss on akkhara
in AS, and below on the Discourse at Madhurft). Thereafter, as

DISCOURSE ON WHAT IS PRIMARY (AGGA/~I~IA-SUTTA)

319

ubiquitously in early texts, the Buddha uses the designation 'brahmin'

to refer to one who lives according to Buddhist values, not to the
members of a particular group. 32 First, he says that 'whoever among
men makes his living by keeping cows . . . is a farmer not a brahman';
then the same thing is said of various 'occupations': craftsman,
merchant, servant, thief, fighting man, sacrificer (one who lives by
porohicca, a reference to the Brahmanical office of purohita) and king.
'Nor do I call (him) a brahman (who is) born in a (particular) womb,
and has his origin in a (particular) mother' (yonija... mattisambhava;
cf. AS # 4). After a long and lyrical evocation of the values which
define a (true, Buddhist) brahmin, each of these occupations is said to
be a matter of 'action', and 'the supreme state of being a brahman' is
said to be attained by the religious life of celibacy; any person who
has reached liberation (in Buddhist terms) 'is Brahmfi (and) Sakka [i.e.
the Brahmanical gods Brahmfi and Indra] to those who know'.
In this text the two youths are first presented as pupils of 'distinguished and wealthy brahmins', in dispute over an issue central to
Brahmanism; at the end they declare themselves lay-followers of the
Buddha. Exactly the same thing is true of the Tevijjd Sutta, in which
the dispute concerns their respective Brahmin teachers, and where the
closing refrain is identical to that of the Vdset.t.haSutta. In AS they are
presented as living with the monks at Sfivatthi, and aspiring to become
monks (themselves). The commentaries to both the Tevijjd Sutta and
AS (Sv 406, 860) connect the three stories about the two youths into
a continuous narrative: after the Vdset.tha Sutta they declared themselves lay-followers; after the Tevijjd they did so again, but thereafter
(after a few days according to Sv 406) took the Minor Ordination to
become Buddhist novices. At the start of AS they are aspiring to
become monks, hoping to take the Major Ordination. (Sv 406 says
that after AS they did so, and attained liberation; cp. Sv 872.) Being
formerly 'adept in the three v e d a s . . , philologists, grammarians, like
our teachers in (vedic) recitation', they would thus be a good audience
both for the references to Vedic hymns and for the 'etymologies' in
AS; as novices they would presumably be becoming familiar with the
Vinaya rules, 33 and thus also a good audience for the references to it
in the parable of origins. (Obviously, in a historical sense, the 'audi-

320

STEVEN COLLINS

ence' for AS was much wider; but if that wider audience were familiar
with the characters of Vfisettha and Bhfiradvfija, they would be appropriate as the represented audience.)
In another text of the Sutta Nipdta, the Discourse on 'Brahmanical
Lore' (Brdhmanadhammika-sutta) there is a discussion at Sfivatthi
between the Buddha and wealthy brahmins, but this time, as the text
insists, they were 'aged, old, elderly, advanced in years, in their old
age'. With the sense of passing time perhaps characteristic of old age,
they ask the Buddha whether 'brahmans n o w . . , live in conformity
with the brahmanical lore of the brahmans of old' (pordn.dnam.
brdhmandnam brdhman,adhamme; cp. AS # 4.1 and # 13.4 on
pordn,a). He replies in the negative, and paints a picture of such old
Brahmanical lore which blends motifs from both Buddhist and Brahmanical asceticism. They had no wealth, but begged their food from
door to door; they lived as celibate students for 48 years, 34 and then
married within their own group without bride-price; those who were
married had sex only at the right time, while 'the supreme brahman
did not indulge in sexual intercourse even in a dream'; 3S their
sacrifices did not involved killing cows. However, they began to covet,
inter alia, the wealth and 'excellent women' of kings, and so began
composing hymns 36 to acquire them. 'And they, receiving wealth there,
found pleasure in hoarding it up' (Sn 306, dhanam laddhd sannidhim
samarocayum: see below on 'storing up [objects of] desire'); they
began the 'ancient mean practice' of cow-sacrifice. Because of this, the
other three social classes were 'split up' (the commentary, Pj II 324,
interprets this to mean that they no longer lived in harmony). This
text, like AS, criticises Brahmins by saying that they have forgotten the
past; both recount a narrative of their degeneration from an ideal. The
two narratives are, on the surface level of a temporal sequence of
actual events, quite different; but when read as parables using stories
of the past to make a contemporary moral point, they complement
each other perfectly well.
I have said that AS (and related texts) presents a Buddhist ascetichierarchical model of society; but it is a complex one with an inner
dynamic of its own. The moral values on which the hierarchy is based
are often, indeed normally, correlated with social status: monks and
nuns are 'above' all laity, including Brahmins and kings9 But this
9

D I S C O U R S E ON W H A T IS P R I M A R Y ( A G G A N N A - S U T T A )

321

correlation is not held to be automatic or intrinsic. AS # 21--6,

concluding the story of origins, tells of the birth of five social 'groups'
(man.d.ald: the four Brahmanical 'classes' and ascetics, samand).
Immediately thereafter, sections r
teach that anyone in any of
five social groups can go to hell, heaven, or attain liberation from
rebirth. Although the ascetic is in a very much better position to
practise the 'right conduct' necessary for liberation, that goal neither
follows automatically from the social status of 'the ascetic-group' nor is
it confined to it. This does not, I think, make a great difference in
practice. Although monastic status is not a necessary condition for
liberation, the celibate life is: it is quite impossible to conceive of the
religious life described in early Pail texts being practised fully by those
sexually active. But it does make a difference in interpretation, in what
the representation of 'society' is trying to achieve. Thus although AS
states twice ( ~ 7, ~ 31) that 'of these four classes.., the arahant is
what is called primary', the ascetic hierarchy is not itself a social
hierarchy with intrinsic values, despite the fact that, as # 8 shows,
even kings do honour to ascetics.
This point is, I think, supported by the only other texts in the Pali
Canon where the sentiments attributed to Brahmins in AS # 3 and
# 4 are found. In Sutta 84 of the Majjhima Nikdya (M II 83--90), the
Discourse at Madhurfi, a king hears that a Buddhist monk whom he
calls 'the ascetic (samana) Mahfikacc~na' is living nearby; he goes to
him and asks what he thinks of the sentiments expressed by Brahmins
(as in AS ~ 3, from 'Brahmins are the best class' down to 'heirs of
Brahma'). Mahfikaccfina replies that this is just 'soinething people say'
(lit. 'an utterance in the world', ghoso yeva.., lokasrnirn.); and in
elucidation he gives four arguments to persuade the king that 'the four
classes are exactly the same' (samasamd). They are: (i) individuals
from any of the four classes can, with enough money, have someone
from any of the four classes as a servant; (ii) good and bad behaviour
lead to heaven and hell, for anyone from any class (this passage
blends elements from AS # 5, # 6, # 27 and # 28); (iii) if an individual from any of the four classes commits crimes such as bu@ary or
adultery, the king punishes him, by execution, banishment, etc. (cp. AS
# 20): in such a case 'the previous designation (sama~d) "ksatriya"
(etc.) has disappeared, and he is reckoned simply as "thief" '; (iv) if an

322

STEVEN COLLINS

individual from any class leaves home for homelessness, and lives
virtuously, the king 'salutes him respectfully, rises up from his seat for
him' (as in AS # 8), supports him materially and affords him protection: in such a case, 'the previous designation "ks.atriya" (etc.) has
disappeared, and he is reckoned simply as "ascetic"' (yd hi ' s s a . . .

pubbe khattiyo ti samafifid, sd 'ssa antarahitd, samano t' eva satikharn

gacchati). Two things are worthy of comment: first, it is a condition of
such respect that the ascetic be virtuous; second, as in the Vdsettha
Sutta, the designation 'ascetic' is no more an ascribed social status
than is 'thief'. Both can be acquired by behaviour, by an individual
from any of the four ascribed Brahmanical classes.
In Majjhima Sutta 93 (M II 147--157), the Discourse with
Assal~yana, the Buddha persuades the eponymous Brahmin to agree
that 'purity belongs to (all) four classes' (cdtuvanni suddhi), in opposition to the notion that Brahmins are the best class, etc. He does this
by means of a long series of arguments, of which the first three are
relevant here. The first is that that Brahmin women give birth in the
normal way, and so Brahmins are not 'born of Brahm~' (the same as
AS # 4: see # 4.2). The third is exactly the same as (ii) in the
Discourse at Madhur~ above, blending language from sections # 5,
# 6, # 27 and # 28 in AS. The second is that 'in Yona, Kamboj~ and
other neighbouring areas' there are only two classes, masters and
slaves, and that these positions can be reversed (a striking anticipation
of Hegel). If we can assume that the narrative voice in texts which use
the same language to make the same points is homogenous, it follows
that AS cannot be thought to offer a myth of cosmogony and society
which takes for granted the universal existence of the four Brahmanical classes, when they are specifically said in the Assal~yana Sutta
to be local and contingent arrangements. When the two texts are taken
together, we must take the story of origins in AS to be a parable
exemplifying a moral truth rather than an account intended to convey
a simple (and single) historical truth.
(iv) notes on the status of my interpretation
It might be useful to offer some remarks on the status of the interpretation offered here. As mentioned earlier, the story of origins, from
section # 10 to some or all of # 21;circulated separately elsewhere in

DISCOURSE ON WHAT IS PRIMARY (AGGAIqNA-SUTTA)

323

the Buddhist tradition, from at least the time of the Mahdvastu and
the M~lasarvdstivdda Vinaya. The Pali commentarial tradition seems
not to have understood the text in the way I do, or if it did has not
said so; the commentary does notice some jokes (see # 4.3) and both
commentary and sub-commentary are aware of one of the references
to Vedic ideas (see # 3.3), but neither mention the Vinaya. For some
readers, perhaps, this in itself might be enough to render what I say an
over-speculative and purely modern reading. Ultimately, the worth of
my interpretation must be judged on its own merits, according to the
weight of the detailed evidence and arguments adduced. Modern
scholars of Classical Greek texts are not necessarily concerned if their
interpretations of them were not shared by commentators in the
intervening centuries. I do think that, in the Pali Buddhist case, there
are reasons for paying great attention to the commentarial tradition
when considering the development of systematic and context-free
religious doctrine, and the way the later tradition understood (and
redacted) its canonical texts from that point of view. But with a text as
context-sensitive as I believe AS to be, perhaps we might accord
ourselves greater interpretive autonomy. I do not wish merely to say
that, as a reader, I have the privilege to respond in any way I like to a
text, and so my interpretation is self-justificatory. I want my reading to
be accorded historical value, to be seen as discovering motifs and
intentions genuinely present in the text and in the minds of its original
composer(s) and (at least some of) its original audience(s). If this is so,
how can I explain the fact that other texts present versions of # 10-21 with apparently straightforward seriousness, and that the commentarial tradition has failed either to notice or to mention the references
to the Vinaya I see in AS? I will sketch out an answer to these
questions, but I do not expect to close the issue decisively -- I am
content to leave the matter open for discussion.
If I am right, AS (that is, some oral ancestor of our written text)
was originally composed in the pre-Mauryan period in Northeast
India, in the socio-economic and political circumstances described in
Part I (ii) and (iii), circumstances discovered by modern historiography.
Although it would be wrong to allege that ancient South Asians had
no 'historical consciousness', it is hardly surprising that memories and
narratives of the past were then deployed, in texts, in ways very

324

STEVEN COLLINS

different from those of modern socio-political history. So perhaps it is

no surprise that the picture which modern historians reconstruct, and
which I have used to elucidate AS, was not of concern to the Buddhist
tradition. Both the Mahglvastu and MVdasarvdstivddin Vinaya, in the
form we have them, are without doubt post-Mauryan (although they
may preserve some materials from earlier times), 37 and are contemporary with what I called earlier the 'tension-filled amalgam' of 'Hindu
orthodoxy', which evolved as a successful response to the challenge
posed by non-Brahmanical groups during the Mauryan empire.
Perhaps in such changed socio-political and cultural circumstances the
situational relevance of AS was lost. I cited earlier the fact of Lucian's
'satires' being treated non-satirically by later writers in Europe. AS
may be a parallel case. Furthermore, while I hold that AS as we have
it is a coherent and continuous whole, with lexical, semantic and
thematic elements common to both the parable of origins and its
frame, there would seem no reason to deny that some comparable
story of origins could have existed separately in the oral culture of
early Buddhism. If so, and if later composers wished to use it in other
ways, they may have adopted the version redacted in AS, without
worrying that the connexions between the origin-story and its flame
therein would be lost. Once the flame-story had been excised, and
with it what I have called the aspect of moral commentary in AS,
there would remain little significance in the references to the Vinaya,
and so they too would have been ignored.
What of the commentarial texts, redacted as we have them in Sri
Lanka in the second half of the 1st. millenium A.D.? Gombrich (92a:
160--1) has argued that the composition of the commentaries was
'separated from that of the suttas not only in time but also in space'
(that is, in South India and/or Ceylon centuries later); this is undoubtedly true of their present redaction, but it may be that some
commentaries preserve much earlier material.3s In the case of the
commentaries on AS, I must simply argue that the fact that they misunderstand and/or ignore many of the references and meanings that I
allege it contains, is in itself evidence that they were not composed in
close temporal and spatial cultural proximity to it. (In the case of
resonances with the Vinaya rules, perhaps they were too obvious to
need mention?) Certainly the cultural circumstances of Theravfida

DISCOURSE ON WHAT IS PRIMARY (AGGANI~A-SUTTA)

325

Buddhism in Sri Lanka and mainland Southeast Asia were very

different from those of early (and, indeed, later) Buddhism in India.
Indian Buddhist texts were produced in a milieu of constant and
endemic ideological plurality: kings, Brahmins, Jains and others all had
their own hierarchical models of social relations. In Sri Lanka,
however, although it would be severely mistaken to assume that the
Mahfivihfirin version of Theravfida which survives was ever the only
and unchallenged version of Buddhism, 39 and although there were
periodically kings who favoured some version of what is now called
'Hinduism', by and large Pali commentaries were produced in a
cultural situation where articulated alternative visions of society were
very much less powerful. Accordingly here, as in Southeast Asia after
Theravfida had been 'established' there by kings, given that Brahmanical ideology was not a significant competitor, Buddhist intellectuals had merely to work out some modus vivendi with kings. Thus
as a system of thought Buddhism became more concerned with
cosmology and cosmogony, in the sense that it became more concerned to give its own ideological grounding of the existing social
'world' than it ever was in India, where it could preserve its stance of
moral commentary on Brahmanism. It would, indeed, also be severely
mistaken to see such later Buddhist texts as proffering a whole and
complete legitimation of social and political life: no transcendental
salvation system can ever be without some tension with mundane
matters, and I prefer to speak, as earlier, of an antagonistic symbiosis
between king-thugs and Buddhist-legitimators. Both the antagonism
and the symbiosis could fluctuate in different times and places. But it
becomes less surprising in such a context that the story of origins in
AS should no longer be read as a parable, and that, since elements
from it were now detached from their original context in AS, 4~ it
should take on the force of a legitimatory myth-charter. Here again,
once the references to Brahmanical ideology and the socio-moral
context of AS had been lost, it is plausible to assume that the allusions
to other Buddhist texts I adduce here would also have been lost, or
ignored.

326

STEVEN

COLLINS

III. The Story of Origins, monastic life and ideals, and the Vinaya
(i) individual verbal reminiscences of the Monastic Code (Vinaya)
Here I simply list the relevant places: the translation notes contain the
textual and linguistic detail.
# 11ft. The first food-stuffs likened to ghee (sappi), cream (navanita)
and honey (madhu): three of the five 'medicines' allowed to monks
and nuns (Nissaggiya Pficittiya 23, et freq.; see below).
# 12. Tasting the 'earth-essence' with the finger: contravenes Sekhiya
rules 52 and 53 (see # 12.2).
# 12. Taking (big) mouthfuls with the hands: contravenes Sekhiya
rules 39, 40, and (possibly) 42 and 46 (see # 12.3).
# 16. Having sex: contravenes Pfirfijikfi 1. Note the use of dpaffati in
# 17 (see # 17.3). 'Away with you and your impurity!' (nassa asuci):
recalls use of ndseti as a technical term for expulsion from the monkhood (see # 16.4 and 5).
# 17. Making houses (agdrdni... kdtum): contradicts the fundamental
symbol of monastic life, 'going forth from home to homelessness'
( agdrasmd anagdriyam, pabbaffd). Such pabbaffd constitutes the (first)
change of status from layperson to (novice) monk, the 'lower ordination'. Houses are said to be made for the purpose of concealing/
covering immorality; cp. the ubiquitious motif 'dwelling in a house is a
constriction.., going-forth is an open-air life'. 41
# 17. Storing food for 8 days: contravenes Nissagiya Pficittiya 23 and
Pficittiya 38. Note the grammatical peculiarity of sannidhi-kdrakam
(see below pp. 328--9).
# 18. 'Setting a limit' (mariyddd): may recall monastic boundaries
(sTmd) (see # 18.2).
# 20. The verb khiyati, 'become angry': found standardly in this sense
only as a formulaic expression in the Vinaya (see # 20.1).
# 21. The term mahdsammata: modelled on monastic appointments
(see Appendix 1).

DISCOURSE ON WHAT IS PRIMARY (AGGAIqI~A-SUTTA)

(ii)

327

the Five Impossible Things

These verbal reminiscences are already enough to suggest that AS is

deliberately alluding to the Vinaya; this is proven, I think, by the
semantic and lexical parallels between AS, the Vinaya Code, and a list
of ascetic ideals found in a number of texts, which give five 'impossible
things' (abhabba-tthdndni), five things which an enlightened monk
cannot commit (so abhabbo pa~ca thdndni affhdcaritum). The similarities are striking, albeit that many of the phrases are common in
other texts. The five things are listed, in only very slightly differing
forms, at D I I I 133 (in a list of nine), 235, M I 523 and A IV 370; I
cite the version at M l 523. The first four are similar in content,
though not in language, to the first four of the Five Precepts, incumbent on all Buddhists; the specific wording of all five recalls directly
various rules from the Vinaya code:
'It is impossible that a monk who is an Arahant, in whom the corruptions have wasted away, who has lived the (holy) life, done what has
to be done, laid down the burden, attained the true goal, in whom the
fetters of existence are destroyed, who is released by right wisdom,
should commit the five crimes.' (This list of epithets is found very
frequently, and appears in AS # 7, 31) They are:
1. 'He cannot intentionally deprive a living thing of life' (sa~cicca
pdnam, fivitd voropetut.n). The wording here is identical to Pficittiya 61,
and almost identical to Pfirfijikfi 3, against killing human beings

(sa~cicca manussaviggaham fivitd voropeyya).

2. 'He cannot take what is not given, intending to steal it' (adinnam
theyyasar.nkhdtam, dddtum). This is identical to Pfirfiiikfi 2. The texts
explain theyyasam,khdtam as referring to the intention to steal (Vin I
46, Kkh 26--7) (but see # 20.2).
3. 'He cannot have sex'
identical to Pfirfijikfi 1.

(methunam. dhammam patisevitum). This is

4. 'He cannot tell a conscious lie' (sampajdnamus6 bhdsitum). This is

identical to Pficittiya 1 (the first ten Pficittiyas are called the 'chapter
on lying', m~sdvddavaggo); it also recalls Pdrdfika 4, against falsely
claiming higher spiritual achievements.

328

STEVEN COLLINS

5. 'He cannot enjoy (objects of) desire, making a store (of them) as
he formerly did when living in a house' (sannidhikdrakam. kdme
paribhufifitum, seyyathdpi pubbe agdriyabh(~to). This recalls both
Nissaggiya Pficittiya 23 and Pficittiya 38: the similarity calls for
extended comment.
(iii) 'making a store ~ the Fall of Mankind and Vinaya infractions
The word kdma can be both subjective and objective, referring to
desire and its objects. While it stands for attachment to anything and
everything in Buddhist psychology, there is clearly an emphasis on
sensual and sexual pleasure. The verb bhuj, used here, can mean to
eat, and also to consume or enjoy in any and every sense. In this
context all these connotations of the word are in play, with the idea
that the householder stores both the actual objects of his enjoyment
(from food to wife) and the psychological propensity to desire them.
The form of the word sannidhi-kdrakam is of particular importance
here. The suffix -kdraka is usually used to refer to a person or process
which makes something, or to the act of making. Here neither sense is
syntactially appropriate, and the word seems to agree with nothing in
the sentence. The commentary to the passage cited (Ps III 234) glosses
it as an absolutive or gerund, sannidhim katvd, 'having made a store',
and states that such a monk cannot eat foods such as sesame, husked
rice, ghee, cream, etc., which he has stored for present consumption,
as he did when a layman enjoying material pleasures (yathd pubbe

gihibhg~to sannidhim katvd vatthukdme paribhu~]ati evam tila-tand,ulasappi-navanitddini sannidhim katvd iddni paribhuhfitum abhabbo). 42
The explanation of sannidhikdrakam as a gerund is found in most of
the relevant commentaries,43 and seems to be historicaly correct.
Edgerton, in BHS Grammar ~ 22.5 and # 35.5. describes what he
calls 'quasi-gerunds' in -akam, and adds that some are found in the
Pail Pfitimokkha. Examples in the Pfitimokkha are Nissaggiya Pficittiya
23, Pficittiya 38 (both with exactly this phrase; see below), and
Sekhiya 18--28. Gerunds in -am are called adverbial by Whitney
(1989: 359--60), n.amul in the terminology of PS_ninian grammar: they
are even rarer in Pall than in Sanskrit, but they are found.44 AS
contains two, in ~ 12 dlumpa-kdrakarn. (see # 12.3), and in # 17 this
word, sannidhi-kdrakam.

