Vous êtes sur la page 1sur 197


Journal of the International

Association of Buddhist Studies
Volume 26 Number 1 2003
In Memoriam Professor Akira HIRAKAWA
by Kotabo FUJIJA......... ...................................................................... 3
Paul M. liARRrSON
Relying on the Dharma and not the Person: Reflection on authority
and Transmission in Buddhism and Buddhist Studies.............. ....... 9
Gleanings from a Comparative Reading of Early Canonical Buddhist
and laina Texts......................... ......................................................... 25
Robert H. SHARF
Thinking through Shingon Ritual.................................. ... ...... ... ........ 51
On the Nikiiya Affiliation of the Srlghaniiciirasaligraha and the
Sphutiirthii Srzghaniiciirasaftgrahatlkii...... ....................................... 97
Can all Beings Potentially Attain Awakening?
Gotra-theory in the Mahiiyiinasutriilarrzkiira ................................... 115
Candrakirti on Digniiga on ......................................... 139
StnlcturalAnalysis of the bSam gtan mig sgron. A Comparison of the
Fouifold Correct Practice in the Aryiivikalpaprave.saniimadhiiralJl and
the Contents of the four Main Chapters of the bSam gtan mig sgron 175
Notes on the Contributors................................................................. 197
The Journal of the International
Association of Buddhist Studies
(ISSN 0193-600XX) is the organ
of the International Association of
Buddhist Studies, Inc. It welcomes
scholarly contributions pertaining to
all facets of Buddhist Studies.
JIABS is published twice yearly, in
the summer and winter.
Address manuscripts (two copies)
and books for review to:
The Editors, JIABS, Section de
langues et civilisations orientales,
Universite de Lausanne,
BFSH 2, CH-1015 Lausanne,
Address subscription orders and
dues, changes of address, and
business correspondence
(including advertising orders) to: .
Dr. Jerome Ducor,
Treasurer IABS,
Section de langues et civilisations
orientales, Faculte des lettres
Universite de Lausanne, BFSH 2
1015 Lausanne-Dorigny
iabs. treasurer@orient.unil.ch
Fax: +4121692 30 45
Subscriptions to JIABS are USD 40
per year for individuals and USD 70
per year for libraries and other
institutions. For informations on
membership in IABS, see back cover.
Web: www.iabsinfo.org
Copyright 2003 by the International
Association of Buddhist Studies, Inc.
Printed in Belgium
Cox Collet
JAINlPadmanabh S.
LOPEZ, JI. Donald S.
MAcDONALD Alexander
SHARF Robert
Professor Akira lIIRAKA WA, known widely not only in Japan but through-
out the world as a respected authority in Buddhist Studies, passed away
of natural causes on March 31, 2002, at the age of 87.
Born in Toyohashi City in Aichi Prefecture on January 21, 1915,
Hirakawa studied as an undergraduate and then graduate student (1939-
1945) at the Department of Indian Philosophy and Sanskrit Philology,
Faculty of Letters, Tokyo Imperial University (now University of Tokyo),
and became Research Assistant of that department in 1946. He was
appointed Associate Professor of the newly established Department of
Indian Philosophy at Hokkaido University in 1950. After teaching for four
years in Hokkaido University, he returned to Tokyo in 1954 to become
Associate Professor of Buddhist Studies at his alma mater. Hirakawa was
granted a full professorship in 1962, a position he held until reaching the
University of Tokyo's mandatory retirement age of 60 in 1975, at which
time he received the title of Professor Emeritus. After his retirement he
taught for 10 years (1975-1985) Buddhist Studies at Waseda University,
Department of Oriental Philosophy, School of Literature. Hirakawa also
served as Chairman of the Directors of the Japanese Association of Indian
and Buddhist Studies for eight years (1983-1991), where he made tremendous
contributions toward the advancement of the Association. In 1993 he was
selected to be a member of the Japan Academy. He went on to become
Chairman and Professor at the International College for Advanced Buddhist
Studies (established in 1996), where in addition to his duties as the director
of research and education, he was responsible for the general administration
of the College. He held this position until passing away.
In a career that spanned over 60 years, Hirakawa brought to fruition
vast achievements in the various fields of Indian, Chinese and Japanese
Journal of the International Association of Buddhist Studies
Volume 26 Number 1 2003
Buddhist studies. Hirakawa was extremely polific in his publications, and
his works attest to the brilliance of his accomplishments. While the list
of his works is too long to enumerate here, particularly worthy of
mention is the Hirakawa Akira chosakushU (The Collected Works of
Akira Hirakawa), 17 vols. (Tokyo: Shunjusha, 1988-2000), on which he
spent 12 of his later years in completing. This work actually comprises
two sub-collections: Bukkyo shiso kenkyii (Studies in Buddhist Thought),
Vols. I-VIII, and Bukkyo no karitsu (The Buddhist Discipline), Vols. lX-
XVII. Most of the main contributions Hirakawa made to the field of Bud-
dhist Studies are included in this collection. I should therefore wish to
briefly summarize its contents below.
Of the first collection, the first two volumes, Ho to Engi (Dharma and
Pratltyasamutpada) (1988) and Genshi Bukkyo to Abidaruma bukkyo
(Early Buddhism and Abhidharma Buddhism) (1991), examine from var-
ious angles the basic teachings of Early and Sectarian Buddhism. The sec-
ond of these includes a revised Japanese version of "An Evaluation of the
Sources on the Date of the Buddha", an English paper presented at the
Fourth Symposium on Buddhist Studies at Gottingen University in April
1988 (published in The Dating of the Historical Buddha ed. H. Bechert,
Part 1, Gottingen, 1991, pp. 252-295).
Vols. ill (1989) and IV (1990), Shoh daijo bukkyo no kenkyii (Studies
in Early Mahayana Buddhism) I, II, are based on an earlier volume of the
same name (1968) with significant additions and revisions. These two
volumes further develop his revolutionary and now famous theory that
Mahayana Buddhism developed out of groups of lay Buddhists centered
around stiipa worship. In order to advance his theory, Hirakawa criticized
both the notion that Mahayana sutras do not represent the words of the
Buddha and the theory that Mahayana developed out of the Mahasiiqlghika
sect; and he sought evidence for his theory in a wide range of literary
sources. The ideas and methodology presented in this work had their
beginnings as early as 1954, in a paper entitled "Daijo bukkyo no kyo-
danshiteki seikaku" (The Historicity of the Mahayana Buddhist Order).
Over the course of the next few decades his theory became widely
accepted in Japan, and the main points of his theory eventually gained the
attention of Buddhist scholars worldwide through the publication in Eng-
lish of "The Rise of Mahayana Buddhism and Its Relationship to the
Worship of Stilpas" (1963). In recent years, however, Hirakawa's theory
has come into question, and has been met with criticism both within and
outside Japan. The majority of these criticisms tend to be based on the
view that Mahayana developed from within the traditional Buddhist sects.
It must be said through that the origin of Mahayana Buddhism is not a
settled matter. Nonetheless, Hirakawa's theory is, as an academic
achievement, worthy of praise for - at the very least - bringing the
matter into question and for ushering in a new phase in the attempts to
answer unresolved problems.
Although it may be what he is best known for outside of Japan,
Hirakawa's accomplishments in Mahayana research are by no means lim-
ited to the above mentioned theory. Of his Collected Works, the follow-
ing are also representative of his immense knowledge of and contribution
to the field of Mahayana studies: Daijo bukkyo no kyori to kyodan
(Mahayana Buddhism: The Order and Its Doctrine) (VoL V, 1989), Shoki
daijo to hokke shiso (Early Mahayana and the Philosophy of Sad-
dharmapUlJcJarika Literature) (Vol. VI, 1989), JOdo Shiso to Daijo-kai
(Pure Land Philosophy and the Mahayana Morality) (Vol. VII, 1990),
Nihon bukkyo to chUgoku bukkyo (Japanese and Chinese Buddhism) (Vol.
The second half of his Collected Works is a compilation of various
studies dealing with the Buddhist Discipline. Vols. IX (1999) and X
(2000), Ritsuzo no kenkyu (A Study of the Vinaya-pitaka) I, II, constitute
a reworking of his doctoral dissertation and first published book of the
same title (1960). Based on the groundwork laid out in these two volumes,
which can be considered his most monumental achievement, Hirakawa
then went on to paint a historical portrait of the early Buddhist order
in Vols. XI (2000) and XII (2000), Genshi bukkyo no kyodan soshiki
(The Structure and Form of the Early Buddhist Order) I, II, which is like-
wise a revised and expanded version of his earlier Genshi bukkyo no
kenkyu (A Study of Early Buddhism) (1964).
The last five volumes are not based on any previous work. Vol. xm
(1998), Bikuni-Ritsu no kenkyu (A Study of the Precepts for Nuns) and
Vols. XIV-XVII, Nihyakugojikkai no kenkyu (A Study of the 250 Precepts
for Monks) I (1993), II (1993), ill (1994), IV (1995), are all devoted to
examining the articles of the and carefully comparing
the extant versions of the various sects. With respect to both detail and
sheer volume, such research had until then not been undertaken in either
Japan or elsewhere, and it is unlikely to be surpassed in the foreseeable
future. It would not bean overstatement to say that Hirakawa's profes-
sional career began and ended with the study of the Vinaya-pitaka.
While the above is a summary of Hirakawa's publications of Vinaya
studies in Japanese, amongst his international contributions in this field,
we may mention the following two works published in India: Shan-Chien-
P'i-P'o-Sha: A Chinese Version by Sanghabhadra of Samantapasiidika,
by P.V. Bapat, in collaboration with A. Hirakawa (Poona, 1970); Monas-
tic Discipline for the Buddhist Nuns: An English Translation of the
Chinese Text of the by A. Hirakawa, in
collaboration with Z. Ikuno and P. Groner(Patna, 1982).
In addition to the works cited above, Hirakawa produced other works
in a variety of fields that deserve mention. Kusharon sakuin (Index to
the 3 vols. (1973-1978), which received the
Japan Academy Award in 1980, is a work which was painstakingly com-
piled in collaboration with his students (S. Hirai, S. Takahashi, N. Haka-
maya, and G. Y oshizu). This landmark reference work continues to be of
great benefit to scholars all over the world.
Furthermore, Hirakawa edited the Bukkyo Kan-bon daijiten (Buddhist
Chinese-Sanskrit Dictionary) (1997) which is based on the U. Wogihara
and N. Tsuji's Kanyaku-taishO bon-wa daijiten (Sanskrit-Chinese-Japan-
ese Dictionary) (1979), though the planning of this work originally began
with a request from the late Professor John Brough. Needless to say it is
a useful contribution to the study of Buddhist texts in Chinese translation.
Hirakawa also authored several general surveys, including a two vol-
ume work entitled Indo bukkyoshi (A History of Indian Buddhism) (1974,
1979), and Indo-Chugoku-Nihon bukkyo tsushi (A Historical Survey of
Indian, Chinese, and Japanese Buddhism) (1977). The first volume of the
former has been translated into English (A. Hirakawa, A HistOlY of Indian.
Buddhism: From Sakyamuni to Early Mahayana, translated and edited by
P. Groner, University of Hawaii Press, 1990). Hirakawa, in addition to
the works of primarily academic interest mentioned above, also produced .
more than 10 books intended for a more general readership, not to men-
tion the numerous academic articles not included in the Collected W orks_.
Of these, more than a few have been translated into other languages such
as Chinese and Korean.
A booklet surnmarizinghis career and listing his work was issued by
the Library of the International College for Advanced Buddhist Studies
to mark the first anniversary of his death. The College is planning to put
out a more thorough version of the same in March 2004, to be published
in the Journal of the International College for Advanced Buddhist Stud-
ies, Vol. Vll: Akira Hirakawa Memorial Volume.
It should also be said that Hirakawa was a warm-hearted man. He was
kind to his students and colleagues, of a humble and friendly disposition,
and was loved and respected as both a person and scholar by people all
over the world. Upon completing his Collected Works, he remarked that
"there still remain many problems, for example that of the basic needs
of the Early Buddhist Order (i.e. food, clothing, and shelter), which require
further consideration, and I would like to deal with them in future inves-
tigations" (preface to Vol. XU). Even in his old age, he never lost his
intense passion for discovering and facing new problems, and one cannot
help but be struck with wonder and admiration for such boundless dedi-
cation. Throughout his writings, when he came across a problem that was
beyond the scope of the work on which he was laboring, he sometimes
just identified the problem and commented that he would take it up at the
next available opportunity. Unfortunately though, we may no longer look
forward to such opportunities. It is left to the rest of us now to solve the
problems remaining in the Buddhist Studies of Akira Hirakawa.
Your Royal Highness, venerable members of the Sangha, colleagues,
friends. Grandmothers are often fonts of folk: wisdom, and one of the
things my grandmother used to say was: Sin in haste, repent at leisure.
I always think of this when I sit down at last to write a conference paper,
take out the abstract sent off six or eight months previously, and see what
I promised in a rash and unguarded moment to do. In this case, I under-
took to reflect on the current state and future prospects of Buddhist Stud-
ies, on the relations - past, present and future - between the Buddhist
Order and the Western university, and on issues of authority and trans-
mission in Buddhism and Buddhist Studies, linking these reflections to
the Mahtipaddasutra and the Catul;pratisaralJasutra. So much for what
the abstract commits me to, but of course while doing all this I should also
avoid the temptation-which increases with age - to pontificate or lay
down the law for everybody else, and finally, I might consider myself
obliged to satisfy the expectations of those of you who have heard me
speak before that I might occasionally say something amusing. I clearly
have much to repent, but no more leisure in which to do so.
lt is customary on such occasions as these to assess what is called "the
state of the field." Attempts to review the special character and problems
of Buddhist Studies as a discipline (or congeries of disciplines) have
already been made at previous meetings of our organisation, especially
the one held in Mexico City in 1994, and a whole issue of our journal was
I This is the lightly edited text of the plenary address presented at the opening session
of the XIIIth Conference of the International Association of Buddhist Studies at Chula-
longkorn University, Bangkok, on 9 December 2002, in the presence of Her Royal High-
ness Princess Maha Chakri Sirindhorn, representatives of the Sangha, and members of the
Journal of the International Association of Buddhist Studies
Volume 26 Number 1 2003
subsequently devoted to publishing the relevant addresses by Professors
David Seyfort Ruegg, Luis Gomez, Jose Cabezon ~ d others.2 Things
would be moving very fast indeed if the field itself had changed to any
significant degree since these colleagues surveyed it. However, nobody
has been shifting the boundary pegs at night while we all slept, no impi-
ous local spirits - to advert to the story of the building of Samye - have
been dismantling the entire structure in the dark so that I have to start all
over again. There have, of course, been a few interesting developments
since 1994 - and I will come to them in due course - but on the occa-
sion of this conference, the first in the new millennium by the Western
calendar, and given this particular audience, I want to take a different
tack altogether, and reflect on the institutional context of our work, and
the ways in which that context influences what we do. For it is true that
these previous reflections, useful as they are, generally pay only passing
attention to the environment in which we operate, 'and the ways in which
it affects our work.
It is after all not only what we do that is important,
but how and particularly where we do it. This is perhaps an unwelcome
gambit: we come to conferences like this partly because they enable us
temporarily to get away from, even to forget, the contexts in which we
work. And even when we are at home, we may not care to spend too
much time reflecting on the material circumstances of our lives as schol-
ars. Yet such reflection is entirely appropriate, especially for those of us
here who are academics. After all, many of us have taken to emphasis-
ing the need to study the physical, material, economic circumstances of
Buddhists past and present, the everyday realities of their lives, in con-
trast to investigating such things as the doctrine, philosophy, logic and
other more abstract and theoretical products of elite Buddhist culture,
which is partly why, for example, we have seen an efflorescence of Vinaya
2 See especially D. Seyfort Ruegg, "Some Reflections on the Place of Philosophy in
the Study of Buddhism," Journal of the International Association of Buddhist Studies
(JIABS), Vol. 18, No.2 (1995), pp. 145-181; Luis O. Gomez, "Unspoken Paradigms:
Meanderings through the Metaphors of a Field," ibid., pp. 183-230; and Jose Ignacio
Cabezon, "Buddhist Studies as a Discipline and the Role of Theory," ibid., pp. 231-268.
See also an earlier presidential address by Ruegg, "Some Observations on the Present and
Future of Buddhist Studies," JIABS, Vol. 15, No.1 (1992), pp. 104-117.
3 Seyfort Ruegg's 1992 article is an exception.
Studies in our field of late. It is only fair then to turn the same spotlight
on ourselves, and examine our lives rather than our texts.
Of course, we are not all inhabitants of academia, university teachers
and graduate students. There are in fact two principal groups of people
gathered here today, the second being members of the Buddhist Sangha.
Thus two great institutions are represented in this room, one, the Sangha,
being now almost two and a half thousand years old, the other, the Western
university, which began life in Europe around the 12th century, being a
comparative novelty, historically speaking. This week, then, we gather
together from monastery and campus, united by a common interest in the
study of Buddhism, even if we may be divided by our ideas about what
Buddhism is. To put it like this involves considerable oversimplification,
I admit. Not all Buddhist scholars are academics, nor do all members of
the Sangha, in its broadest sense, live in monasteries. Furthermore, there
are many who possess dual citizenship, who are both Sangha-members
and academics. But you will permit me, I trust, to distinguish the two
institutions and their members for the purposes of this lecture. What I
want to address today is the way in which they have come together in' the
field we know as Buddhist Studies. For it is certainly not the case that the
Sangha simply provides academia with its object of study, that being the
only connection between them. Far from it.
Earlier this year I spent a term visiting an American university well
known for its contributions to Buddhist Studies. In the Department of
Asian Languages and Literature at that university, on the door of the pho-
tocopying room, next to the mailboxes, where it would have to be seen
many times each day, was a large poster from the Office of Student Affairs
proclaiming the message: "Honouring and respecting our differences and
similarities." This is a small example of the fatuous nonsense circulating
in our institutions of higher learning. In this case the entirely reasonable
summons to respect other people's differences has been engulfed by such
a pious concern for inclusiveness that even their similarities cannot go
unhonoured and unrespected. Nobody must be left out, nobody is unwor-
thy of honour and respect, and all must answer the call to provide them,
even if it is now meaningless.
Well, although there are many differences between them, the Sangha
and academia also share many similarities. Some of these similarities are
what you might call generic. Recently in the contex( of another (as yet
unpublished) paper I had occasion to reflect on the 18 ways in which
Buddhaghosa in his VisU(ldhimagga says a monastery can be unfavourable
to meditation practice, to the development of concentration.
The 18 fea-
tures include - to paraphrase Buddhaghosa lightly - too many admin-
istrative tasks, frequent distractions from students, constant official meet-
ings (sangha-kamma), too much construction activity, too many people
coming and going for their own purposes or wanting things from you
(worse, he says, when the place is famous), and the need to deal with
fractious or incompatible colleagues. I am not the first person to observe
that Buddhaghosa's comments about the monastery apply just as well to
the modern university,s indeed I doubt that anybody in this room could
read the passage and fail to see the likeness. One could say that all insti-
tutions in which people gather to live and work togetlJ,er naturally display
certain family resemblances, but I think there is more to it than this, that
there are ways in which, because of their special orientation, the monastery
and the university campus share particular features not so evident in other
institutions, that there is, in short, some kind of deeper connection between
their respective enterprises. Given the pedigree of the Western university
and its relations with the Church, and the traditional role of the Buddhist
monastery in Asia as a major centre of higher learning, this is perhaps
hardly surprising.
In the past, these two institutions developed along their separate lines,
but more recently, for a century or so, the Sangha has in many ways also
been appropriating some of the structural and attitudinal features of West-
ern academia, a process which is seen most clearly in the emergence of
Buddhist universities. This is nothing new: it began in Japan about a cen-
tury ago, and in that country the number of Buddhist universities and
colleges is now quite considerable. Many of them have a long history,
others have only recently arrived on the scene, so the process is clearly
4 See Ven. Bhikkhu NiiI;lamoIi, trans., The Path of Purification (Visuddhimagga) by
Bhadantiicariya Buddhaghosa (Seattle: BPS Pariyatti Editions, 1999), pp. 118-121.
5 See, e.g., Florin Deleanu, "A Preliminary Study on Meditation and the Beginnings
of Mahayana Buddhism," Annual Report of the International Research Institute for
Advanced Buddhology, Vol. 3 (1999), p. 84.
ongoing. Sri Lanka saw similar developments about 50 years ago, as did
other parts of the Buddhist world. Now the Chinese Sangha - or to be
more precise, the Taiwanese Sangha - is entering this phase with especial
vigour. Some months ago, for example, Hsi Lai University in Los Angeles
announced in the newspapers that its progress towards official accredita-
tion had proceeded to the point where it could offer doctoral degrees.
This is one example among many. Everywhere the educational opera-
tions of the Sangha are appropriating the modes of discourse ofthe West-
ern university. At the same time, growing numbers of Buddhist monks and
nuns have been taking degrees at Western universities in Buddhist Stud-
ies, and in this way a convergence of approaches has continued to unfold.
This is not without its occasional problems, because of different cultural
presuppositions and ways of doing things, or divergent understandings of
what education and scholarship are about, as any Western academic who
has supervised Asian Sangha-members as graduate students can testify.
All relationships involve conflict and compromise, these are no different,
and in most cases the problems can be worked through.
However, behind all these more day-to-day difficulties looms a bigger
systemic problem. The system into which the Sangha has been busily
integrating itself is arguably in a state of collapse. The Western univer-
sity is, if not actually in ruins, to advert to the title of the relevant study
by Bill Readings,7 at least in a critical structural condition, to the extent
that all of us should now be warning our graduate students that they enter
it at their own risk. The problems of the system differ from country to
country, but many trends are universal, they are simply mixed in varying
proportions. The list includes sinking government funding, rising costs,
burgeoning administrative superstructures, rampant managerialism, the
6 Hsi Lai University, located at Hsi Lai Temple in Hacienda Heights, Los Angeles, is
part of the educational arm of the Buddha Light International movement, whose base is
Foguangshan in Taiwan.
7 The University in Ruins (Cambridge & London: Harvard University Press, 1996).
For a more readable and less theoretically overburdened treatment of the same issues, see
Cary Nelson & Stephen Watt, Academic Keywords: A Devil's Dictionary for Higher Edu-
cation (New York and London: Routledge, 1999). The latter work restricts itself largely
to the North American situation, but much of the ground it covers will be familiar to those
growth of an all-pervasive accounting mentality, in<;:reasingly intrusive
and record-gathering, more intense competition for resources,
increasing insistence on "relevance" (however that is defined), and a
decline in the morale of the academic profession and of its status in soci-
ety. The university as we knew it - if we ever really knew it - is dis-
appearing, and scholars as disinterested seekers after truth, motivated only
by intellectual curiosity, leading the life of the mind at a leisurely pace
with all material needs taken care of, have long since left the building, to
be replaced by corporation workers like any other, members of "cost cen-
tres" rather than departments or disciplines, monitored and reported on
by people ever alert to the ratio of outputs to inputs, and thus continually
worried lest the satisfaction ratings provided by their "customers" or
"clients" (who used to be called students) drop below 3.5 (or whatever
the magic figure is) on the scale, and chronically appre1!ensive about their
continuity of employment. Tenure is an island whose surface area is
shrinking as the tide of casualisation rises, and at the lower levels tenured
academics with a long-term commitment to their institutions are increas-
ingly replaced by mobile staff on fixed-term contracts or by underpaid
graduate students whose exploitation is one of the major scandals of
higher education. We are all familiar with the general picture. There may
be some who have risen so high into the Brahmaloka of academia that
they believe the destruction of the system will not affect them, but even
though the view from the upper stories can still be quite good, most of
us are faced with the effects of all these changes on a daily basis. All this
has been well said by other people, so there is no need for me to dwell
on it further, even if I can warm to this topic to the point of meltdown.
I am not here to give you a tirade on the decline of the university, com-
posed in equal parts of romantic nostalgia and peevish frustration. Such
a mixture tastes extremely sour; you would not want to imbibe it. But I
do want to reflect on how all this impacts on what we do.
One of the most obvious consequences is that we now have less time
for scholarship, to say nothing of the serene and untroubled state of mind
necessary for the prolonged periods of intense concentration which cer-
tain work requires - as Buddhaghosa was no doubt aware. But since the
pressure to raise publication rates grows ever stronger, the result is an
increase in quantity accompanied by a decline in quality, and I am sorry
to say that this is as apparent in our field as it is in others.
Among the
many new activities diverting us from what are now known as our "core
business operations" of teaching and research, at the upper levels the
professor has added fund-raising to the list of tasks needing to be per-
formed. It is no longer a case of protecting disciplinary territory within
the institution, as the pie is cut up and shared out: now there is often no
pie at all, and professors must forage outside the walls to maintain their
Here, in the matter of fund-raising, the linkages between
academia and the SaiJ.gha and between Buddhist scholars and Buddhist
believers become ever more important, for it is in donations from believ-
ers and institutional links with Asian Buddhist groups that Buddhist Stud-
ies is finding some of the means to survive in this more competitive
environment. The global network of visiting chairs in Buddhist Studies
funded by the generosity of Mr Yehan Numata is the most prominent
example, but there are many others, including the endowment of posi-
tions, the provision of scholarships, the funding of conferences, the sub-
vention of publications, and the underwriting of various projects to digi-
tise the canons, where the generosity of Buddhist donors has been
instrumental to the progress of Buddhist Studies in the academic envi-
ronment. It is perhaps an irony that Buddhist scholars should be required
to tum for funding to those whom they study, and, like members of the
SaiJ.gha, take to mendicancy. But here one can easily overstate the anal-
ogy, since there is no spiritual value ascribed to the process, begging is
not embraced as a means to greater humility or to the conquest of pride
and egotism, but seen merely as a necessary evil. And we will be seeing
a lot more of fund-raising, along with all the other allegedly necessary
8 This decline in quality is also a reflection of specific work practices. Pressure to pub-
lish in the world of academia conspires with cost-cutting by publishing houses to produce,
courtesy of the use of word-processing technology, books which frequently add little or
nothing to our knowledge and understanding of Buddhism. These same books, having
been printed from camera-ready copy (yet another burden transferred to the shoulders of
academics), betray no sign of an editor's hand, and are therefore often longer than they
need to be, and riddled with mistakes.
9 One telling indication of this trend is the announcement by the American Academy
of Religion of a summer workshop entitled "The Entrepreneurial Chair: Building and
Managing your Department in an Era of Shrinking Resources and Increasing Demands,"
to be held at Georgetown University, Washington, D.C., 19-21 June 2003.
evils of maintaining high teaching loads, writing repor:s, generating plans
and mission statements, engaging in research assessment exercises, and
so on.
We all know that these things are only means to an end, and most of
us have a pretty clear idea what that end is, even though, caught up in the
interminable planning and the tedious reviewing now deemed necessary
for quality assurance (or quality management) in our universities, we might
be forgiven sometimes for losing sight of it. What is that end? It is the
preservation, generation and transmission of knowledge, in our case about
the Buddhist religion, and it is something more than that as well, it is a
kind of practice to do with that knowledge. And this is analogous, I would
contend, to the purpose of the Sangha, the institutional core of the reli-
gion we study, a purpose the Sangha has been pursuing now for almost
two and a half millennia: the transmission and of the dharma.
Its members too have had occasion to reflect on the problems that some-
times arise in pursuing that purpose, and some of these reflections have
crystallised into sacred writ. One example of this is the well-known
CatuJ}pratisaralJ-asutra, the Sutra of the Four Refuges or Four Reliances,
as studied in an important paper by Etienne Lamotte.1O Many of you will
be familiar with this short text, which presents guidelines for the inter-
pretation of tradition. It maintains that when assessing teachings which
have been passed down, one should rely on four things: on the dharma
itself rather than the person (pudgaZa) teaching it, on the meaning (artha)
rather than the letter (vyaftjana), on the sutras of explicit or definite mean-
ing (nftiirtha) rather those those which require further interpretation
(neyiirtha), and on direct knowledge (jftiina) rather than discursive
sensory consciousness (vijftiina).l! Some of these terms are far from
10 See Etienne Lamotte, "La critique d'interpretation dans Ie bouddhlsme," in Annuaire
de l'Institut de Philologie et d'Histoire Orientales et Slaves, Vol. 9 (Brussels, 1949), pp. 341-
361. This has been translated into English by Sara Boin-Webb as "The Assessment of
Textual Interpretation in Buddhism," in Donald S. Lopez, ed., Buddhist Hermeneutics
(Honolulu: University of Hawaii Press, 1988), pp. 11-27. The same material is usefully
recapitulated in the more comprehensive survey by Ronald M. Davidson, "An Introduction
to the Standards of Scriptural Authenticity in Indian Buddhism," in Robert E. Buswell, ed.,
Chinese Buddhist Apocrypha (Honolulu, University of Hawaii Press, 1990), pp. 291-325.
11 The core of the sutra text, as cited by Yasomitra in his Sphuttirthti Abhidhar-
makosavytikhyti (ed. Wogihara Unrai, Tokyo, 1932-1936), p. 704, runs as follows:
straightforward, and I have skated over the difficulties with these rendi-
tions of them, but in attempting to assess the applicability of these guide-
lines to the work of academia I will try to pick some of the problems up.
First, however, there is a general principle here worth noting, applicable
to all the pratisaralJas, which is this: in each pair, one term does not
completely cancel out the other, but is accorded priority over it.
For example, in "the dharma and not the person," it is not that persons
are unimportant in teaching, just that they are and should be ultimately
secondary to what is taught. Indeed our greatest successes as teachers
come from inspiring students to want to emulate us, and in more than just
our scholarship, but our primary aim should be to cultivate in them a
relationship not with ourselves, but with knowledge, and a passion for it
so intense that eventually they surpass us in its pursuit. Our supreme
achievement as teachers is to be eclipsed rather than replicated by our stu-
dents, even though our personal example is hardly unimportant to the
process. Yet any standards or values we pass on should lie outside our
own persons.
The second pratisaralJa, the primacy of artha over vyafijana, appears
to be the easiest to assimilate, used as we are to the distinction between
the letter and the spirit, and so may appear to need no further comment.
But we should reflect on the importance of the letter before we rush to
accord priority to the spirit. One of our hardest tasks as academics is to
act as custodians of language, to inculcate in our students a concern for
clarity of formulation and elegance of expression. In an increasingly visual
culture this becomes ever more difficult, and so the Sangha is not alone
in preserving, as it often does, a language which is not the common tongue
of the day. I am not here to lament declining standards of spelling, but
to observe that if the vyafijana is confused and unclear, the artha is likely
to be so as well,u Yet there is no doubt that the meaning, the spirit is pri-
mary, and we should be teaching our students - and attempting ourselves
catviirfmiini pratisaraT}iini I katamiini catviiri I pratisaraT}aJ?l na
I pratisara7JaJ?l na vyaiijanaJ?l1 nftiirthasiltraJ?l pratisaraT}aJ?l na neyiirthaJ?l I
jfziinaJ?l pratisaraT}aJ?l na vijiiiinam I.
12 The point is made in Lamotte, "Textual Interpretation," p. 12.
13 So too Lamotte, "Textual Interpretation," p. 14, in noting that though the letter is to
be subordinated to the spirit, it is still important.
- always to go beyond the letter of whatever we are dealing
with, in order to arrive at what is truly meaningful, or what has purpose.
Artha in Sanslcrit has many senses, and the sense of purpose, or benefit,
is surely in play here (as it is in the treatment of the four reliances in cer-
tain Mahayana texts).14 .
The distinction between sutras of definite or explicit meaning (nitiirtha),
and those whose meaning is implicit, or needs to be drawn out or inter-
preted (neyiirtha) is perhaps the most difficult to apply to the academic
operation, given its specific and technical reference in Buddhist hermeneu-
tics,15 but at the same time we are familiar enough with the need to read
evidence in more than one way and the dangers of an excessive literal-
ism to see how it might be applied to our own work, if only loosely, by
analogy. We all need to know what to take seriously, or au pied de la
lettre, and what not. Thus while the nftiirthalneyiirthp distinction may
appear at first sight to be relevant only to a closed system which main-
tains an orthodox position and therefore needs to come to terms with pro-
nouncements from authoritative sources ostensibly at odds with that posi-
tion, it can at the same time be read more generally as a summons to
employ a certain degree of hermeneutical sophistication when dealing
with so-called authorities. Yet academia supposedly thrives on open-ended
interpretation, questioning and doubt, and ought never to privilege clear,
unambiguous and definitive statements of truth, since it does not see that
there is one truth, or one way of expressing it, to which everything else
has somehow to be made to conform. On this point, it seems, the Sangha
and academia must part company, or at least agree to differ.
Finally, as for the primacy of direct knowledge over discursive sen-
sory consciousness, jiiiina over vijiiiina, one thing it does do is point up
the inadequacy of "consciousness" as a translation for vijiiiina. I take it
14 Indeed, in the restatements of the four pratisaralJas in Mahayana terms found in
the Bodhisattvapifakasutra and the the formulations of the
vyafijanalartha distinction often make no sense unless one understands artha as "purpose"
or "intent."
IS See for example David Seyfort Ruegg, "Allusiveness and obliqueness in Buddhist
texts: Sa![1dha, sal'{ldhi, sal'{ldhyii and abhisal'{ldhi," in Colette Caillat, ed., Dialectes dans
les litteratures indo-aryennes (Paris: College de France, Institut de Civilisation Indienne,
1989), pp. 295-328.
to mean that it is better to have direct knowledge of something, or direct
understanding of it, than merely to know about it. For the academic
study of Buddhism I think this pratisaral}a is of the essence, in that it
reminds us that information about Buddhism, no matter how much we
accumulate of it, is no substitute for understanding, for cognition in its
strongest sense. This raises a number of thorny questions, not least the
contentious insider-outsider issue, which are of general relevance to the
entire academic enterprise. Information, after all, can be accessed these
days so much more easily, and in vast quantities, through the internet,
but what is important is the ability to know what to do with it, which
no amount of surfing the web can ever impart. That ability, the ability
to think critically, weigh evidence, evaluate arguments, exercise judge-
ment and so on can only be acquired through the kind of training which
universities impart, at least when they are doing their job properly.
As valuable as they are, we can pile up editions of manuscripts, trans-
lations of texts, and ethnographic studies until they reach the height of
Mt Sumeru, yet we may still be no closer to understanding Buddhism.
For that to happen, scholars need to leave the campus and enter the
monastery, in one way or another, they need to look real Buddhists in
the eye, otherwise we run the risk of the Buddhology of idealisation, or
the Buddhology of contempt, of admiring or deploring an abstraction of
our own making.
You see how hard it is after all to avoid the temptation to be pre-
scriptive, but in fact I have merely been taking my cue from the
Catu}JpratisaralJasiltra, which is unashamedly so. Its guidelines bear
on the transmission and interpretation of a tradition of knowledge, in
this case of the dharma, but that dharma, as is well known, is both
teaching in the form of text and practical realisation (ddanii and
adhigama). This is equally true of Western academia, at least in its
.ideal form, in which what is supposed to be transmitted is not infor-
mation - that is in a sense merely the carrier for something else -
but the critical spirit of free inquiry and rigorous intellectual honesty,
a species of practice, in short, which needs to be realised in actual
All very well, I hear you thinking, a fine piece of exegetical whimsy,
but what about the real world in which this practice unfolds, with all the
stresses and challenges which I mentioned before. Eyen granted that this
is our purpose, how is its pursuit affected by the environment in which
we are situated? Well, there's no doubt that this environment is having
effects on current trends in our field. I've already mentioned some of the
less fortunate ones, now it's time to tum to the positive side of the ledger.
Consider, if you will, the smallness of our field. To give you an illus-
tration: by way of experiment, I went through the list of 190 participants
posted on the website for this conference and counted 71 people I knew
personally, over one third of the total (this doesn't include the people
I know who haven't come, who would easily push the figure over the
hundred mark). I am sure many of you would arrive at a similar result.
We are comparatively few, and scattered all over the globe, even in the
most unlikely places, like New Zealand, and it is therefore not surprising
that we seek each other out, cultivate relationships with each other, and
maintain them assiduously. In this matter of global linkages we are of
course encouraged by our institutions, each one of which vaunts its inter-
national excellence, excellence being one of the most popular and most
meaningless buzzwords of the modem university. In fact, we may not all
be excellent, but we are all international, and becoming more so. This can
be seen very clearly in the current tendency in our field towards collabo-
rative or group work, with the collaborations or groups in question often
spanning national boundaries. There are many reasons for this. One is the
fact that it is becoming less likely that a single scholar possesses all the
necessary skills and abilities, especially linguistic, required for the kind
of work we do, and this is especially true of the philological side of our
field, where this trend is most noticeable. Here the death of Jan Willem
de Jong in January 2000, shortly after our last conference, marks the pass-
ing of an era in the field, since he was in many ways an exemplary fig-
ure. To the best of my knowledge, he never collaborated with anyone on
any thing. 16 Despite his wide influence, he was sufficient to himself.
The multilingual erudition he commanded, however, is increasingly rare,
not because of any diminution in natural talent, but because the social and
institutional conditions that fostered it and permitted its operation no longer
16 That is to say, in the bibliographies of his writings for the years 1949-1997 published
in the H okke bunka kenkyu not a single co-authored piece is listed.
obtain. So we pool our resources, with two; three, or more scholars doing
the work that previously one might have done alone. We come as yet
nowhere near the'sciences in multi-authored pieces of work, where the list
of contributors sometimes seems as long as the paper itself, but 'fie are
making a modest start. This is not at all a bad thing, since it enables dif-
ferent perspectives to be brought to bear on the material. Indeed, there is
also something very fitting about it, since, as far as Buddhist philology is
concerned, the texts we work on were often generated in this way, as
group projects. It thus seems appropriate, for example, that the Chinese
translations of Buddhist texts, which were produced by teams, should now
be studied, edited and translated by teams. In both cases the teams were
and are international, providing yet another example of how we mimic the
supposed object of our study. We might take a positive view of this trend,
emphasising its undoubted benefits, but at the same time it also reflects
the globalisation of knowledge and its production, in which scholars are
becoming detached from their home bases, able to be deployed anywhere
and everywhere. In a casualised academic workforce, this is not always a
cause for self-congratulation, nor is the fact that we now move about so
much more, as our graduate students often find to their cost.
I expect we will see more such international co-operation, and expect
too that increasingly it will bring the Sangha and academia closer together
in collaborative undertakings. In a way it provides a solution to the prob-
lem of authority in our field, insofar as the agreement by groups of schol-
ars as to what is worthwhile to work on, to devote time and resources to,
helps to determine the directions which our work and our field as a whole
take. It is perhaps not too far-fetched to derive a lesson from the Mahii-
'paddasutra here, that for any tradition to continue, in the absence of a
single personal source of authority, it must ideally be sanctioned by a
"formally constituted community" and should at the same time be con-
sistent with what has gone before. I? Like the Catu!;pratisarw;asutra, this
17 The Mahiipaddasutra, versions of which are found in the Dfghanikaya, the Aizgut-
taranikaya and in various other sources, holds that members of the Sangha should accept
teachings as authentic if they receive them from one of four great authorities (rnahapadda),
namely (1) the Buddha himself, (2) a Sangha of elders, (3) a group of elder monks
specialising in the transmission of Dharma (i.e. Satra), Vinaya or Maq-ka (interpreted either
as proto-Abhidharma lists or as the or (4) a single elder specialising in the
set of prescriptions is designed for a situation where authority is in dis-
pute, and where there may be serious disagreements about what the
dharma actually is. We all know there have been plenty of these in the
history of Buddhism. As the Mahiipaddasutra suggests, t:4e values of
the tradition cannot be dispensed with, no matter who says so: another
way of stressing the primacy of the dharma over the person.
But that of
course is to set up a standard that may well shift over time. Certainly, what
counts as authoritative transmission in Buddhist Studies is now rather
more vigorously contested than it used to be. It was much easier to deter-
mine in former times, when the philological approach was dominant, and
editions and translations of texts could be judged with relative ease as
accurate or flawed, good or bad, on the grounds of a scholar's knowledge.
of the relevant languages and mastery of the canons of textual criticism.
In such circumstances a polymath like de Jongcould set himself up as
gatekeeper, and with his reviews determine who was worthy of admission
and who was not, like a Buddhological equivalent of Cerberus. Nobody
could take his place in that capacity these days, not because of any lack
of erudition, but because there are now simply too many gates to guard.
The dominance of philology is a thing of the past.
Philology itself, however, is certainly not dead, although reports of its
demise regularly corne to our ears. Indeed, there is a continuing need for
it, and it is flourishing quite strongly at the moment, stimulated by the dis-
covery during the last decade of large quantities of Buddhist manuscripts
in Afghanistan and elsewhere. The interest these fmds generate, judging
by the panels at our last conference in Lausanne in 1999 and no doubt at
this one, and the capacity audiences they draw, show that research into
transmission of these texts. But teachings heard from any of these authorities should only
be accepted if they are also in agreement with the Sutra and Vinaya, i.e., with existing
scriptural tradition. For a detailed discussion and references see Etienne Lamotte, "La cri-
tique d'authenticite dans Ie bouddhisme," in India Antiqua (Leiden: E.J. Brill, 1947),
pp. 213-222. An English version, again by Sara Boin-Webb, appears as "The Assessment
of Textual Authenticity in Buddhism," in Buddhist Studies Review, Vol. 1, No.1 (1984),
18 This is accentuated in certain Sanskrit formulations of the text, which add the pro-
viso that received teachings should also be consistent with the way things are (dharmatiirrz
ca na vilomayati). Clearly an understanding of the way things are is determined to a large
extent by existing tradition. .
Buddhist texts has hardly become a marginal activity. I for one certainly
hope that this is not the swansong of Buddhist philology, for even with-
out the newly discovered manuscripts, there is still a great need for this
kind of work. So much literature remains unexplored, and so many edi-
tions and translations made a century or more ago are still being heavily
used, even though they now show clear signs of being seriously deficient
and in need of replacement.
But that said, our field is increasingly diverse. Many of us approach our
work with methods and theoretical tools drawn from anthropology, soci-
ology, feminist studies, cultural studies, literary theory and so on. And we
work in different areas, as can be seen in many of the panels offered at
this and previous conferences: Buddhism in the West, environmental
issues, gender, ethics (particularly as applied to the tougher moral prob-
lems of our age, like euthanasia, abortion, violence and conflict). Such
diversity, in a field as small as ours, is admirable. But, as diverse as our
methods and areas of interest may be, we should ideally continue to be
able to talk to each other, to pool our resources, and make common cause
in a much harsher and more materialistic environment where the study of
Buddhism, however it is defined or pursued, may be questioned as an
unaffordable luxury, and in which it may well become much more diffi-
cult to produce scholarship of quality. It will certainly be harder to do so
if we do not all help each other to work with the needs of the subject and
not the enhancement of our own CVs in mind (dharma, not pudgala), to
make meaningful contributions rather than simply swelling the word-
count (artha, not vyafijana), and to foster useful understanding rather than
merely amplifying the buzz of information (jfiana, not vijfiana). Try as
I might, I am unable to work the nitarthalneyartha distinction in here, but
since the Western academic system thrives on drawing things out, on
meanings which require interpretation, this is hardly surprising.
In many respects, as I've attempted to show, the world of Buddhist
Studies is rather similar to the world it takes as its object of enquiry, and
these similarities are far from superficial or accidental. Both the Sangha
and academia are decentralised institutions engaged in passing down a tra-
dition and a practice of knowledge, and both are no strangers to internal
disagreement about what that knowledge should be (as is indicated in the
Sangha's case by the very existence of the texts I've referred to). Both
the Sangha and academia are also institutions susta41ed by the economic
surpluses of society, in which people are afforded the leisure and the
means to pursue objectives which many outside simply do not understand
or see the point of. Monks and nuns, like academics, have throughout the
history of Buddhism been regularly denounced as parasites, and have just
as enthusiastically been supported by the societies in which they lived.
This support could never be guaranteed, it had to be continually renego-
tiated and carefully cultivated. And yet, despite the unreliable and at times
even hostile nature of its social matrix, the Sangha is now halfway through
its third millennium, and it is still very much alive. It has changed a great
deal during its history, but it remains recognisable. Whether the Western
university, faced as it is with similar challenges to its existence, will last
as long or as well is not so clear. Its demise is predicted by many, but
similar predictions have been made in the past about the decline and
disappearance of the Sangha. Somehow we are always in mapp6,
the pascimakiila or last days of the Dharma, but somehow the institu-
tions survive, and they survive of course precisely by changing.
Our field too is part of that process of change. As we look forward to
the next thousand years, the next century, even the next decade, the oppor-
tunities and challenges are unpredictable, and it would be foolish of me
to attempt a forecast. It is clear enough, however, that we are engaged in
an ongoing relationship and an ongoing conversation, both of which are
centred on something which is itself in perpetual flux, Buddhism. But
lack of identity does not mean lack of continuity, and hopefully we can
continue to negotiate the increasingly closer relationship between acade-
mia and the Sangha, and the increasingly diverse conversation about Bud-
dhism, to our mutual benefit, with a sense of the two long traditions which
stretch back into the past behind us. In doing so it may not be possible,
or even necessary, to honour and respect all our differences and similar-
ities, but it certainly may pay us to know what they are, and even occa-
sionally to rise above them, and look each other in the eye, in mutual
Following many other scholars, it is proposed, in this paper, to consider
some parallelisms or similarities, in beliefs and customs, that can be seen
to exist in Buddhism and Jainism. Naturally, since the XIXth century,
such questions have been investigated more than once
. Nevertheless
attention can be drawn to various interesting details that have come to the
fore in the last decades, but risk being completely ignored in the present
circumstances, when we all are eager to know more concerning the recent
discoveries of Buddhist documents, that have been so remarkably pre-
sented in 1999, in Lausanne, during the XIIth International Conference
of the International Association of Buddhist Studies, and again in
Bangkok, at the XIIIth International Conference.
As far as ancient Buddhism in particular is concerned, Ludwig Alsdorf
has emphasised that "C'est ... Ie bouddhisant pour qui la connaissance du
jainisme et la comparaison des deux doctrines peuvent etre d'une grande
importance ... les memes conditions leur ont donne naissance, elles ont de
1 This is an enlarged version of the presidential address delivered at the 12th Interna-
tional Conference of the International Association of Buddhist Studies at Lausanne in
August 1999.
2 Cf. among others the survey by Ernst Leumann, Buddha und Mahiivfra. Die bei-
den indischen Religionsstifter, Miinchen 1922 (Untersuchungen zur Geschichte des Bud-
dhismus 6). His student Walther Schubring in tum drew attention to "the consonance with
the tipitaka and the anga of the Buddhists" of Mahavlra's teaching, which, in the
. Svetambara canonical texts, is called duviilas'anga ga/;ti-pitjaga, "the basket of the
teacher(s) containing 12 Angas", or, more generally, niggantha-piivayalJa (cf. Pali
piivacana, the technical name of the Buddha's predication according, in particular, to
AggavaIpsa's Pili grammar, cf. Helmer Smith, Saddanfti. La grammaire palie d'Aggavarrzsa
IV.l, Lund 1949, p. 1130 Cf. Schubring" Die Lehre der Jainas. Nach den alten
Quellen dargestellt von.", Berlin und Leipzig 1935 (GIAPhA IIL7) / The Doctrine of the
Jainas. Described after the old Sources by ... , Translated from the revised Gennan edition
by Wolfgang Beuden, Delhi ... 1962) 37 [= Lehre / Doctrine].
Journal of the International Association of Buddhist Studies
Volume 26 Number 1 2003
nombreux traits communs, a ce point qu' on a pris recemment 1'habitude
en Inde d' opposer leur civilisation monacale et asc6tique, que l' on qua-
lifie de a la civilisation "brahmanique"3. As a matter of
fact, the versatile scholar P.S. Jaini, in the Preface to his Collected Papers
on Jaina Studies and Collected Papers on Buddhist Studies explains:
"Seven papers in the Buddhist Studies volume appear under a sectional
heading of Buddhism and Jainism. These are primarily based on Bud-
dhist material but include also a number of Jaina sources. Seven papers
in the volume on Jaina Studies are also relevant to Buddhist studies.
They demonstrate the interdependent nature of these two traditions and
stress the need for exploring them together"4.
Such a comparison is all the more natural as the two spiritual teach-
ers, the Jina Mahavrra and the Buddha Gautama are more or less con-
temporary - a point that has easily been deduced from the sutras of both
their communities, and is regarded as practically certain by scholars
including those who, in recent years, have reexamined "The Dating of the
Historical Buddha"5. Further, the two Masters stem from neighbouring
3 Les etudes jaina. Etat present et taches futures. Conferences par ... , [Paris] College
de France, I?65, p. 3. Alsdorf observed that less attention has been yaid to the Digambara
than to the Svetiirnbara church: this is mainly due to the fact that Svetiirnbara documents
have been more easily available. Hence, in most cases, the present paper also will mostly
refer to the latter (though, thanks to several prominent Digambara scholars' efforts and pub-
lications, their achievements are now better known).
4 The preface is almost identical for the two volumes, Delhi 2000, 2001, p. xiv. Compare
Jacobi's Preface to his translation of the Ayarailga Sutta (p. viif.): "The insertion of a Jaina
text in the publications of the Pilli Text Society will require no justification in the eyes of
European scholars. For them all Jaina documents would have an interest of their own,
even if they did not throw a light on the times, or the moral and intellectual world, in which
Buddha lived. But it is possible that Buddhist subscribers, who aid our labours by their
accession to the Pil.li Text Society, and by the interest they show in it, might take umbrage
at the intrusion, as it were, of an heretical guest into the company of their sacred Suttas.
Yet if they look him attentively in the face, they will fmd there many traces that will inter-
est them strongly, though they may not come to like them. The NigaI,ltha Nataputta was,
it is true, an opponent, if not an enemy, of Gotama the Buddha. Still he was one of his
contemporaries; and in the writings handed down amongst his successors and followers
there are treated many of those questions and topics for which the superior genius of Bud-
dha found the solutions which still form the tenets of the Buddhist in Burma,
Siam, and Ceylon ... "
5 Cf. Heinz Bechert (ed.), The Dating of the historical Buddha / Die Datierung des his-
torischen Buddha. Parts 1-3. Symposien zur Buddhismusforschung IV/I-3 (Abhandlungen
der Akademie der Wissenschaften in Gottingen, Philologisch-Historische Klasse, Dritte
kingdoms and from comparable families, at a time when these
social groups seem to have developed "an independent world-view ...
which was opposed to many brahmanic ideas rooted in ritualistic think-
ing"6. They moreover appear to have boldly vindicated their rights and
status against the brahmanic claims to superiority: both the Buddha and
the Jina are regarded as having embodied the "sramanic" ideals, as illus-
trated in many of their pamphlets, where attacks are repeatedly made
against the Vedic animal sacrifice and the violence it involves, as well as
against the social hierarchy that is upheld in the brahmanical cast-system:
J.q;atriyas and brahmaIfas are contrasted in many Buddhist and Jaina poems
with "the true brahmin" and with "the true sacrifice" which is internal
and purely spirituaF.
1. Did the J aina attitude towards the brahmanic system even harden at
some time? Perhaps this could be deduced from details that, in the Sve-
tambara canonical tradition, surround Mahavfra's prebirth. In an old text,
the Ayiiranga-sutta, it is reported how the future Vardhamana "first took
the form of an embryo in the womb of Devanandi, wife of the BrahmaIfa
Then ... the compassionate god (Indra), reflecting on what was
Folge, Nr. 189, 194,222), Gottingen 1991, 1992, 1997. Cf. the critical review by D. Seyfort
Ruegg, "A new publication on the date and historiography of the Buddha's disease
(nirviil}a): a review article" (BSOAS 62.1, 1999, p. 82-87); see the conclusion: "a time
frame between 420-350 B.C. emerges as most likely" (p. 86); contra Alex Wayman, Indo-
logica Taurinensia 23-24 (1997-98), p. 205-216, who prefers the "long chronology".
6 Cf. Hans-Peter Schmidt, "Ahirp.sa and Rebirth", in Inside the Texts, Beyond the Texts.
New Approaches to the Study o/the Vedas. Edited by Michael Witzel (HOS Opera Minora
2), Cambridge 1997, Department of Sanskrit and Indian Studies, p. 207-234 (p. 219), refer-
ring, in particular, to the comprehensive study by Paul Horsch, Die vedische Giithii- und
'Sloka-Literatur, Vorstufen der indischen Seelenwanderungslehre, Asiatische Studien 25,
Bern 1971. - But H.-P. Schmidt also points to the necessity of taking into account the
multiple aspects of the Vedic culture, and the gradual interiorization of the ritual.
7 Cf., e.g., the 25th and the 12th lessons of the Uttarajjhiiyii, respectively on the "true
sacrifice", and on the "mum" HarikeSa, of svapaka descent: the latter has been compared
with the PaJi Miitaflga-liitaka (cf. Michihiko Yajima, "A Note on Uttarajjhaya 12 and
PaJi Matailga-Jataka", CASS Studies 5, University of Poona, Pune 1980, p. 179-185, ubi
8 In another important canomcal text, it is recorded how MahavIra himself once declared
to his chief disciple that his real mother was the briihrnaI)I Devananda: Deviinandii miihal}/
mama ammagii, aharrz l}a,!l Deviil}andiie miihal}fe attae, Viyiihapannatti IX 33 (ed. JAS I
the established custom (with regard to the birth of TIrthakaras) removed
the embryo from the southern brahmanical part of.:. Kundapura to the
northern part of the same place ... , lodged the fetus in the womb
of Trisala..., wife of the Siddhartha"9. Another siltra, the
lil:wcariya, explains the reason: "the following ... idea" had occurred to
Sakra: "It never has happened, nor does it happen, nor will it happen
that Arhats ... in the past, present or future should be born in low fam-
ilies ... beggars' families ... or brahmanical families. For indeed, Arhats ...
are born in high families, noble families, royal families .. 10". Then he
entrusted the task of removing the embryo to "Ha.riJ;legamesi, the divine
commander ofthe foot troops", who perfectly executed the orderll. This
prebirth episode is unknown to the Jaina Digambara tradition. It is
nevertheless famous, for it is represented on a Jaina relief found in Mathura
and is often depicted in J aina manuscripts 13 , where Ha.riJ;legamesi is shown
on his delicate mission, respectfully bowing to, and transporting Var-
dhamana's embryo. In any case it is significant of the Jainas' old, lasting
and unflinching opposition to the brahmanic hierarchical order.
To a certain extent, this episode has a counterpart, viz. in the
Pali Nidanakathii. While he prepares for his rebirth on earth, and looks for
the suitable country, etc., and family in which to be reborn, it occurs to
the Great Being that it is unsuitable for Arhants, etc., to be reborn in mean
families; but it seems there was no fundamental objection to Buddhas
being reborn in brahmaI).a as well as in kulas. Nevertheless,'
9 Cf. Ayar II 15, Jacobi's translation, SBE XXII, p. 190; Jacobi's ed. p. 121.23-122.13
(JAS 2.1 734 f.): Usabhadattassa miihal)assa ... Deviil)aT[ldiie miihal)fe ... kucchiT[lsi
gabbhaT[l vakkante ... - tao l)aT[l ... al)ukampantel)aT[l devel)aT[l jfyaT[l eyaT[l ti kattu ...
diihil)a-miihal)a-Kul)tjapura-saT[lnivesiio uttara-khattiya-Kul)tjapura-saT[lnivesaT[l Niitiil)aT[l
khattiyiil)aT[l Siddhatthassa khattiyassa Tisaliie khattiyiil)fe ... kucchiT[lsi gabbhaT[l siiharati.
10 Lives ofthelinas, Jacobi's translation, SBE XXII p. 223 ff.; ed. Jacobi 16 ff.: tae
l)aT[l tassa Sakkassa ... ayam eyiiritve ... saT[lkappe samuppajjitthii: Una eyaT[l bhityaT[l, na
eyaT[l bhavvaT[l, na eyaT[l bhavissaT[l jaT[l l)aT[l arahantii ... anta-kulesu vii ... bhikkhiiga-
kulesu vii miihal)a-kulesu vii iiyiiiT[lsU vii iiyiiinti vii iiyiiissanti vii. EvaT[l khalu arahantii ...
ugga-kulesu vii bhoga-kulesu vii ... riiiTJl)a-kulesu vii ... iiyiiiT[lsU vii 3.
11 Ibidem, 22-30.
12 Ascribed to the period, cf. V.P. Shah, Studies in laina Art, Banaras 1955,
p.ll (referring to Biihler, EI II, p. 11ff.).
13 Cf. Iyotindra Jain and Eberhard Fischer, laina Iconography I, Leiden 1978 (Icono-
graphy of Religions XIII.12), p. 4ff., plate N.
following the general consensus of the time, it is the family that
is actually chosen
. Were the Buddhists more conciliatory than the Jainas?
Or did they consider the matter to be irrelevant? Be that as it may, there
is no doubt that, in the suttas, e.g. the Amba{tha-sutta, the superiority of
the khattiyas, that of the Sakka princes in particular, is vividly vindi-
. Thus, by comparing the Buddhist and the Jaina traditions, the
modern reader can get a better glimpse of the ancient disputes, and see
how they were liable to rise and to subside. Can they ever be extinct?
Not long ago, it was observed by a respected Jain scholar that "in Jain-
ism, the Sramm:ta replaces the Brahman in the caste hierarchy, leaving no
truly defined station for the latter. The Jina or his mendicant disciple may
be called miihalJ-a metaphorically, but he is certainly not a Brahman in the
sense of a member of the classical Brabmm:ta varlJ-a" 16.
2. Even comparisons that, at first sight, would seem to be far-fetched
might prove helpful in solving some vexed questions. The Jaina doctrine
is repeatedly said to be very conservative and to have preserved archaic
features - among others the theory of the "colours of the souls", the leSyii
(lessii) doctrine
According to it the souls are supposed to radiate a par-
ticular lustre which, in fact, is indicative of their spiritual level. Follow-
ing its defilement by karman, or, more accurately, by the karmic matter,
the soul (jiva) is black blue (nfla), grey (kiipota), or yellow (pita),
lotus-pink (padma), luminous white (sukla) ... , so that six soul-types are
14 Ja I 49 [so read]. 21-25 (quoted in P.S. Jaini, The Jaina Path of Purification, Berke-
ley, Los Angeles, London, 1979 [= Path], p. 7, n. 9): "Buddha nama vessa-kule va sudda-
kule va na nibbattanti, loka-sammute pana khattiya-kule va brahmal}a-kule va dvfsu yeva
kulesu nibbattanti, idani ca khattiya-kulaT(lloka-sammattaT(l, tattha nibbbattissami" ...
15 D I 87-110.
16 P.S. Jaini, "The Pure and the Auspicious in the Jaina Tradition", in Purity and Aus-
piciousness in Indian Society, ed. John B. Carman and Frederique A. Marglin, Leiden
1985 (International Studies in Sociology and Social Anthropology XLIII), p. 84-93.
Did comparable, albeit different, claims inspire Arnbedkar and those Indians who, in
the course of the XXth century, encouraged conversion to Buddhism?
17 Cf. Schubring Lehre / Doctrine 18; 97; P.S. Jaini, Path, p. 114. On the etymo-
logy and meaning of lessa / leiya, Jacobi, SBE 45, p. 196 n. 2; Kyoshu Tsuchihashi,
"On the literal meaning of lesya", Indologica Taurinensia XI (1983), p. 195-202. - See,
among others, Viyahapannatti I 2 (Deleu p. 76, ubi alia; etc.); Pannaval}a, chapter 17
(JAS 9.1, p. 274-303); Uttarajjhiiya, chapter 34; Tattvarthasutra 2.6, etc.
thus defined. This teaching has been scrutinized more than once, and,
given the fact that, according to Jainism, karman is a material substance,
it has been supposed to reflect "primitive conceptions" 18. On the other
hand it has also been remarked that the Jaina theory is not totally isolated:
"The notion of several soul-types, each with an identifying color. .. may
have been a common belief among various sramaJ]a groups in ancient
times" 19.
As a matter of fact, it appears tohave been accepted by the A.jlvikas,
who, as stated by the Buddha, distinguish six classes of mankind
(abhijati)20. As far as the Jainas are concerned, they, explicitly or implic-
itly, consider these colours to be either spiritual, psychic (bhava-idya)
or material, physical (dravya-Idyafl. The latter are said, in particUlar,
to characterize the three / four main categories of gods. Their colours
are black, further blue and grey, as far as the groups of infernal
deities are concerned, yellow for the luminous divinities of the middle
18 Cf. Schubring, Lehre / Doctrine 18.
19 P.S. Jaini, Path p. 114 n. 26. The colours of the three gUIJas of the Sfu1lkhya natu-
rally come to mind; various other comparisons have been suggested, see Willem B. Bollee,
Studien zum Siiyagmj.a, Die Jainas und die anderen Weltanschauungen vor der Zeitwende,
I, Wiesbaden 1977 (Schriftenreihe des Siidasien-Instituts der Universitat Heidelberg 24),
p. 144 ff., ubi alia.
20 See A.L. Basham, History and Doctrines of the A.jfvfkas. A vanished Indian Religion,
London 1951, p. 139, 243ff., referring, in particular, to A III 383 f. (Sv I 162). He con-
vincingly concludes: "The Ajrvika system of spiritual colours is a general classification
of humanity according to creed or occupation, while that of the Jainas classifies man's psy-
chic development and virtue ... It seems ... probable that the two systems of colour classi-
fication are derived from a common body of ideas which was widespread among ascetic
groups in the days of the Buddha." According to the Buddhists, the Ajrvikas teach a
supremely white group (comprising Nanda Vaccha, Kisa Sarpkicca, Makkhali Gosiila),
a white category (containing the Ajrvikas and AjlvikinIs), a green one (the householder clad
in white robes, the disciple of the acelakas), a red one (nigaIJthas who wear a single gar-
ment), a blue one (bhikkhus who live as thieves, believers in karma), and a black abhijati
(whose members live by violence). The Ajlvika enumeration, which refers to the six con-
stituents of the society, appears to partake of both the Jainas' (supra) and the Buddhists'
(infra) scheme.
21 Cf. PannavaJJa chapter 17.2 (ed. JAS I p. 279 ff.); further, Viyahapannatti XIl5.3;
the notes ad Tattvarthasiitra 4.2, by Sukhlalji and N. Tatia (referring to the Svopajiia-
22 Six distinctive colours are also attributed to the 24 irrtharpkaras: the majority, 14,
are golden, 2 are yellow, the 8th and 9th are white, the 6th and 12th are red, the 19th and
world, yellow, pink, white as far as the. gods of the upper world are
The Buddhists- did not share such a belief in soul-Iesyas, that would evi-
dently have been incompatible with their doctrinal tenets. But they also
made use of colours as identifying marks: this is how, in particular, they
distinguished the components of the social groups, whether divine or human.
In the M ahaparinibbalJa-sutta, the Buddha draws the attention of the monks
to the clothes and ornaments of the troup of the Licchavis, whom he
describes as being formed of four groups, each characterized by one colour,
viz. black, yellow, red, white. He adds that this colourful procession is, on
earth, an image of the Tavatirp.sa gods
. It has been convincingly argued
that this fourfold Buddhist division results from the early adaptation to the
fourfold varlJa system of the Indian society of a prehistoric Indo-European
scheme: India appears to have transformed an older tripartite functional
classification, that can similarly be traced in Rome, where such coloured
symbolism is also seen to be in use
In this connexion, it is noteworthy
that the Pili commentators specify that the gods' colours are purely sym-
bolic, it is "not their natural colour" (na tesal'fl pakati-valJlJa .. . )25. But these
colours serve to distinguish different categories in an organic whole
The above set of Jaina leSyas could thus be seen as a sort of syntheti-
cal representation, referring both to the metaphysical equality and simi-
larity of all the jfvas, and, at the same time, to the various aspects of the
transmigrating jfva, to the complexity of the existent; thus they remind
us, ultimately, of the "two fundamental principles of life" taught by the
TattvarthaSutra: "that of spiritual and physical symbiosis and that of
cause and effect" (through karma)27. To sum up, thanks to the above
21st blue, the 20th and 22nd black, cf. Mrs. Sinclair Stevenson, The Heart of Jainism,
Oxford University Press 1915 (Indian edition 1970), p. 312-314; A. Guerinot, La religion
djafna. Histoire, Doctrine, Culte, Coutumes, Institutions, Paris 1926, p. 100ff.
23 See D II 96.5ff.
24 Thus, to distinguish those who participate in a race, d. Georges Dumezil, La cour-
tisane et les seigneurs colores et autres essais, Paris 1983, p. 17-27.
25 Cf. Sumangalavilasinf, 1096-97, quoted by J.J. Jones, in the notes to his translation
of the Mahiivastu I, London 1949 (PTS, SBB XVI), p. 214 n. 2.
26 For identifying colours in the Epics and Hinduism, see V.M. Bedekar, ABhORI
1968, p. 329-338; W.B. Boilee, ad Suy 2.1, p. 145, ubi alia; T. Goudriaan, Maya divine
and human, Delhi [1978], ch. 4, "Bewildering colours".
27 Cf. N. Tatia's Introduction to his translation of TS, p. xix.
"sramaJ.la" records, the modern reader can get a glimpse of an antique
ideology and of old mental tools and methods