DISCOURSE ON WHAT IS PRIMARY (AGGAI~IqA-SUTTA)

329

Nissaggiya Pficittiya 23 forbids storing the five medicines for longer

than 7 d a y s : . . , sappi, navanitam. , telam, rnadhu, phdnitam. : ldni

pat iggahetvd sattdhaparamam, sannidhikdrakam paribhuFtjitabbdni," tam

atikkdmayato, nissaggiyam, pdcitliyam, 'after ghee, cream, oil, honey
and molasses have been accepted, they can be eaten after keeping
them in store for seven days at most; if someone lets (seven days)
pass, this is an offence requiring expiation with forfeiture'. Pficittiya
38 forbids storing alt foods: yo pana bhikkhu sannidhikdrakarn

khddaniyam, vd bhojaniyam vd khddeyya vd bhuFzjeyya vd, pdcittiyam,

'if a monk eats hard or soft food which he has stored, there is an
offence requiring expiation'; an amendment given immediately after
the rule (Vin IV 87) states that storing food for up to seven days (here
using a standard gerund nidahitvd) is allowable (cp. also Vin I 209,
with the gerund form -kdrakam.). The motif of the ideal monk not
storing food was well-known; it appears, for example, at Sn 924, in
one of the earliest Buddhist poems known to us: 'having received
(something) he would not make a hoard [na sannidhirn kayird] of food
and drink, eatables and clothes'. (Cp. Sn 306 concerning the [true]
brahmin, cited on p. 320.) Note that the commentary cited above, Ps
III 234, used 'husked rice' (tand.ula) to exemplify the things an ideal
monk cannot store. Monks, of course, are presented with husked rice
by their donors, cooked or uncooked -- it is one of the foods which
monks are allowed to acccept and take into the monastery to be kept
and cooked up to a week later at Vin I 211 -- as were the beings in
AS before storing began. In AS # 17, it is after storing has reached 8
days that 'powder and husk covered the grain' ('grain' =tand. ula).
Thus storing rice in the parable has no consequence until one being
goes over the seven day Vinaya limit. Storing food beyond that time,
of course, is allowable for a layperson; the fifth impossible thing
specifically states that the monk cannot now make a store of (objects
of) desire 'as he formerly did when living in a house'. In AS # 16 and
# 17, two sections which could be separated differently, as does
Walshe's (87) translation, or run together to form a singIe section,
there is a close connexion between sex, living in houses, storing food,
and the subsequent need for rice-cultivation.
The conclusion seems to me inescapable that both the 'five impossible things' and AS are deliberately recalling the Vinaya rules, both in

330

STEVEN COLLINS

language and content. The rules should have been familiar to all
monks and nuns, as the Pdtimokkha was (supposed to be) recited
every fortnight in the presence of all members of each monastic
community. The five infractions of the Vinaya impossible for an ideal
ascetic are thus precisely the stages in the Fall of Mankind from
celibacy to civilisation, in sections # 16--20. Each and every event in
the degeneration of beings is in some way related to the monastic
order, its ideals and its Code:
# 12 eating with a finger and then in handfuls: contravenes Sekhiya
rules (as above, p. 326).
# 13--5 'pride in appearance' (van. n.a): cp. the class- (van. n.a-)pride
and related denigration of Buddhist monks attributed to Brahmins at
the start of the sutta, which had been abandoned by Vfisettha and
Bhfiradvfija in their intention to become monks (cf. IV below on
keywords).
# 16 having sex: (Impossible Thing no. 3) contravenes Pdrgtjika 1:
methunam dhamrnam patisevim, su; beings 'expelled from the sam. gha'
because of this: nassa (referring to ndsand).
# 17 storing rice for more than 7 days: (Impossible Thing no. 5)
contravenes Nissaggiya Pdcitiya 23, Pdcittiya 38): yato te. . . sattd
sannidhikdrakarn sdlirn upakkamim, su paribhuhjitum. , atho kano pi
tand. ularn pariyonandhi, thuso pi tan. d.ulam pariyonandhi. . .
# 19 theft: (Impossible Thing no. 2) contravenes Pdrdfika 2: sakam
bhglgam parirakkhanto ahftataram bhdgam, adinnam ddiyitvd paribhuhfi. Note that the text specifies that he kept his own portion (of
rice) while taking another's, which has not been given; this underscores that it counts as an intentional theft, not simple carelessness
about ownership.
# 19 lying: (Impossible Thing no. 4) contravenes Pficittiya 1: musdvddo
pa~fiayati.
19--20: Impossible Thing no. 1 (cf. Pficittiya 61 against killing any
living thing, and Pdrdfika 1, against murder) is not described directly
as AS, but may be inferred. It is first adumbrated by the violence of
the beings in r 19, and then -- by implication -- by the legitimate

DISCOURSE ON WHAT IS PRIMARY (AGGANIC,IA-SUTTA)

331

punishment given by the first king in # 20. The text does not specify
that capital punishment is involved, but this is assumed by the
commentarial tradition: Sv 870 glosses the phrase khfyitabbam
khiyeyya, inter alia, as hdretabbam hdre)ya, simply 'remove (whoever)
has to be removed'; the sub-commentary, DAT III 59) explains,
somewhat gingerly, sattanikdyato nihdretabbam., '(whoever) is to be
removed from the world of beings'. It has almost always been an
accepted part of a king's function throughout Buddhist and Indian
history to execute criminals, and this of course makes Buddhist moral
ambiguity about them automatic and unavoidable. 45 In AS, sex and
storing rice are called in ~ 18 pdpakd akusald dhammd, 'bad,
unwholesome things'; in # 19 and # 20 theft, lying and violence are
called 'bad'; the term datjddddnam., 'taking up the stick', refers here to
the beings' violence in # 19, but dan.d.a is also a standard term for
royal punishment; in # 22 these three things, and the now apparently
legitimate royal activity of pabbdjana, 'banishing', are called 'bad,
unwholesome things', seemingly with the approval of the narrative
voice (see # 20.2 on royal punishment for theft, and # 22.1).
IV. The structure of AS, and key-words
Part of the unity and coherence of AS is achieved by the repetition of
certain key words, often with deliberate plays on their various senses.
The words are agga, aggahha, set(ha, the prefix brahma-, and vanna.
(For the close relationship between the first four of these terms, see
# 7.2, 7.3, and # 9.2.) I accept Gombrich's (92a: 169--70) analysis of
aggahha as an adjective formed by the ending -hha added to agga in
the sense of 'first'; aggahha thus means, in his rendering, 'primeval' or
'original'. I think there is also a deliberate play on words here with
agga in the sense of 'best', found in # 7 and # 31 (see # 7.2).
Fortunately for a translator, the English 'primary' can also have the
same two senses: the first two meanings listed in the Oxford English
Dictionary are 'of the first order in time or temporal sequnce; earliest,
primitive, original' and 'of the first or highest rank or importance; that
claims the first consideration; principal; chief'. In the title of the sutta I
render the word 'what is primary', as I see it deliberately catching both
senses; the grammatical nature of the phrase aggan.~ akkhdyati in # 7

332

STEVEN COLLINS

and # 31 suggests the English 'what is primary' there too (see # 7.2),
albeit that to refer to a person as 'what is primary' is a little ungainly
in English.
The word settha means 'best', and is used repeatedly in the text by
various people of various things (see below); the prefix brahma- can
be used with the same meaning in Pali, and this allows puns on the
name of the Brahmanical god Brahmfi (see # 9.2, # 32.1). It is not
surprising, in a text proposing an (ascetic) hierarchy, that there should
be so many words, so often repeated, for 'best', etc. The most polyvalent term in AS, and the most frequently repeated, is vanna. On its
first appearance it is used by Brahmins to refer to their social class; as
Gombrich (92a: 163, cf. 168) points out, it can also mean 'colour',
'complexion', 'good looks'; I have sometimes rendered it 'appearance'
(see # 11.3 and 13.1). All these senses are used as the text goes along;
but in each case echoes of the other senses are also in play.
I set out here the structure of AS as I see it, showing where and
how the key-words occur. I label the first two parts 'Story of the
Present' and 'Story of the Past' to evoke the use of the same terms in
the structure of Jfitaka narratives. I see the organisation of AS as in
this sense analogous to that of Jfitaka tales, although it is not, of
course, presented as such.

1. Story of the Present ( # 1--9)

# 1--7

Conversation with Brahmins about Brahmins

set.tha -- 'best', in # 3, # 4, # 7, used by Brahmins of their
class (vanna); used by the Buddha of the Dhamma
in # 7
agga --- 'best' ('what is primary') used by the Buddha of the
Arahant in # 7. (The logic of # 7 proves there to
be a close link between settha and agga: see # 7.3)
Brahmd: used as the name of a god in # 3, # 4, # 7

# 8

King Pasenadi and the Buddha

settha = 'best' used by the Buddha of the Dhamma (twice)
Note the epithets used for the Buddha and Pasenadi (see
# 8.1 and 3)

# 9

Buddhist ascetics superordinate to kings and Brahmins

DISCOURSE ON WHAT IS PRIMARY (AGGAlqI'~A-SUTTA)

333

(For continuity with # 8 re: kings, see g~9.1 on Sakyaputtiyd)

Brahmd: used both as the name of a god and, punningly, as
a prefix in the sense of 'best' (see # 9.2)
Dhamma, said to be 'best' in g~7, g~8, replaces god
Brahmd in the series of epithets used of the Buddha
2. Story of the Past ( # 10--26)
# 10--17 From immaterial celibacy to household life, food-storage
and agriculture
aggafifia = 'primary' (agga ~ 'first') in /~ 13, ~r 15, g~ 16
vanna ~1 'colour' (of foodstuffs) in # 11, # 12, # 14
--2 'appearance' (by itself and in compounds) in
# 12, # 13, # 14, # 15, # 16 (total of 29 times):
echoes of 'class' as used in g~3, # 4, # 7
# 18

Recapitulation, and appearance of private property

# 19--20 From private property to crime and punishment

tadagge --- 'from this beginning' ( agga ~ 'first')
# 21--26 Etymologies for the 4 Brahmanical classes, called 'groups'
(mand. ala); renouncer-Brahmins, royal cities, the ascetic
group
agga~ia -----'primary', used 8 times in # 21--25
settha: used of the Dhamma in ~ 21, J/24, # 25, g~26, and
(pejoratively) of Brahmins in # 23
. ,

# 27--32 Conclusion
# 27--30 Morality, Rebirth and Release the same for all social groups
(repeats sentiments of g/5--7)
# 31

The Arahant is what is primary (repeats end of # 7

verbatim)
agga -- 'best' (perhaps a hint of the other sense: see # 31.2)
settha ~ 'best', used of the Dhamma
. .

# 32

Brahmfi Sanamkumfira's (the Ever-Virgin's: see ~r 32.1)

Verse

334

STEVEN COLLINS

settha = 'best', used of ksatriyas vis-a-vis the human world,

of 'the person endowed with wisdom and right
conduct' vis-~t-vis the universe as a whole
Brahmd: used in a characteristically Buddhist sense (see
#32.1)
This outline can show the overall structure of the text, and the
continuity of vocabulary; it cannot give a feel for the logic of the
narrative as it moves along in a coherent sequence. The story of
origins, then, far from being an extraneous and disconnected insertion,
as has been alleged, is intimately tied to the focus of the text as a
whole. The immediate transition from # 9 to # 10 is effected by the
replacement of the Vedas by the Dhamma (the Word of the Buddha)
and of Brahmfi by the Buddha, who 'has the best body', and 'is the
best' (brahma-kdyo, brahma-bh~to) (see gt 9.2).
I shall not take space here to recapitulate the argument of the Introduction as a whole, nor to set out in detail the correspondences
between Gellner's vision of agrarian society and the themes of AS. If
they are not obvious already, they will become so on re-reading.
In the translation which follows, as said earlier, I have aimed at
accuracy rather than elegance. I have ommitted some, but not all
repetitions in the text, in deference to modern English prose style.
This does of course slightly alter the flavour and pace of the narrative,
but this will not matter for my purpose here, which is exegetical.

NOTES TO THE INTRODUCTION

I am grateful to various friends and colleagues for helpful comments on an earlier
draft: at Chicago, Wendy Doniger, Paul Griffiths and Sheldon Pollock made both
minor and major suggestions for revision; Phyllis Granoff provided additional data
and suggestions. Roy Norman read it with characteristically generous care, making a
number of corrections and useful remarks. I am especially grateful to Patrick Olivelle
for forcing me to think much harder about the issue of how to treat the 'satire' in AS
than I had done, and for patiently discussing at length with me a number of other
questions. Anyone who reads Richard Gombrich's (88) book and (92a) article will see
how much I have learned from him, despite differences of emphasis in our readings of
AS. I imagine they would all still disagree with some, perhaps much of what I say.
Responsibility for the interpretation given here is entirely mine.
Since I present this article as an improvement on the Rhys Davids's translation, I

DISCOURSE

O N W H A T IS P R I M A R Y

(AGGANIqA-SUTTA)

335

should like to dedicate it to the memory of T. W. Rhys Davids, a giant on whose

shoulders we stand.
J The argument is stated at greatest length in (88), from which page references are
given in the text; versions of it appear in a number of Gellner's works: see, e.g., the
summary version in (83) Chapter 2, 'Culture in Agrarian Society'.
2 In this article I use Sanskrit rather than Pali terms for the four Brahmanical social
classes, as they are best known in English in that form; I used the anglicised
'brahmin(s)' except when quoting from others, where I retain their spelling.
a Gellner's phrase 'conspicuous self-exile' recalls Peter Brown's work on the late
antique 'holy man', and the 'rituals of disengagement' which make possible his
particular social role(s). See Brown (71) and (78). I have tried before, in Collins (88),
to apply this analytical perspective to Buddhist monasticism.
4 Erdosi (88: 112).
5 Summary accounts can be found in Ghosh (73), Sharma (83) and Erdosi (88); the
latter contains a judicious and helpful discussion of the problems of dating both
Brahmanical and Buddhist texts, and of relating both to the archeological data
currently known.
6 See Bechert (91), and other forthcoming volumes on the subject under his
editorship.
7 Erdosi (88: 16).
8 Gombrich (88: 49--59) offers an elegant summary.
9 See Olivelle (92), Introduction section 2, and (93) Chapter 2.2--3. Earlier van
Buitenen (81:12) had sketched out a siwAlar idea.
10 For this political history, see esp. Erdosi (88: 118--50).
1i They are: the majority of the Digha, Majjhirna, Sam.yutta and Aliguttara Nikdya-s,
and the Sutta Nipdta, with perhaps some other short texts from the Khuddaka Nikdya.
The Abhidhamma has always been accepted to be late; recent evidence is tending to
suggest that the version of the Vinaya we have is a later redaction, although it too
contains no reference to imperial formations.
i2 Where such a figure is mentioned, as in the well-known Cakkavattisihandda Sutta,
which was redacted next to AS in the Digha collection, and which shares certain
narrative motifs with it, there are two interpretive options. Either one decides such
passages were redacted after the Mauryan empire; or one follows Gombrich (88: 82),
who writes 'the representation of one's own king as a world-ruler of untrammelled
power is a commonplace of the ideology informing Vedic ritual It was an institutionalised fantasy'; therefore, it could have existed before the realisation of large-scale
empires. In either case, such passages cannot be read as containing depictions of the
pre-Manryan historical world. This is not the case with AS.
13 See Olivelle (93: Chapter 2.2 note 85) who cites Ghosh (73: 19--21) and Erdosi
(88: 126).
i4 See Olivelle (91), and other literature cited there.
~5 (78); see also Burghart (85).
16 See Erdosi (88: 17--8, 118).
17 Such literature was in large part written by brahmins. Indeed, as Patrick Olivelle
reminds me, we should note another ambiguity, or divergence of emphasis within
Brahmanism, in addition to that between ascetics and non-ascetics discussed in the
text below: that is, between liturgical and other texts which place the brahmin at the

336

STEVEN COLLINS

top of the social hierarchy, and other texts -- including even the Manusmrti -- which
make the king the highest.
18 (92: 22--3), referring to Heesterman (85).
19 The first to suggest that the story of origins must have been a separate text was
Edmnnds (04: 207--9); more recently Schneider (57) and Meising (88) have taken a
similar approach.
20 I am here influenced by recent trends in Homeric scholarship; see, e.g., Macleod
(82), esp. pp. ix and 37--40, and Griffin (80), esp. pp. 12--5.
21 I use this term in the sense proposed by Ramanujan (91).
22 See also Sarkisyanz (65). Much of the heat, and much of the point, can be
removed from the debate between Tambiah and Carrithers (77), (87), (93: chapter 7),
when one distinguishes what AS and the figure of mahdsammata might be in their
earliest form, and what the 'myth of origins' and Mahfisammata (sic, now become a
proper name for an individual) became in the later tradition.
23 Some seem to have taken it as historically factual, pointing to what they thought
was the 'natural', pre-political, 'we-Contract' condition of the newly-discovered
American Indian tribes as empirical evidence; others seem to have recognised the
allegorical nature of the story, while still according it explanatory and legitimatory
value. See Lessnoff (86).
24 He was writing of the poet Andrew Marvell (32: 255). I am grateful to Gananath
Obeyesekere for introducing me to this passage (in a talk entirely unconnected with
AS), and for kindly tracking down the precise reference.
25 Manuel and Manuel (79: 16, 80, 103, 229, 343).
26 ibid., p. 1 on More, and passim.
27 I think it would be interesting to study the Brahmanical tradition of dharma-ddstra
from this 'o/eu-topian' perspective; but that must await another occasion, and perhaps
another scholar.
28 See, e.g. KRN Coll. Pap. II # 43 on the Sabhiya Sutta.
29 Kahrs (83) has argued that in Sanskrit, such 'etymologies' are to be understood in
relation to the Brahmanical view of it as a language with a special and direct relation
to reality: thus, the more meanings perceivable in a word, the more it tells us about
the world. RFG cites Kahrs, arguing that the etymologies in AS are deliberately
parodying this view. I am not sure that this point is decidable. As I hope to show in a
future article, while there is some evidence that early Buddhism did have a view of
language as purely conventional, by the commentarial period the form of Middle
Indo-Aryan we call Pali (for the texts, the language of Magadha) was seen as having a
privileged, epistemologically direct relation to reality, if not the ontological status
accorded to Sanskrit in the Brahmanical MimSmasfi school.
30 See Bronkhorst (83); and for the a-historicism of Sanskrit Deshpande (85).
31 Translations from the Sutta Nipata, here and of the Brdhmanadhammika Sutta, are
by Norman (92).
32 For a list of examples of Brahmanical terminology used in a (new) Buddhist sense,
see KRN Coll. Pap. 1V # 99.
33 Although novices were not allowed to attend formal-ritual recitations of the
Pdtimokkha ('in a seated assembly', nisinna-parisd, Vin I 135), as 'co-resident pupils'
of a preceptor, they learnt to recite it (Vin I 47: cp. H o m e r [51:62 n. 7] and CPD s.v.

uddisdpeti.
34 The word is komdrabrahmacariya: see AS # 31.2 and # 31.3.
32 See Collins (forthcoming) on the significance of this in Buddhist monasticism.

DISCOURSE

O N W H A T IS P R I M A R Y

(AGGANlqA-SUTTA)

337

36 The phrase is manthe ganthetvd; cp. AS # 2 3 and #23.1 on ganthe karontd.

37 See Winternitz (33: 247) for elements of the former as late as the 4th. century
A.D. Lamotte (88: 657) dated the latter to the 4th--5th. centuries A.D.; Gnoli (77:
xix--xxi) disagreed, preferring an earlier date, at the time of Kaniska. This, notoriously, is not known exactly, but is usually thought to be lst.--2nd, century A.D.
38 See KRN Coil. Pap. I: 156. RFG (ibid.), acknowledging that 'in nuce the commentarial tradition goes back to the first generations of Buddhists in northern India', cites
Trantmann's work on Dravidian kinship patterns in the commentaries to show the
later provenance of at least those sections. But Norman argues (pers. comm.) 'since
North India was Dravidian and Munda before it became Indo-Aryan, I see no reason
to doubt that the Buddha's family were in fact Dravidian, and it could have been that
the Sakyans were Dravidian speakers until only shortly before the time of the Buddha.
There is, therefore, no need to assume that the commentaries which deal with such
things were composed in South India'. Thus, he says, while some parts of the
commentaries clearly were composed in South India and/or Sri Lanka, 'I think one
has to treat each piece of commentary on its merits, with the possibility that
everything is old unless it can be proved otherwise'.
39 Oil this see, inter alia, Collins (90) and the literature cited there.
40 Both the Dipavam.sa (Chap. 3) and Mahdvamsa (Chap. 2), roughly contemporaneous with the final fixing of the commentaries in Pali, make Mahfisammata a
king at the origin of the Sakyan family; and in the later tradition the figure of the 'first
king' Mahfisammata occurs in many kinds of text, from legal to ritual, as well as in
inscriptions. Tambiah (89) lists some of these later sources; I hope to add more in a
future publication.
41 Sn 406, translation from Norman (92: 44); cp. D I 63 et freq.
42 The reference to vatthu-kdma, 'material objects of desire' here uses a standard
division of '(objects of) desire', into material and mental; the latter is kilesa-kdmd,
literally, '(objects of) desire (consisting) in defilement(s)'. Unenlightened monks are as
prone to this as householders; given that the description is of an enlightened monk, in
whom there can be neither form of desire, the reference must be to such a person's
behaviour -- that is, he is being said specifically not to store things such as food.
43 Sp 710. Kkh 76, Sv 913. The text of A IV 70 in the PTS edition reads sannidhikdrake kdme, but this would seem to be a scribal error, perhaps by someone who did
not recognise the gerund and altered the word to agree with kdme; there is a v.L
-kdrakam, which is also the reading in the commentary (Mp IV 169--70), which
glosses sannidhim katvd.
44 For examples, see K. R. Norman (92: 229) ad Sn 773. Norman writes (pets.
comm.): 'I think that Whitney is correct in seeing that narnuls are nouns used in the
accusative as adverbs, e.g. "eating food in a himp-making sort of way" [refering to
dlumpa-kdrakam in AS # 221.
45 There is at least one exception: Agoka, after the massacre of the Kalifigans and his
subsequent remorse, if KRN Coll. Pap. I # 26 is right. In the story of Siri Sam.gha
Bodhi in Sri Lanka, given in Mhv XXXVI 80ft., although he did not execute
criminals, he had dead bodies brought to be burnt publicly in their place; the telling
of the story at Att 20 says that he ruled adand,ena asatthena, 'without punishment,
without weapon(s). See also the Mfigapakkha Jfitaka, Ja VI 1--30, discussed by
Gombrich (88: 70).