3. Considerable signifiCance being attached, in ancient India, to behav-
iour and discipline, it is not surprising that conduct has been, and remains,
of vital concern in Buddhism
. The subject has naturally led to numer-
ous comparisons between Buddhism and Jainism, and between the latters'
monastic laws and certain prescriptions formulated in the early Brah-
manic literatures, e.g. concerning "non injury": H.-P. Schmidt recalls
how "the renouncer (sannyiisin) or wandering ascetic (pravriijaka,
parivriijaka) is subjected to the strictest rules of ahif!1sii", how "rules
similar to those for the sannyiisin apply to the viinaprastha, the hermit in
the forest"3o. But these prescriptions concern individuals and definite cir-
cumstances, not a whole, well organized, community. On the contrary,
at an early age, the Buddha and the Jina succeeded in bringing their fol-
lowers together and organizing comparative large, long lasting saf!1ghas,
united by clear codes of conduct. So doing, they naturally borrowed var-
ious rules and models accepted in the Indian society, viz. those that had
been set by the Brahmanic ascetics, as demonstrated more than a century
ago: H. Jacobi recalled how "Professor Weber has pointed out the near
relation existing betweeen the five great vows of the Jainas and the five
cardinal sins and virtues of the Buddhists; and Professor Windisch has
compared the Jaina vows (mahavrata) with the ten obligations of the
Buddhists (dasasil)"; on the other hand Jacobi emphasized that "it can
be shown however, that neither the Buddhists nor the Jainas have in this
regard any claim to originality, but that both have only adopted the five
vows of the Brahmanic ascetics (sa.I!IDyasin)"31. Such is the general sit-
uation; nevertheless it is remarkable that the Buddha and the Jina did
28 Compare, e.g., the similar composition of two canonical treatises, following the
increasing number of topics, the Buddhist AnguttaraNikiiya, and, among the Svetfunbaras,
the ThiiIJanga (in fact a common composition device, cf. that of a Saiva manual edited by
Bruno Dagens, Le florilege de fa doctrine ivai"te - Saiwigama-paribhCi:jlimafijarf de Veda-
jfiiina, edition critique, traduction et notes, PondicMry 1979).
29 As testified by several papers presented in the 1999 Lausanne Conference.
30 H.-P. Schmidt, 1. c. p. 210.
31 H. Jacobi, SBE XXII, p. xxii f., ubi alia; quoted in O. von Hiniiber, A Handbookof
Piili Literature, Berlin New York 1996 (Indian Philology and South Asian Studies 2
succeed in their organizing efforts. At the same time, the fact should not
be minimized that an important process of methodical reflexion and redac-
tion took place in both communities, resulting, in particular, in the com-
position of the Buddhist (included in the Pali Vinaya)32, and,
as far as the Jainas are concerned, of the Svetambara Chedasfitras.
The formation of the Theravada Patimokkhasutta has recently been
minutely investigated, and it has been shown how inherited material has
been fundamentally reshaped and formulated anew, so as to result in a
rationally and aesthetically well balanced law code
. In his essay
Das Patimokkhasutta der Theravadin, O. von Hiniiber develops the views
he had already expressed in A Handbook of Pali Literature, and shows
how "the legal structure of the Patimokkha is quite obvious. The rules are
arranged in such a way that the severest offenses are named first and the
lightest. .. are placed at the end. The textual structure, on the other hand,
shows that the Patimokkha must have developed over a certain period
before it was shaped by some redactor(s) to its present form"34.
As far as the Jainas are concerned, they have elaborated a list of ten,
or nine, atonements (payacchittas, prayascittas)35 that include, apart from
[= Handbook] 18, with notes. - As far as MahavIra is concerned, he is regarded as hav-
ing accepted, completed and perfected, the rules set by his predecessor (infra).
32 As is well known, the Pili Patimokkha "is a set of 227 rules for bhikkhus and 311
for bhikkhunfs", K.R. Norman, Piili Literature. Including the Canonical Literature in
Prakrit and Sanskrit of all the Hfnayiina Schools of Buddhism, Wiesbaden 1983 (A His-
tory of Indian Literature. Edited by Jan Gonda VII 2), p. 18. Also see The Piirimokkha.227
Fundamental Rules of a Bhikkhu, with Introduction by Phra Sasana SobhaJ.la (Suva<;l<;lhano).
Translation of the PaIi by Ven. Nal).amoli Thera, Bangkok 2535/1992.
33 Cf., recently, o. von Hintiber, Handbook 15-21; Idem, Das Piitimokkhasutta der
Theraviidin. Seine Gestalt und Entstehungsgeschichte. Studien zur Literatur des Theravada-
Buddhismus IT. Stuttgart 1999 [=Piitimokkhasutta]; Idem, "Nochmals tiber das Patimokkha-
sutta. Anmerkungen zu K. Klaus: "Zur Entstehung des Patimokkhasutta der Theravadin",
WZKS XLV (2001), p. 41-58.
34 Cf. o. von Hintiber, Handbook 18; and the table, in von Hintiber's Piitimokkha-
sutta, p. llf.; also Norman, I.c. p. 18f.: 1. piiriijika ("Defeat", 4 rules), 2. sarrzghlidisesa
("Formal meeting", 13 rules), 3. aniyata ("Undetermined", 2), 4. nissaggiya-piicittiya
("Forfeiture", 30), 5. suddhika-p. ("Expiation", 92 rules), 6. piiridesanfya. ("Confession",
4 rules), 7. sekhiya ("Training", 75 rules), adhikarm;a-samatha ("Legal questions",
7 rules).
35 Cf. Uttarajjhiiya 30.31: payacchittarrz tu dasaviharrz; Uvaviiiya (ed. E. Leumann)
30; cf. also the DigambaraMalacara 5. 164f. (10 payacchittas: payacchittarrz ti ... dasa-
viharrz tu); but TattviirthaSatra 9.2lf. (9 priiyascittas).
(1-2) "confession" and repentance (aloymyi, a/ocana; paejikkamalJ-a,
pratikramalJ-a), etc., such sanctions as (6) ascetic exerCises (tava, tapas)36,
further (7-8) partial or radical suppression of religious seniority (cheya,
cheda; mUla), ultimateiy (9-10) demotion and total exclusion from the
sarp.gha (alJ-avatthappa, anavasthapya; paraficiya, paraficiya). The latter
has naturally been compared with the Buddhist parajika
On the other
hand, the Jainas, besides the Ayaraliga-sutta (the ftrst sfitra of the fust sec-
tion of the Svetfunbara canon) that teaches right conduct, have devoted a
section of their canon to the enumeration of the faults and expiations pos-
sibly incurred by the monks and nuns: the name, Cheyasutta (Chedasutra),
apparently borrows that of the seventh prayascitta (supra). This section
includes seven treatises, traditionally referred to as the Dasa-Kappa-
VavaMra. Thus this ancient threefold dvandva refers (i) to the ancient
"Ten (books)", the last of which, the PajjosavalJ-a-kappa (or Samayari)
collects prescriptions for the right monastic conduct during the rainy sea-
son. The above dvandva further refers to the two important sfitras con-
cerned (ii) with the "Rules"(kappa, Sk. kalpa) prescribed for the lives of
the monks and nuns (niggantha [nirgranthaj, nigganthl; also bhikkhu;
infra), and (iii) with the "Procedures" (vavaMra, Sk. vyavaMra). The fust
twenty sfitras of the latter also feature in the twentieth and last chapter of
the next Cheyasutta, the Nislha-sutta
Viewed as a Cheyasutta, the Nis
seems more or less to aim at a systematic and comprehensive reorgani-
zation and continuation of the Kappa-Vavahara codes
. But, according
36 In the Svetambara tradition, tava, tapas (often interpreted as "fasting"), appears to
have replaced the so-called parihiira: the latter, that is prescribed in the Kappa- and
Vavahiira-suttas, consisted in the temporary isolation from the sal'{lgha (infra). The Digam-
bara list reads: ... tava chedo mUlal'{l pi ya parihiiro c'eva saddahaJpi, Muiiiciira S.16S;
TS 9.22: tapas-cheda-parihiiropasthiipaniini.
37 Already by Sylvain Levi, "Observations sur une langue precanonique du boud-
dhisme", JA 12.2 (1912), p. 49S-S14 (p. S03ff.). - On the piiriijika, cf. infra.
38 On this title, see W. Schubring, Doctrine S1, Vav p. 9 niseha, "prohibition" x
nislhiyii, "place for study").
39 On the composition of the Nis, Doctrine Sl, Drei Chedasutras p. 92. Like K-Vav,
it is concerned with the parihiira. It offers lists of transgressions and sanctions reaching
successively from one to six months, liable to be reduced or not: uddesa 1, 1 month with
no reduction; udd. 2-S, 1 month, liable to be reduced; udd. 6-11,4 months, no reduction;
udd. 12-19, 4 months, liable to be reduced udd.20: up to 6 months. - According to
Schubring, though apparently well balanced, the detail of the Nis is chaotic!
to the Svetfunbara tTadition, the Nis, before being an independant treatise,
had served as the last "appendix" (cilIa) of the Ayarangasutta (supra)40.
To conclude, there can be little doubt that the Cheyasutta section of the
Canon has been submitted to a deliberate, protracted, process of remod-
. But, whereas the Buddhist fInally appears as defmitely
well planned, the Jaina Chedasiltras are seen, so to say, as still in the
process of rearrangement. .
In this matter the Buddhists' approach appears to have been much bolder
than that of most of their contemporaries. Indeed, the example had been set
by the Buddha who, having experienced, and discarded, the ways of the
practitioners of meditation as well as ascetic training, had attained the Bodhi
all by himself, had discovered the "Four Noble Truths" and taught the
"Noble eightfold Path" which avoids the extremes of pleasure and self-
torture. Though more conservative, MahavIra nevertheless can also be
regarded as a successful reformer and organizer: having fIrst accepted the
dharma preached by his predecessor Parsva, that was characterized by four
, he soon replaced it by the "dharma of the five great vows, with
<confession and> repentance included" , or "including meditation "43, thus
insisting on the ethical and spiritual aspect of his message. Assuredly, it is
not to be denied that, in contradistinction to the Buddha's "Middle Path",
the Jina's dharma lays more emphasis on the benefits to be derived from
, but it should be kept in mind that tavo, tapas, in Jainism, is
said to be twofold, both external and internal. The latter includes expiations,
good behaviour, service to others, study, meditation, abandonment (of all
activity, so as to remain in a motionless position and meditate)45.
40 On this restructuring, cf. Jacobi, SBE XXII, p. L; Schubring, Vay p. 8.
41 Schubring also notes how, in the K / Vay, the niggantha- and bhikkhu-suttas tend
to be specialized in different subjects, Vay p. 5ff.
42 All Trrthmpkaras, except the 1st and 24th ones, are said to haye preached the caujjama
dharnrna, Thfu.! 266, 692; Doctrine 16. - For a different interpretation, P.S. Jaini, Path,
p. 17, ubi alia.
43 It is known as the paiica-rnahavvaiya sa-parjikkarnal}a dharnrna, cf. Doctrine 16;
Viy XX 8; JAS ed. p. 877; Deleu p. 256; or sa-bhiival}a dharnrna, Thfu.! 693.
44 Cf. E. Leumann, Buddha und Mahiivfra, passim, opposing their names (p. 17ff.), their
goals and means ("Askese und Samyak", p. 22ff.), etc.
45 Cf. Doctrine 179, ubi alia; see the lists, in Ernst Leumann, Das Aupapatika Sutra,
Leipzig 1883 (AKM 8.2), p. 40ff.: payacchitta, vil}aya, veyavacca, sajjhiiya, jhiil}a,
4. The Jaina Smpgha has always been fourfold, being composed of lay men
and women, and of companies of ascetics, either men or women. The asce-
tics were to be totally "free from all ties", whether external or internal, and
hence were technically caned ni( g)gantha, ni( g)ganthl in Prakrit, nirgrantha,
-I in Sanskrit. They are also named bhikkhu, bhikkhw}1 (cf. Pali bhikkhu,
bhikkhunf). But, unlike the early Theravada Buddhists, the Jainas composed
no special section for their nuns in their canonical law books - though,
naturally, some rules were completed, or added with the bhikkhw}-ls in
v i e ~ . On the other hand, though in religion as in society women are always
dependent, their presence seems to have been readily accepted at least in
the Svetambara Sarp.gha, where they have always been, and still are, by far
in the majority47. The Svetambaras even consider that the 19th Trrtharpkara,
Malli, was a woman
The Digambaras, though, were not as tolerant4

In religion, the general rule for the Buddhists and Jainas is to live as
members of a given group (the gafJa or gaccha of the Jainas). But the
old Svetambara disciplinary texts mention exceptions (whether momen-
tary or definitive) that are sporadically alluded to or discussed in the
46 Concerning the Buddhists, cf. U. Hiisken, "Die Legende von der Einrichtung des
buddhistischen Nonnenordens im Vinaya-Pitaka der Theravadin" (ubi alia), Studien zur
Indologie und Buddhismuskunde, Festgabe des Seminars fur Indologie und Buddhis-
muskunde fur Professor Bechert, hrsg. von R. Griinendahl, J.-U. Hartman, P. Kieffer-Pillz,
Bonn 1993 (Indica et Tibetica 22); Idem, Die Vorschriften for die buddhistische Non-
nengemeinde im Vinaya-Pi{aka der Theraviidin, Berlin 1997 (Monographien zur indischen
Archaologie, Kunst und Philologie 11). - It has been suggested that the Buddha's hesi-
tation concerning the ordination of women was not due to personal reluctance, as he was
broad-minded, but to the desire not to hurt the feelings of his contemporaries, not to go
against the normally accepted behaviour. Nuns in early Buddhism have recently been the
subject of several papers, e.g. by Peter Skilling, also in recent issue of nABS, 24.2 (2001),
an issue precisely on "Buddhist Nuns".
47 Doctrine 20, ubi alia, in particular Viy IX 33, concerning the conversion of
Devananda, who is entrusted to Ajja-CandaI).a. - The Jaina tradition mentions no episode
comparable to MahaprajapatI's request to be ordained as a nun, and the rebuff she and her
companions first had to suffer from the Buddha. Or is it significant that the frrst Jaina
schism is ascribed to Jarnali, the husband of Mahavrra's daughter?
48 With this belief compare the story of "Gautarna's last Female Incarnation", cf. the
two articles by P.S. Jaini, reproduced in his Collected Papers on Buddhist Studies, Delhi
2001, chapters 22, 23. - The above data would tend to show that the position of women
has been a disputed subject, at least in srarnanic circles, cf. Doctrine 16.
49 Cf. P.S. Jaini, Path, p. 39f., on the position of women, one of the "Points of
Controversy between the Two Major Sects".
corresponding commentaries. Exceptionally, some religious are seen to
be. "apart from the flock", or "indifferent towards it"5o. In such cases,
they are mentioned under three headings: the jinakalpikas (Amg. jilJakap-
piya), the parihiira-visuddhikas (Amg. parihiirakappa-tthiya), the yathii-
landa-(pratimii-)kalpikas (Amg. ahiilandiya). Following the observance
called yathiilanda-pratimii (which is particularly obsolete), the niggan-
tha imposes upon himself, among other things, time limits: hls quest in
one particular area must not exceed five days. If submitted to the pari-
hiira-kappa penance, he lives, for a limited time (theoretically from one
to six months) separated from his gaccha: he is gaccha-niggaya,
As for the jilJakappiyas I jinakalpikas, they conform to
Mahiivrra's standards, as recorded in the accounts of his last years 52.
According to this "rule" (kalpa), ascetics go about naked, .have no bag-
gage, observe severe penances, and, in particular they stay apart from the
galJa and are constantly alone. This description reminds us of the
Pratyeka-Buddhas, mentioned in Jainism as well as in Buddhism
. It also
reminds us of the ascetic behaviour known thanks to the famous Buddhist
poem that extolls the khaqga-vi!fiilJa-kalpa. Transmitted in several Bud-
dhist traditions, whether in Pali
, Buddhist Hybrid Sanskrit or Gandhari
Prakrit, it "espouses the virtues of solitude"55. The meaning of the com-
pound has been abundantly discussed, and the refrain in which it is used
(pali: eko care khaggavisiilJa-kappo) has been variously translated:
"one should wander alone like the rhinoceros" or "one should wander
50 Ardhamagadhi gaccha-niggaya, niravekkha, cf. Caillat, Expiations, p. 52ff.
[cf. n. 62].
51 Cf. Expiations, p. 52f.; p. 171ff.
52 Especially as summarized in Ayiiranga 1, Uvahiir;a-suyaT[!, "The Pillow of Right-
eousness" (Jacobi), "Die Uberlieferung vom Fasten" (Schubring).
53 Cf. the four Pratyekabuddhas in the Pilii Jiitaka III 381.16*f., and in the Jaina Uttara-
jjhiiyii 18.46 (cf. Jacobi, Ausgewiihlte Erziihlungen in Miihiirii-rtrf, 1886, p. 34): Karakar;cji1
Kalingesu Paftciilesu ya Dummuho / Namf riiyii Videhesu Gandhiiresu ya Naggal.
54 Sn 35-75.
55 Solitude is also praised in many other passages, e.g. in the Thera- and Therf-giithiis,
cf. Th 6, 31, 41, etc.; 49: na me taT[! phandati cittaT[!, ekatta-nirataT[! hi me, "Amidst the ...
cries of the birds, this mind of mine does not waver, for devotion to solitude is mine" (trans-
lation K.R. Norman). - Also compare the "arar;ya-dwellers" (examined by Sasaki Shizuda
at the XIIIth Bangkok IABS Conference, December 2002, quoting a number of previous
alone like the rhinoceros horn" .. 56. Though the Pali commentaries under-
stand -kappa as meaning csadisa, "like", an explanation that seems to
have been often accepted, it would seem preferable, considering the above
Jaina testimony, to retain the full meaning of the substantive kappa,
kalpa, "usage, practice" (ifc.: "following the regulations or rule")57,
thus, for "following the habits of the rhinoceros"58.
Such an animal comparison is not surprising in India, and would not be
exceptional in a Buddhist context
: the fifth stanza of the same "Rhi-
noceros Sutra" compares "an understanding man" with "a deer which
is not tied up" and "goes wherever it wishes in the forest for pasture"60;
and the Dhammapada recommends, "if one does not find a zealous com-
panion ... , one should wander alone like a matailga naga elephant in the
forest,,61. As for the Jainas, the commentaries of some Chedasutras liken
56 For the references to, and summaries of the numerous discussions on the meaning
of the compound, see K.R. Norman, "Solitary as Rhinoceros Hom", Buddhist Studies
Review 13.2 (1996), p. 133-142; Richard Salomon, A Gandhiirf Version of the Rhinoc-
eros Sutra. British Library Fragment 5B, Seattle and London 2000. Concerning
"the Meaning of Khagga-visa'.la I p. lOff., he decides "not entirely with-
out doubts, to understand the primary sense of the refrain of the verses of the G1indhiirI
text, eko care as 'one should wander alone like the rhinoceros', with
the proviso that the other possible sense, 'one should wander alone like the rhinoceros
hom,' is by no means ruled out and in fact may have been understood to be equally and
simultaneously valid" (p. 14). - For the association of the Rhinoceros Siitra with the
Pratyeka-Buddhas, or "solitary enlightened ones", Idem, ibidem, p. 8, ubi alia.
57 Cf. M. Cone, A Dictionary of Pali, S.v., 2.(i) (m.) a rule, ... a practice; CPD III, s.v.
4kappa, m., usage, praCtice; also 7kappa, mfn. following the regulations or rules (of a reli-
gious community).
58 Salomon (p. 11) refers to Norman's comparison of the Pali simile with a prose pas-
sage of the J aina Kalpa-sutra (Jacobi's "Jinacaritra" edition 118): khaggi-visa'.laJ?1 [ sic]
va ega-jae (i.e. [Jacobi's translation], "single and alone like the hom of a rhinoceros"),
"where the neuter form" -visa'.laJ?1 "proves that it means 'rhinoceros hom' and not 'rhi-
noceros'''. But, in the Jinacaritra, the prose passage is followed by a summary in the arya
metre (even pada): vihage khagge ya bharuJ?14e ("a bird, a rhinoceros, and BharuI).Qa",
Jacobi's translation) which would tend to show that the comparison is with the animal (or
59 Cf. the siJ?1ha-nada, or lion's roar of the Buddha, etc. Also see Jeannine Auboyer,
Le trone et son symbolisme, quoting Jean Przyluski, on "Le symbolisme du pilier de Sar-
nath", p. 488. - C. Rhys-Davids, "Similes in the Nikayas", JPTS p. 52-151.
60 Sn 39 (K.R. Norman's translation): migo arafzfzamhi yathii abaddho I yen' icchakarrz
gacchati gocaraya, I vifziiu naro... .
61 Dhp 329 (K.R.Norman's translation): no ce labetha nipakaJ?1 sahayaJ?1 I ... eko.care
matang' araiifze va nago = Ja III 488.20ff.
the young b h i k ~ u to the m[ga, the more senior monk to the vuabha, the
master to the sif!lha: these comparisons are conspicuous in the ritual of