338

STEVEN COLLINS

INTRODUCTION TO THE TRANSLATION

In the translation I follow the section divisions made by J. Estlin
Carpenter in his edition of the text for the Pali Text Society (11). This
is not because I find them especially good -- other, perhaps better
divisions of the text are possible (see, e.g., # 17.1) -- but to facilitate
comparison between my translation, previous translations, and the text
itself. The text was reprinted in Meisig (88), but with an unfortunately
complex textual apparatus; see K. R. Noman's (89) review.
There have been previous translations of AS into European
languages. The first, to my knowledge, was of sections # 10--21 by
Edmunds (04); next were two into German by Franke (1913) and
Neumann (1918). I have not gone into differences between my
rendering and theirs. The most widely-known translation, as mentioned in the Introduction, is that of T. W. and C. A. F. Rhys Davids
(1921). This was republished in 1981 under the name of Trevor Ling,
who claimed that his version was 'a fairly extensive re-translation' (p.
xxiv); in fact for the most part it simply modernises RhD's English,
and introduces some inaccuracies of its own. Walshe's recent (87)
version is better, but it seems clear to me (for example from the
extensive use made of RhD in the notes) that like Ling Walshe
depends almost entirely on RhD for his understanding of the text. I
have made this translation not primarily for the sake of offering
another English version but to give structure and coherence to my
notes and interpretation.
TRANSLATION
# 1. Thus I have heard. At one time the Blessed One was living in
the palatial monastery built by Migfira's mother in the Eastern Park
outside Sfivatthi. At that time Vfisettha and Bhfiradvfija, aspiring to
become monks, were living with the monks there. One evening the
Blessed One rose from his solitary meditation, went outside the
monastery, and was walking back and forth in its shade, in the open
air.
# 2. Vfisettha saw that the Blessed O n e . . . was walking back and
f o r t h . . , in the open air, and said to Bhfiradvfija:

DISCOURSE ON WHAT IS PRIMARY (AGGANfqA-SUTTA)

339

'Friend Bhfiradvfija, here is the Blessed One . . . walking back and

forth . . . in the open air. Come, friend, let's go to him: perhaps we
may get a chance to hear a Dhamma-talk from the Blessed One
himself.'
'Alright, friend', agreed Bhfiradvfija; and so Vfise.ttha and Bhfiradvfija
went up to the Blessed One, greeted him, and walked back and forth
together with him.
# 3. Then the Blessed One addressed Vfisettha (and Bhfiradvfija): ~
'Monks, you were (both) born brahmins, in brahmin families, (but) you
have gone forth from home to homelessness, (leaving your) brahmin
family. Surely brahmins (must) revile and abuse you.'
'Indeed, sir, brahmins revile and abuse us fully, completely, with the
(sort of) abuse one would expect (from them).'
'How do they abuse you . . . ? '
'Sir, brahmins say "The brahmin is the best class (vann. a), (any) other
class is inferior. The brahmin is the fair class, (any) other class is
dark. a (Only) brahmins are purified, not non-brahmins. Brahmins are
Brahmfi's own sons, born from his mouth, 3 born of Brahmfi, produced
from Brahmfi, the heirs of Brahmfi. You here have left the best class
and gone (over) to an inferior class, since you have become wretched
shaven-headed (pseudo-)ascetics, members of some sect, 4 (no better
than) offspring of our Kinsman's ii.e. Brahmfi's] feet. It is not good, it
is unseemly, that you have left the best c l a s s . . . [and] have become
. . . offspring of our Kinsman's feet." That is how they revile us . . .
with the (sort of) abuse one would expect (from them).'
# 4. 'Surely, monks, the brahmins are not recalling the past 1 when
they say [this]. Brahmin women, (the wives) of brahmins, are seen to
mensturate, become pregnant, give birth and give suck; and (so) these
brahmins who say: "the brahmin is the best c l a s s . . , brahmins are
born from Brahmfi's m o u t h . . , heirs of Brahmfi", are (in fact) born
from vaginas. 2 They are slandering Brahmfi, telling lies, and producing
demerit?
# 5. Monks, there are these four classes: ksatriya (warriors/kings),
brahmin (priests), vaiiya (farmers, merchants) and gfidra (servants). 1

340

STEVEN COLLINS

[Take the case where] a particular ksatriya is a murderer, a thief,

misbehaves in sexual matters, tells lies, speaks maliciously, harshly and
frivolously, is envious, malevolent and holds wrong views: in this way,
monks, those things which are unwholesome, blameworthy, not to be
followed, unworthy to be called noble (and known to be all these
things), dark with (karmic) result, and censured by the wise, are seen
here in a particular ksatriya. [The same thing is then repeated of
individual members of the other three classes.]
# 6. [Take the case where] a particular k.satriya refrains from murder,
theft [and all the other things mentioned in # 5]: in this way, monks,
those things which are wholesome [and the converse of what is said in
# 5] . . . fair result and praised by the wise, are seen in a particular
ksatriya. [The same thing is then repeated of individual members of
the other classes.]
# 7. Given that bad and good things, 1 what is censured and praised
by the wise, occur mixed up together in these four classes, wise men
will not tolerate it when brahmins say "the brahmin is the best class
9 brahmins a r e . . , heirs of Brahmfi". Why? Of these four classes,
monks, he who is a monk, an Arahant, in whom the corruptions are
destroyed, who has lived the (holy) life, done what was to be done,
laid down the burden, attained the true goal, in whom the fetters of
existence are destroyed, who is released by Right Wisdom -- he is
properly called what is primary 2 among them, not improperly [as are
the brahmins]. For the Dhamma is the best (thing), monks, in this
world, both in this life and in the future. 3
# 8. (The fact) that Dhamma is best in this world and in the future,
monks, can be understood by the following illustration. The king of
Kosala, Pasenadi, realises "the ascetic Gotama, who is unsurpassed, 1
has gone forth from a Sakyan family". The Sakyans are (now) vassals of
King Pasenadi of Kosala, monks: they (have to) fall at king Pasenadi's
feet in obeisance, salute him respectfully, rise up from their seats for
him, do him homage with hands together. Now, all this kow-towing
which the Sakyans do before king Pasenadi, king Pasenadi does before
the Tathfigata, 2 (thinking) 'Indeed the ascetic Gotama is well-born,

DISCOURSE ON WHAT IS PRIMARY (AGGAIqI~A-SUTTA)

341

while I am ill-born; the ascetic Gotama is powerful, I am not; Gotama

is charismatic, I am ugly [or: of bad appearance, dubban, n.o]; he has
great authority, I have little'.3 It is in reverence, honour, worship and
respect for this Dhamma that king Pasenadi falls at the Tathfigata's
feet in obeisance, salutes him respectfully, rises up from his seat for
him, and does him homage with hands together. By this illustration,
monks, one can understand that Dhamma is the best in the world, in
this life and in the future.
# 9. You, monks, are from various castes, of various names, from
various clans and various families, and (yet) you have gone forth
from home to homelessness. When asked who you are, acknowledge
that "we are sons of the Sakyan".1 When anyone has a faith in the
Tathfigata which is firm, rooted, (well-)established and which cannot
be disturbed by (any) ascetic, brahmin, god [or: king, deva], Mfira,
Brahmfi, or anyone in the world, that person can fittingly say "I
am the Blessed One's own son, born from his mouth, born of the
Dhan-~ma, produced by the Dhamma, the heir of Dhamma". Why?
(Because) these are epithets of the Tathfigata: "he who has Dhamma
for a body", "who has the best body", "who is Dhamma", "who is the
best". 2
# 10. Eventually, after a long time, monks, it comes to pass that the
world contracts; as it contracts, usually1 beings devolve as far as the
Abhassara world. There they remain for a long time, made of mind,
feeding on rapture, providing their own light, moving about in the air,
glorious. Eventually, after a long time, monks, it comes to pass
that the world evolves; as it evolves, usually beings die from their
Abhassara-bodies and come to this world. (Here) they remain for a
long time, made of mind, feeding on rapture, providing their own light,
moving about in the air, glorious.
# 11. At that time there is nothing but water, (all) is darkness, (just)
deep darkness. It is not possible to discern the moon or sun, the
twinkling stars, night or day, months or half-months, seasons or years,
men or women. Beings just have the name 'beings'. 1 Then (on one
such occasion) an earth-essence spread out on the waters. 2 It appeared

342

STEVEN COLLINS

in the same way as (does) the spreading out (of skin) on top of boiled
milk-rice as it cools down? It had colour, 4 smell and taste; its colour
was like sweet ghee or cream, its taste like fine clear honey. 5
# 12. Then, monks, a certain being, greedy by nature, 1 thinking 'what
can this be?', tasted the earth-essence with his finger? As he tasted the
earth-essence with his finger he was pleased, and craving came upon
him. Other beings imitated that being, tasting the earth-essence with
their finger(s). They too were pleased, and craving came upon them.
Then, monks, these beings started to eat the earth-essence taking (big)
mouthfuls of it with their hands? As they did so, their self-luminosity
disappeared. When their self-luminosity disappeared, the moon and
sun appeared; when the sun and moon appeared, the twinkling stars
appeared; when the stars appeared, night and day appeared; when
night and day appeared, the seasons and years appeared. Thus far,
monks, did the world evolve.
# 13. Those beings, monks, spent a long time eating the earth-essence,
living on it as their food. According to how (much) these beings ate,
so (much did) their bodies become hard, and good and bad looks
became known. 1 Some beings were good-looking, others ugly; those
who were good-looking despised those who were ugly: "we are betterlooking than they, they are uglier than us!' Because these (beings),
proud and arrogant by nature, 2 were proud of their appearance, the
earth-essence disappeared. When it had disappeared, they came
together and lamented "Look (aho), the (earth) essence (rasam.) (has
disappeared), look . . . the essence" so nowadays, when people have
tasted something good they say "Oh the taste, oh the taste!" (aho
rasam. 3 They recall the original, primary word(s), but they don't
understand what they mean. 4
# 14. Then, monks, when the earth-essence had disappeared, a
fragrant earth appeared for those beings; it appeared (suddenly) just
like a mushroom. 1 It had colour, smell and taste; its colour was like
sweet ghee or cream, its taste like fine clear honey. Then they started
to eat the fragrant earth. Those beings, monks, spent a long time
eating the fragrant earth, living on it as their food. According to how

DISCOURSE ON WHAT IS PRIMARY (AGGANlqA-SUTTA)

343

(much) these beings ate, so to an even greater degree did their bodies
become hard, and good and bad looks become known. Some beings
were good-looking, others ugly; those who were good-looking despised
those who were ugly: "we are better-looking than they, they are uglier
than us!" Because these (beings), proud and arrogant by nature, were
proud of their appearance, the fragrant earth disappeared. When it
had disappeared, a (kind of) creeper appeared. 2 It appeared Iike a
kalambukfi plant. 3 It had colour, smell and taste; its colour was like
sweet ghee or cream, its taste like fine clear honey.
# 15. Then they started to eat the creeper. Those beings, monks,
spent a long time eating the creeper, living on it as their food. According to how (much) these beings ate, so to an even greater degree did
their bodies become hard, and good and bad looks become known.
Some beings were good-looking, others ugly; those who were goodlooking despised those who were ugly: "we are better-looking than
they, they are uglier than us!" Because these (beings), proud and
arrogant by nature, were proud of their appearance, the creeper
disappeared. When it had disappeared, they came together and
lamented 'we've had it, the creeper has given out on us!'. 1 So nowadays, when people are touched by some hardship, they say 'we've had
it, it's given out on us!' They recall the original, primary word(s), but
they don't understand what they mean.
# 16. Then, monks, when the creeper had disappeared, there
appeared for those beings rice, growing without cultivation; it was
without powder, (already) husked, sweet-smelling and ready to eat. 1
Whatever they gathered in the evening for their evening meal, in the
morning had grown back ripe again; whatever they gathered in the
morning for their morning meal, in the evening had grown ripe again:
(the work of) harvesting was unknown. 2 Those beings, monks, spent a
long time eating the rice which grew without cultivation, living on it as
their food. According to how (much) these beings ate, so to an even
greater degree did their bodies become hard, and good and bad looks
become known. The female parts appeared in a woman, and the male
parts in a man; 3 the woman looked at the man with intense, excessive
longing, as did the man at the woman. As they were looking at each

344

STEVEN COLLINS

other with intense longing passion arose in them, and burning came
upon their bodies; because of this burning, they had sex. When the
(other) beings saw them having sex, some threw earth (at them), some
threw ashes, 4 other cow-dung, (saying) "Away with you and your
impurity, away with you and your impurity!''s "How could a being do
such a thing to another being?" So nowadays, people in certain areas,
when a bride is being led out, throw dirt, ash or cow-dung. They recall
the original, primary (actions), but they don't understand what they
mean. 6
# 17. Monks, what was thought improper at that time is nowadays
thought proper. 1 At that time the beings who took to having sex were
prevented from entering either small or larger settlements 2 for a
month or two. Since those beings were excesssively intoxicated at that
period of time by (this) immorality,3 they took to building houses to
conceal it. 4 Then a certain being, lazy by nature, thought "Well! Why
am I troubling myself gathering rice for my evening meal in the
evening, and (again) for my morning meal in the morning? Why
shouldn't I gather it just once, for both evening and morning?" And he
did so. Then another being came up to him and said "Come, being,
let's go to gather rice". "There's no need! I've gathered rice just once
for both evening and morning." The (second) being thought "that
seems (a) good (idea); my friend", 5 and imitated him by gathering rice
just once for two days. Then another being came up to the (latter)
being, and said "let's go to gather rice". "There's no need! I've
gathered rice just once for two days". The (third) being thought "that
seems (a) good (idea)", and imitated him by gathering rice just once
for four days. Then another being came up to the (latter) being, and
said "let's go to gather rice". "There's no need! I've gathered rice just
once for four days". The (fourth) being thought "that seems (a) good
(idea)", and imitated him by gathering rice just once for eight days.
Because these beings took to eating rice which they had stored up, 6
powder and husk then covered the grain, cutting without regeneration
and harvesting became known; and the rice stood in clumps. 7
# 18. And then, monks, the beings came together and lamented "bad
things a have appeared for us beings; we were formerly made of mind

DISCOURSE ON WHAT IS PRIMARY (AGGAIqNA-SUTTA)

345

. . . [they recount their degeneration, concentrating on the succession

of food-stuffs, and omitting the aetiology of sex and houses. Each
disappearance is said to be "because of the appearance of bad,
unwholesome things among u s " ] . . , and now the rice stands in
clumps. Let us now divide up the rice, and set up boundary-lines". 2
And so, monks, they divided up the rice and set up boundary-lines.
# 19. Then, monks, a certain being, greedy by nature, while keeping
his own portion (of rice), took another portion without its being given,
and ate it. (Other beings) grabbed him and said "you have done
something bad, being, in that you kept your own portion but took
another portion without its being given, and ate it. Don't do such a
thing again! .... Alright', he agreed. [This happens a second and third
time, after which] they hit him, some with their hands, some with clods
of earth, some with sticks. From this moment on, monks, stealing,
accusation, lying and punishment became known. 1
r
Then, monks, those beings came together and lamented "bad
things have appeared for us beings, in that stealing, accusation, lying
and punishment have become known; what if we were to appoint one
being to criticise whoever should be criticised, 1 accuse whoever should
be accused, and banish whoever should be banished? 2 We will (each)
hand over to him a portion of rice". Then, monks, those beings went
to the one among them who was most handsome and good-looking,
most charismatic and with greatest authority 3 and said "come, being,
(you) criticise whoever should be criticised, accuse whoever should be
accused, and banish whoever should be banished; we will (each) hand
over to you a portion of rice. He agreed [and did as they asked]; they
(each) gave him a portion of rice.
21. "Appointed by the people" [mahdjanena samrnato], monks (is
what) mahdsammata (means): 1 'mahfisammata' was the first term (for
the ksatriya class) which appeared. 'Lord of the Fields' [khettdnam
pati] is what khattiya means: 2 'khattiya' was the second term (for the
ksatriya class) to appear. 'He brings joy to others [paresam...
rafijeti]3 according to Dhamma', is what rdjd ('king') means: 'rajfi' was
the third term (for the ksatriya class) to appear. This was the birth of

346

STEVEN COLLINS

the ksatriya-group, (along). with the original, primary term(s); of just

these beings, no others, of similar (beings), not dissimilar, properly
and not improperly.4 For the Dhamma is the best (thing), monks, in
this world, both in this life and in the future.
# 22. Then some beings, monks, thought "bad things have appeared
for us beings, in that stealing, accusation, lying, punishment and
banishment have become known; let us keep away from (these) bad,
unwholesome things. 1 [And they did so.] "They keep away from bad,
unwholesome things" [pdpake... bdhenti], monks, (is what)
brdhmand means: 'brfihmana' was the first term (for the brahmin
class) to appear? They made leaf-huts in the forest and meditated in
them; without coals or smoke (from a cooking fire), pestle (and
mortar) set down, 3 they went into villages, towns and royal c i t i e s 4 in
search of food, in the evening for their evening meal, and in the
morning for their morning meal. 5 When they had got their food, they
went back again to their leaf-huts in the forest to meditate. (Other)
humans saw this and said "these beings have made leaf-huts in the
forest and meditate there; without coals or s m o k e . . , they go back
again to their leaf-huts in the forest to meditate". 'They meditate'
(jhdyanti), monks, (is what) 'those who tend a (sacrificial) fire'
(jhdyakd) (means):6 'jhfiyaka' was the second term (for the brahmin
class) to appear.
23. Some of these beings, monks, were unable to maintain (the life
of) meditation in forest leaf-huts; they went to the outskirts of villages
and towns and lived there making (up) textsJ (Other) humans saw
them and said "These beings are unable to maintain (the life of)
meditation.., they live there making (up) texts. They do not meditate".
'They do not meditate' (na... jjhdyanti), monks (is what) Students
[or: Reciters] (of the Veda) (ajjhdyakd) (means)? 'ajjhfiyaka' was the
third term (for the brahmin class) to appear. It was considered a lesser
thing at the time, monks; nowadays it is considered the best. This was
the birth of the brahamin-group (along) with the original, primary
term(s); of just these beings, no others, of similar (beings), not dissimilar, properly and not improperly. For the Dhamma is the best
(thing), monks, in this world, both in this life and in the future.

DISCOURSE ON WHAT IS PRIMARY (AGGAIqlqA-SUTTA)

347

# 24. Some of these beings, monks, practised sexual intercourse and

took up occupations of high repute; I 'they practise sexual intercourse
and take up occupations of high repute (vissuta), monks (is what)
vessd means. The term 'vessa' (thus) appeared. This was the birth of
the vaigya-group, (along) with the original, primary term; of just these
beings, no others, of similar (beings), not dissimilar, properly and not
improperly, For the Dhamma is the best (thing), monks, in this world,
both in this life and in the future.
# 25. The beings who were left over led cruel, mean lives; 1 'they live
cruel, mean lives' (Iudddcdrd, khudddcdrd) (is what) suddd (means).
the term 'sudda' (thus) appeared. This was the birth of the g~dragroup, (along) with the original, primary term; of just these beings, no
others, of similar (beings), not dissimilar, properly and not improperly.
For the Dhamma is the best (thing), monks, in this world, both in this
life and in the future.
~r 26. There came a time, monks, when a ksatriya disapproved of his
own tasks, and left home for homelessness, in order to become an
ascetic. 1 A brahmin, too, disapproved of his own tasks and left home
for homelessness, in order to become an ascetic. So too did a vaigya
. . . and a ifidra . . . in order to become an ascetic. This was the birth
of the ascetic group, monks, from these (other) four groups; 2 of just
these beings, no others, of similar (beings), not dissimilar, properly
and not improperly. For the Dhamma is the best (thing), monks, in
this world, both in this life and in the future.
# 27. A ksatriya, monks, who misbehaves in body, speech and mind,
and who holds wrong views, because of (his) wrong views and acquisition of (bad) karma is reborn, after the break-up of the body in death,
in a bad destiny, in hell. 1 A brahmin, m o n k s . . , a vaigya.., a/fidra
. . . an ascetic who misbehaves.., is r e b o r n . . , in hell.
# 28. A ksatriya, monks, who behaves well in body, speech and
mind, and who holds right views, because of (his) right views and
acquisition of (good) karma is reborn after the break-up of the body
in death, in a good destiny, in heaven. A brahmin, m o n k s . . , a vai@a

348

STEVEN COLLINS

. . . a gfidra . . . an ascetic who behaves well . . . is reborn . . . in

heaven.
# 29. A ksatriya, monks, who does both (good and bad) with his
body, speech and mind, and whose views are a mixture (of good and
bad), after the break-up of the body in death, experiences both
happiness and suffering. A brahmin, monks . . . a vaigya . . . a gfidra
. . . an ascetic who does both . . . experiences happiness and suffering.
# 30. A ksatriya, monks, who is restrained in body, speech and mind,
by means of (his) cultivation of the seven things which pertain to
Enlightenment attains nirvdna in this life. A brahmin, m o n k s . . , a
vaigya . . . a gfidra . . . an ascetic who is restrained . . . attains nirvdna
in this very life.
# 31. Of these four classes, monks, he who is a monk, 1 an Arahant,
who has lived the (holy) life, laid down the burden, attained the true
goal, in whom the fetters of existence are destroyed, who is released
by Right Wisdom -- he is properly called what is primary 2 among
them, not improperly [as are the brahmins]. For the Dhamma is the
best (thing), monks, in this world, both in this life and in the future.
# 32. The Brahmfi Sanarp. kumfira, 1 spoke this verse:
For those who rely on clan, the ksatriya is the best in this
world;
(but) the person endowed with wisdom and (good) conduct is
the best in the whole universe. 2
This verse was well-sung by the Brahmfi Sanarp. kumfira, monks, not illsung, well-spoken not ill-spoken, endowed with meaning not without
meaning: I approve of it. I too, monks, say:
For those who rely on clan, the ksatriya is the best in this
world;
(but) the person endowed with wisdom and (good) conduct is
the best in the whole universe.'
The Blessed One said this. Vfiset.tha and Bhfiradvfija were pleased,
and rejoiced in the Blessed One's words.

DISCOURSE ON WHAT IS PRIMARY (AGGAIqlqA-SUTTA)

349

NOTES TO THE TRANSLATION

# 3.1. The text has Vdset.fham in the accusative singular here; and the
Buddha's words (here and throughout) use the vocative singular
Vdseftha; but the pronouns and verbs are in the plural (some mss.
have the vocative plural Vdset.fhd). As RFG 164--5 suggests, since the
Buddha in # 9 says to his audience ' y o u . . . are from various castes
...', from that point on he must be taken to be addressing these two
individuals as part of a wider audience. I translate the vocatives simply
and neutrally as 'monks'.
# 3.2. Although the colour-symbolism for the four varna-s (brahminwhite, ksatriya-red, vaigya-yellow and gfidra-black) was not based on
skin-colour, the opposition sukko/kanho vaqno here does also refer to
fair/dark skin colour (see RFG 163), and simultaneously to pure/
impure social and moral status. Here the terms are being used from a
Brahmanical perspective, but the same opposition is found regularly to
describe good and bad karma and its results from a Buddhist viewpoint (as in # 5, # 6 and # 7 below). The brahmins' claim is that fair
skin and social standing are signs of moral (karmic) superiority, as the
next sentence states.
# 3.3. Putt6 orasd mukhato jdtd. Puttd orasd is literally 'children
(born) from the breast', hence commonly in Sanskrit and Pali '(one's)
own', 'legitimate'. Although this is often a dead metaphor -- it is
commoilly used in Buddhist texts of monks and nuns as the Buddha's
'spiritual sons and daughters' (see CPD s.v.) -- it is possible that the
narrative voice here intends an irony in juxtaposing it with mukhato
jdtd, 'born from the mouth'. The latter is, as RFG 163--4 (cf. 167)
states, a reference to the well-known Vedic motif connecting brahmins
with Brahmfi's mouth and speaking the Veda, found most famously in
Rig Veda X.90. The commentary here does not mention this; but the
sub-commentary does. Although RFG's text implies that the subcommentary cites the birth-place of all four classes in the precise
manner of that hymn, this is not the case (it speaks only of brahmins,
DAT III 46); but I think we can assume that both commentary and
sub-commentary are well aware of the Vedic motif in its entirety. The

350

STEVEN COLLINS

commentary to another occurrence in the Digha Nikdya (D I 90) of

the phrase used here of ascetics (mundake... bandhupdddpacce; see
# 3.4), gives a version of the motif adapted to include ascetics:
'brahmins come out from (nikkhantd) Brahmfi's mouth, Ksatriyas from
his chest, Vaigyas from his navel, gfidras from his knees and ascetics
from the heels of his feet' (Sv 254). Moreover, it refers to all four
Vedas by name at Sv 247, and on the same page cites the 'etymology'
for ajjhdyaka given in AS # 23 (see 23.2). The editors and redactors
of the Digha commentary and sub-commentary were clearly aware of
the metaphor as a live image of Brahmanical cosmogony. Accordingly,
when the commentary to AS here explains that what is meant is that
they stay in the breast (for a while) growing, and then come out from
the mouth, it must be taken to be ironic (although it also prepares for
the corresponding metaphorical explanation of the phrase used of
Buddhist monks in g~9: see # 9.2). The irony in the text of AS
consists in suggesting that brahmins posit two different and incompatible bodily 'places of birth' (cf. below in # 4 the Buddha's insistence on the real bodily facts), one of which the usual Vedic motif says
is the birthplace of ks.atriya-s.
# 3.4. Both 'wretched' and '(pseudo-)' render the -ka suffix in
samanaka, which can indicate a pejorative intent and/or an implication of falsity, as the commentary and sub-commentary recognise. Cp.
Sn p. 21 and p. 80 (transl. Norman [92: 14, 48], where Pj II 175
glosses sarnanabhdvam figucchanto, at 302 na so samano sarnanapatir@akattd samanako ti. For ibbha as 'member of a (particular)
sect' see KRN Coll Pap. I 77--81; cf. the discussion in Caillat (74).
# 4.1. At this point in the sutta the Buddha's remark about not
recalling the past (or mis-remembering it, assarantd, an uncommon
word) is unexpected, and seems obviously chosen to introduce something else. 'The past' (pordnam.) the Buddha intends to evoke is, first,
the brahmins' immediate past (their birth), and second, their ancient
scriptures and the cosmogony recounted in them. See further on
pasavanti in # 4.3 and on anupadanti/anussaranti in # 13.4.
# 4.2. Yoni can, of course, also refer to the uterus, of womb, and

DISCOURSE ON WHAT IS PRIMARY (AGGANTI~A-SUTTA)

351

thence generally 'origin', in various senses. Here, however, in view of

the Buddha's insistence on the physical facts of reproduction (which is
clearly intended to confront Brahmanical concerns with purity), it
seems preferable to adopt a more vivid term (see also # 4.3).
# 4.3. This is a rich sentence, I think, with a number of ramifications.
First, one should ask why are the brahmins said to 'slander' Brahmfi?
The verb abbhdcikkhati (Skt abhy-d-khyd) means 'to accuse falsely': of
what are the brahmins falsely accusing Brahmfi? The answer would
seem to be that since they are in fact yoni-ja when they say they are
Brahma-ja, they are by implication calling Brahm/t a vagina. Sv 862 -a (partial) verbal precursor of Magritte's painting Le Viol -- says that
they slander Brahmfi because if what they said were true, Brahmfi's
breast (ura, whence orasa: see # 3.3) would be the womb of brahmin
women and his mouth their urinal passage. Siegel (87: 207) uninhibitedly renders the commentary's remark 'the mouth of Brahmfi is a
brahmin woman's cunt'. The brahmins are guilty of such slander
explicitly in their words as cited in # 3 and repeated in ~ 4; but
perhaps the verb pasavati might be taken to suggest, implicitly,
another context. I have rendered it here 'produce'. (A literal sense is
'to beget', and one might see a contrast between the brahmin women
who produce children and brahmin men who beget demerit.) This
translation derives the word from (Skt) pra-s~, from st~ (or su) to
beget, whose present stem in Vedic and classical Skt. is -suva or -sava.
There is no reason to dispute this; but perhaps a sophisticated
audience of the sutta, knowledgeable about things Brahmanic, might
also catch echoes of the verb pra-su, from su to press out or extract
(Soma juice for the Vedic sacrifice). This verb is well-known not
merely for its importance in the sacrifice, but as the standard paradigm for Skt. verbs of the fifth class, which all educated brahmins
learnt. Although its present stem is -sunu, there are other forms
derived from this root, verbal and otherwise, which do use -sara,
notably the noun prasava, which is the standard term for the act of
pressing out Soma juice. I do not suggest that this could be more than
a nuance, but perhaps the word is meant also to suggest that when
brahmins press out their Soma, reciting Vedic hymns which (according
to the Buddha's understanding of their remarks here) mean that

352

STEVEN COLLINS

Brahmfi is a 'cunt', all they are doing is 'pressing out' demerit (by what
we would now call blasphemy).
# 5.1. As stated in note 2 to the Introduction, I give transliterated Skt
forms for the four classes. I assume that most readers of this translation will be familiar with the scheme. Leaving the terms untranslated
will also recall to mind that they do not denote 'priests', 'warriors' or
'farmers' in any and every context, but only people with those occupations as defined and represented by Brahmanical ideology. (You can't
be a ks.atriya in Kansas.) The placing of the ksatriya class first, before
brahmins, is common in Buddhist texts and clearly intentional (see
also # 32).
# 7.1. Literally dark and white dhamma-s, states of mind, phenomena, etc.
# 7.2. The last phrase is so tesam aggam akkhdyati dhammen' eva no
adhamrnena. 'Properly' for dhammena clearly cannot catch all the
nunaces of the term, nor catch the overlap with dhamma in the next
sentence: 'according to the Dhamma' (sc. the Buddha's teaching) is
also possible, as are many other renderings. In the phrase aggam.
akkh@ati, aggam (both 'first and best': see Introduction p. 331) can
be taken as neuter substantive, in place of an adjective in agreement
with the subject (thus RhD 'declared chief among them'; this is a
commentarial gloss, Mp III 74, etc.; of. CPD, PED s.v. akkh~yati). If
the verb is taken to mean not simply 'is declared', but 'is proclaimed,
acclaimed, praised', then aggam may be taken adverbially, 'in front', 'in
the first place', etc. (as PED s.v. agga seems to do). The sense is the
same in both cases: I choose the former alternative, and translate 'what
is primary', to bring out the connexion with aggaftfia, both in the text
and in the sutta's title, and below in # 9.1, although no doubt more
elegant renderings are possible in particular contexts. This phrase is
found frequently, used with terms in all genders (see CPD, PTC s.v.
akkhSyati); at A III 36 it is found with the singular verb used of a
plural subject; at A V 59--60 aggam, akkhgffati is found first with king
Pasenadi, and then aggam., with plural verb, is used of the Abhassara
gods, in a passage closely resembling AS # 10. (See # 9.1 below.)

DISCOURSE ON WHAT IS PRIMARY (AGGAlqiqA-SUTTA)

353

Commentaries (e.g. Ps I 136, Mp I 124, I 114, It-a 102) say that 'the
word agga is seen in (the senses of) beginning, top, a part, and what is
best (agga-saddo ddi-koti-koffhdsa-set.thesu dissati): examples of the
first and second are ajjatagge, beginning from now, henceforth, and
atigulagga, finger-tip (I omit the third, which raises difficulties irrelevant
here); the fourth is exemplified by the phrase Tathdgato tesarn aggarn
akkhdyati, 'of these the Tathfigata is called what is primary'. The word
is also common in the compound etadaggarn., 'primary among these',
as in A I 23--6, the etadagga-vagga, which gives a long list of the
Buddha's disciples whom he declares 'primary' or 'pre-eminent' in
different accomplishments. This vagga is well-known to Buddhist
tradition, since the commentary attaches extensive biographical details
about each figure named.
# 7.3. That there is an explicit connexion between the uses of agga
and set.t.ha here is proved by the logic of this section. The Buddha
asks why wise men do not tolerate the brahmins' claim to be the best
class, settho vann.o; and answers by stating that the Arahant is called
aggarn, and that the Dhamma is seCt.ha.
# 8.1. Anuttaro. The commentary reads anan~ard (agreeing with
Sakya-kuld) in place of anuttaro here, and glosses antara-virahitd,
attano kulena sddisd, 'without (social) distance, like his own family';
that is, from the same social level. Antara means an interval or
distance, in a variety of semantic domains (see CPD s.v.). CPD s.v.
anantara accepts the reading and comments 'prob. -- .next in rank';
RhD also accepts the reading but interprets the word geographically,
'neighbouring', commenting that 'the c o n t e x t . . , does not call for such
a word as anuttaro'. I do not see why: Pasenadi, whose words these
are, is said to revere Gotama, and anuttara is a standard flattering
superlative in Pali and Sanskrit. In a text which proposes a Buddhist
ascetic hierarchy, it is no surprise when a king is made to say that the
Buddha 'is unsurpassed'. But both readings make sense.
g/8.2. The text repeats all the words for 'obeisance... (etc.)' twice
over; my translation both summarises and attempts to catch the tone
of the Buddha's insistence on the physical realities of subordination.

354

STEVEN COLLINS

The Sakyans had only recently been conquered by Pasenadi: perhaps

the emphasis here is meant to suggest that their obeisance is not freely
given.
# 8.3. The last two terms used of the Buddha (which are also used of
the first king in # 20) are pdsddikataro and mahesakkhataro. 'Charismatic' for the former is obviously a rather tendentious rendering; I do
not use the word in the technical sense current in the sociology of
religion, following Max Weber, but in the more diffuse everyday sense
in which a charismatic person is one has some special, indefinable
force to his or her personality. The word literally means 'with most
pdsdda', a term suggesting a wide range of possible qualities, many of
which do concern religious values and virtues: included in its connotation are the ideas of brightness, clarity, pleasantness, joy, serenity and
faith. To describe someone as pdsddika is to say both that he or she
has these qualities and that he or she inspires them in others. (For a
translation of pdsddika as 'pious' see K. R. Norman [69: 45, 87, and
89, for Th 432, 927 and 949].) Mahesakkhataro is taken to mean
'with most {sa' (see PED s.v. mahesakkha; BHSD v.s. mahes.dkhya), a
term from root ~g, is to have power, authority, capacity, etc.
While it may be surprising to have Pasenadi refer to himself as ugly
(dubban. n.a), the text's putting such a sentiment in the mouth of a king
both prepares for the use of vanqa as 'appearance' in # 13, where
part of the foolishness of beings is to be concerned with differences in
van.n.a, and also aligns the king with brahmins in that both are concerned with vanna rather than behaviour viewed from the Buddhistascetic perspective.
# 9.1. There is a play on the compound Sakyaputtiyd here, which
means both 'Sons of the Sakyan', in the sense of monks in Gotama's
Order, and also 'Sons of the Sakyans': the monks, unlike their lay
'brothers', do not have to pay obeisance to king Pasenadi in the way
described in # 8. Note that the overlap between the reference to the
Sakyans here and in # 8 makes this section # 9 a simultaneous response by the Buddha to assertions of 'primacy' by kings ( # 8) -- note
the possible double meaning of deva in the text -- and by brahmins
( # 3). The substitution by the Buddha of himself (~ Dhamma) for

DISCOURSE ON WHAT IS PRIMARY ( A G G A N N A - S U T T A )

355

Brahmfi/the Vedas at the end of the section leads directly to the

substitution of his 'cosmogonic myth' for the Vedic in # 10ft.
This is perhaps the best place to cite A V 59ff. (see # 7.2), a
,assage which is relevant to several points made in the Introduction
~d these notes. In it the Buddha is made to list to his audience a
~,,:~ries of things which are 'called what is primary' (aggam akkhdyati).
The first three are as follow:
(i) 'Monks, as far as the Kfisis and Kosalans (extend), as far as
there is (territory) conquered by king Pasenadi of Kosala, there king
Pasenadi is called what is primary. But king Pasenadi alters, monks, he
changes (with time); the learned noble disciple who sees this turns
away from him, he is disinterested in what is primary -- how much
more so in what is lesser' (pageva hinasmim).
(ii) 'Monks, as far as the ten-thousand world-system (extends),
there Great Brahmfi is called what is primary. But Great Brahmfi
alters, monks, he changes (with time). The learned noble disciple who
sees this . . . '
(iii) 'Monks, there is a time when the world contracts; as it cont r a c t s . . . [= parts of # 10 below]. When the world has contracted the
Abhassara gods are called what is primary. But the Abhassara gods
alter, monks, they change (with time). The learned noble disciple
who sees t h i s . . . ' After a long list of other things 'called what is
primary', the end of the sutta says that for the Buddha 'what is
primary among (all) these (things) (etadaggar.n)is nirvdna-in-this-life:
that is, Arahantship.
A number of points are worthy of comment. Here Pasenadi,
Brahmfi and the Abhassara gods are given in the same sequence as
AS # 8--10, in a clearly analogous thematic context. The mention of
the Kfisis recalls that Pasenadi had conquered them, as well as the
Sakyans: on the one hand, this suggests (as can in fact be readily
supported from elsewhere) that 'Pasenadi the Conqueror' was a wellknown figure in Buddhist texts; but, on the other hand, the explicit
limitation of his conquest (only as far as his territory extends is he
primary) renders his primacy a matter of local significance (he is no
cakkavatti), just as the Discourse with Assalfiyana rendered the four
Brahmanical classes local and contingent arrangements (see Introduction p. 322). As said in the Introduction, one cannot simply assume

356

STEVEN COLLINS

that all audiences wouldcatch allusions from one text to another; but
this passage from the A~guttara Nikdya suggests that to those familiar
with Buddhist texts, the conjunction of Pasenadi, et al., with a discussion of 'what is primary' in AS would not be unique or unusual. (For
a discussion of the geography depicted in the early texts, see Rhys
Davids (03: 23ff.), and for a convenient map of Pasenadi's territory,
Cousins (84: 284).)
# 9.2. I explain these four epithets as follows: (i) 'he who has
Dhamma for a body' is, as RFG 165 points out, a way of expressing
an idea found elsewhere, that the significance of the Buddha is his
enunciation of the Truth. Thus, famously, at SIII 120 'he who sees me
sees the Dhamma; he who sees the Dhamma sees me'. (Paul Harrison
[92] has shown that the word dharmakdya in early MahfiySna sfitras is
often to be understood, as here [see the commentary Sv 865], as a
possessive adjective, a bahuvri~, agreeing with 'the Buddha', and not
as a substantive set in apposition.) (ii) The phrase 'who has the best
body' translates brahmakdyo, but cannot retain the pun with 'has/is
Brahmfi's body' (i.e. replaces Brahmfi in the Brahmanical view as the
source for scripture, 'father' of the monks, etc.). The word brahma is
standardly used to mean 'best' in compounds; here Sv 865 remarks
that in the first epithet the Buddha's body is called dhamma; and then
states that 'dhamma is said to be brahma-/Brahmgl in the sense of
'best' (dhammo hi setthat.t.hena Brahma ti vuccati). (iii) In 'who is
Dhamma' (dhammabh(tto) the element -bh~to functions simply as an
identity marker, as it does also in (iv) 'who is best' (brahmabMtto); but
in the latter there is a further play on 'has become Brahmfi', in the
sense of being the Buddhist replacement/parallel for Brahmfi in the
Brahmanical view. The epithets dhammabhfita and brahmabh(tta are
found in a number of other places, always explained similarly (e.g. M I
111 with Ps II 76; S IV 94 with Spk II 389). BrahmabhVtta can be
applied to arahants, as at S III 83 (which also says they are best in the
world, ete agg(~ ete setthd; Spk II 282 glosses setthabh~tta).
The commentary here (Sv 865) is quite certain that the sutta is
carefully arranged, with section # 10 following logically after # 9.
Earlier (Sv 861 on # 3) it had explained orasd mukhato jdt~ in the
Brahmanieal view (see note # 3.5) as 'having been in the breast they

DISCOURSE ON WHAT IS PRIMARY (AGGAIqI~A-SUTTA)

357

come out by the mouth'. Here it gives the Buddhist parallel, keeping
alive what is otherwise a dead metaphor: the person with 'firm faith' is
'an own son, born from the mouth through being established in the
Paths and Fruits by means of the sound of the Dhamma coming out
from [the Buddha's] mouth, which had been in his chest'. When the
Buddha is said to 'have Dhamma for a body' it explains 'the Tathfigata
thought out the Word of the Buddha (recorded) in the Three Pitakas
in his heart and brought it out with his voice'. At the start of the next
section, it states: 'thus far the Blessed One has made clear the argument which refutes (the Brahmanical view of) the best; now in another
way, in order to [do the same thing] he says [the opening phrase of
# 10]' (ettavatd Bhagavd set.tha-cchedaka-vddarn dassetvd iddni
aparena pi nayena set.tha-cchedaka-vddarn eva daseturn . . . ddim dha).
# 10.1. The profusion of imprecise time-words, hoti kho s o . . .
samayo yam. kaddci karahaci dighassa addhuno accayena, and
yebhuyyena, 'usually', 'as a rule', might suggest that there is a studied
vagueness about the cosmogony here.
# 11.1. See RFG 166--7 for reminiscences of Brahmanical cosmogony in this section, as now found in Rig Veda X.129 and
Brhaddranyaka Upanis.ad 1, 2.
# 11.2. The narrative now changes to the aorist, from the generalising
present tense used previously. The compound rasa-pathavi was
translated by RhD 'savoury earth'; as PED s.v. rasa remarks, this is
"not quite clear'. It would seem to be a karmadhdraya, 'essence-earth'
or 'earth-essence'; the Sanskrit versions of this story have pr.thivi-rasa
(Mhv I 339; see Jones' note to his translation vol. 1 p. 286; Msv Vin
p. 7); AS r 13 has the two members of the compound separately,
rasdya pathaviyd. It is also possible to take it as a p(trva-nip&a, a
compound whose members appear in reverse order. Pathavi-rasa is
found elsewhere with the sense of an 'essence of earth' which, along
with the ~essence of moisture' (@o-rasa) is necessary for plant growth
(e.g. S I 134, A I 32, V 213, Spk I 250, II 84, Pj II 5--6). This is what
distinguishes fertile soil from ash or stone. The nature of the gradual
evolution here, the step-by-step move from immaterial to material

358

STEVEN COLLINS

existence, is carefully handled by the text, but obscured by RhD's

rendering. At this point there is (material) water, but what 'spread out'
on it was a more ethereal 'essence'.
~/11.3. This comparison refers not to the physical appearance of
the 'earth-essence' itself, but to its manner of becoming visible. It
'appears', i.e. looks like, has the same colour as, sweet ghee or cream

(sd ahosi van.n.a-sampannd . . . seyyathd pi ndrna sampannam, vd sappi,

sarnpannam, vd navaMtam. , evam. -van. n.d ahosi). The verb 'appeared'
here, pdtur ahosi, means 'to come into appearance', 'to become
visible'. (Cf. below # 14.1 on the comparison with the growth of
mushrooms, and # 14.2 on the kalambukd plant.)
# 11.4. The commentary (Sv 866) tells us that it was like the

kanikg~ra flower, which is yellow (this no doubt suggested by the

further comparisons with cream and honey).
# 11.5. The word khudda (not in PED in this sense: see PTC and
Childers) can refer both to honey and to the bees which make it. The
three foods with which the earth-essence is compared, along with oil
(tela) and molasses (phdnita), comprise the five kinds of medicine
(bhesajja) allowed to monks, which can be eaten after noon and stored
up for up to a week (see Introduction p. 329). At D I I I 71, the
Cakkavatti-sihangtda Sutta, another tale whose reference to 'historical'
reality is as complex as that of AS, it is said that these five things in
the future will disappear when life on earth has degenerated to its
lowest level, and a kind of raw grain (kudr~saka) will be 'the best'
(aggam.... bhavissati). The commentary there (Sv 853) says that the
five things are 'the primary (i.e. best) tastes in the world' (loke aggaras(mi); and the sub-commentary (DAT III 38) explains that they
are 'pre-eminent' in the sense of being sweet and in that of being
medicine. Here in AS, of course, they are agga because they are
chronologically primary. As part of a standard set of four requisites
(parikkhdra) for monks, these medicines would have been very
familiar to the monastic imaginaire.
12.1. Lola-jdtiko; there is a reminiscence of jdti as caste here: as
ever, the relevant 'caste' distinctions for the Buddha are moral ones.