5. Confession plays an essential role in Jainism as well as in Buddhism

According to the old Jaina disciplinary books, it leads the transgressor from
the avowal to the expiation of the fault. The process includes: (I) the dec-
laration of the fault, (2) the repentance, (3) the gUilt which he feels in his
own conscience, (4) his self-reproach in the presence of the guru, (5) the
repudiation of the sin, (6) the total purification, (7) the finn purpose of
amendment, (8) the performance of the appropriate atonement
ing to a later text, one has to remove all "darts" (sal/a, satya) or unconfessed
faults in order to acquire superior knowledge and supreme perfection

Similarly, in several passages of the Vinaya concerning lay or religious
trangressors, it is underlined that confession will result in spiritual
progress: "in the discipline of the noble, this is growth: whoever having
seen a transgression as a transgression, confesses it according to the rule,
he attains restraint in the future", vut/,ghi h' esa ... ariyassa vinaye yo
accayaf!l accayato disva yatha-dhammaf!l patikaroti ayatif!l saf!lvaraf!l
The appropriate behaviour of the CUlprit is detailed e.g. in
the development concerning Pacittiya VI: the lay follower, "saluting the
62 For references, cf. S.B. Deo, History of Jaina Monachism, Poona 1956 (Deccan Col-
lege Dissertation Series 17), p. 226; Colette Caillat, Les Expiations dans Ie rituel ancien
des religieux jaina, Paris 1965 (Publications de l'Institut de Civilisation Indienne 25)
[= Expiations. Revised English edition: Atonements in the Ancient Ritual of the Jaina
Monks, Ahmedabad 1975 (L.D. Series 49)], p. 31, 47, 15U., ubi alia.
63 For I. Duncan M. Derrett's views on "Confession in Early Buddhism", cf. Baud-
dhavidyasudhiikara/:t, Studies in Honour of Heinz Bechert on the Occasion of His 65th
Birthday, edited by P. Kieffer-Ptilz und I.-U. Hartmann, Swisstal-Odendorf 1997 (Indica
et Tibetica 30), p. 55-62. Reference could also be made to the recent book by Kuo Li-ying,
Confession et contrition dans Ie bouddhisme chinois du Ve au Xe siecie, Paris 1994 (EFEO,
Monographies n 170).
64 Kappasutta 4.25; Vavahiirasutta 1.35: bhikkha ya annayara7[l akicca-tthiilJa7[l sevitta
icchejja aloettae, ... aloejja paejikkamejja nindejja garahejja viu{{ejja visohejja akaralJayae
abbhutthejja ahii'riha7[l tavo-kam'ma7[l payacchitta'rz paejivajjejja. Cf. Caillat, Expiations,
65 Mahiinisfhasutta 1.16.
66 Yin I 315.18ff. = II 126.18ff. = 192. 20ff. (Translation I.B. Homer).
feet of the venerable Anuruddha with her head, spoke thus to the vener-
able Anuruddha: 'Honoured sir, a transgression has overcome me, in that
I acted thus, foolish, misguided, wrong that I was. Honoured sir, let the
master, acknowledge fot me the transgression as a transgression for the
sake of restraint in the future,,67. Such assertions can be compared with
the conclusive sentence of the text introducing the "recitation of the
Rule", according to Venerable Nfu:1amoli' s text and translation of the
Patimokkha: "false speech in full awareness has been pronounced by the
Exalted One to be a thing obstructive (to progress); therefore any actual
(undeclared) fault should be declared by a bhikkhu who remembers to
have committed it and who looks for purification. To have declared it is
for his good", sampajiina-musiiviido kho . .. antariiyiko dhammo vutto bha-
gavatii. Tasmii saramiinena bhikkhunii iipannena visuddhapekkhena santi
iipatti iivikatabbii, avikatii hi'ssa phiisu hoti
In this conclusive phrase,
the adjective phiisu, a Middle Indo-Aryan derivative related to Pali phas-
seti (Sk. sparSayati), "to cause to touch, bring into contact", "to touch",
retains its full meaning: it indicates the transformation of the fault which,
thanks to the avowal, has become exactitude, truth, hence reaches, and
leads to (the spiritual goal)69.
Such an asseveration is best understood in the light of the historical and
prehistorical confession doctrine. The latter has been reexamined recently
by Calvert Watkins 70, who refers to Indo-European data, several Vedic
passages (and Sylvain Levi's remarks on the subject71): "by the verbal
act ... of confession the sin itself becomes exactitude, reality, truth: Vedic
67 Vin IV 18. 32ff. (translation I.E. Homer): ... ayasmato Anuruddhassa padesu sir-
asa nipatitva ayasmantaf!1. Anuruddhaf!1. etad avoca: accayo maf!1. bhante accagama yatka
balaf!1. yatka mulhaf!1. yatka akusalaf!1. ydhal?! evam akasif!1..
68 The Pafimokkha. 227 Fundamental Rules of a Bhikkhu ... Translation of the Pali by
Ven. Nfu:1amoli Thera, Bangkok 2535/1992, p. 66f. cf. Vin 1103.8-11.
69 Cf. Journal Asiatique 1960, p. 41-55; K.R. Norman, Journal of the Oriental Insti-
tute, Baroda XI (1962), p. 32-34.
70 Calvert Watkins, "On Confession in Slavic and Indo-European", in Calvert Watkins,
Selected Writings, ed. Lisi Oliver, Innsbruck (lnnsbriicker Beitrage zur Sprachwissenschaft,
Bd. 80), 1994, IT p. 602-621 (first published in Studies in Honor of Horace G. Lunt, ed.
by E. Scatton et al., Folia S1avica 2.1-3, Columbus (Ohio) 1978, p. 340-359).
71 Sylvain Levi, La doctrine du sacrifice dans les BrdhmaiJas, Paris 1898 (2eme edition
1966), p. 158, quoted by C. Watkins: "L'aveu retablit les faits; il ne rep are pas moralement
la faute, ilIa fait dispara'itre, en effet, puisque l' acte et la parole sont des lors conformes".
satyam", "the acknowledged existence of the transgression reestablishes
exactitude, reality, truth"n.
Further, Watkins emphasizes that confession is intimately bound up
with the 'Act of Truth' (Sanskrit * satyakriya) 73. Examples of the latter in
Buddhist literature have been recently examined in a study that shows the
vitality of this belief in Buddhism, in any case if one is to judge by the
many occurrences quoted especially (though not exclusively) from Pali
. The formulated Truth, conjuring up, as it does, an essential char-
acter of the performer, has an infallible issue. Similarly, an essential lie
will entail the direst consequences. Both situations (first the negative,
then the positive one) are enacted in the 12th lecture of the Svetambara
UttaraJJhiiya. The hero is Harikesa Bala,
"(1) born in a family of svapakas; he became a monk and a sage, ... who
had subdued his senses.
(3) Once on his begging tour, he approached the enclosure of a Brah-
manical sacrifice ...
(7) 'Who are you, you monster? .. go, get away ... '
(8) At this turn, (a) Y a k ~ a . . . spoke the following words:
(9) 'I am a chaste sramaI).a ... 1 have no property ... and do not cook my
food. 1 have come for food ...
(10) 1 subsist by begging; let the ascetic get what is left of the rest ... '
(11) - 'We shall not give you such food and drink ...
(16) This food and drink should rather rot, than we should give it you,
(18) Are here ... no teachers with their disciples, who will beat him ... and
drive him off?'
(19) On these words of the teachers, many ... rushed forward, and they
all beat the sage with sticks, canes, and whips.
72 Cf. C. Watkins, ibid., p. 613, 617; p. 616 notes "the efficacy of the act of confes-
sion in ancient India".
73 Ibidem, p. 614.
74 Toru YAGI, "Once again on the Forms of Oath in Classical India (Ill): in Connec-
tion with saccakiriya-", Bulletin of the Cultural and Natural Sciences in Osaka Gakuin
University, Nos. 43-44, Osaka, December 2001, p. 47-90 (ubi alia); p. 59ff.: "(II. Three
types of the Act of Truth)"; p. 60, "the asseveration of truth". Also see Michael Witzel,
"The case of the shattered head", in Festschrift Wilhelm Rau, Studien zur lndologie und
Iranistik 13/14 (1987), p. 363-415, ubi alia (see p. 383 n. 39; 41Of.).
(20) At that turn king Kausalika's daughter, Bhaqra, ... appeased the
angry youngsters.
(21) 'He is the very man to whom the king ... had given me, but who ...
has refused me.
(22) He is that austere ascetic, of noble nature, who subdues his senses
and controls himself.'
(25) Appearing in the air ... the Asuras beat the people. When Bhadra saw
them with rent bodies, spitting blood, she spoke again thus:
(26) 'You may as well dig rocks with your nails ... as treat contemptu-
ously a bhikkhu ...
(28) Prostrate yourself before him for protection ... if you want to save
your life and your property ... ' "75
It will have been observed that the situation brought about by the brah-
rnins who made false statements concerning Harikesa is reversed thanks
to the intervention of a witness, who, moreover, is intitled to make a
*satyakriyii. By uttering a superior (! metaphysical, ultimate, eternal)
truth, she contributes to the restoration of the right order of the society
and of the world (she restores [ta).
Similarly, according to the Buddhists, because Devadatta pretends to
be, or tries to be considered as, the supreme sage, superior to the Bud-
dha, he signs his death sentence, that will be executed some way or other:
blood spurts from the apertures of his face, etc
The Buddha himself,
when he dismisses some brahrnaI).as' claim to a status superior to the
75 Uttarajjhaya 12, Jacobi's translation. For philological remarks and corrections
(inserted infra), see L. Alsdorl, IIJ 6 (1962), p. 128-133 (=Kleine Schriften, ed. A. Wez1er,
Wiesbaden 1974, Glasenapp-Stiftung 10, p. 243-248):
Hariesabalo nama asi bhikkhuji'indiyo (1) II 'samalJo aharrz sarrzjao bambhayarzvirao ...
I annassa atthti iha-m-agao mi (9) II ... sesavasesarrzlabhau tavassf' (10) II 'na u vayarrz
erisam anna-palJarrz I dtihtimu tujjharrz (11) II ke ettha ... I eyarrz khu dalJq.elJa phalelJa
hanta ... khalejja jo lJarrz?' (18) II ... tattha bahU kumara I dalJq.ehi vittehi kasehi c' eva
samagaya tarrz isi talayanti (19) II ranno tahirrz Kosaliyassa dhuya Bhadda ttL. I ... kud-
dhe kumare parinivvaei (20) II 'dinna mu ranna ... /. .. jelJ' amhi vanta isilJa sa eso' (21)
II te ... (A)sura tahirrz tarrz jalJii talayanti I te bhinna-dehe ruhirarrz vamante pasittu Bhaddii
ilJa-m-ahu bhujjo (25) /1 'girirrz nahehirrz khalJaha ... I ... je bhikkhul?l avamannaha (26) II
szselJa eyarrz saralJarrz uveha ... I ja! icchaha jfviya . . .' (28) II
76 On the numerous accounts of Devadatta's crimes and fate, cf., e.g., Ma1alasekera,
Dictionary of Pali Proper Names I, s.v. (p. 1107ff.); A. Bareau, Recherches sur la
Tathagata's, indirectly explains why Devadatta met such a gruesome fate:
"the brahmin of Verafija spoke thus to the lord:
'I have heard, good Gotama, that the recluse Gotama does not greet brah-
mins who are worn, old, stricken in years ... ; nor does he stand up or ask
them to sit down. Likewise, ... that the revered Gotama does not greet brah-
mins who are worn ... ; nor does he greet them or stand up or ask them to
sit down. Now this, good Gotama, this is not respectful.' .
'Brahmin, I do not see him in the world of devas including the Maras,
including die Brahmas, including recluses and brahmins ... , whom I should
greet or rise up for or to whom I should offer a seat. For, brahmin, whom
a tathagata should greet or rise up for or offer a seat to, his head would split
asunder.' "77
In all the above examples, the transgression does not concern just some
individua1(s), but endangers the whole social group. Hence it is funda-
mentally heinous, and has to be dealt with appropriately, viz. by the com-
plete annihilation of the danger.
Bearing this general conceptual context in mind, it might be worth-
while to reconsider once more the pdrdjika rules as taught in thePra-
They have lent themselves to repeated comparisons both with
prescriptions detailed in the Buddhist Vinaya itse1f78 and with prescrip-
tions valid among the Brahmanic and Jaina ascetics (supra). The technical
biographie du Buddha dans les Sutrapitaka et les Vinayapitaka anciens ... III, Paris 1995
(pEFEO 178), p. [239 ff.] = BEFEO 78 (1991), p. 105ff. (his death, p. [246] = p. 112).
77 Vin III 1.22-2.13: atha kho Veranjo briihmal}o yena bhagavii ten' upasarrzkami,
upasarrzkamitvii bhagavatii saddhirrz sammodi ... Eka-m-antarrz nisinno kho Veranjo
briihmal}o bhagavantarrz etad avoca: "sutarrz m' etarrz, bho Gotama, na samal}o Gotamo
briihmal}e Fl}l}e vurjrjhe mahallake ... abhiviideti vii paccuttheti vii iisanena vii nimantetfti.
Ta-y-idarrz bho G., tatth' eva, na hi bhavarrz Gotamo briihmal}e jil}l}e v. m .... abhiviideti ...
nimanteti. Ta-y-idarrz bho Gotama na sampannam evii ti. - Nahan tarrz, briihmal}a, pas-
siimi sadevake lake samiirake sabrahmake ... , yam aharrz abhivadeyyarrz vii paccuttheyyarrz
vii iisanena vii nimanteyyarrz. Yarrz hi, bnihmal}a, Tathiigato abhiviideyya vii paccutfheyya
vii iisanena vii nimanteyya, muddhii pi tassa vipateyyii ti" . Translation, I.B. Homer, Book
of the Discipline I p. 2f. (q.v. for concordances, and similar assertions in Theravada liter-
ature). "The Shattered Head Split ... ", a well-known Vedic motive, has many later paral-
lels, cf. Michael Witzel, l.c., p.381ff., 5, for references to "early Buddhist texts"; also to
Mahiivastu (ed. Senart 3, p. 114.12: nastica so satvo vii satva-kiiyo vii yasya Tathiigate
pratyupqsthihante na saptadhii murdhnarrz na spahaleyii). Also see Stanley losIer, in Bul-
letin d'Etudes Indiennes 7/8 (1989-1990), p. 97-139.
78 Cf. O. v. Hiniiber, Patimokkhasutta, p. 24ff. Also cf. the 4 akaral}lyas, p. 41ff.
term piiriijika has been translated as "Defeat" by LB. ,Horner, a transla-
tion that has been widely accepted. It has generally been admitted that the
Buddhist pariijikas have been rearranged on the model of the Buddhist
sllas, the moral "habits" or precepts, of which the counterparts are also
prescribed for the Brahmanic and Jaina ascetics, though in a different
order. Among others, the Svetambara Dasaveyiiliya-sutta prescribes the
abstention (1) from injuring any [living] being, (2) from false speech, (3)
from taking that which is not given, (4) from sexual acts
. The fact that,
in the Buddhist list, false speech is not the second but the fourth item is
evidently intriguing, all the more as telling a conscious lie again recurs
as the first of the piicittiya transgressions. But the latter is comparatively
trivial, whereas the object of the fourth parajika is fundamentally differ-
The exceptional nature of the 4th piiriijika did not escape I.E.
Horner8l. She remarks: "The first three Parajika rules are levelled against
the breach of a code of morality generally recognized among all civilised
communities: against unchastity, against the taking of what was not given,
and against the depriving of life ... The curious fourth Parajika, concerned
with the offence of 'claiming a state of further-men' (uttarimanussa-
dhamma), seems to have been fashioned in some different mould, and to
belong to some contrasting realm of values. It is by no means a mere
condemnation of boasting or lying in general, for it is the particular nature
of the boast or the lie which makes the offence one of the gravest that a
monk can commit..
". As a matter of fact, it is exactly comparable to
Devadatta's attempts to supplant the Master, to control, and ultimately
79 Cf. Dasaveyaliya-sutta, ed. Ernst Leumann, p. 615: paIJtiivayao veramalJaIJI ...
musavayao veramaIJalJl... adinn' adaIJao veramaIJalJl.. mehuIJao veramaIJalJl
(cf. Schubring's translation).
80 Cf. O. v. Hiniiber, Patimokkhasutta, p. 45 (ubi alia).
81 Yin III 90.32**-91.2**: yo pana bhikkhu anabhijanalJl uttarimanussa-dhammalJl
attupanayikalJl alamariya-ftanadassanalJl samudacareyya iti janami iti pa,ssamfti, tato
aparena samayena samanuggahiyamano va ... evalJl vadeyya: ajanam evalJlavuso avacalJl
janami, apassalJl passami, tucchalJl musa vilapin ti, ayalJl pi parajiko hoti asalJlvaso ti
82 BD I, p. xx-xxv. She adds: " ... the boast of having reached some stage in spiritual
development, only attainable after a long training in the fixed and stable resolve to become
more perfect, and to make the potential in hinl assume actuality". A complementary inter-
pretation is proposed here. O. v. Hiniiber's suggestion that the parajilcas could have been
arranged following a decreasing order of gravity does not seem convincing.
destroy the Sa:qlgha and the Doctrine. Hence the fourth piiriijika natu-
rally entails the religious death of the transgressor.
Could the exceptional character of the transgression explain why
"falsely claiming a state of further men" occupies the fourth rank in the
parajika list, whereas avoiding false speech is mentioned as the second
vow of the Brahmanic and Jaina ascetic
? But perhaps there is more to
it. For there seems to be some affinity between speech, language, and the
number "4". O. v. Hiniiber points to the fourfold expansion musiiviida +
pisu[lii viicii pharusii viicii samphappaliipa
, merging in the cattiiro
vohiirii, "the noble usage, noble mode", mentioned in the Sal(lgltiSutta
Further, in the Jaina AyiiraftgaSutta the rules concerning speech (bhiisa-
jiiya) are dealt with in the fourth lecture of the second section. It is stated
that "For the avoidance of these occasions to sin, a mendicant should
know that there are four kinds of speech: the first is truth; the second is
untruth; the third is truth mixed with untruth; what is neither truth nor
untruth, nor truth mixed with untruth, that is the fourth kind of speech:
neither truth nor untruth. Thus I say"86. The prominence of the number
"four" in developments concerning speech is striking and reminds us of
the four pada-jiitiini in Patafijali's M a h i i b h i i ~ y a Paspasii, commenting
upon the catvdri paddni in which vdc is measured according to RS 1.164:
in the world-view of the Vedic poet, only one quarter of speech is used
in every day language, whereas the other three quarters, dealing with eso-
teric, secret Truth(s), remain hidden to ordinary men. Given this general
83 Cf. Charles Malamoud, in Cuire Ie monde, Paris 1989, p. 137-161 (in the chapter
"Semantique et rhetorique dans la hierarchie hindoue des 'buts de l'homme'''), the con-
siderations on "Quatre egale trois plus un" [4=3+1], where it is observed that in a four-
fold scheme, the quarters are not equal: "Ie quatrieme element complete, ou bien englobe,
ou bien encore transcende les trois premiers" (p. 142).
84 D I 4.13-29. - Compare Manu 4.138: satya,!! bruyat priya,!! bruyant na bruyat
satyam apriya,!! / priya,!! ca nan[ta,!! bruyat, "Let him say what is true, let him say what
is pleasing, let him utter no disagreeable truth, and let him utter no agreeable falsehood"
(Biihler's translation, SBE 25). Also "Prohibited speech and subhasita in the Theravada
Tradition", Indologica Taurinensia XI1 (1984) p. 61-73.
85 Dill 232. 7f. - Cf. Patimokkhasutta p. 27f.; IT XI1 (1984) p. 67f.
86 Ayar II 4.1.4: bhikkhu jal}ejja cattari bhasa-jayai,!!, ta,!! jaha: sac cam ega,!!
paq,hama,!! bhasa-jaya,!!, bfya'!! mosa,!!, taiya'!! sacca-mosa,!!, ja,!! n' eva sacca,!! n' eva
mosa,!! n'eva sacca-mosa,!!, a-sacca-mosa,!! ta,!! cauttha'!! bhasa-jayal]l, se bemi; trans-
lation H. Jacobi, SBE 22 p. 150 (n. 2: "The fIrst, second and third cases refer to asser-
tions, the fourth (asatyam[iia) to injunctions").
context, it can be surmized that, by pushing the offence 9f falsely "claiming
a state of further men" to the 4th rank of the piiriijika series, the P iitimokkha
warns that such a pretence should not be considered inconsequential: atten-
tion is drawn to the pregnant potency of such utterances
, to the fact that such
deceptive speech and imposture in fact endanger the Community, will set it
into chaos, hence finally entail the destruction of the Sarp.gha.
The Jainas apparently do not enter into such considerations: they are
more matter of fact, as can be seen in the Ayiiranga (2.4, supra), or the
DasaveyiiUya (chapter 7). The latter states that the monk "should not say
that he will explain all, really all: a thoughtful [monk] should in all cases
make a precise [and] complete report". The chapter concludes: "[He
who] speaks after consideration, controls his senses well, has overthrown
the four passions, [and] is without [worldly] support, purges [his soul] of
the dirt resulting from previous evil deeds [and] is sanctified in this world
and the next. Thus I say"89.
When faced with the same or similar problems, the Buddhists and the
Jainas produced more or less comparable or divergent answers, as these
had to fit into different systems. It is manifest that both Buddhism and
Jainism have preserved a considerable amount of antique beliefs, cus-
toms, phrases ... 90 On the other hand it is no less evident that they have
87 It is therefore proper to distinguish this heinous offence from the false, abusive or
slanderous speech for which pacittiyas 1-3 are prescribed. Compare the distinction made
between killing a human (manussa) and another living being (pal}a), respectively sanctioned
by parajika 3, and by pacittiya 61 (cf. O. v. Hinuber, Ptitimokkhasutta, p. 40).
88 It will also be remembered that, in the Brahmanic tradition, correct speech has more
than once been considered to be of religious value, cf. L. Renou, Histoire de la langue san-
skrite, Lyon 1956, p. 6: "L'idee de la grammaire comme instrument de purification est
presente dans Ie plus ancien commentaire grammatical, la Paspasa du M a h a b h ~ y a , comme
a travers toute la Mimfupsa" (quoted IT XII, p. 71 n. 53, ubi alia). - For South-East Asia,
see F. Bizot I F. Lagirarde, La purete par les mots, Paris ... 1996, EFEO (Textes boud-
dhiques du Laos).
89 Dasav 7.44,57 (Schubring's translation): 'savval!l eyal!l vaissami, savvam eyal!l' ti
no vae I al}uvfi savval!l savvattha eval!l bhasejja pannaval!l II 44 II parikkha-bhasf
susamahi'indie cauk-kasayiivagae al}issie I sa niddhul}e dhutta-malal!l pure-karJal!l, ara-
hae logam il}al!l taha paral!lll 57 ti bemi.
90 Concerrring khaejga-viijal}a-kalpa (supra, 4 and n. 56), Prof. K.R. Norman points
out that the word kharJga is ambiguous, and may mean both "sword (hom)" or "rhinoc-
eros". In the fil}acariya passage, therefore, it may signify that "the hom is solitary" or
"the rhinoceros is solitary" [personal letter, January 2003].
transformed the old legacy, forged new conceptual frames and schemes,
invented original rules, procedures and structures, that aimed at promot-
ing the spiritual' as well as the material welfare of the group as a whole
and of each of its members individually.
BUDDHIST TEXTS (editions and abbreviations as in A Critical Fiili
NidiinaKathii, cf. liitaka I p. 1-94; The Story of Gotama Buddha ... Trans-
lated by N.A. Jayawickrama. Oxford 1990 (PTS).
Piitimokkhasutta: The Piitimokkha. 227 Rules of a Bhikkhu ... Translation
of the Pali by Ven. Nanamoli Thera, Bangkok 2535/1992.
Aciiriiftga / Ayiiraftga: The Ayararrzga Sutta of the C;vetambara lains.
Edited by Hermann Jacobi, London 1882 (Pali Text Society). - Trans-
lated from Prakrit by Hermann Jacobi, Oxford University Press 1884
(reprint Delhi, etc., 1964, laina Sutras I (Sacred Books of the East 22).
Ayiiradasiio: see Chedasutra.
Chedas(ttra / Cheyasutta: Drei Chedasutras des laina-Kanons,
Ayiiradasiio, Vavahiira, Nisiha, bearbeitet von Walther Schubring. Mit
einem Beitrag von Colette Caillat, Hamburg 1966 (ANISH 11).
Dasaveyiiliya: Dasavaikiilika-sutra und -niryukti, nach dem Erziihlungs-
gehalt untersucht und hrsg. von Ernst Leumann, ZDMG 46 (1892), p. 581-
663. The Dasaveyiiliya Sutta ... translated with Introduction and Notes, by
Walther Schubring, Ahmedabad 1932.
Dasii-Kappa-Vavahiira: see Chedasutra.
JAS: Jaina .Agama Series, Bombay 1968 +.
linacarit(r)a 11iIJ-acaria: The Kalpasutra of Bhadrablihu ed .... by
Hermann Jacobi, Leipzig 1879 (AKM 7.1). - Translated from Priikrit by
Hermann Jacobi, "Lives of the finas", SBE 22, 1884 (repr. Delhi 1964).
Kappasutta: Das Kalpa-satra. Die alte Sammlung jinistischer
Monchsvorschriften. Einleitung, Text, Anmerkungen, Ubersetzung ...
von Walther Schubring, Leipzig 1905 (Indica 2).
Mahiinisihasutta: Studien zum Mahiinisiha. Kapitell-5 von Jozef Deleu
und Walther Schubring, Hamburg 1963 (ANISH 10).
Malacara: Eine Digambara-Dogmatik. Das fonfte Kapitel von Vattakeras
Malacara hrsg., iibersetzt und kommentiert von Kiyoaki Okuda, Wies-
baden 1975 (ANISH 15).
Nislha-sutta: See Chedasatra. - Ed. Walther Schubring, see Vavahiira-
PannavaIJ-a: PaIJ-IJ-avaIJ-asutta1!l, Ed. Muni PUIJ.yavijaya, Dalsukh Malvat}ia,
Amritlal Mohanlal Bhojak, Bombay 1969, 1971,2 vol. (JAS 9).
Tattvartha Satra, That which is, UmasvatilUmasvaml. Translated with
an introduction by Nathmal Tatia, San Francisco, London, Pymble 1994
(Institute of Jainology).
Uttarajjhayii: The Uttaradhyayanasutra. Edited by Jarl Charpentier,
2 vol., K!I>benhavn, 1921-22 (Archives d'Etudes Orientales 18).
- Translated from Priikrit by Hermann Jacobi, Oxford 1895 (SBE 45),
(repr. Delhi 1964).
Uvaviiiyasutta: Das Aupapatika Sutra, erstes Upanga del' faina. 1. Ein-
leitung, Text und Glossar. Von Ernst Leumann, Leipzig 1883 (AKM 8.2).
Repr. Nendeln 1966.
Vavaharasutta: See Chedasatra. - Ed.: Vavahiira- und Nisiha-sutta.
Hrsg. von Walther Schubring, Leipzig 1918 (AKM 15.1). - Vavahiira,
in Drei Chedasatras des faina-Kanons (supra).
Viyiihapannatti: ViyahapaIJ-IJ-attisutta1!l: Part I, Ed. Bechardas J. Doshi,
Bombay 1974, Part II, III, Ed. Bechardas J. Doshi, assisted by Arnritlal
Mohanlal Bhojak, Bombay 1978, 1982 (JAS 4).
Further: Iozef Deleu, Viyiihapannatti (Bhagavai). Th.e fifth Anga of the
Jaina Canon. mtroduction, Critical Analysis, Commentary & Indexes,
Brugge 1970 (Rijksuniversiteit te Gent (Werken uitgegeven door de Fac-
ulteit van de Letteren en Wijsbegeerte, 151. Aflevering).
Preamble: Ritual Meanings
There must be readers who are shocked, angry, or depressed at the thought
that ritual (not to mention religion and even language) is not only complex
but also meaningless. I am not a bit sad about it. I prefer a thing, like a per-
son, to be itself, and not refer to something or somebody else. For all we
know life itself may be meaningless.
Frits Staal
In 1979 Frits Staal, a Sanskritist who specializes in Vedic ritual, pub-
lished an article in which he proclaimed ritual to be devoid of meaning.
Staal's argument, subsequently developed in a number of publications
is at flrst glance deceptively simple: when we ask about the meaning of
a ritual we seek an explanation in language. Such an explanation will
always involve a conceptual reduction, in that we seek to transpose the
lived complexity of a ritual performance to a verbal formulation. Ritual,
according to Staal, resists such reduction by its very nature. Ritual is
"pure activity" (Staal 1979a: 9); it is a "discipline engaged in for its
own sake, which cannot therefore be thus reduced .... Basically, the irre-
ducibility of ritual shows that action constitutes a category in its own
right" (Staal 1983: 1.16).
A draft of this paper was presented at the symposium "Matrices and Weavings: Expres-
sions of Shingon Esoteric Buddhism in Japanese Culture and Society," held at the Uni-
versity of Hawai'i, August 31 to September 2,2002. I would like to thank the participants
of that conference, especially Torn Eijo Dreitlein of Koyasan University. for their com-
, ments and advice. I am also indebted to Poul Andersen, Phyllis Granoff, Charlie Orzech,
; and Elizabeth Horton Sharf for their comments and suggestions.
1 See, for example, Staal 1979a; 1979b; 1983: 2.127-134; and 1990.
Journal of the International Association of Buddhist Studies
Volume 26 Number 1 2003
Staal's claim that ritual is meaningless is on the corollary
claim that ritual is antecedent to language. A variety of non-human species
display ritual behavior, and ritual may well have been commonplace
among Homo sapiens long before the advent of language and culture.
(Staal argues that language is actually an outgrowth of ritual in general
and ritualized vocalizations - the precursors of mantras - in particu-
lar.) Besides, individuals often acquire competence in a rite before they
learn what, if anything, the rite signifies. Scholars are then wrong to
assume that there are symbolic meanings running through the minds of
ritualists and that such meanings constitute the sine qua non of ritual per-
formance. According to Staal, people do rituals simply because they have
been taught to do so, often from an early age.
If ritual is meaningless - if it does not refer to a domain of meanings
extrinsic to ritual action itself - then popular theories such as "rituals
enact myths," "rituals reflect social structures," or "rituals inculcate val-
ues and norms" are misguided as they confound the historical and logical
relationship between ritual and meaning. Besides, says Staal, those who
hold that rituals enact myths, encode social structures, or impart collec-
tive norms, fail to explain why anyone would want to use ritual for these
tasks when words would serve just as well if not better (Staal 1979a: 7;
1990: 123).
Ritual, according to Staal, is behavior - acts and sounds - that is gov-
erned by rules. The rules constitute a "syntax" allowing the creation of
infinitely malleable recursive structures not unlike those of language.
But unlike language, ritual has no semantics; the acts and sounds that
constitute ritual interact without reference to meaning (Staal 1990: 433).
Ritual is then not so much like language as it is like dance, about which
Isadora Duncan famously proclaimed: "If I could tell you what it meant,
there would be no point in dancing it" (Staal 1979b: 120).
Needless to say, Staal does not deny that individuals do ascribe mean-
ings to rituals. His point is that such meanings are secondary or super-
fluous and hence tell us little about the transcultural (not to mention cross-
species) phenomenon of ritual per se. Following the earlier observations
of Arnold van Gennep, Staal notes that rituals may be transmitted through
time with little or no change, despite changes in the meanings ascribed
to them. If a rite remains the same irrespective of shifts in meaning, then
meaning cannot be intrinsic to the rite. "In the development of our con-
cepts and theories of ritual it is only a small step from 'changing mean-
ing' to: 'no intrillsic meaning' and 'structural meaning,' and from there
to: 'no meaning'" (Staal 1979a: 11). Moreover, if the goal of ritual were
the conveyance of meaning, then ritual would admit change in so far as
this or any other goal was well served. Thus rituals lack not only mean-
ing but also a purpose or goal.
One reason that the absence of visible or otherwise detectable results causes
[the ritualist] no concern is that large rituals are ends in themselves ....
The rites have no practical utility and have lost their original function, if ever
they had one. The ritualists perform them not in order to obtain certain ends,
but because it is their task. Lack of practical utility, incidentally, is a char-
acteristic that ritual shares with many of the higher forms of human civi-
lization.1t may be a mark of civilization. (1983: 1.18)
Pushing this argument to its logical conclusion, Staal closed his origi-
nal1979 article with a passage, used as an epigraph above, suggesting that
just as ritual is meaningless, so too is religion, language, and even life itself
(1979a: 22). In doing so, Staal unwittingly revealed his hand: he had stip-
ulated the conditions for ascribing "meaning" such that they can never be
met. Staal will only admit meanings that are both invariant and intrinsic
to the phenomena under investigation. But this is to ignore the insight, fun-
damental to linguistics and semiotic theory, that meaning does not reside
within a sign. Rather, meaning emerges from the complex cultural system,
determined in part through social interactions, that marks a particular phe-
nomenon as a "sign" in the first place. A signifier is meaningful only as
a point in a set of relations. And since meaning never resides in the "thing
itself," meaning must always be extrinsic, contingent, and variant