D I S C O U R S E ON W H A T IS P R I M A R Y (AGGA1N~lqA-SUTTA)

359

# 12.2. This, I suggest, is a reference to Sekhiya Rules 52 and 53, at

least in the way they were understood by the commentarial tradition.
The former reads na hatthanillehakam bhu~jissdmi, 'I shall not eat
licking (my) hand(s); and the latter na pattanilIehakam, bhufzjissdrni, 'I
shall not eat licking (or wiping) the bowl'. Both Sp 894 and Kkh
150--1 explain the former by saying that one must not lick one's
hand, even to the extent of a finger (atigulimattam. pi); but it is
permissible to eat with one's fingers, in the sense of taking solid foods
and putting them in the mouth. (Note that the 'earth-essence' is not yet
'solid food'.) The latter is explained as meaning that one must not
wipe the bowl even with one finger (eka/zguliyd pi), nor lick it with the
tongue (the verb is nillehitabbo in both cases). (See also next note.)
# 12.3. f~lumpa-kdrakam.. This is an adverbial, n.amul gerund. I take
the picture here to be of greedy beings, full of craving, grabbing big
handfuls of the 'earth-essence' in both hands and stuffing them in their
mouths. Note that the plural hatthehi is used: this, I think, means that
each being used both hands. In the previous scene, when the beings,
imitating the first miscreant, tasted the earth-essence 'with their
finger(s)', the singular ariguliyd was used distributively: each one used
one finger. If in this scene only one hand were used (in the normal
South Asian manner of eating), the singular hatthena would have been
possible. I see here an implicit reference to the manner of eating
prescribed in Sekhiya rules 39, 40, and perhaps also 42, and 46 (Vin
II 214 and IV 194--6; commentary at Sp 893, Kkh 150). Sekhiya 39
prohibits mouthfuls which are too large (atimahantam. kabalanO; 40
directs monks to make round mouthfuls (parimandalan.~ dlopam), not
long (dfgham.). (ftlopa and kabala are synonyms [see CPD s.v. dlopa];
texts and commentaries here either alternate between dlumpa- and
6lopa-, or explain the former by the latter: Sv 866, DAT III 53, Vista
417 = HOS ed. XIII 44, Vism-.t 606.) Sekhiya 42 prohibits putting the
whole hand into the mouth when eating, 46 stuffing one's mouth so
that the cheeks are swollen (avagan.d.a-kdrakam., another adverbial
gerund; commentaries explain 'like a monkey'). RhD rendered
dlumpakdrakam. 'breaking off lumps of it', perhaps following Sv 866,
p#jda-pind, e chinditvd, 'breaking (sc. dividing) it up into morsels'; but
an actual physical act of 'breaking' only makes literal sense if rasapathavi is already solid earth. The verb chindati in such contexts does

360

STEVEN COLLINS

not have to refer literally to a physical act of breaking; and if I am

right to see rasa-pathavi as an ethereal, not-yet-fully-material 'earthessence', the word -kdrakam must refer rather to the beings' manner
of eating. The namul form was called an adverbial gerund by Whitney,
since 'the accusative of a derivative nomen actionis in a, used
adverbially, assumes sometimes a value and construction so accordant
with that of the usual gerund that it cannot well be called by a
different name' (1889: 359--60: see Introduction note 42). One might
then render dlumpa-kdrakam. 'in the manner of someone taking
mouthfuls'. The point is not so much the precise physical action as the
excessively greedy manner of eating. (The connotation of greed and
savagery might also be seen in the fact that dlopa can also mean
'plunder'; Vism-t 906, commenting on this passage, suggests this when
it glosses vilopam katvd: vilopa usually means 'plunder' of robbers or
kings.)
The linguistic parallels in this note and # 12.2 might seem to
modern readers not entirely immediate and obvious. But the rules of
monastic deportment, usually ignored by western scholarship, are in
practfce extremely important and immediate in the day-to-day experience of the monastic order. Monks and nuns -- especially young ones
in training -- are likely to have been very familiar with the rules
specifying the correctly restrained manner of eating, and so a monastic
audience would be more alert to the significance of descriptions of
greedy consumption than are modern western readers.
# 13.1. I take the relatives yathd yathd.., tathd tathd distributively.
The Skt. versions make the point explicit: those who ate more became
of bad varna, those who ate little were good-looking, varnavant (Mhv
I 341 bahu.m/alpam.; Msv Vin p. 9 alpataram./prabh~tataram.). Spence
Hardy's version of the story (1880: 67), based on a Sinhalese paraphrase of the sutta, says 'the difference in their complexions was
increased, in proportion as the brahmas partook of it with more or
less avidity' (on the use of brahmas (sic) here see # 32.1 below).
Kharatta means hardness, (RhD 'solid'), but also harshness (MW
'coarser'); the term refers both to an increased materialization of the
previously immaterial beings in general, and also to a differentially
greater coarseness, ugliness, amongst them.

DISCOURSE ON WHAT IS PRIMARY (AGGAIqI'~A-SUTTA)

361

I translate the text's van.n.a-vevaqnatd as a dvandva, 'good and bad

looks', so that the next sentence is to be taken as expanding the point
(cf. Msv Vin p. 9 vam. advirndtratd). RhD here has 'variety [in ~r 14
'difference', # 15 and ~ 16 'divergence'] in their comeliness', taking it
as a tatpurusa.) The commentaries, however, read and explain vannavevajjatd (Sv 868, DAT III 56--7), 'loss' or 'diminution of vaqn.a'. If
one accepts the reading -vevann. atd, there is a pun here with van.n.a in
the sense of class. Vevannatd can mean 'discoloration', but this refers
to becoming pate, which does not sit well with the light-brahmin/darkothers difference in appearance relevant here. Mon.W gives 'heterogenousness, diversity' for Skt. vaivarn.ya, and in this sense a tatpurusa
would have almost exactly the same sense as a dvandva: 'diversity in
appearance'. S.v. vivarna he gives 'low, vile', 'outcaste', citing no
Sanskrit text; Phyllis Granoff informs me (pers. comm.) that this is a
common meaning of the term in the Pf~rdna-s. (BR do not give these
senses, although they, followed by Mon.W,. do cite Sanskrit sources for
the sense 'of mixed caste'.) PED s.v. vevanniya, and Mon.W s.v.
vaivarnika take these terms to mean 'being an outcaste'. Edgerton,
BHSD s.v. vaivarnika disagrees, probably rightly: Ja III 394 uses
vevanniya of the change in appearance brought on by old age; A V
87, 210 use the term of someone who has become a monk, but the
commentary (Mp V 38) says that by becoming a monk one acquires
vevan,niya in body, by cutting off hair and beard, and in relation to
monastic 'requisites', by giving up luxury. In these two cases the vann.a
in question is clearly appearance rather than social status. But given
the intensive playing with the various notion(s) of vann.a in this and
following sections there may well be an echo of 'bad-casteness' in the
term: the 'pride in vanna' (van.ndtimdna) here attributed to beings
before the emergence of the four classes is obviously meant to evoke
also the contemporary class-pride of brahmins which began the whole
narrative.
# 13.2. Again the last element of the compound here, -jdtikd, recalls
the sense of jdti as 'caste'.
# 13.3. The witticism here, that the expression first used as a lament
for the lost earth-essence is now an exclamation of delight when

362

STEVEN COLLINS

something tastes good, is perhaps underscored by the fact that the

verb anutthunim,su, translated as 'lamented', from anu-stan, could also
be derived from anu-stu, to praise.
13.4. Tad eva pordn,am. aggahham, akkharam anupatanti [v.ll].
anupadanti and anussaranti] na tv en' assa attham djdnanti. (For
aggahha see Introduction p. 331.) In the PTS text the first verb is
anupatanti, which would mean to follow after or come upon; CPD
prefers the v.1. anupadanti (found also in the commentary, Sv 868).
This can mean simply to follow, but has a more specific and relevant
sense, of repeating words after a teacher. This would fit the context,
since the sentence opposes the akkhara, the word(s) of the expression
aho rasa, to their attha, meaning. (This opposition is commonly
expressed by vyah]ana and attha; cp. Msv Vin p. 9 aksara-padavya~jana). There would be then an implied reference to brahmins,
who repeat the words of their ancient texts, without understanding
their real meaning. This is true also of another v.l., attested in
Burmese sources, anussaranti. I prefer to read this latter, since it picks
up directly the words of # 4 above, that the brahmins 'do not recall
the past', pordn,am assarantd; here they do 'recall the original, primary
word(s)' (pordn.a m . . . anussaranti) but don't understand them. The
commentary (Sv 868) glosses aggahham, akkharam as lokuppattivam.sa-katham, 'the account of (the) history (of)/from the arising of
the world', which seems inappropriate here; but it is repeating the
gloss given for por@am, on its first appearance in # 4, pordn,akam.
agga~ham, lokuppattim, cariya-vam,sam. (Sv 862); the point being that
although in general brahmins do not remember the true account of
'cosmogony', when they say this phrase they do recall one small part
of it (without understanding). The word akkhara is used in a number
of ways in this sutta; in all of them there is an opposition, explicit or
implicit, between what words/actions are taken to mean and what
their pre-history or real meaning is ('real' here in the sense appropriate to a parable).
# 14.1. MW gives, s.v. parpata(ka) the meanings 'medicinal plant' and
(from lexicographers) 'fragrant earth'. Of these two senses, the latter
seems preferable, if one accepts the interpretation of rasa-pathavi

DISCOURSE ON WHAT IS PRIMARY (AGGAI~/qA-SUTTA)

363

given above. As with the comparison to the skin on cooling milk-rice

in # 11 (see # 11.3), the comparison to a mushroom here elucidates
the manner of (sudden) appearance of the 'fragrant earth', not the
nature of what appeared. RhD's translation captures this, but some
authors who rely on RhD write as if it were a mushroom which
appeared (but see RFG 172 on soma, and next note on the p~tilatd
creeper.) Phyllis Granoff (pets. comm.) cites pappataka from the
Middle Indic dictionary Paiasadda-mahannava as 'a kind of roti, like
the modern Pfipad ['papadom'], a crispy fried thing made of lentils or
wheat. The other meaning is dried earth that looks like such a roti.
This is a nice contrast with the rasa pathavi. The rasa pathavi is moist,
while this is dried up. That follows the cosmology in most Hindu texts
which go from a watery existence to grosser solid forms'. Whether it
means this or 'fragrant earth', bhflmi-pappataka represents a further
stage in the solidification process of the beings' foodstuffs: 'rapture',
(immaterial/liquid) earth-essence, fragrant earth (or something solid),
creepers growing along the ground (or tubers: see next two notes), and
finally rice (in # 16ff.), first uncultivated and then cultivated.
# 14.2. The text here has badalatd, with a v.1. bhaddalatd, 'excellent
creeper'; this is given also by the commentary in exegesis of the word,
with a v.1. for the latter padalatd. The latter is an alternate reading
given in ross. of Vis 418 (= HOS. ed. XIII 49); Vism-t 907 says this is
the Galoci creeper, which is also known as p&ilatd, 'striking creeper',
which seems inappropriate -- although there is a striking similarity
with the p~tika creeper mentioned by gayana, as reported by RFG
172 (cf. Mon.W s.v.). The Skt. versions read vanalatd, 'forest creeper'.
Mon.W, s.v. bhadralatd, latd and mddhaff, says that Skt. bhadralatd is
Gaertnera Racemosa, also called the Mfidhavi creeper, which is a
spring flower, 'bearing fragrant white flowers' (see next note).
# 14.3. It is not entirely clear to me what plant this refers to. (For
this note I have used the dictionaires, Cone's translation and notes to
the Vessantara Jdtaka (in Cone and Gombrich [77]), and the botanical
works cited in the latter.) The most likely candidate is Convulvulus
repens (also given as reptans), a plant which grows along the ground
(cf. Ja VI 536 nicekalambakd, 'low convulvulus'), although it can

364

STEVEN COLLINS

become a creeper growing up trees. At Ja VI 534 kalambaka plants

grow alongside 'white water-lilies' beside a lake. Given what Mon.W
says about bhadralatr, etc. (see previous note), this might be taken to
mean that it has white flowers; the text here says that the van.n.a of the
creeper, as the earth-essence and fragrant earth, was like sweet ghee
or cream -- something between white and pale yellow. Elsewhere in
the Vessantara, however, 'dluka and kalamba (roots)' are among
things which Maddi 'digs up' (khanantd, Ja VI 578). The same phrase
is found at Ja IV 371, and the same list of edible plants at Ja IV 46,
where the commentary glosses kalamba as a kind of root (tdlakanda);
the phrase also occurs at Ap 145, where Ap-a 417 comments that all
four plants, names are 'names of kinds of roots' (kandajdffna.m namdn'
eva). Given the context -- beings feeding on and in the earth -- it
might seem better to take this as a root or tuber; but given that here,
as before, the comparison is with the way the creeper appeared (pdtur
ahosi) not with what it looked fike (van.n.a), either is possible. But in
either case, the scene imagined is very much 'at ground level'. RhD
rendered this 'bamboo', perhaps influenced by the gloss at Sv 869
n.gd.ikd (cf. D A T III 57 n.dl.ivalli). PED is wrong, I think, to state that
kalambaka at Ja VI 534 (not 535) is 'the Cadamba tree': and also
wrong to cite Ja IV (not VI) 290 for kalamba-rukkha. There is
constant confusion in ross. between kalamba and kadamba; the latter
is a tree, quite different from Convulvulus repens or whatever the
'root' kalamba is. Ja IV 290 has a v.1. kadappa, which is perhaps an
error for kadamba. Msv Vin 10 seems also to have made this confusion: kadambakdpuspam; Mhv 341 has kalambukd, which Jones
renders as 'bamboo', perhaps following RhD.
15.1. This rendering of the difficult phrase ahu vata no, ahdyi vata
no is taken from R F G p. 169. Ahdyi is an aorist passive from hd, to
leave, whose passive takes the sense 'decline', 'come to an end', etc.;
the v.1. apdyi is an aorist from apeti, to go away.
# 16.1. Akan. o, athuso.., tand.ula-phalo. Kano is a fine red powder
under the husk; thusa is the husk or chaff which grows around rice
and corn; tand.ula refers to husked rice, (PED 'ready for boiling'), so a
cumbersomely literal rendering of tan.d.ula-phala would be 'having a

DISCOURSE ON WHAT IS PRIMARY ( A G G A I q ~ A - S U T T A )

365

fruit which consists in husked rice'. Compare the same terms used at
Mhv V 30, of the paradisial conditions at the time of Agoka's consecration, and Anfig v.27, at the time of the future Buddha Metteyya.
# 16.2. n@addnam, pafi~dyati. The commentary (Sv 869), followed
by RhD, takes apaddna in the literal sense of 'cut', and interprets this
to mean that the signs of cutting from the previous gathering were not
to be seen. I prefer to see this phrase as referring to another in the
long list of things which were not yet known, as the aetiology of
contemporary everyday life unfolds; thus I take apaddna to refer to
the whole business of harvesting _rice with husks, in opposition to the
carefree 'gathering' of ready-husked grains described here.
# 16.3. In this section instead of another repetition of vanna-pride,
the narrative switches directly to the aetiology of sex. The version in
Buddhaghosa's Visuddhimagga (418 -- XII 50--1) provides an
interesting transition (found in abbreviated form in Spence Hardy's
Sinhalese version (1880: 67) and in the 19th. century text NitiNighand.uva: see Appendix 2). After the rice appears so too, spontaneously, do cooking utensils and fire; although the rice thus cooked
is like jasmine flowers, and needs no sauces, when the beings eat this
'gross' (ol.drika) food, urine and excrement appear in them -- presumably the previous foodstuffs were too ethereal for such a result -and openings appear in the body to let it out, like wounds or sores
(van.a-mukhdni). Thus maleness and femaleness (purisa-, itthi-bhdva)
arise: a somewhat ironic idyll of 'domestic life in the olden days'. (The
body and its orifices are regularly compared to a boil with nine holes:
e.g. A IV 386--7, Pj I 46, Vism 196 = VI 93.) There seems no point
worrying here, as does the commentary, about how a man and a
woman could have existed before the appearance of the parts (it says
this is due to previous karma); it is that appearance which constitutes
them as differently gendered. R N ) has here 'in the female appeared
the distinctive features of the female, and in the male those of the
male. Then truly did woman contemplate man too closely, and man,
woman.' This rendering suggests either that all (? some) members
of each gender acted thus, or perhaps that the terms are generic:
'Woman' and 'Man'. But there is no reason for either assumption; all

366

STEVEN COLLINS

the transgressions in the 'Fall' are committed first by an individual (or

individual pair, as here), and then by others.
# 16.4. I follow the commentary's gloss of set.t.hi as 'ashes' (chdrikam.,
Sv 869). KRN Coll. Pap. IV: 163--4 derives the word, which is not
in PED, from Skt. *~isti (from root gis to remain), and renders it
'remainder, dregs', suggesting that it might have meant liquid dregs,
and that the gloss in Sv was a guess.
# 16.5. Nassa asuci, nassa asudti, literally 'perish, (disappear, become
lost) impure (one)'. The phrases are in the singular: I take them
distributively, applying to each person of the couple. Phyllis Granoff
(pets. comm.) suggests that there might be here a reference to a
Brahmanical custom at weddings (and at the consecration of images:
she cites Visn.udharmottara Pf~rdna III, 100 v.1) where substances
such as earth, dung or ash would be used for purification. In this case
nassa asuci would mean 'depart impurity' (asuci here in the neuter
rather than masculine and feminine). Although speculative, I am
attracted by the suggestion, and my translation attempts to retain
traces of both these senses.
The causative form of nassati, ndseti, apart from the meaning 'cause
to disappear, destroy', found in various contexts, has a specialised
sense in the Vinaya, where it means 'expel' from the sam.gha (wrongly
explained in PED and PTC as 'atone'): Vin I 85--6 89, 173, 1II 33,
40, etc. Commentaries distinguish different forms of 'expulsion'
(ndsan6): Sp 582, 1014--5, Kkh 127--8, etc. See also Homer (38:
xxvii), (51:109 n.1). Vin I 85--6 concerns various sexual offences
which either prevent ordination or entail expulsion; in one case a
pan. d.aka -- usually translated 'eunuch' but better, as Zwilling (92:
204) explains, 'homosexual' -- attempts to have sexual relations with
monks and novices, and is told by them to 'be off', 'depart' (nassa,
vinassa, Homer's translation); elsewhere the commentary to the Vinaya
explains the second term by the first, adding that the sense is 'go
where we cannot see you' (Sp 871), which nicely fits AS here: in ~ 17
people build houses to conceal this 'immorality'. The parallel between
the beings' explusion from their community in AS, and the explusion
of members of the sam.gha is very neat; and the reversal of the normal

DISCOURSE ON WHAT IS PRIMARY (AGGAI~IqA-SUTTA)

367

values of the wedding ceremony (especially if one adds Granoff's

suggestion about the use of these three substances for purification) is a
particularly deft touch.
f/16.6. It is odd to find akkhara used of actions; RFG 171 must be
fight in seeing this as 'the levelling process typical of oral transmission', and/or as having ousted some other word. The general sense is
the same here, of course; people repeat something without understanding its original meaning.
# 17.1. (A)dhamma-samrnata; dhamma, as always, is polyvalent. The
commentary (Sv 869) takes this to refer to the act of throwing earth,
etc., and not to having sex (followed by MW who transposes the
sentence to # 16; see his note 836 ad loc.). Although this nicely
captures the reversal of values in the marriage ceremony itself (DAT
II 59 specifies this as vivdha-kamma), both the restriction in scope of
the comment and the transposition seem unnecessary. It might make
sense to put the first three sentences of # 17 in ~/16, where they
would highlight the connexion between sex and housing, but they work
just as well in r
where the beginning of human settlement arises as
the immediate consequence of sex, and thence the storing of food
and the need for agriculture. Perhaps best, both sections could be
coalesced into one. It scarcely needs to be remarked that calling
(agrarian) human civilisation adhamma is rather extreme.
# 17.2. Gdmam vd nigamam, vd, lit. 'villages or towns'. Since houses
have not yet been invented, there is a slight anachronism here if one
takes the words in their usual sense. Not every element of the evolutionary process is described explicitly; some elements, like coming
together in separate settlements, are included by ellipsis.
# 17.3. Asaddharnrne ativelam pdtavyatam @ajjimsu. PED derives
pdtavyatam (also written pdtaby-) from pdteti, to cause to fall, destroy,
and translates 'downfall'. CPD s.v. asaddhamma cites the phrase with
pdtabbatam., from pivati/pibati to drink: 'the fact of (its) needing to be
drunk', thus 'intoxication'. Ps II 371 gives pivitabbatam yathdrucim.
paribhufijitabbatam, dpajjanti, 'they reach (the state of) intoxication, of

368

STEVEN COLLINS

needing to enjoy (pleasures, kdme) whenever they want' (cp. Mp II

369). The commentary to AS here (Sv 869) glosses sevitabbatam.,
'addiction'; the sub-commentary (DAT Ill 59) gives paribhufijanakatam (v.1. preferable, -janatarn.), 'the fact of enjoying', and then gives
explanations for the different syllables of the word, finally glossing
yathdrucirn, paribhufijim, su, 'they enjoyed (it) at will'.
The verb dpajjati can simply mean to reach or attain (a state); but it
is commonly used in Buddhist monastic law in the sense of 'to
commit' an infraction (whence the technical term dpatti, a fault).
# 17.4. MW, rather quaintly, says that RhD is wrong to assume that
the houses are for the purpose of concealment, and takes them as
being for shelter. I take the point to be this: at the moment when the
beings are under the intoxication of sex (tasmim. samaye here refers to
that particular time; the more usual narrative-historical 'at that time' is
tena samayena, as at the beginning of the previous sentence), they
cannot control themselves; thus in order to stop repeated exile from
the settlements, they build houses. Note that the first 'offence', in
# 16, is committed by one couple; given the reference to houses
(p!ural) here, one must assume that in # 17, again by narrative ellipsis,
the practice has spread to others. (It is true, of course, that Skt. gr.ha
can be used in the singular or plural for 'house'.)
# 17.5. Evam kira bho sddh5 ti. This phrase occurs at the end of the
sentence; RhD took it as being said after the second being had
gathered rice: 'so much, they say, will about do'. But this ignores bho,
a vocative of friendly or respectful address; and 'they say' for kira is a
little odd at this point in the narrative. So far, only one being has had
the idea of storing rice, and the second being is the first to have
thought of storing two days' worth: the necessary amount can hardly
have been received wisdom. I assume that, regardless of positioning in
the sentence, the phrase is said by the second being, either overtly to
the first, or mentally, addressed either to the first being or to himself,
giving his (the second being's) implicit rationale. It is also possible,
as K. R. Norman suggests (pers. comm.), that kira simply has an
emphatic meaning here.

DISCOURSE ON WHAT IS PRIMARY (AGGANlqA-SUTTA)

369

# 17.6. Sannidhi-kdrakam sdlim. (see Introduction pp. 24--8). For the

Buddha's superiority to other ascetics on the grounds of his not
making a store, see D I 6 with Sv 81--4.
# 17.7. I take the point to be that rice-cultivation is now necessary,
and so the rice, which previously grew wild, now stands in planted
groups or lines.
# 18.1. The word is dhammd: states of mind, phenomena, etc.
# 18.2. Mariyddam. is glossed by the commentary as simam, the word
for a monastic boundary. (I take both text and commentary as using
generalising singulars.) K this refers to physical boundaries, then this is
definitely another reference to monastic fife; but the word mariyddd,
although it can refer to the boundary of a field (e.g. Vin III 50, A IV
237--8), can also refer to a system of rules: e.g. A III 227, Vism 15 =
I 41 (referring to the rules for marriage, etc., of ancient brahmins,
pordn, am brdhmandnam mariyddd). Perhaps connotations of both
senses are present. The Skt. versions are clear: Mhv 346 gdliksetrdndm

sired nayensuh.. I imam bhagavantdndm ~dliksetram imam dtmakam;

Msv Vin p, 14 ksetrdni mdpivantah., simam baddhavantah., marydddm.

sthdpitavantah. .
# 19.1. 'From this moment on' is tadagge, 'from this beginning' (cf. Sv
870 tam aggam, katvd). Note that not only theft and lying, but also
accusations and punishment (normal parts of legal procedure) are seen
as 'bad things', as the next section states. See further # 20.2 and 22.1.
# 20.1. Sammd-khfyitabbam. khiyeyya. The verb khfyati is said by
PED to be the passive of khayati, meaning to waste away, but it is
difficult to see how this could be so semantically. The word is used in
this sense almost solely in the Vinaya, where it standardly occurs in a
group of three verbs, ujjhdyati khiyati vipdceti, said by CPD (s.v.
ujjhdyati) to be 'a stock group of three near synonyms' meaning 'to
become indignant or irritated, grumble, murmur, complain, protest'.
The contexts in which the verbs are used show that they must have

370

STEVEN COLLINS

this meaning, despite the difficulties of derivation (cf. Sp 296, where

khiyanti is glossed as avan.n.am. kathenti, 'they speak dispraise').
# 20.2. Pdrdfikd 2 explicitly links theft with royal jurisprudence and
punishment: 'should any monk with intent to s t e a l . . , take what is not
given, in the kind of theft for which kings arrest the criminal and beat
him, imprison him or banish him, saying "this is a criminal"...' (yo

pana bhikkhu. . . adinnam theyyasafikhdtarn, ddiyeyya, yathdNpe

adinnadgme rdjfio coram gahetvd haneyyum, vd bandheyyum, vd
pabbdfeyyum, vd, coro t i . . . ) . The original story in the Vinaya (Vin III
41ff.) concerns king Bimbisfira of Magadha; perhaps the reference
simply stayed on in the text of the Pdtimokkha; but perhaps the rule
means to outlaw only the kind of theft for which 'a monk could be
subjected to royal jurisprudence, punishment and publicity (defined
in the text as stealing anything worth a pdda or five mdsaka-s). The
compound theyya-sam, khdta (see Introduction pp. 327, 330) means
'counted as theft'; traditional exegesis takes this to mean 'counted (by
the criminal) as theft', i.e. intentional. The mention of kings here
suggests another possible meaning: 'counted (by the king, publicly) as
theft' -- that is, an act of 'taking what is not given' has to be actually
or potentially a public matter, liable to bring disrepute and difficulties
for the Order as a whole, if it is to be important enough to justify
explusion from it. But regardless of the history or precise sense of the
rule, this is the wording it has in the Vinaya with which all monks and
nuns would be familiar. The exegesis of 'beat' at Vin III 47 includes,
as here, beating with hands or sticks.
# 20.3. For 'most charismatic' and 'with greatest authority' see # 8.3.
The first three adjectives here are found elsewhere, where the translation 'most charismatic' for pdsddikatara would be less appropriate.
# 21.1. The etymologies in # 21--5 were the subject of studies by
Kuhn (12) and Schneider (54). More than one analysis of the term
mahdsammata is possible: for detailed argument see Appendix 1. The
most significant point concerning mahdsammata which previous
writers have failed to grasp, in my opinion, is that since all of the
other words etymologised in these sections are standard terms apply-

DISCOURSE ON WHAT IS PRIMARY (AGGAlqlqA-SUTTA)

371

ing to all members of the relevant social group; it seems necessary to

find a meaning or meanings of mahdsammata where this is also the
case. In addition to the nirukti given in the text, the use of mahdsammata plays on three meanings, all of which I assume would be caught
by an educated audience. It is: first, an adjective applied to all
members of the ksatriya class, meaning 'considered to be great',
'regarded as a nobleman'; second, an adjective applied to all ksatriyas,
meaning 'highly esteemed'; third, a noun, an ironic description
denoting the first person appointed in ~ 20, the first king, and means
'the Great Appointee'. The irony here lies in the fact that this sense of
-sammata recalls its use as a standard term for monastic appointments: the first king of the primitive 'community' is just like a
monastic functionary. In all senses the word is used as a description or
title, not a proper name as it later became.
# 21.2. On this nirukti see now Gombrich (92b), who finds here a
reference to another Brahmanical motif.
# 21.3. I read paresam, here, with some mSSo, the commentary, and
Vism 419 = XIII 54. Once again, the precise range of connotations of
dhammena is difficult to determine, and impossible to convey in a
translation. Mahflbhdrata 12.59.127 has rahfitdg ca prajdh, tena rdjeti
iabyate.
# 21.4. The second half of this sentence, which is printed separately
in the PTS edition, must, I think, be taken to refer to the ksatriyas; the
point being that different groups in society are differentiated by
function, and the 'original, primary terms' applied to them for this
reason. RhD took it to refer to the beings en masse, translating 'their
origin was from among those very beings, and no others; like unto
themselves, not unlike': as if the point were that the ksatriyas were no
different from the other beings. (MW follows suit, slightly further
elaborated to 'like ourselves, no different'.) While this is not an
inappropriate sentiment in the context, I do not think the syntax
supports such a rendering. The words here are in the genitive case:
# 26 has 'thus was the birth of the samana-group, from these four
(other) groups' (imehi. . . catfthi mand. alehi samana-mand, alassa

372

STEVEN COLLINS

abhinibbatti ahosi), where the phrase 'from the (other) four groups is
in the ablative. Although the genitive case can usurp the functions of
other cases, there seems no reason why the ablative should not have
been used in # 21 also, if RhD's interpretation had been intended.
As in # 7.2. my translation does not try to catch the overlap
between the uses of dhamma here and in the next sentence.
# 22.1. Note that banishment, one of the tasks of the first king, is
included in what is bad and unwholesome. There is no reason to
assume that the narrative voice here intends any criticism of the
sentiment.
# 22.2. This etymology for brdhmana is common elsewhere: e.g. Dh
386, Sn 519. It would have had more immediate sense in a dialect
where brdhmana appeared as bamhana or bambhana; it seems that
such a form lay behind Pali brdhmana, which was restored as a
Sanskritism: see, e.g., KRN Coll. Pap III 131--2, 239--40.
# 22.3. The PTS text has pann.a-, which must be a mistake (or
alternative) for the reading panna-, which is found in the commentary
and glossed as a past participle of pat, to fall. RFG 172--3 suggests
reading sanna-musald, '(with) pestle set down', here, seeing in this
word and in vftadhumd yet another allusion to Brahmanical texts, this
time to law-books. But there are problems with this. As he notes, the
Brahmanical dharma-rule applies these words to those from whom the
ascetics beg, not the beggars themselves; they should beg from houses
where the pestle has been laid down and there is no more smoke
(from the cooking fire). The texts RFG cites read sanna-; the version
of the rule in the Brdhm@d.a Purdna (I, 7, 177) reads vipanna- (the
corresponding passage of the Vdyu Purdna has dsanna-). I cite these
latter two texts from Kirfel (27: 99). Patrick OliveUe suggests (pers.
comm.) that the parallel with the dharma-rule can be maintained by
reading (whatever the first part of the compound) -musale, a locative
case meaning 'when the pestle has been set down'. One would have to
make a sentence break after jhdyanti ('they meditated'), and read all
three next words in the locative to preserve the full parallel.

DISCOURSE ON WHAT IS PRIMARY (AGGAlqlqA-SUTTA)

373

# 22.4. Again, the narrative alludes by ellipsis to a further stage of

evolution. Where gama and nigama in # 17 were used for 'settlements' before the use of houses, here they can be taken to refer to
familiar kinds of village and town known to the audience of the sutta.
Moreover, 'royal cities', in the plural, suggests by implication that we
are now some time away from the first king, and that the society
envisaged here is not ruled by a single monarch, but by a plurality of
local chiefs.
# 22.5. Here the old brahmins recreate the time before storing food,
in # 16. They are thus like (ideal) Buddhist monks, who beg for their
food and store it only minimally.
# 22.6. It would seem that all translators, from the time of the
Chinese versions cited by Meisig (88: 146--7) through to modern
versions, have missed the joke here, taking jhdyakd from jhdyat?, to
meditate: thus RhD 'the brooding one'. PED s.v. jhdyaka (Stede rather
than Rhys Davids, I assume, unless he changed his mind) renders 'one
who makes a fire', presumably deriving the word from/hdyati 2, to
burn. This is surely fight. It is a much more obvious general term for
brahmins than 'meditators', and it gives the nirukti some point, which
it otherwise lacks: contemporary (non-ascetic) brahmins tend their
sacred fire, whereas the true brahmins of old, like Buddhist monks
and the ideal brahmin-renouncers described in the law-books, abandoned their fire for a life of mendicancy and meditation. Mette (73:
33), who also sees in the word-play a criticism of Brahmin sacrificial
priests, suggests that it derives from Skt. ydjaka, 'sacrificer' (fi'om root
yaj), which becomes jdyaga in Prfikrit; the word-play is thus with
jhdyaga > jhdyaka in Pali. The joke would be much the same in
either case.
# 23.1. Although the text of RhD has 'make books' for ganthe
karontd, this is an anachronism; they quote and translate the commentary's gloss in a footnote: tayo Vede abhisa~kharontd c'eva vdcentd
ca . . . 'compiling the three Vedas and teaching others to repeat them'.
Vdcentd here could also mean simply 'reciting'. The choice of the

374

STEVEN COLLINS

verbs karoti in the sutta, and abhisatikhroti in the commentary are

both incompatible with the (later) Brahmanic view of the Veda as
apauruseya, without an author (human or divine), although it cannot
be proved that either text is expficitly contesting that claim. There may
also be a pun with ganthe in the sense of 'fetters'. Sheldon Pollock
suggests (pers. comm.) that we might take the verb here simply as
'doing': thus 'doing texts' means performing them, reciting them, etc.
# 23.2. This, of course, is the witticism everyone has seen and agrees
on: a-jjhdyaka (which would be Skt a-dhydyaka) is a pun on affhdyaka
(Skt adhy-dyaka), 'student/reciter (of the Veda)'.
# 24.1. RhD has 'various [vissa] trades'; the text has vissutakammante with the v.ll. vissu/visu/visum. Visum is a fairly rare word
found in post-canonical texts, usually in the distributive visum visum,
separately, individually. Vissuta (Skt. vi~ruta) is common, meaning
famous or celebrated. There is clearly no certainty as to the correct
text or interpretation here (or in the next section with sudda). My
translations are defended below and in # 25.1: but nothing vital for
the interpretation of AS as a whole hangs on them. Schneider (54:
587--9) says that the reading vissuta here 'makes so little sense that
one would have to emend it, even if it were certain to be the oldest
and best reading of the mss. known to us'. I think one can make good
sense of it. KRN Coll. Pap. I 256--8 also rejects vissuta-, saying that it
must have come into the text from the commentary (Sv 871), where
the text is given as vissu-kammante, glossed as vissute uggate kammante; these are exemplified by cowherding and trade, to which DAT
III 61 adds agriculture. (He points out that these are three main
occupations of vaigyas found in Brahmanical law-books from the time
of Manu.) Uggata means 'exalted' or 'high', and Norman comments
that this is given in exegesis by the commentary 'despite the fact that
the occupations listed are certainly not uggata'. But a well-known
passage in the Vinaya (Vin IV 7) divides 'work' (kamma) into 'low'
(hfnam.) and 'high' (uggata), and lists under the latter precisely agriculture, trade and cowherding (the same Pali terms). Norman prefers
to read vissu-, which he derives from Skt. vedman, and translates the
compound 'domestic tasks'. If this is right, then presumably the point

DISCOURSE ON WHAT IS PRIMARY (AGGANIqA-SUTTA)

375

is that unlike ksatriyas and brahmins, vaigyas have no specifically

public function.
I retain vissuta-, not because I feel certain of it or find it especially
important, but in order to tie this nirukti to what I see as the spirit of
AS as a whole. The phrase kammante payojesi usually refers to
activities which require money for capital outlay (e.g. D I 71, UI 66).
This is obviously necessary for trade, and I assume (the verb is
causative, 'set going' or 'cause (others) to undertake') that the cowherders and agriculturalists in question are not so much the actual
labourers as those who own or manage the herds and land. (See
Gombrich [88: 55--7] for the social standing of the majority of the
Buddha's lay-following.) Early Brahmanical yarn.a-theory, before the
time of Manu, had been first formulated among agricultural communities, but looked down on agriculture as an occupation for
brahmins. Moreover, it had had difficulty accomodating merchants
into its scheme: they were eventually, by the time of Manu as has been
seen, assigned to the vaigya class (see Gombrich ibid.). It is wellknown that early Buddhism, on the other hand, was closely connected
to cities and trade. Thus one might interpret AS here as calling vaigya
occupations 'of high repute' in relation to this context, as a deliberate
rejection of contemporary Brahmanical attitudes.
The juxtaposition of the references to sex and occupation might
seem strange (logically, one assumes, the ksatriya-group are not
thought to be celibate, nor presumably are the backsliding 'NonMeditator' brahmins; and it would have taken some generations to
produce 'villages, towns and royal cities'); but it is characteristic of the
keen sociological awareness of ascetics to point out the mutual
necessity of production and reproduction. Smith (89: 274--9) states
that in Brahmanical cosmogony, vaigyas are associated with natural
fecundity, which may also be relevant here. To judge from Meisig's
translation (88:150--1 and notes) textual confusion is also to be seen
in the Chinese translations.
# 25.1. As RhD says in a note, both the readings and the word-play
present difficulties. I follow the reading which has both ludddcdrd and
khudddcdra in this sentence as well as the next, so as to make this
section verbally parallel to the others. RhD interpreted the first word

376

STEVEN COLLINS

to mean 'live on hunting', but ludda, 'cruel', is connected with a

number of activities involving bloodshed, mostly in later texts (e.g. Pp
56, Ja II 153--4, Pj I 73, Vbh-a 228, 259, Vista 245 = VIII 68). The
point perhaps is the opposition between the nature of vaigya activities,
'of high repute', and the 'mean' nature of gfidra lifestyle (khudda
means small, inferior: English 'mean' captures the sense in this context
quite well). The nirukti here, in this form, is less semantically pointed
than the others; RhD calls it 'a mere jingle of rhymes'. The Chinese
versions cited in Meisig (88: 152--3) are of no help.
26.1. 'Disapproved' is garahamdno, quite a strong word; 'his own
tasks' is sakam dhamman, which is surely intended to recall the
Brahmanical idea of svadharma, 'own-duty', the idea that particular
classes and castes have determinate duties in society, a notion always
in some tension with a positive evaluation of ascetic renunciation,
most obviously in the Bhagavad Gitd. The commentary here (Sv 871)
says that this ksatriya realised that 'it is impossible to attain purity
(sujjhitum.) merely by lifting up the white umbrella (of kingship)'.
# 26.2. The language here is exactly parallel to that of previous
sections on the four Brahmanical classes, but without any etymology,
and so without the phrase '(along) with the original, primary term(s)'.
The commentary says 'the ascetic group is not (a) separate (class)'
(samana-mandalam ndma visum n'atthi), and that this sentence is
spoken 'because it is not possible to attain purity through caste (but
rather) purity comes from one's own right conduct'. Thus the four
other groups 'give way to' (anuvattanti) the ascetic-group; 'on coming
to the ascetic-group and fully practising right conduct they attain
purity'. It is in order to make this point clear, it says, that the Buddha
goes on to give the teaching which starts in the next section. (See the
discussion of the Madhurgl and Assaldyana Sutta-s in the Introduction,
pp. 3 2 1 - - 2 ) .
# 27.1. The text gives four standard synonyms for 'hell': apdyam,

duggatim, vinipdtam., nirayam.