In claiming that the thing-in-itself has no meaning, Staal has uncovered
hot the meaninglessness of the thing itself but rather the semiotic logic
that renders meaning possible in the ftrst place
. Rituals trade in signs that
" , 2 Some might argue that there is one case in which we can talk of "fixed meaning,"
namely, as a defming feature of the class of proper nouns. Yet Derrida, for one, questions
even this restricted sense of "fIxed meaning" (Derrida 1985) .
. 3 Or, one might say that Staal has simply reaffirmed the Wittgensteinian insight that
l:ile abstract "thing-in-itself" is a piece of philosophical nonsense. For a critique of Staal
similar to my own, see Andersen 2001: 162-163.
don't possess meaning so much as they invite meaning. To speak of the
meaning of a rite one must adopt a particular perspective - situate one-
self in a particular world of discourse - and different perspectives yield
different meanings. As anthropologists have noted since the time of Tylor,
even participants in one and the same rite will hold various and often
conflicting interpretations of the event, and the interpretations will change
over time
. Moreover the "emic" accounts of ritual participants will dif-
fer dramatically from the plethora of "etic" readings offered by histori-
ans of religion, sociologists, anthropologists, or psychologists. But here
ritual is surely no different from any other cultural product, including
works of art and literatures.
Some of what is conveyed in a particular ritual performance may indeed
be difficult if not impossible to convey in words. Even then it may be mis-
leading to label these elements "meaningless." When Isadora Duncan
says, "If I could tell you what it meant, there would be no point in danc-
ing it," she does not mean that dance is meaningless but rather that she
is incapable of putting into words that which she puts into dance. Staal
has every right to stipulate that meaning only be predicated to properly
formed linguistic utterances - to restrict meanings to language. But such
a stipulation renders the rest of his argument tautological. Few would
quibble with the claim that ritual is constituted not by language so much
as by action.
This criticism aside, Staal does make several important points. Ritual
activity qua activity is indeed difficult to translate into words. Moreover,
4 "The old and greatest difficulty in investigating the general subject [of idolatry 1 is
this, that an image may be, even to two votaries kneeling side by side before it, two utterly
different things" (Tylor 1920: 2.168-169).
5 Staal's argument was, I suspect, inordinately influenced by the archaic nature of the
rituals he was studying. The Agnicayana is an ancient rite consisting largely of mantric
utterances in Vedic, the meaning of which is inaccessible to most of the participants.
The archaic character of the rite, and the fact that so much of the liturgical content is gib-
berish to the actors, may account for its seeming invariance across generations. But even
then Staal likely overestimates resistance to change. It is precisely because the rite has
only been practiced in fits and starts over the last hundred years that the Brahmins abide
so closely to textual authority. (What other authority can they call upon, now that the
authority of received tradition has been compromised?) Even then, many significant alter-
ations were made in the performance Staal observed, including the use of plant offerings
instead of goats.
rites have lives of their own, independent of the symbolic and mytho-
logical associations that may be ascribed to them. Finally, appreciation
of the symbolic and mythological world of ritual does not in and of itself
account for the obsessive, rule-bound character of ritual action. Adepts
may spend years acquiring competence in elaborate and physically ardu-
ous rites the historical origins and symbolic associations of which remain
obscure to them. To castigate such adepts for their "ignorance" would
only betray our own.
Shingon Ritual
That ritual is resistant to conceptual reduction and discursive appropria-
tion has posed a particular problem for modem Shingon exegetes. Shingon
apologists, like their counterparts in other religious traditions, have felt com-
pelled to respond to modem rationalistic and scientific critiques of religion
in general and ritual in particular. This has led some writers to ignore or
downplay elements of the tradition considered "unscientific" or "magical"
in favor of Shingon teachings deemed properly philosophical, psychologi-
cal, spiritual, or aesthetic. But this has not been easy, given that sacerdotal
ritual lies at the heart of the Shingon tradition. Ritual performance was essen-
tial, of course, to virtually all schools of Buddhism throughout Japanese
history, but other schools have had an easier time reinventing themselves in
the light of late nineteenth- and early twentieth-century modernist mores.
Apologists for Zen, for example, insisted that "true" Zen eschews ritual
altogether in favor of unmediated spiritual experience, while Pure Land
exegetes recast their tradition in theological terms strikingly similar to Protes-
tantism: Pure Land, we are told, is a doctrine of divine grace predicated on
faith in an all-compassionate being6. Even Tendai and Nichiren partisans have
gotten into the act: sectarian introductions to these traditions invariably fore-
ground doctrine and cosmology at the expense of ritual practice.
Some Shingon exegetes tried adopting similar strategies. They pro-
duced books on Shingon that simply ignore ritual practice altogether7, or
6 For an alternative view of medieval Japanese Pure Land see Dobbins 2001.
7 Minoru Kiyota's book Shingon Buddhism: Theory and Practice (1978), for example,
,despite the title, is all theory and no practice. The English study of Kukai' s works by
that depict Esoteric (mikkyo ritual as a means toward inculcating
inner transformation and mystical experience
. the intellectual
genealogy of categories such as "mystical experience," such claims are
always a bit suspect (Sharf 1995, 1998), but in the case of Shingon they
are especially so.
The sticking point is not the absence of a sophisticated body of doc-
trine or theology through which to frame Shingon ritual. On the contrary,
Shingon doctrine is conceptually rich and partakes of a certain literary and
aesthetic elegance that may well appeal to modem sensibilities. One might
cite, for example, the notion that the phenomenal world is the theophany
of the dharmakiiya-buddha Mahavairocana (Dainichi *- 8), or the related
doctrine that the enactment of the "three mysteries" (sanmitsu -=1$, the
ritual performance of the body, speech, and mind of the deity) gives
tangible form to the practitioner's primordial identity with the divine
Such tenets resonate, at least on the surface, with popular Western con-
ceptions of mysticism and the "perennial philosophy."
The problem for Shingon modernists is not doctrine. Rather, it is that
doctrine is patently secondary to a complex set of ritual procedures that
constitute the core of the monastic curriculum. The early popularity and
rapid growth of the Shingon lineage in Heian Japan was due to its
possession of the exalted eighth-century "Tantric" rites that Kukai
(774-835) brought back from the Tang capital. These rituals constituted
a world unto themselves, and while their connections to "normative"
Buddhist teachings were not always salient, they carried the imprimatur
of celebrated Indian Buddhist masters. Most important, those with access
to this ritual technology were promised the power to defeat their ene-
mies, end droughts and famines, cure disease, and attain exalted states on
the Buddhist path.
(Indeed, this is one reason why so many recent attempts to define
"Tantra" have failed. To date, virtually all attempts begin by identifying
the conceptual foundations - the soteriology, cosmology, metaphysics,
Yoshito Hakeda (1972), while learned and important, focuses exclusively on doctrine while
remaining mute on the subject of rituaL
8 Matsunaga 1989: 25-27, 1990: 27. See also Toganoo 1982b: 23; Ishida 1987: 27; .
Yamasaki 1988: 123; and the discussion in Sharf 200lb: 193-195.
9 On the sanmitsu see Hizoki KDZ 2.40.
or what have you - supposedly common to Tantric traditions across
Asia. Even in so"called polythetic definitions of Tantra, the extended set
of salient characteristics is comprised entirely of symbols and concepts
- what Staal would classify as "meanings" - rather than of ritual imple-
ments, gestures, sounds, and procedures. Yet, if it makes sense to talk
about a pan-Asian phenomenon of Tantra at all - and this is a big "if"
- then I believe it is better approached not in terms of thought ["mean-
ings"] but of practice ["actions"]. If the term Tantra has any cross-cul-
tural referent, it is to a body of technological expertise comprised of cer-
tain powerful tools - mantras, mudriis, icons, altars, esoteric implements
including ceremonial weapons, and so on - and the arcane procedural
knowledge necessary to wield them. This technology could be, and appar-
ently was, appropriated by diverse religious traditions and transmitted
independent of any theoretical or doctrinal overlayyo
The fact that Shingon apologists may experience difficulty in recasting their
ritual practices in an acceptably modem or rational light need not concern
scholars who stand outside the tradition. There is no shortage of theoretical
models and conceptual strategies on which scholars might draw. They could
adopt a comparative approach, for example, noting structural similarities
between Shingon and non-Shingon traditions. Think of the striking parallels
between Shingon ritual and the traditions of shamanism, spirit mediumship,
and possession that are so widespread throughout Asia. In each case an ini-
tiated master engages in an occult performance through which he or she
comes to personify or embody a divine being. The performance endows the
shaman or ritual master with the deity's power and authority by virtue of
which the perforiner is able to intervene in worldly and otherworldly affairs
Comparativists might step back even further and view Shingon ritual
under the rubric of "sacrifice" a la Henri Hubert and Marcel Mauss
10 On the nature and status of Tantric Buddhism in China see Sharf 2002: 263-278.
II It is unlikely that the structural parallels observed between East Asian shamanism
and Buddhist Tantra are entirely accidental. They may well be the product of shared ances-
try or cultural diffusion and borrowing. Edward Davis, for example, shows how non-
Buddhist ritual masters (jashi I*ffrJi) in Song China employed Tantric techniques (including
mantra and mudra) to invoke their "guardian spirits" who were then used in rites of exor-
cism (Davis 2001: 49). On the connections between Buddhist Tantra and East Asian spirit
possession see also Strickmann 2002: 198-218.
([1898] 1981). In Shingon, as in all sacrificial traditiops, particular goods
are purified through consecration or aspersion rites and then offered to
powerful supernatural beings in exchange for some boon
. Or, follow-
ing van Gennep, scholars might foreground the initiatory, ascetic, or trans-
formative dimensions of Shingon ritual, placing Shingon under the broad
rubric of rites du passage. Shingon rites thus create a liminal situation in
which the officiant is transformed, at least temporarily, from one social
or sacerdotal state to another.
Such broad theoretical models, under the rubrics of shamanism, sacri-
fice, liminality, or what have you, mitigate some of the arcane "otherness"
of Shingon ritual by framing it as an instance of a larger transcultural
and transhistorical human phenomenon. But this conceptual gain comes
at a price: these theories impose a set of foreign categories and concerns
that obscure as much as they reveal. Moreover, as grand narratives they
tend to reduce the distinctiveness, complexity, and internal coherence of
the particular tradition at hand.
In this paper I will focus, instead, on an expository narrative that orig-
inates not from without but from within the Shingon tradition. This is the
so-called guest-host paradigm, according to which all major Shingon rit-
ual practices (Sk: s,ridhana) are structured as feasts or entertainments for
visiting deities, wherein the practitioner assumes the role of host (shujin
3:.A), and the main deity takes the position of honored guest (hinkyaku
daihin This narrative, familiar to all Shingon priests, is of
considerable antiquity and is believed to bespeak the ancient Indian prove-
nance of the rites. The guest-host paradigm is used in both traditional
commentaries and contemporary sectarian tracts to explain individual pro-
cedures and to relate them to the overall structure of the rite, lending nar-
rative coherence to the whole
12 Precisely because such a notion of sacrifice is so broad, some would limit the term
sacrifice to situations involving the slaughter of a sacrificial victim; on the concept of sac-
rifice see esp. the discussion in Heesterman 1993: 7-44. For a brief discussion of Shingon
ritual in the context of sacrifice see Payne 1991: 88.
13 On the guest-host structure (daihin geisha no keishiki see Takai 1953:
109-110, 117; Toganoo 1982b: 45-46; Yamasaki 1988: 162; Strickmann 1989: 16-17;
and Payne 1991: 88-89. The guest-host structure is also widely used in Tibetan exegesis
of Buddhist Tantra, but that topic lies outside the confines of the present paper.
While this paradigm does account for the general structure of Shingon
practices, it also engenders certain ambiguities and contradictions.
These ambiguities, I will suggest, may shed light on features of the Shin-
gon ritual system that relate to the early development of Buddhist Tantra.
Thinking through the narrative will also allow us to revisit Staal's thesis
concerning the relationship between ritual actions and meanings. Before
turning to this narrative, however, it is necessary to say a few words about
the Shingon monastic curriculum.
Shingon Training
Anyone wishing to become a Shingon priest today must undergo a
sequence of four initiations known collectively as the Shidokegy6
or "four preliminary practices of liberation." Each of the practices is cen-
tered around the invocation of a particular buddha, bodhisattva, or other
divine being (known as the honzon :<fs:# or "principal deity")14 and his
or her retinue. The rite proper takes from two to five hours to complete
and is repeated three times a day in the context of an extended ascetic
. In addition to the central rite, the priest undergoing Shidokegy6
training performs a variety of auxiliary practices, including daily visits to
neighboring shrines and temples, ancestral rites for lineage patriarcj:ls,
. offerings to hungry ghosts, and so on, leaving little time for meals or rest.
If done in a traditional manner, the Shidokegy6 sequence requires over
one hundred days to complete, whereupon the practitioner is eligible for
consecration (kanjo Sk: as a Shingon "master" (ajari
Sk: acarya). This consecration authorizes the priest to perform
Esoteric rituals on behalf of others.
All Shingon rituals and ceremonies are organized as a sequence of
smaller liturgical procedures that typically consist of an incantation
14 The term honzon (C: benzun) is likely derived from Tantric sources, but it lost its
explicitly Tantric overtones rather early and came to be used by all sects of Japanese Bud-
dhism. On this term see esp. the Benzun sanmei *-l':::11!< chapter of the Mahiivairocana-
siitra (T.848: 18.44a-b); Hizoki, KDZ 2.30; Mikky6 Jiten Hensankai eds. 1983: 2068b-
c; Mochizuki 1933-36: 5A697b-4698a; and Goepper 1979.
15 The three daily performances, each of which is called a "single sitting" (ichiza gyob6
,1!1H9$), are known respectively as "early night" (shoya fJJ1Jt), "late night" (goya 1f1Jt), and
"mid-day" (nitchil 8'1').
(a mantra, dhtiralJ-l, hymn, etc.) accompanied by a hand gesture (mudrii)
and a guided contemplation (kanso iI.;W). The four Initiations that com-
prise the Shidokegyo - namely the JUhachido + (eighteen meth-
ods), Kongokai practice), Taizokai n&iiiJf.. (matrix-
realm practice), and Goma (fire ceremony)16 - consist of hundreds
of such segments of varying duration and complexity. Of these hundreds,
modern commentators regard three specific segments that usually appear
in each of the Shidokegyo rites as the soteriological core of the practice.
These three units - "interpenetration of self [and deity]" (nyuga-ga'nyu
A:fJt:fJtA, #51), "formal invocation" (shonenju #53), and "sylla-
ble-wheel contemplation" (jirinkan fifBiI., #55) - unite respectively the
body, speech, and mind of the practitioner with the body, speech, and
mind of the principal deity (honzon) of the rite. As such, they constitute
the ritual instantiation of the "three mysteries" (sanmitsu), a cardinal
Shingon doctrine that affirms the identity of practitioner and buddha!7.
Traditional Shingon ritual manuals, known as shidai (sequential
programs) or giki {i!\lJL (ritual regulations), often list only the names ofthe
dozens of procedures that comprise the rite. With less common procedures
the manuals may include mnemonic aids such as the pronunciation of the
mantras (in Siddham script, Chinese characters, and/or the Katakana syl-
labary), the text of liturgical hymns and recitations, directions on how to
form certain mudriis, and diagrams to help in the contemplations. In any
case, the manuals presume a vast store of ritual knowledge on the part of
the practitioner. The more elaborate rites such as the Taizokai and Goma
consist of hundreds of such procedures, many of them of considerable
complexity .
Traditionally, these manuals were not printed but were hand-copied
and transmitted from master to disciple. Thousands of such manuals sur-
vive in temple archives, and hundreds have made their way into modern
printings of Buddhist and Shingon canonical collections. A comparison
of the manuals quickly reveals a host of small but notable alterations in
16 Prior to undertaking the Iiihachido, the practitioner must complete the Raihai kegyo
(preliminary prostration practice), which today takes from one to four weeks to com-
plete. As such, the modem Shidokegyo sequence is actually organized into five segments.
I7 See, for example, the discussion in Toganoo 1982b: 66.
the liturgies: elements are added or removed, recitations and contempla-
tions are modified, and some manuals include short expository directions
and comments - interlinear notes that threaten to enter the liturgy proper
- that may represent the "oral transmission" (kuden OIJ) of a particu-
lar master or lineage. The Shingon ritual tradition, now many centuries
old, was conservative but - pace Staal - by no means invariant.
These differences gave rise toa profusion of independent lines (ryu 1m)
that differed in the details of their ritual performances and exegeses
There was no opprobrium associated with amending the rites; all bona fide
Shingon masters (ajari) were sanctioned to alter the rites as they saw fit.
Two related reasons are given for this authority: (1) masters were regarded
as spiritually advanced and ritually sanctified beings whose interpretations
of the rites reflected their inner wisdom; (2) more practically, there was no
single authoritative Chinese or Japanese textual source for the rites on which
the Japanese could draw. There was, rather, a profusion of sanctioned texts
and teaching lineages, a situation readily acknowledged by the tradition
Annen (841 ?-915?), for example, notes that the reason there were so
many differences in the ritual transmissions brought back to Japan by Ennin
I1!H= (794-864) was that he studied under eight different teachers20.
At the same time, scholars should not exaggerate the differences
between Shingon initiatory lineages. While these lines did compete for
prestige and patronage, in the end the variations in ritual performance
are relatively minor and rarely affect the rites' underlying structure
18 According to tradition, there are twelve major Shingon initiatory lineages, six asso-
ciated with the Ono line 1j\lJ!fmE and six associated with the Hirosawa line However,
there are dozens of sub-lineages, the two most important today being the Chuin-ryii. '1'l!it:mE
now dominant on Mount Koya and the Sanboin-ryii. '=:'!ifll!tmE stemming from Daigoji lIJIMl9'
in Kyoto. For a full discussion of the complex relationship between the various lineages
see Toganoo 1982a: 239-266; 1982b: 33-40; and Takai 1953: 25-58.
19 The absence of a single authoritative ritual text and the freedom of an iiciirya to
interpret and alter the ritual as he pleases is discussed in KakuchO's l\!iti1 (960-1034) San-
mitsu shOryoken .=:It:FJ>'iSf1lii (T.2399: 75.633c14 ff.) and Taizokai shOki JI/liM'I-;ti!9 (T.2404:
75.806cl ff.); see also Todaro 1986: 114.
20 Taizokai daihO taijuki T.2390: 75.54a22; Todaro 1986: 114.
21 The same can be said for the differences between Shingon mikkyo writ large (Tomitsu
*It:) and Tendai mikkyo (Taimitsu tl'lt:). The sequence in which Tendai priests perform the
Shidokegyo is slightly different (in Tendai the Taiz6kai rite precedes the Kongokai), but
the overall narrative structure and most of the individual procedures are identical.
This structure is rooted in a subset of eighteen ritual procedures known
as the juhachido or "eighteen methods" which constitute a latticework
around which are hung dozens if not hundreds of additional elements.
Jiihachido is also the name of the first of the four Shidokegyo practices,
and it is through this extended rite that a Shingon priest comes to acquire
a basic understanding of the ritual system
(In this paper I use lower case
italics ["juhachido"] to refer to the original sequence of eighteen proce-
dures, and capitalized unitalicized script ["Jiihachido"] to refer to the full
Shidokegyo rite.) The guest-host narrative is captured in the root proce-
dures of the eighteen methods.
History and Structure of the Eighteen Methods
The origin of the eighteen methods is far from clear. The DaishOten
kangi soshin binayaka hO a work preserved in
fascicle nine of Kfikai's Sanjujo sakushi contains what some
Shingon scholars believe to be the earliest record of the rite23. But the de
facto locus classicus is another roughly contemporaneous text, the ]uhachi
geiin (C. Shiba qiyin). Tradition holds that this text, extant only
in Japan, is the work of Kfikai' s teacher Huiguo ;!;:W; (746-805) as
recorded by Kfikai himself, but little is known with certainty about the
provenance of the work
The same is true of virtually all of the ritual
manuals attributed to Kfikai, including his three other eighteen-methods
22 There are a variety of rites structured around the eighteen-methods sequence that can
be used for the Shidokegy6 Jilhachid6 perfonnance. These ritual fonns, such as the Nyoi-
rinb6 are referred to as "ritual sequences established on the eighteen methods"
(juhachido date no shidai see Toganoo 1982b: 47-49. The Nyoirinb6 serves
as the Jiihachid6 in the Sanb6in-ryil, using a manual derived from the Shonyoirin Kanjizai
bosatsu nenju shidai by Geng6 5C* (914-995); see the appendix to
this paper.
23 The Sanjujo sakushi is traditionally considered a work by Kiikai dating to his years
in China (Toganoo 1982b: 44; Ono 1932-36: 4.86).
24 In his ]uhachido kuketsu (1225-1304) writes that Kilkai received
the eighteen methods from Huiguo (T.2529: 79.71c9ff). Most modem Shingon scholars
accept this position and view the ]uhachi geiin as Kilkai's record of Huiguo's instructions
(Takai 1953: 116-117). The ]uhachi geiin is reproduced in both the Taish6 canon (T.900),
where it is attributed to Huiguo, and the KobO Daishi zenshU (KDZ 2.634-645);
where it is attributed to Kilkai.
manuals: the ]uhachid6 nenju shidai ]uhachid6 kubi
shidai + and Bonji juhachido /\J!!. While they are all
likely early Shingon compositions, they may well postdate Kiilcai's death.
The sequence of eighteen root procedures appears to be a Japanese sys-
tematization of a ritual pattern found in group of related Chinese manuals
associated with Amoghavajra. The texts most commonly mentioned as
sources for the eighteen methods are the Wuliangshou rulai guanxing
gongyang yigui (T.930), Guanzizai pusa ruyilun
yuqie ill (T.1086), Guanzizai pusa ruyilun niansong
yigui (T.1085), and Dabao guangbo louge
shanzhu mimi tuoluoni jing *_ (T.1005a). In his
Sangakuroku Kukai mentions the first two as the basis for his own
mhachido manuals (Takai 1953: 111-116), but all these texts were familiar
to KUkai and all share a common ritual structure. I will have occasion to return
to these texts below. Here I will only note that there is some question as to
whether or not the Chinese texts, and the ]uhachi geiin itself, contain the seg-
ment known as the "contemplation of the sanctuary" (dojokan )i:l$}ill, #31),
an important rite traditionally included among the list of eighteen

As mentioned above, the telTIl "juhachido" refers to (1) a skeletal struc-
ture of eighteen procedures that was incorporated into more complex rites
such as the Taizokai, Kongokai, and Goma; and (2) a full-fledged rite in
itself that, in its mature fOlTIl, consists of some seventy or eighty discrete
procedures. The latter JUhachido rite, typically with Nyoirin Kannon
(CintamaJ).icakra Avalokitesvara) as the principal deity, was
incorporated into the Shingon monastic curriculum by the end of Kukai' s
lifetime. Kukai's Shingon denju saM for example, mentions
the Juhachido as one of six practices mandatory for all Shingon priests

25 Takai 1953: 117-118. The d6j6kan is also known as the "Tathagata fist mudrii"
(nyorai ken 'in The tenn d6j6kan is rare in Chinese texts; one of the few relevant
instances is found in the Sheng huanxitian shifa lll11iXg3<A11;, a ritual manual ascribed to
'the somewhat obscure Tang monk from the "western regions" Prajiiacakra
(also known as Zhihuilun the reference is found in T.1275: 21.325b5). However,
the provenance of the work is unclear, and it may well be a Japanese apocryphon.
0026 These six rites are (I) Kechien-kanjo *lilii<i1Iilli (consecration establishing a bond with
the deity); (2) Iilhachido; (3) Issonbo -#11; (single deity practice); (4) Kongokai;
(5) Taizokai; and (6) Goma. These are followed by the Koka-kanjo and Jumyo-
kanjo I:'l!llililli initiations (KDZ 4.417ff.; see the discussion in Toganoo 1982b: 25-26).
And a court document dated 2-23-835, shortly before Kukai's death,
records that the Hihachido was to be included among' practices compul-
sory for monks seeking ordination among the annual ordinands (nen-
bundosha fF:5J\.J3t1lf)27.
The Shidokegyo sequence was becoming standardized, with the
JUhachido as the initial rite, by the time of Kakuban (1095-1143) who
writes in his ShOjuhO shokan that he received the Hihachido
at age eighteen, the Kongokai and Taizokai at nineteen, the Koka-kanjo
three times between the ages of twenty-two and twenty-seven,
and lastly the DenbO-kanjo a total of eight times
Today all
Shingon (as well as Tendai) priests begin with the Hihachido, although
the principal deity of the rite differs depending on the monk's initiatory
lineage (ryu). Priests in the SanbOin-ryli =jt{llJ'GmE use Nyoirin Kannon as
the principal deity while CMin-ryu I=j:lllJ'Gt'fE priests direct the ritual to
Dainichi Nyorai :kBJ!D* (Mahavairocana). Even then, the differences
between the rites as actually performed are relatively few and far between.
The overview of the original eighteen-methods structure that follows
is based primarily on the manuals ascribed to Kukai, notably the ]uhachi
geiin, ]uhachida nenju shidai, and ]uhachida kubi shidai
I have also
consulted a number of medieval texts on the juhachida, including the
]uhachida kuketsu (T.2529) by Raiyu (1225-1304) and
the ]uhachi geiin gishaku shOki (T.2475) attributed to
Jojin 11::i:m (fl. 1108), both of which provide considerable commentary30.
I have subdivided the eighteen methods into "six procedures" (roppa
27 Study of the TattvasaJ?1graha and Mahiivairocana siitras was also compulsory as
was facility with sJu5myo chanting. See Todaro 1985: 104; on the system of annual
ordinands see Abe 1999: 39-40.
28 Toganoo 1982b: 26. On the complex problem of the historical origins of the
Shidokegy6 system see the discussions in Toganoo 1982b: 26-27; Takai 1953: 74-75;
and Ueda 1986: 55-58.
29 There are any number of modem accounts of the eighteen methods. The most com-
prehensive accounts in Japanese are Toganoo 1982b: 44-53,286-318; Takai 1953: 109-
216; Tanaka 1962: 95-151; Oyama 1987: 67-143; and Ueda 1986: 102-207. In English
see Miyata 1984; Miyata and Todaro 1988: "Eighteen Rites"; and Payne 1991: 207-227.
30 There is some question about the authorship of the liihachi geiin gishaku shaki.
The Shoshii shosho roku lists the author as J6jin of Kiyomizu-dera m7.l<'i'f, but
other traditions attribute it to Shink6 of Kojima-dera INll,'i'f, or the Tendai monk Annen
(Kamata et a:l. eds. 1998: 734).
:f\!), following a popular medieval mode of analysis
. (The numbers
given for the individual rites correspond to their place in the modem
Iiihachido sequence provided in the appendix.)
(1) Procedure for Adorning the Practitioner (shOgon gyoja hO
Jf:t.!Hr11f!, goshin b6 Traditionally, this section includes the fIrst
five of the eighteen methods, all of which serve to purify, adorn, and pro-
tect the practitioner, rendering him or her a suitable host The practitioner
begins by anointing his/her body with incense (zuko #6), followed
by a sequence of three rites - the buddha family assembly (butsubu san-
maya #9), lotus family assembly (rengebu sanmaya
#10), and vajra family assembly (kongobu sanmaya
#11) - that call upon the deities of the three assemblies
to empower (kaji and purify the practitioner. Then one protects the
body (goshin by donning armor (hiko f1!{1fl, kongo katchii
#12); the mantra for this rite invokes Agni, and the sequence is said to
protect the host from all manner of natural disasters, demons, and evil
(2) Procedure for Binding the Realm (kekkai hO Now one pre-
pares one's abode - i.e., the sanctuary - for the deity. First, one secures
the sanctuary firmly to the earth by driving a vajra pillar through the
practitioner's seat to the center of the earth (jiketsu :!1l1*S, jikai or
kongoketsu #29). The four sides of the perimeter are secured
(shihO ketsu JIg:1J*S, #30) by erecting an indestructible vajra wall (kon-
gosho The roof has not yet been sealed to allow for the descent
of the deity.
(3) Procedure for Adorning the Sanctuary (shOgon doj6 hO
According to virtually all medieval texts, this section consists of two
rites: the "contemplation of the sanctuary" (or "contemplation of the
31 On the six procedures see the Juhachi geiin gishaku shi5ki, T.2475: 78.115c19-27;
Oyama 1987: 69; and Takai 1953: 110. I have also consulted the Taizokai nenju shidai
yoshuki (SZ 25) by GeM *_ (1306-1362). In this work GOM analyses the
structure of the Taizokai rite into "seven limbs" (shichi shi t5i:), drawing on Annen's Kon-
gokai daihi5 taijuki (T.2391: 75.170b8-1O); the seven are (1) gyogan rrllli,
(2) sanmaya =:lIIollll, (3) jojin (4) dojo (5) shi5seilhi5sei (6) kuyo lJt3t, and
(7) nenju $,ilj. This breakdown of the ritual is based, according to Gehe, on fascicle seven of
the Mahiivairocana-sutra (T.848: 18:45a ff.); see the discussion in Toganoo 1982b: 54.
locus of enlightenment," dojokan #31)32 and the "universal offer-
ings of the great sky-repository" (dai kokazo futsa kuyo
ilE, #32). As mentioned above, the Jahachi geiin does not explicitly men-
tion the dojokan - a rite that involves the visualization of the principal
deity of the rite. However, there is a "contemplation" (so that is clearly
Next imagine that in the middle of the altar there is a lion throne set on top
of a great eight-petaled lotus blossom. On the throne is a seven-jeweled tower
bedecked with colorfully embroidered banners and jewel-covered pillars
arrayed in rows. Divine garments are hung about and it is surrounded by fra-
grant clouds. Flowers rain down everywhere and music plays. Jeweled ves-
sels hold pure water, there is divine food and drink, and a mani gem serves
as a lamp. Having performed this contemplation intone the following verse:
"Through the power of my own merit, the power of the Tathagata's grace,
and the power of the dharma-realm, I dwell in universal offerings."
: ..tffllilTJilio

"&J.iJ.'1MI::iJ, 11fllHlHliHi.33
Note that this contemplation from the Jahachi geiin makes no reference
to the presence of the principal deity. TIlls is striking, since the dojokan
found in all later manuals, including the Jahachido nenju shidai attrib-
uted to KUkai (KDZ 2.621), foregrounds the appearance of the principal
deity in his jeweled palace. The following liturgy, used in the SanbOin-
ryu IUhachido, is typical:
Form the "tathagata fist mudrii" .... Contemplate li!l$ as follows:
In front [of me] is the syllable ah (J: aku). The syllable changes into a pala-
tial hall of jewels. Inside is an altar with stepped walkways on all four sides.
Arrayed in rows are jeweled trees with embroidered silk banners suspended
from each. On the altar is the syllable hrfh (kiriku) which changes and
becomes a crimson lotus blossom terrace. On top is the syllable a (a) which
changes and becomes a full moon disk. On top is the syllable hrfh (kiriku),
32 The tenn d6j6 is used as a translation of the Sk. bodhimar;ga, the seat upon which
a buddha sits at the time of his enlightenment. Its use for the sanctuary - the site of prac-
tice - is thus metaphorical, and the title of this section might well be translated "adorn-
ment of the seat of enlightenment." See Takai 1953: 117-118, and Oyama 1987: 101-
33 T.900: 18.782cll-17. By the time of the compilation of the liihachid6 nenju shidai
attributed to Kiikai, the verses at the end had become separated from the d6j6kan and
appear in a recurring unit called the "three powers" (sanriki :=:':IJ; KDZ 2.620).
and to the left and right there are two triih (taraku) syllables. The three syl-
lables change and become a vajra jewel lotus. The jewel lotus changes into
the principal deity, with six arms and a body the color of gold. The top of
his head is adorned with a jeweled crown. He sits in the posture of the Free-
dom King (Jizai {) !31:Ex), assuming the attribute of preaching the dharma.
From his body flow a thousand rays of light, and his upper torso is encir-
cled by a radiant halo. His upper right arm is in the posture of contempla-
tion. His second right arm holds the wish-fulfilling gem. His third right arm
holds a rosary. His upper left arm touches the mountain [beneath him].
His second left arm holds a lotus blossom. His third left arm holds a wheel.
His magnificent body of six arms is able to roam the six realms, employing
the skillful means of great compassion to end the suffering of all sentient
beings. The eight great Kannons and the innumerable members of the Lotus
realm assembly surround him on all sides