# 31.1. One might see the presence of 'monk' here, given what has
just been said, as 'the levelling process typical of oral transmission'

DISCOURSE ON WHAT IS PRIMARY (AGGANNA-SUTTA)

377

(RFG 171): the sentence from # 7, with the standard list of titles and
epithets, is simply repeated verbatim. But as argued in the Introduction, the presumption is so strong that the celibate life aimed at
nirvdna translates into monastic status that almost anyone who is 'an
A r a h a n t . . . ' , etc., will in practice be a monk. Most mss. here omit

vusitavd.
# 31.2. It is not impossible that there is here a slight echo of the
other sense of agga, 'first'. The pre-lapsarian beings did not commit
any of the Five Impossible Things (see Introduction); they lived as a
celibate community (sam.gha) and were not (yet) capable of the Vinaya
infractions which constitute the Fall/Evolution of Mankind. By this
stage of the sutta, an audience sensitive to the multiple puns and
cross-references might well see a sense in which the Arahant recreates
that state, but here without the possibility of lapsing. There would then
be even more point in ending with a verse from 'the ever-virgin
Brahmfi'. There would also be the implication that monastic celibacy is
a kind of return to a 'Golden Age', an idea common in many traditions. Other elements from AS are found in Buddhist depictions of
future utopias (e.g. Anfig v.27), but the parallel 'past Golden Age =
future Utopia' is not exact, and the point here is a matter of nuance
only.
# 32.1. This verse is often cited and ascribed to Brahmfi Sanamkumfira; its incorporation here is significant for various reasons. First, the
name recalls the Buddhist view that the term Brahmfi does not refer
to a Creator God, as in the Brahmanical view with which the Sutta
began; it is not even the proper name of a single supernatural, but
designates a class of beings who live in the Brahma-worlds, above the
Divine Worlds (devalokd) in which 'gods' live. Second, Sanamkumfira
means 'ever a youth'; I translate it 'Ever-Virgin' because in Brahmanical
legend this was the name of one of Brahm~'s sons, who remained
perpetually celibate. Like the true brahmins of old (see note # 22.1)
he practises the celibate, 'divine' life (komdra-brahrnacariya, A III
224ff.; cp. Introduction p. 320 and note 34). The name can be given,
as a title, to any great ascetic (see Mon.W.s.v.), and so it is appropriate as a conclusion to a sutta extolling an ascetic point of view.
Third, as Buddhaghosa explicitly points out in his re-telling of AS at

378

STEVEN

COLLINS

Vism 417 = XIII 44, the Abhassara-world in which the myth starts is
a Brahma-world, and so the mention of this figure recalls the prelapsarian beginnings of the tale (the Sinhalese version given in Spence
Hardy (1880: 66--8) uses the word 'brahmas' in place of 'beings'
throughout its telling of the story of origins). Finally, since here the
Buddha cites this verse from a Brahmfi as 'well-spoken' (subhdsita)
and then recites it as his own words (it is also ascribed to the Buddha
at S II 284), he replaces the god Brahmfi and the Vedas of the
Brahmanical tradition with his own 'revelation', the Word of the
Buddha; just as he had done in # 9 (see # 9.2). I take this connexion
between opening and closing frame to be deliberate.
~r 32.2. I assume the word loke is understood, agreeing with devamdnuse; the world of gods and men here encompasses all the supernatural worlds, and it goes without saying that no being in a world
below the human could be 'best'. The phrase jane tasmim, which I
render here as elsewhere 'in this world' refers to the human world;
jana, like loka, can refer to the world and/or its inhabitants.
ABBREVIATIONS
Abbreviations for Pali texts are those of C P D (see E p i l e g o m e n a to vol. 1, and later
additions); texts cited are PTS editions, unless otherwise stated.
AS = Aggah~a Sutta
B H S D = Buddhist Hybrid Sanskrit Dictionary, F. E d g e r t o n (53)
B R = O. Boehtlingk and R. R o t h (1855--), Sanskrit W6rterbuch
Childers = R. C. Childers (1875), A Dictionary of the Pali Language
C P D = Critical Pali Dictionary
H O S = H a r v a r d Oriental Series
K R N Coll. Pap. = K. R. N o r m a n , Collected Papers vols. I--1II.
M o n W = Monier-Williams (1899), A Sanskrit-English Dictionary
M s v Vin = Mddasarvdstivddin Vinaya, ed. R. Gnoli (77)
M W = translation of AS in M. W a l s h e (87)
P E D = The Pali Text Society's Pali-English Dictionary, T. W. Rhys Davids and W.
Stede ( 1 9 2 1 - - 5 )
P T C = Pali Tipitikam. Concordance, ed. F. W. W o o d w a r d et al. ( 1 9 5 6 - - )
PTS = Pali Text Society
R F G = R. F. G o m b r i c h (92a)
R h D = translation of A S in T. W. and C. A. F. Rhys Davids (21)
Skt = Sanskrit

DISCOURSE ON WHAT IS PRIMARY (AGGAlXII~IA-SUTTA)

379

APPENDIX 1: ON THE WORD MAHfISAMMATA

(i) on the form and meaning of the compound
Mahd- means 'great' and sammata is the past passive participle of the
verb sam-man, to consider, esteem (highly), appoint (see below).
Schneider's only comment on this etymology (54: 576) is the odd
suggestion that 'perhaps' this is one of those occasions where the
usually faulty method of associating semantic with phonetic similarity
hits on the right answer. (He says this is true as well of etymology 5,
for jhdyakd, but I think this is also mistaken: see # 22.6.) He refers in
a note to the earlier paper by Kuhn (12); but that writer (p. 216) had
merely stated his disagreement with BR's rendering of mahdsammata
as 'hochgeehrt', and his approval of Childers' 'Great Elect'. Later
Buddhist traditions, and modern scholars, have taken mahdsammata
(hereafter m-s) to be the proper name of the individual appointed in
# 20, the First King. (In later texts and inscriptions he is taken to be
the Buddha Gotama in a former life, and/or founder of the lineage of
both the gfikyan family and of various later kings: see note 40 to the
Introduction.) I think this is not the case in AS itself, although there
must be a reference to the individual appointed in # 20, partly for
obvious contextual reasons, and partly because of the linguistic echo
between sammanneyydma in ~ 20, 'let us appoint' and sammato in
~ 2 1 . The other seven etymologies in # 21--5 are of terms which
were familiar words, and which can be used of all members of the
relevant class or 'group'; the three etymologies in # 21, given in
parallel, precede the statement that 'this was the birth of the ksatriyag r o u p . . , of just the beings, no others . . . ' (in the plural). It seems
necessary to find a sense or senses of m-s which make it a familiar
word, so that the nirukti analysis can have some purchase, and which
can make it apply simultaneously to the individual in # 20 and to all
members of the ksatriya-group.
M-s can be interpreted according to the semantic values of its two
terms, and the possible syntactic relationships between them. Mahd can
be an adjective meaning 'great', 'important', etc., and an adverb meaning
'greatly', 'highly'. Sammata derives from the verbal root sam-man:
O) man, with or without the prefix sam-, can simply mean 'think',
'consider', 'regard (as)';

380

STEVEN COLLINS

(ii) man and sam-man can mean 'think highly of', 'esteem';
(if) the prefix sam- means 'together', and so sam-man can mean
'think together with', 'agree on', 'consent to', 'appoint' (for the last see
below on monastic appointments).
The relevant syntactic relationships possible between the parts of
compounds with -sammata, are three:
(i) X-sammata can be a karmadhdraya, a descriptive determinative,
meaning 'thought to be X', 'regarded as X'. This is by far the most
common use. Some examples are: (a)dhamma-s., 'thought (im)proper'
(DIII 89, i.e. AS # 17; see # 17.1); hfna-s./settha-s., 'considered a
lesser thing/the best' (D HI 94, i.e. AS # 23), ratana-s., 'regarded as
valuable' (Vin IV 163, etc.), atthi-s., 'considered to exist' (Patis-a 635:
the whole phrase is yam. bhikkhave atthi-sammatam loke panditdnam
aham 'pi atthi ti vaddmi; the PTS edition does not print atthi-s, as a
compound, which is a mistake); sddhu-s., 'considered good' (D I 48,
etc.); sacca-s., 'considered true' (S IV 230); mitga-s., 'thought to be
dumb' (Sn 713); sukha-s., dukkha-s., 'thought of as happiness,
regarded as miSery' (Sn 760, transl. Norman [92: 88]), kdya-s., (a
collection of bones) 'regarded as a/the body' (Ps I 260, Spk III 190,
Vbh-a 354). Commentaries in such cases standardly explain the
compound as X ti sammato X-sammato. M-s in this analysis would be
'considered to be great', mahd ti sammato mahdsammato. This is quite
appropriate for a king: kingly self-descriptions do tend to use the
word; and as a substantive mahd can mean 'a great or nobleman'
(Mon.W s.v.). It seems to me certain that this meaning is in play here:
the sense is 'considered to be great', 'regarded as a nobleman'. This
sense of the word mahd is the familiar element which makes m-s
parallel to the better-known terms khattiya and rdja.
(ii) X-sammata can be a tatpursua, a dependent determinative,
meaning '(highly) regarded by X', 'esteemed by X'; the nirukti in AS
pretends that this is the case, seeing mahd- as a short form of
mahdjana. This seems rather unlikely; the very unlikelihood is part of
the ironic point, I take it. Whatever kingly hierarchy AS assumed or
was reacting to (see Introduction), I doubt that the king's 'being
appointed by the people' formed part of it. Examples of other compounds with this form are: loka-s., 'esteemed by people', 'famous' (Ja I
49, Vism 120 --- IV 10: the HOS reading is to be preferred to PTS's

DISCOURSE ON WHAT IS PRIMARY (AGGAlqNA-SUTTA)

381

lena-s.), bahujana-s., 'esteemed by (many) people' (Sv 269), tittha-s.,

'thought much of by heretics' (Ud-a 74). Vbh-a 9 gives two interpretations of hina-s, at Vbh 2: hfnam ti loke sarnmatam, hinehi vd
sammatam g~ttha-bhakkehi g~tho viya, 'considered lesser by people, or
esteemed by lesser (beings), like dung by dung-beetles'.
(iii) it could also be a Karrnadhdraya, with the first part either in an
adverbial sense, 'greatly', or as an adjective 'great': this is the analysis
favoured by modern lexicographers (thus BR 'hochgeehrt', Mon.W
'highly honoured', Childers 'the Great Elect', followed by RhD) and
translators. The word mahd- is certainly regularly used in this way,
but I have not found many examples of compounds with sammata
which have a comparable adverbial or adjectival first member. One
might be at Mhv-t 454, where the simple word sammato at Mhv
XXIII 68, used to mean 'esteemed', 'famous', is glossed sddh~ ti sutt.hu
samrnato, 'highly esteemed as "(a) good (man)" '. The last two words
are printed in the PTS edition separately; they could be read together
as a compound. But although other compounds of this kind with
-sammata are infrequent, this seems the only analysis of m-s which
can make the obvious connexion with the use of sammanneyydma in
# 20. (The version at Vism 419 = XIII 54 underlies the point by
repeating the descriptive aorist sammannimsu, 'they appointed', after
the optative suggestion 'what if we were to appoint' as in AS, directly
before the sentence with m-s.) Thus in this sense m-s as 'greately
esteemed' is an adjective applying to all ksatriyas, and a substantive,
'the Great Appointee', applying to the individual in # 20.
In conclusion: in AS m-s is used in all three of these senses. Sense
(ii), which is linguistically improbable, is given in the text as a nirukti.
In sense (i) it is an adjective applied to all members of the k.satriya
class, meaning 'considered to be great', 'regarded as a nobleman',
which takes up existing (self-)descriptions of kings as 'great'. In sense
(iii) it is an epithet applying differently both to all ksatriyas and to the
first person appointed in # 20, the first king. In the latter sense it
resembles a title, such as Prime Minister or President, rather than a
proper name, as it later became. (See further [ii] below on the 'first
king' as functionary, and [iii] on the word as denoting a role rather
than a person.)

382

STEVEN COLLINS

(ii) on sammata and related terms in the Vinaya

The verb sam-man and its derivatives, including the noun sammuti,
are of course used inmany different ways in Pali texts. They are,
however, employed as standard, recognisable technical terms in two
specific contexts: (a) in connexion with the idea of 'worldly convention', and (b) in the Vinaya, as terms of monastic organization and
self-government. In relation to (a) sammuti is contrasted in postcanonical texts with paramattha, as 'conventional' rather than 'ultimate'
truth, language or teachings (see Collins [82: pp. 154--5]). I cannot see
any direct relevance of this in AS. A related sense, however, may be
in play. From the time of late-canonical texts (Nidd II 173, Vbh 422:
here the PTS edition reads sammati-deva) onwards, one finds regularly a tripartite division of kinds of devd, 'gods': these are sammutidevgt, 'gods by convention', that is, human kings, upapatti- (or
uppatti-)devd, 'gods by birth', i.e. the normal kind, and visuddhi-devd,
'gods by purification', referring to Arahants and Buddhas. (For
references see CPD s.v. upapatti-deva, and for discussion KRN
Coll.,Pap II nos. 31, 44.) Although most uses of these terms are
probably later than AS, Norman has argued that the idea of sammutideva is implied in an early text (M II 213), by the words ucce
sammatam kho etam . . . lokasmim, yadidam atthi devd ti, 'it is firmly
accepted in the world that devas exist' (ibid., Norman's transl, p. 7).
That the connexion between this sense of sammuti and m-s. was
recognised can be seen from Jfi I 132: Mahasammata-kdlato pat.fhdya
lokena devd ti sammatattd rdjardjakumdrgMayo sammutidevd ndma,
'because from the time of Mahfisammata kings and princes have been
regarded by the world as 'gods' (there are) gods by convention'. (See
section (iii) on sdmafiha-ndma.) This attitude towards kings' 'divinity',
once again, must have contrasted with kings' self-understanding; later
audiences may perhaps have seen an indirect echo of the philosophical
sense of sammuti, in that everything 'conventional' is, 'ultimately', of
secondary importance.
Sense (b) seems to me certainly evident in AS. In the Vinaya, 'apart
from the consent of the (relevant group of) monks', ahhatra bhikkhusammutiyd, is a standard qualification included in the description of
infractions of the Rules; sammuti appears at the end of a number of

DISCOURSE ON WHAT IS PRIMARY (AGGANIZ,/A-SUTTA)

383

compounds, denoting that 'agreement', 'consent', or 'permission' is

being accorded to or for certain things in certain circumstances. Some
examples are: santhata-sammuti (Vin III 228), 'rug-permission',
allowing a monk to acquire a new blanket before six years (the period
for which monks are required to keep rugs before acquiring a new
one); ticfvarena avippavdsa-sammuti (Vin III 199, cp. Vin I 109),
'permission to be (regarded as) not separated from the three robes',
granted to sick monks in certain circumstances (i.e. they may in fact
be without them); dat.~d.a- and sikkd-sammuti (Vin II 133--4), 'permission to use a stick, and a string', for sick monks to walk and hold their
bowls; sikkhd-, vut.t.hdna-, and vut.th@ana-sammuti, 'permission to
train, be ordained and ordain (others)', given to women and nuns as
they progress from novices to nuns of twelve years standing (e.g. Vin
IV 319, 320, 330). In these compounds the first member is an object
or action, as it is also in sima-sammuti, 'agreement to (or setting) a
boundary' around the area within which monks are to hold communal
ritual and legal gatherings (Vin I t 06ft.). Forms of sam-man, including
standardly sammata, are found in contexts very close to that in the
AS, when a monk or group of monks is 'consented to', 'agreed on', in
the sense of being appointed to a certain task. Some examples are:
monks appointed to oversee the reception, storage and distribution of
robes (Vin I 283ff.), to oversee certain legal questions decided by less
than a full community (Vin II 95), to remove silver wrongly acquired
and kept by another monk (Vin HI 238), to check on sites which
might be used for a monk's hut (Vin III 150), to assign lodgings and
distribute meals (Vin III 158), to give advice and exhortation to nuns
(Vin IV 50). The standard form in such cases is sammato samghena
itthanndmo bhikkhu, 'a monk called such-and-such is appointed by the
community' (to be robe-receiver, etc.). (An idea of the ubiquity of
sammata in the Vinaya can be quickly gained by looking at the
number of entries for the term, as well as for sammuti, in the index to
its commentary, Sp vol. 8., pp. 1669--70) The usage is so familiar that
the simple words a/sammata are used at Vin I 94 to designate 'un/
appointed' 1TIonks with no reference to any process of selection. Similarly, forms of sam-man are used of the appointment of Mahfikassapa,
Ananda and Upfili to recite the Dhamma and Vinaya at the First
Council (Sp 13, Sv 11, Mv III 9, 10, 31); and in the Dipavamsa (XII

384

STEVEN COLLINS

20, 22) Mahinda is appointed by the monastic community to lead the

expedition to Ceylon (bhikkhusamghena sammato). Mhv XXIX 59,
63 uses the word of a minister appointed by a king, and of the date
selected for a certain task to be performed.
It seems to me quite certain that this term, used of the appointment
of the first king in AS, mahdsammata, would have evoked in the
minds of a monastic audience the process of appointment with which
they were readily familiar, on an everyday practical basis: that of their
own officers or functionaries. The first person appointed with the
function of 'criticising, accusing, and banishing' miscreants, therefore,
was a 'Great Appointee'. It need not be repeated at length here how
such an attitude to the function of kings must have contradicted kings'
self-descriptions. It is true, indeed, that we have no evidence before
Agoka of any kingly self-description. If von Hiniiber (90) is right, this
would not be surprising: he argues that writing did not exist in India
until A~oka's time or shortly before it; so no inscriptional discourse
could have existed. There was no body of oral-textualists to preserve
kingly discourse as brahmins, Buddhist and Jain monks preserved
theirs. Later kings, as Burghart (78) and others have shown, depicted
themselves either as incarnations of God on earth or at least his most
important and most favoured worshipper and representative (see
Fuller [92: 106--54] for an overview). Even Agoka, albeit that he used
Buddhist rhetoric (with whatever degree of sincerity and/or opportunism), always referred to himself as devdnampiya, 'beloved of the
gods' (assuming that this is what the compound means). It seems a
plausible assumption, therefore, that other kings, especially those less
influenced by Buddhist ideas, would have used similar or even
stronger self-descriptions of their relationship to divinity. The account
of kingship in AS denies this, deriving the king's authority extrinsically
from his function and not intrinsieally from his person (despite the
qualities of 'charisma' and 'authority' attributed to the being in # 20).
(iii) on Pali categories of 'names'
As has been stated, later Buddhist tradition took m-s to refer without
irony only to one individual, and used it as a proper name. In speaking of proper names and/or titles, descriptions, etc., one is of course
using English categories. There is no term in Pali which is used only

DISCOURSE ON WHAT IS PRIMARY (AGGAI~NA-SUTTA)

385

to mean 'proper name', and a brief discussion of Pali categories here

may be useful.
In AS m-s is called, as are a variety of other linguistic and nonlinguistic items, as akkhara (see Introduction Part 1~ [ii], ~ 13.4 and
# 16.4): Sv 870 glosses this as sahkhd, sarnah~d, pa~atti, vohdro, all
terms which mean conventional words, categories, concepts, etc. The
Skt. versions have samjhd (Mhv 348, Msv Vin p. 15). Buddhaghosa's
re-telling of the AS story in the Visuddhimagga (419 = XIII 54) says
that the first king (= the future Buddha Gotama) 'was known by
(these) three names' ( s o . . . ti tihi namehi pahfidyittha). The semantic
range of ndma is, of course, much wider than that of 'proper name'; at
Pp-a 175 proper names are classed under paccatta-pa~hatti, 'concept(s) for an individual'. A number of later texts (Moh 110, As 390-2, Patis-a 306, Sadd 879--80) offer a four-fold classification of ndma:
these, with examples, are:
(i) sdmahha-ndmam. (Sadd 879 has sama~hd, with sdmafiha as v.l.,
cf. sdrnahfia- at 880; see below), 'a general name', such as mahdsammata (the only example given in any of the texts);
(ii) gun.a-ndmam, 'a name (indicating) qualities', such as 'preacher'

( dhammakathika );
(iii) kittima-ndmam, 'a made-up name', such as any personal name
decided on by parents and relatives, such as 'Tissa' or 'Phussa'; and
(iv) opapdtika-ndmam., 'a spontaneously-occurring name', such as
'moon' or 'sun' (canda, suriya) which are the same in every age, and
whose referents 'create their own names as they arise' (attano ndmam
karontd uppaffanti, As 932; Mob 111 has pavattanti, 'occur'). Mob
110 and As 392 also class under this fourth heading Buddhist techical
terms such as r@am., vedand, etc., and nibbdna: these are called at
Vista 441--2 = XIV 25, Vbh-a 387 sabhdva-nirutti, translated by
lqanamoli (75: 486--7; 91: 127) as 'individual-essence language' and
'natural language'. The other three kinds of ndma are simply the
function of process of name-giving (ndma-karan. a-kiccam Mob 112,
ndma-gahana-kiccam As 392). (There are other classifications of
ndma at Sadd 878--80, Vism 209 = VII 54, Sp 122, Spk I 95, Pj I
107, which do not help the present discussion.)
Sdmafi~a is derived from samdna, 'equal', 'shared', 'general', etc. In
both Pati and Sanskrit, samdna can be opposed to visesa (Skt. vi~esa)

386

STEVEN COLLINS

'specific' (thus at Pj II 548); in Abhidhamma contexts, all dhamma-s

are said to share the general characteristics (sdmahfia-lakkhana) of
being dukkha, anicca and anattd while individual dhamrna-s have
their own specific characteristics (paccatta- or sabhdva-lakkhana) (e.g.
Vbh-a 150, Vism 607 = XX 3, Spk I 177, II 291--2; for a use of
visesa in this sense see L. S. Cousin's note in Iq~amoli [91:300 n.
12]). Thus the most obvious way of taking sdmahha-ndma here would
be 'general term/description' as opposed to 'specific (proper) name'.
(Maung Tin [20: 499] translates sdmahha-ndma as 'name given on a
special occasion', which is rather loose.) The distinction between
'general vs. particular' is made more complex by developed Buddhist
cosmology, where the same thing happens at the start of each cosmic
era: thus all four texts cited, Moh, As Patis-a and Sadd give the explanation pathamakappiyesu mahdjanena samannitvd t.hapitattd mahasammato ti, 'through being established by the people's appointment in
(each of) the first a g e s . . . ' Hence, for these texts dealing with the
meta-linguistic analysis of referring-terms, although m-s denotes a
specific individual in each cosmic age, the name is a 'general name'
referring to his role rather than being a 'made-up' name for a person
such a Tissa or Phussa. It is like 'President of the United States' rather
than (in 1993) 'Bill Clinton'. When other texts (see note 40 to the
Introduction) refer to King Mahfisammata, they are, to borrow a
distinction from modern analytic philosophy, using the word m-s
rather than mentioning it meta-linguistically; being unconcerned with
cosmic eras other than our own, they can employ it simply and
directly, in the same way as any ordinary (proper) name or definite
description for an individual.
Some further complexity seems to have been caused by the similarly between sdmahha (Skt. sdmdnya) and samahhd (Skt. samdj~d),
meaning 'name or 'designation'. PED s.v. sdmahha, derives the word
from samdna, but translates the compound as 'a name given by
general assent', perhaps because of the phrase mahdjanena sarnannitvd
thapitattd used in all four of these texts. At Sadd 879 the reading
samahha-ndma is explained by tam hi samahfidya pavattam, ndma (cp.
Sn 648 and Pj II 470--1), 'this name occurs by the agreement of people
acknowledging him (as the holder of the role)'. Here sama~hdya is a
gerund from sam-d-jhd.

DISCOURSE ON WHAT IS PRIMARY (AGGAI~IqA-SUTTA)

387

APPENDIX 2: IS THERE A 'SOCIAL CONTRACT' IN AS?

The account of the origin of kingship in AS is often described as a
form of social contract theory in modern histories of Indian social and
political thought. Two influential examples, from many, are Basham
(63: 83--4), and Spellman (64). In my view this is correct, but the
phrase needs to be defined precisely; and the conceptual issue of what
ldnd of contract the AS story envisages must be separated from the
question of whether the text and its 'theory' have ever been used to
make a political argument by actual historical agents. With almost no
exceptions (see below), the answer to the latter question seems to
be no.
According to Lessnoff's (86: 4--5) review of social contract
theories in medieval and early modern Europe, a distinction may be
drawn between two types:
Broadly speaking, two models of social contract can be distinguished... In one, the
parties are the people and their ruler or rulers -- here there are just two parties,
though at least one party, the people, is a collectivity.In the other, the parties are, so
to speak, the building blocks of civil society, in some cases conceived of as lesser
social entities, in others as individual citizens. We thus have bilateral contracts
between people and ruler, multilateral contracts between lesser social bodies, and
multilateral contracts between individuals.
The former of these two models was the first to arise, dating at least
from the eleventh century. Such theories began to appear mainly
through a desire to justify insurrection against an unjust ruler, and to
differentiate the human, artificialist nature of the secular power
invested in the E m p e r o r from the sacred and divine power invested in
the Pope. (Europe too had its competing hierarchical models.) In the
earliest version of this model, the given form of social order is
accepted without question to be that of a monarchy; the issue is to
differentiate the monarchical power accorded 'by contract' to the
E m p e r o r from the sacred imperiurn divinely innate in the Church and
Pope. It is not until Salamonio, in the early sixteenth century, writing
not in relation to a contest between Pope and E m p e r o r but in the very
different situation of the emerging Italian city-states, that one finds a
republican contractarianism, where the Prince of a city is seen only as
the chief magistrate, subject like all others to the Law which he

388

STEVEN COLLINS

administers. Going "back to some ideas of the ancient Greek Sophists,

Salamonio held that all men by nature are equal, but that the conventions of social life -- established by contract between citizens -introduce differences in political power and authority. Later writers,
particularly Hobbes, Locke and Rousseau, develop and elaborate this
second model in various more or less individualist ways, both to
justify different forms of political power (including monarchy), and to
derive the very idea of society from an 'Original Contract'.
The story of the appointment of mahdsammata in AS is clearly not
analogous to the second of Lessnoff's two models, whether between
individual citizens or between different social groups. It does seem
reasonable, however, to see the Buddhist story as analogous to the
first category, a bilateral contract between the people as a collectivity
and their ruler. Such contractual theories are found on occasion also
in Brahmanical texts. (For references, see Lang [92].)
As to the second issue mentioned above, there seems little or no
evidence that any historical agents have used AS and/or its contractual 'theory'. As Lang (92) shows, on at least one occasion later
Buddhist theoreticians used the AS story of the king's appointment in
a text representing an argument with a king on the proper dharma for
kings. Dewaraja (72: 208--10) states -- without presenting much
evidence, and certainly none that AS was invoked -- that 'although the
kings of Ceylon were no more elected than their Indian brethren were
divine, the influence of the elective theory was clearly seen in the
Kandyan period' (18th. century), when on a number of occasions a
'monk election' was held, in which a ritual consultation was held with
certain higher-ranking people. But Dewaraja herself presents evidence
(ibid.) that Kandyan kings also used the language of divine kingship.
In Burma, as for example Sarkisyanz (65: 10ft.) shows, the tradition
of legal texts (dhamma-sattha) often incorporated a reference to
Mahfisammata in its introduction, sometimes (but not always) including other motifs from AS such as the idea of the king's appointment;
but there seems no evidence that this was more than an inherited
textual motif. More work needs to be done on this issue, but for the
moment I think I am justified in saying that for the most part (and
pending the production of actual empirical evidence to the contrary)
although the figure of Mahfisammata did have a lively and varied

DISCOURSE

ON WHAT

IS P R I M A R Y

(AGGAIqlqA-SUTTA)

389

after-life, in practice the contract theory of kingship remained unused.

In Robert Lingat's posthumous work (89) Royautds Bouddhiques, he
agrees with Romila Thapar's (61: 147) assertion that Agoka's model of
kingship, was that of the 'father of the people' rather than their elected
leader; but he claims that this is not, as Thapar had said, contrary to
'the theory of early Buddhism', as 15reached in AS:
L'aggahfia suttanta, en effet, ne pr6tend pas faire la thdorie du pouvoir: c'est un texte
cosmogonique et il faut surtout y chercher une th6orie de l'origine de f h o m m e et des
groupes sociaux (van.qa, varn.a). Dans aucun pays bouddhique, le roi n'est consider6
simplement c o m m e l'61u de son peuple, et surtout comme tenant ses pouvoirs de cette
61ection. Le module bouddhique du roi n'est pas Mahdsammata, 'le Grand Elu', ce
sont le bodhisattva et le souverain universel ou cakravartin. (op. cir. p. 27)

Although I would not call AS 'a cosmogonic text' tout court, I think
Lingat is right to deny that it had a 'theory of power' which was used
by anyone. He cites (p. 97) modem anthropological data from Hocart
(50:51), who reported that Sinhalese villagers believed that Mahfisammata had assigned duties to particular social groups. The Nit#
Nighand.uva, a Sinhala digest of 'traditional law' drawn up in the first
part of the 19th. century for the British colonial government after its
conquest of the Kandyan kingdom, opens with a version of the AS
story, including the 'election' of the king; but it introduces this as an
account of the origin of castes, replacing brahmins, vaigyas and gfidras
with castes indigenous to Sri Lanka (LeMesurier and Panabokke
[1880: 4--5]). This is a small example to support the general claim I
made in the Introduction (p. 325), that in later times in Sri Lanka and
Southeast Asia, where (Theravfida) Buddhist-legitimators had few if
any real challengers in the provision of conceptual coercion, their
'antagonistic symbiosis' with king-thugs tended to be more symbiotic
than antagonistic (though never wholly so), and they were more prone
than were their Indian predecessors to give an ideological grounding
of the existing social 'world'. Thus the AS story becomes an account,
and legitimation, of the existing order of castes. There seems to have
been little if any use of the contract theory to justify defiance of a king
perceived to be unjust, as had been the case with the rise of such
theories in Europe.

390

STEVEN COLLINS
WORKS CITED

Basham, A. L. (63) The Wonder that was India. (2nd. ed.). New York: Grove Press.
Bechert, H. (ed.) (91) The Dating of the Historical Buddha. (Symposien zur Buddhismusforschung IV, 1). Abhandlungen der Akademie der Wissenschaften in
G6ttingen. Philologisch-Historische Klasse Nr. 189. G6ttingen: Vandenhoeek &
Ruprecht.
Boehtlingk, O. and Roth, R. (1855--) Sanskrit WOrterbuch. St. Petersburg.
Bronkhorst, J. (81) 'Nirukta and Astfidhyfiy~: their shared presuppositions', IndoIranian Journal 23: 1--14.
Brown, P. (71) 'The Rise and Function of the Holy Man in Late Antiquity', Journal of
Roman Studies 61: 80--101.
Brown, P. (78) The Making of Late Antiquity. Cambridge: Harvard University Press.
Buitenen, J. B. van (81) The Bhagavad Gitd in the Mahdbhdrata. Chicago: University
of Chicago Press.
Burghart, R. (78) 'Hierarchical models of the Hindu Social System'. Man, n.s. 13:
519--536.
Burghart, R. (85) Introduction to R. Burghart and A. Cantlie (eds.) Indian Religion.
London: Curzon Press.
Caillat, C. (74) 'Pfili ibbha, Vedie ibhya-', L. S. Cousins et al. (eds.) Buddhist Studies
in Honour of L B. Homer. Dordrecht: Reidel.
Carrithers, M. B. (77) 'Review Article: World Conqueror and World-Renouncer: a
study of Buddhism and Polity in Thailand against a historical background by S. J.
Tambiah, Journal of the Anthropological Society of Oxford 8: 95--105.
Carrithers, M. B. (87) 'Buddhist without History', Contributions to Indian Sociology
21: 165--8.
Carrithers, M. B. (93) Why Humans Have Cultures. Oxford: Oxford University Press.
Childers, R. C. (1875) A Dictionary of the Pali Language. London: Trfibner.
Collins, S. (82) Selfless Persons: imagery and thought in Theravdda Buddhism.
Cambridge: Cambridge University Press.
Collins, S. (88) 'Monasticism, Utopias and Comparative Social Theory', Religion 18:
101--35.
Collins, S. (90) 'On the Very Idea of the Pall Canon', Journal of the Pali Text Society
XV: 89--126.
Collins, S. (forthcoming) 'The Body in Theravfida Buddhist Monasticism', in S.
Coakley (ed.) The Body: Comparative Studies in Spiritual and Devotional Practice.
Cambridge: Cambridge University Press.
Cone, M. and R. F. Gombrich (77) The Perfect Generosity of Prince Vessantara.
Oxford: Oxford University Press.
Cousins, L. S. (84) 'Buddhism', in J. Hinnells (ed.) A Handbook of Living Religions.
London: Viking.
Critical Pali Dictionary (1924--), ed. V. Trenckner et al. Vols I and II; vol. III fasc.1.
Copenhagen: Munksgaard.
Davids, T. W. Rhys (1899) Dialogues of the Buddha. Part I. Sacred Books of the
Buddhists vol. II. London:
Davids, T. W. Rhys (03) Buddhist India. New York: G. P. Putnam's sons.
Davids, T. W. and C. A. F. Rhys (21) Dialogues of the Buddha. Part III. Sacred Books
of the Buddhists vol. VI. London: Routledge and Kegan Paul.

D I S C O U R S E ON W H A T IS P R I M A R Y ( A G G A I ~ ' N A - S U T T A )

391

Deshpande, M. (85) 'Historical Change and the Theology of Eternal Sanskrit',

ZeitschJqfi fiir vergleichende Sprachforschung.
Dewaraja, L. (72) A study of the political, administrative and social structures of the
Kandyan kingdom of Ceylon, 1707--60. Colombo: Lake House.
Edgerton, F. (53) Buddhist Hybrid Grammar and Dictionary. New Haven: Yale
University Press.
Edmunds, A. J. (04) 'A Buddhist Genesis', The Monist 14: 207--214.
Eliot, T. S. (32) Selected Essays. New York: Brace.
Erdosi, G. (88) Urbanisation in Early Historic India (Oxford: BAR International
Series 430).
Estlin Carpenter, J. (11) Digha Nikdya vol. III. London: Unwin.
Franke, R. O. (13) D?ghanikdya. G6ttingen: Vandenhoeck and Ruprecht.
FuLler, C. J. (92) The Camphor Flame: Popular Hinduism and Society. Princeton:
Princeton University Press.
Gellner, E. (83) Nations and Nationalism. Ithaca: Cornell University Press.
Gellner, E. (88) Plough, Sword and Book. Chicago: University of Chicago Press.
Ghosh, A. (73) The City in Early Historical India. Simla: Indian Institute of Advanced
Study.
Gnoli, R. (77) The Gilgit Manuscript of the Sa~ghabhedavastu. Part 1. Rome: Istituto
Italiano per il Medio ed Estremo Oriente.
Gombrich, R. F. (88) Theravdda Buddhism: a Social History from Benares to
Colombo. London: Routledge and Kegan Paul.
Gombrich, R. F. (92a) 'The Buddha's Book of Genesis?', Indo-Iranian Journal 35:
179--191.
Gombrich, R. F. (92b) 'Why is a Khattiya called a Khattiya? The AggaFzha Sutta
revisited. Journal of the Pali Text Society XVII: 213--4.
Griffin, J. (80) Homer. Oxford: Oxford University Press.
Gunawardana, R. A. L. H. (79) Robe and Plough. Tucson: University of Arizona Press.
Harrison, P. (92) 'Is the Dharma-kdya the Real "Phantom-Body" of the Buddha?',
Journal of the International Association of Buddhist Studies 15: 44--94.
Heesterman, J. (85) The Inner Conflict of Tradition. Chicago: University of Chicago
Press.
Hintiber, O. yon (90) Die Beginn der Schrifi und friihe Schriftlichkeit in lndien.
Abhandlungen der Geistes- und Sozialwissenschaftlichen Klasse 11. Mainz:
Akademie der Wissenschaften and der Literatur.
Hocart, A. M. (50) Caste. New York: Russell and Russell.
Horner, I. B. (38) Book of the Discipline Part One. Sacred Books of the Buddhists
vol. XIV. London: Luzac.
Homer, I. B. (51) Book of the Discipline Part Four. Sacred Books of the Buddhists
vol. XIV. London: Luzac.
Jones, J. J. (49) The Mahdvastu. Vol. I. Sacred Books of the Buddhists vol. XVI.
London: Luzac.
Kahrs, E. (82) 'Yfiska's use of Kasmdt', Indo-Iranian Journal 25: 231--7.
Kirfel, W. (27) Das Purdna-paFwalaksana. Bonn, Schroeder.
Kuhn, E. (12) Festschrifi Vilhelrn Thomsen. Leipzig: Harrassowitz.
Lamotte, E. (88) History of Indian Buddhism. Louvain: Peeter's Press.
Lang, K. (92) ~
and Candrakirti on the Dharma of Kings'. Asiatische Studien
XLVI 1: 232--243.

392

STEVEN COLLINS

LeMesurier, C. J. R. and T. B. Panabokke (80) Niti-Nighanduva. Colombo, Herbert.

Lessnoff, M. (86) Social Contract. Hampshire: Macmillan.
Ling, T. (81) The Buddha's Philosophy of Man. London: Dent.
Lingat, R. (89) Royeaut~s Bouddhiques. Paris: Editions de l'Ecole des hautes &udes
en sciences sociales.
Macleod, C. W. (92) Homer: Iliad Book XXIV. Cambridge: Cambridge University
Press.
Manuel, F. and F. Manuel (79) Utopian Thought in the Western World. Oxford: Basil
Blackwell.
Maung Tin, P. (20) The Expositor. Vols. I, II. London: Unwin.
Meising, K. (88) Das S~tra yon den vier Stdnden. Wiesbaden: Harrassowitz.
Mette, A. (73) Indische Kulturstifiungsberichte und ihr Verhiiltnis zur Zeitaltersage.
Akademie der Wissenschaften und der Literature. Abhandlungen der Geistes- und
Sozialwissenschaftlichen Klasse, 1. Wiesbaden: Franz Steiner Verlag.
Monier-Williams, M. (1899) A Sanskrit-English Dictionary. Oxford: Clarendon Press.
I~finamoli Bhikkhu (75) The Path of Purification. Kandy: Buddhist Publication Society
(3rd. ed.).
lq~namoli Bhikkhu (91) The Dispeller of Delusion. Vol. II. Oxford: Pali Text Society.
Norman, K. R. (69) Elders' Verses L London:. Luzac.
Norman, K. R. (89) Review of Meisig (88). Journal of the Royal Asiatic Society, 351-352.
Norman, K. R. (90--93) Collected Papers. Vol. I (90) vol. II (91), vol. III (92), vol. IV
(93). Oxford: Pali Text Society.
Norman, K. R. (92) The Group of Discourses. Vol. II. Oxford: Pali Text Society.
Neumann, K. E. (18) Die Reden Gotamo Buddhos. Miinchen: R. Piper.
Obeyesekere, G. (9) The Work of Culture. Chicago: University of Chicago Press.
O'Flaherty, W. Doniger (76) The Origins of Evil in Hindu Mythology. Berkeley:
University of California Press.
Olivelle, P. (91) 'From Feast to Fast: Food and the Indian Ascetic', in J. Leslie (ed.)
Rules and Remedies in Classical Indian Law (Panels of the VIIth World Sanskrit
Conference vol. IX). Leiden: E. J. Brill.
Olivelle, P. (92) Sam.nydsa Upanis.ads. New York: Oxford University Press.
Olivelle, P. (93) The Airama System. New York: Oxford University Press.
Pfili Tipitika.m Concordance (56--) ed. F. L. Woodward et al. London: Luzac (Pali
Text Society).
Ramanujan, A. K. (91) 'Three Hundred Rdmdyanas', in P. Richman (ed.) Many
Rdmdyanas. Berkeley: University of California Press.
Rewatadhamma (69--77) (ed.) Buddhaghosdcariya's Visuddhimagga with Paramatthamahjusd ofBhadantdcariya Dhammapdla. Vols I--II. Varanasi: Varanaseya
Sanskrit Vigvavidyfilaya.
Reynolds, F. (72) 'The Two Wheels of Dhamma', in B. L. Smith (ed.) The Two Wheels
of Dhamma. Chambersburg: American Academy of Religion.
Sarkisyanz, E. (65) Buddhist Backgrounds of the Burmese Revolution. The Hague:
Martinus Nijhoff.
Schneider (54) 'Acht Etymologien aus dem Aggafifia Sutta', Asiatica: Festschrifi
Friedrich Weller. Leipzig: Harrassowitz.
Schneider (57) 'Ein Beitrag zur Textsgeschichte des Aggahha Suttanta', Indo-Iranian
Journal 1: 253--285.

D I S C O U R S E ON W H A T IS P R I M A R Y ( A G G A N I q A - S U T T A )

393

Sharma, R. S. (68) Aspects of Political Ideas and Institutions in Ancient India. Delhi:
Motilal Banarsidass (2nd. ed.).
Sharma, R. S. (83) Material Culture and social Formations in Ancient India. Delhi:
Macmillan India.
Siegel, L. (87) Laughing Matters. Chicago: University of Chicago Press.
Smith, B. K. (89) 'Classifying the Universe: Ancient Indian Cosmogonies and the
Varna System', ConMbutions to Indian Sociology 23, 2: 241--60.
Smith, B. K. (92) 'Canonical Authority and Social Classification: Veda and Varna in
Ancient Indian Texts', History of Religions 32, 2: 103--25.
Spellman, J. W. (64) Political theory of ancient India: a study of kingship form the
earliest times to circa A.D. 300. Oxford: Clarendon Press.
Spence Hardy, R. (1880) A Manual of Buddhism. London: Northgate. (2nd. ed.)
Tambiah, S. J. (76) World Conqueror and WorM-Renouncer: a study of Buddhism and
Polity in Thailand against a historical background. Cambridge: Cambridge
University Press.
Tambiah, S. J. (89) 'King Mahasammata: the first ldng in the Buddhist story of
Creation and his persisting relevance', Journal of the Anthropological Society of
Oxford 20: 101--122.
Thapar, R. (61) Agoka and the Decline of the Mauryas. Oxford: Oxford University
Press.
Walshe, M. (87) Thus Have I Heard. London: Wisdom.
Whitney, W. D. (1889) Sanskrit Grammar. Cambridge: Harvard University Press.
(2nd. ed.)
Winternitz, M. (33) A History of Indian Literature. Vol. II Buddhist Literature and
Jaina Literature. Calcutta: University of Calcutta.
Zwilling, L. (92) 'Homosexuality as Seen in Indian Buddhist Texts', in J. Cabezon
(ed.) Buddhism, Sexuality' and Gender. Albany: State University of New York
Press.
Dept. o f S o u t h A s i a n Languages a n d Civilisations
University o f Chicago