Commentaries typically interpret the dojokan as the moment in the nar-
rative in which the practitioner first establishes contact with the deity,
visualizing him in his divine abode
But then the reference to the "sanc-
tuary" (dojO) in the title to this section ("adornment of the sanctuary")
is ambiguous: is the sanctuary being adorned that of the practitioner or
the deity? (The term "d6jo" is most commonly associated with the site
of practice - an earthly chapel - yet according to the narrative, the
deity has not yet arrived on the practitioner's altar.)
There are a few ways to account for the anomalous nature of this
ffthachi geiin segment. One possibility is that the ffthachi geiin preserves
an early tradition that, in contrast with later manuals, remains closer to
the narrative logic of the guest-host structure. The scene is not the prin-
cipal deity's abode at all but rather the sanctuary being readied for the
deity's imminent arrival. The altar is imagined as the site of the jeweled
palatial tower, with various offerings (flowers, water, food, music, light)
laid out and ready for the god's descent.
Alternatively, the ffthi:1chi geiin may be intentionally ambiguous as to
the location of the divine altar: two of the important sources for the
34 Miyano and Mizuhara 1933: Nyoirin 13-14; cf. Ozawa 1962: Jahachido 51-53.
The Chiiin-ryii dojokan for the Jiihachid6 is similar except for the identity of the princi-
pal deity.
35 Takai explains that the advanced practitioner will contemplate the dojo as within his
own mind, but the novice practitioner must begin by viewing the dojo as outside of him-
self (1953: 161).
jilhachido sequence - the Wuliangshou rulai guanxing gongyang yigui
and Guanzizai pusa ruyilun niansong yigui - contairi contemplations at
this point in the sequence that, while not called "contemplation of the
sanctuary" (C: daochang guan), are still close to the later Japanese
dojokan in that they invoke the figure of the principal deity in his heav-
enly palace
From a doctrinal rather than narrative perspective, the ambi-
guity is felicitous, since the term dojo refers to the locus of enlightenment
itself. From this perspective there is no difference between the "abode"
of the deity and that of the practitioner - they are ultimately coextensive.
Yet another possibility is that the lilhachi geiin was not intended to
serve as a ritual manual at all but rather as a template on which manuals
for individual rites might be drafted. As such, the ambiguity concerning
the site to be visualized is created by the omission of any descriptive
details associated with a specific deity and his divine abode; one was
supposed to "fill in the blanks" later on.
(4) Procedure for Inviting [the Deities into the Sanctuary] (kanjo hO
The practitioner dispatches a jeweled vehicle for the deity
(hOsharo #34). The deity and his retinue are beckoned into the
vehicle (shO sharo #35), whereupon they descend into the sanc-
tuary and are welcomed by the practitioner (geishO #36). Early com-
mentaries note that the practitioner should imagine the carriage as adorned
with jewels and Indian in appearance

(5) Procedure for Binding and Protecting [the Sanctuary] (kechigo hO
Horse-headed Wisdom King (Bat6 my66 Bat6 Kannon
11!ru:ii!1t #39), a wrathful incarnation of A valokitesvara, is stationed out-
side the sanctuary to guard the precincts
The roof is then covered with
an impregnable vajra net (kongo rno #40), and a wall of flames
is established around the perimeter (kain .1<.1lJ'G, #41). The sanctuary is now
sealed off from the outside, making it safe from all malevolent forces.
(6) Procedure For Making Offerings (kuyo hO This is the final
and culminating section of the original eighteen methods. First, pure water
36 See T.930: 19.69a27 ff. and T.1085: 20.205a7 ff. respectively.
37 See, for example, luhachi geiin gishaku shoki, T.2475: 78.121b20.
38 In later texts, notably those associated with the Chiiin-ryii, the wrathful deity G6zanze
my66 (Trailokyavijaya-raja) is used instead.
is provided to wash the deity's feet (akaIUHfm, #43), and lotus seats are
set out for the deity to sit upon (rengeza #44). The luhachi geiin
ends with a section simply called "universal offerings" (jukuyo
in which the practitioner imagines five offerings, each limitless and bound-
less as the clouds or the sea: (1) powdered incense, (2) flower garlands,
(3) burnt incense, (4) food and drink, and (5) light. In the luhachi geiin
these five are accomplished together with a single mantra, bui: later man-
uals will specify a separate rite for each (go #46). They will
also add offerings of music (enacted by ringing a bell, shinrei #45),
hynins (san #47, #48), and so on.
The commentaries explain the content and function of each of the offer-
ings in terms of ancient Indian protocol for receiving and feting an hon-
ored guest. When the deity arrives the host first washes the deity's feet,
as was supposedly the custom in India (although the practice was not
unknown in East Asia as well)39. Commentarial discussions of the sort of
seat to be offered as well as the form and function of the five offerings
(incense, garlands of flowers, and so on) similarly draw on East Asian con-
ceptions of Indian etiquette
. This is all in accord with the theme of the
rite as a great Indian-style feast (kyoyo Takai 1953: 109-110).
The Invocation Procedures
The enumeration of the eighteen methods and six procedures, as well
as the ordinances of the luhachi geiin, end here with the universal offer-
ings41. The host has prepared his or her abode, summoned the guest, and
provided a sumptuous meal and entertainment. Yet Shingon exegetes
agree that the center piece of the rite lies in what follows, namely, the
"invocation procedures" (nenju ht5 As noted above, in contem-
porary Shingon this section consists primarily of three elaborately scripted
contemplations: the "interpenetration of self [and deity]" (nyuga-ga'nyu,
39 See, for example, ]ilhachi geiin gishaku sh6ki, T.2475: 78.l22b29-c8, which com-
ments at length on the qualities and symbolic significance of the water offered to the deity.
40 Ibid. T.2475: 78.l22c9-l23b17.
41 The ]ilhachi geiin gishaku sh6ki acknowledges that this is where the eighteen meth-
ods ends, and explains that since the procedures that follow differ according to the iden-
tity of the principal deity they are not recorded (T.2475: 78.123b17-19).
#51), "formal invocation" (shOnenju, #53), and "syllable-wheel contem-
plation" (jirinkan, #55). The historical, structural, and'doctrinal analyses
of these three rites lies beyond the scope of this paper. Suffice it to say
that they are said to unite the body, speech, and mind of the practitioner
with the body, speech, and mind of the principal deity. Shingon com-
mentators are quick to note, however, that the trope of "union" is yet
another upiiya; the rites of the three mysteries do not bring about this
union so much as they give form to it. In other words, the practitioner has
always been one with the deity; the rites of the three mysteries merely
enact, express, or realize this primordial state of affairs

Each of the rites of the three mysteries is punctuated by a short seg-
ment known as the "empowerment of the principal deity" (honzon kaji
#52, #54, #56), consisting of the recitation of the three mantras
of the principal deity. The three-mysteries sequence, interspersed by this
empowerment, constitutes the core of the "invocation procedures" (nenju
hO) and the heart of all Shidokegyo rituals.
The invocation procedures initiate an abrupt and somewhat dramatic
shift in the liturgical narrative. The guest-host scenario is temporarily
suspended, and the ritual takes a decidedly soteriological and "yogic"
turn as the practitioner is instructed to "enter meditation" (nyu ja AlE)
or "enter samiidhi" (nyu zanmai A-=llJK). This results in a two-tiered and
somewhat incongruous structure that is sometimes explained by refer-
ence to the history of Buddhist Tantra.
Buddhist Tantra, we are told, emerged from a deliberate attempt to
appropriate popular non-Buddhist Vedic or Brahmanic rites. Yixing -iT
(683-727), in the Goma chapter of his Dapiluzhe 'na chengfo jingshu
says that the Mahayana fire ritual was based on its
Vedic counterpart in order to convert followers of the Vedas to Bud-
dhism (T.1796: 39.779a19-21). "Buddha created this teaching out of his
desire to convert non-Buddhists and allow them to distinguish the true
from the false. Thus he taught them the true Goma .... The Buddha him-
self taught the very foundation of the Vedas, and in that way manifested
42 A comprehensive description of these rites along with an analysis of their symbol-
ism and doctrinal significance can be found in Takai 1953: 192-206, and Ueda 1986: 168-
182. See also the discussion in Sharf 2001b: 183-187.
the correct principles and method of the true Goma. This is the 'Buddha
Veda' f?llij!:WE. '?43
While the Buddhist Goma may resemble the Vedic one, Yixing insists
that only the Buddhist version leads to real knowledge and salvation.
To this end Yixing repeatedly distinguishes between the "outer Goma"
(wai humo which is the manifest physical performance of the
rite, and the inner Goma (nei humo which takes place in the prac-
titioner's mind. The Vedas teach the outer Goma alone; the Buddhists,
in contrast, understand the deeper significance and symbolism of the
Goma and thus perform both inner and outer at once. The Goma fire, for
example, is correctly understood by Buddhists to be the purifying wisdom
of the Tathagata (T.1796: 39.662b7-13).
Buddhist polemics aside, this historical or pseudo-historical account
does have a certain explanatory elegance. It seems plausible that the rise,
popularity, and increasing status of non-Buddhist Tantric ritual in fifth-
and sixth-century India led Buddhist practitioners to appropriate the new
ritual technology. Buddhist scholiasts legitimized the appropriation by
reinterpreting the rituals (after the fact?) according to hoary Buddhist
principles. On the one hand, specific elements in the liturgies were
explained as symbols for Mahayana teachings (Goma fire = Buddha wis-
dom). On the other hand, the entire ritual performance was rationalized
as a skillful means for manifesting one's intrinsic buddha-nature and
realizing the bodhisattva vows. The constant refrain running throughout
Yixing's commentary, and indeed all East Asian Esoteric exegesis, is that
the practitioner must envisage his or her body as the body of the princi-
pal deity44. In a single stroke the guest-host narrative of the Indianpuja
rites dedicated to a bewildering menagerie of deities is rendered a mere
upaya for the realization of inherent buddhahood.
The early Chinese manuals - supposedly translations from Indic orig-
inals - lend further support to this theory. Recall that the locus classi-
cus for the eighteen methods, the ]uhachi geiin, abruptly ends at the con-
clusion of the offering section, prior to the more soteriologically oriented
43 T.1796: 39.780bll-15; see the discussion in Toganoo 1982b: 85-86.
44 See, for example, T.1796: 39.582a26, 688c5, 688cl2-13, 701a6, 752b21, 781c27-
29, and so on.
"procedure for invocation." There is, as it were, no r90m for the invo-
cation/contemplation procedures in this stripped down version of the guest-
host paradigm. Most of the Chinese precursors of the eighteen methods
do, however, continue past the guest-host narrative; the offering section
is followed by a wide miscellany of samiidhis, discernments (guan iJl.),
contemplations (nian recitations (niansong and so on, all of
which foreground traditional Mahayana doctrine and soteriological goals.
The offerings in the Wuliangshou rulai guanxing gongyang yigui, for
example, are followed by a series of mantra recitations and guided con-
templations centered around Arnitabha and A valokitesvara
. These prac-
tices are said to induce a samiidhi wherein the practitioner's body becomes
indistinguishable from the body of the deity (T.930: 19.71a28-29).
The power of this samiidhi, claims the text, will bring about the eradica-
tion of defilement, allowing the practitioner to attain the highest level of
rebirth in the Pure Land at death.
The Guanzizai pusa ruyilun niansong yigui, another text that closely fol-
lows the jiihachid6 structure, has a similar series of recitations (song santan
jie contemplations (siwei of the principal deity, mantras,
and dhiiralJl following the offerings (T.1085: 20.206a18 ff.). These prac-
tices culminate in the repeated recitation of the principal deity's mantra such
that the "mind comes to rest in the samadhi of the principal deity" (206b22).
The text goes on to promise those who practice the rite three times a day free-
dom from defIlement, the attainment of wisdom, perfection of samiidhi, a
vision of the deity; and so on, "just as it says in the scriptures" (206c3-5).
The meditative exercises described in these Tang manuals are invari-
ably framed in terms of traditional Mahayana doctrine and soteriology.
At the same time, the meditations and recitations seem less mechanical
or scripted than the concatenation of mantra/mudrii units comprising the
. earlier guest-host sequence. TheTang texts grant the practitioner greater
latitude and flexibility in his or her approach to the invocation proce-
dures, a flexibility redolent of more traditional Buddhist meditative prac-
tices (bhiivanii). The ad hoc quality of the invocation procedures in the
45 The offerings are accomplished with the "Offering of the Great Wish-fulfilling Gem"
(guangda bukong mani gongyang mudrii and accompanying dhdralJi (T.930:
19. 70b6ff.).
early manuals is further evidence of an underlying "two tier" structure
to the rites.
As the rites developed in Japan, the invocation procedures following
the offerings became more routinized, and by the medieval period they
had come to assume the set form still practiced today: three discrete rites
corresponding to the three mysteries. It is unclear, however, exactly
when and how this transformation occurred. Two of the three-mysteries
rites - the all important "interpenetration of self [and deity]" and
"syllable-wheel contemplation" - do not appear in any of the Chinese
sources for the eighteen methods. Indeed, they rarely appear in Chinese
texts at all, and when they do they are not identified with the specific
"mysteries" of body and mind
Nor are they found in the JUhachid6
manuals attributed to Kukai. They do appear, however, in several other
ritual manuals ascribed to Kukai, suggesting an early date for the estab-
lishment of these rites as set pieces in the Shingon curriculum
Even then,
with few exceptions, the early manuals do not construe the "interpene-
tration of self [and deity]" and "syllable-wheel contemplation" as part of
the three-mysteries sequence

In fact, the various invocation procedures found in the early manuals
attributed to KUkai appear relatively fluid, with considerable variation
from rite to rite much like their Chinese prototypes. One frequently comes
46 For a rare appearance of the phrase ruwo woru A'a'aA in the Chinese Buddhist canon
see the lingangdingjing dayuqie himi xindi famen yijue (T.1798:
39.813b15-17). Among the few Chinese references to "contemplating the syllable wheel"
(guan zilun see the.Achu rulai niansong gongyang fa (T.921:
19.15cll, 19c12), and Yixing's Dapiluzhe'na chengfo jingshu (T.1796: 39.689c9).
47 Mention of nyuga-ga'nyu is found, for example, in the Senju Kannon gyohO shidai
T=Fij!!'/flTI*:bzm (KDZ 2.552), liM kongo nenju shidai (KDZ 2.567;cf. 2.580;
4.787), Mujin shOgonzo shidai (KDZ 4.506), Taizo hizai shidai J!/lillliill'l'E:bzm
(KDZ 4.616; cf. 4.659), Taizokai unji shidai lIiliillWII'14::b<:m (KDZ 4.694), Shugokyo nenju
shidai (KDZ 4.768), and Shari hO (KDZ 4.772). The jirinkan is found
in the Kongokai dai giki (KDZ 4.487), Taizo hizai shidai (KDZ 4.617; cf. 4.659),
Taizokai unji shidai (KDZ 4.694), Shugokyo nenju shidai (KDZ 4.769), and Shari hO (KDZ
48 Two notable exceptions are the sequences found in the Taizo hizai shidai (KDZ
4.616; cf. 4.659) and Taizokai unji shidai (KDZ 4.694). In both cases the sequences are
almost identical to the one found in later medieval manuals, such as the manuals by Gengo,
that form the basis of the modem rite. This may be evidence of a relatively late date for
the Taizo hizai shidai and Taizokai unji shidai.
across units such as "formal recitation" (shanenju) anq "entry into med-
itation" (nyu ja .:A..5E)49, yet the specific liturgical content of these segments
was not yet systematized .. Moreover, there are several instances in which
the specific contemplations now associated with the "interpenetration of
self [and deity]" and "syllable-wheel contemplation" are not Identified
as discrete units under those headings, but simply appear as part of other
liturgical units 50. It likely took several generations for the rites to crys-
tallize into their current forms, but given the problems dating the extant
manuals it would be difficult to determine precisely how this happened.
In any case, the available evidence of the Chinese and early Japanese
manuals suggests that the two-tiered structure of the Shidokegyo rites is
the result of a complex historical evolution, in which the "Vedic" guest-
host narrative was both legitimized and confounded by the superimposi-
tion of an explicitly "Mahayana" bhiivana segment. This bhiivana seg-
ment was originally less structured and routinized than the rites of the
guest-host narrative, which is not surprising given the greater antiquity of
the guest-host rites. In any case, the guest-host sequence was reinter-
preted in light of the bhiivana segment, transforming the entire rite into
an extended meditation on, statement about, or performance of one's
inherent buddha-nature. One might view this transformation as concep-
tually elegant and clever, or clumsy and confusing, depending upon one's
point of view. Needless to say, such value judgments were not germane
to traditional exegetes. These exegetes were, however, forced to confront
the confusions that arose from the imposition of two somewhat discor-
dant narratives.
The Dispersed Invocations
The structural ambiguity of the rite comes to a head in the "dispersed
invocations" (sannenju #57), the segment that follows the three
mysteries in the Shidokegyo liturgies. The dispersed invocations are fol-
lowed by the "latter offerings" (go kuya a ritual sequence that
49 See, for example, the Mujin sMgonzo shidai (KDZ 4.506); and SaM shidai j1'iSzll';
(KDZ2.499, cf. KDZ 2.523).
50 See, for example, the SaM shidai (KDZ 2.503-504).
resumes the guest-host narrative and brings the ritual to a close. If only
by virtue of their position in the rite, the dispersed invocations serve to
negotiate the gap between the soteriological program of the three mys-
teries and the guest-host narrative that returns in the "latter offerings."
The latter offerings consist of many of the same procedures found in
the eighteen methods, except that they are performed in reverse order and
often abbreviated. Thus it begins with the five offerings and proceeds to
offerings of water, music,hymns, and so on. Following these offerings
there is a dedication of merit (eko #64), a standard element appear-
ing near the close of all Buddhist rites. The sanctuary is then unsealed:
the encircling flames and the vajra net are removed, Horse-headed
Wisdom King is relieved from sentry duty, the vajra wall is withdrawn,
and the principal deity is dispatched back to his abode (hakken #67).
The rite closes with a repeat of a few of the apotropaic procedures that
opened the performance (goshin bO). The final offerings serve as a denoue-
ment of the guest-host narrative, running the sequence in reverse.
The dispersed invocations that precede the latter offerings consist of a
group of mantras that are repeated anywhere from seven to one thousand
times each. The specific mantras used vary depending on the rite, the lin-
eage (ryu), and the principal deity, although three of the mantras - Bud-
dha-locana (Butsugen Mahavajracakra (Dai kong6rin
and (Ichiji kinrin Ichiji chOrinn6
- always In the Chliin-ryu JUhachid6 for example,
in which l\1ahavairocana is the principal deity, the mantras used in the dis-
persed invocations consist of (1) Buddha-locana, repeated twenty-one
times; (2) Garbhakosadhatu Mahavairocana, repeated one hundred
; (3) Vajradhatu Mahavairocana, repeated one thousand times;
51 The Sanskrit reconstruction of the Buddha-locana mantra, which appears at the begin-
ning and end of the dispersed invocations, is: Namo bhagavat-u:flJl:fa om ruru sphuru
jvala-ti:ffha-siddha-Iocani sarviirtha-siidhanfye sviihii. Miyata and Todaro render this into
English as "Homage to the of the Bhagavat! Om Speak! Speak! Fill up! Radiate!
Remain! Oh, the gaze of the accomplished ones! May you accomplish all aims! Sviiha"
(1988: "Eighteen Rites," 31).
52 For an account of why these three mantras are always present see Raiyu's Usuzoshi
kuketsu T.2535: 79.178b27-cll.
S3 Reconstructed as: Om a vi ra hUm kham ("Om," followed by five seed syllables).
(4) the Four Buddhas - (Ashuku-nyorai Rat-
nasarp.bhava Amitayus (Arnida-nyorai
*), and Amoghasiddhi (Fukii-joju repeated one hundred times
each; (5) Vajrasattva (Kongo satta repeated one hundred times;
(6) Trailokyavijaya-raja (Gozanze myoo repeated twenty-
one times; (7) Mahavajracakra, repeated seven times; (8)
repeated one hundred times; and (9) Buddha-Iocana, repeated
seven times
. These mantras need not be accompanied by any specific
contemplation; the manuals say only that the practitioner repeats the
mantras using the rosary with hands forming the "preaching the dharma
mudrii" (seppo no in
The mantras of the dispersed invocations, like most Japanese mantras,
consist of Japanese pronunciations of Chinese transliterations of Sanskrit
invocations, making it difficult for most priests to discern the semantic
content (if indeed there is any) of the underlying Sanskrit phrases. (Con-
temporary training manuals and scholarly commentaries often provide
Sanskrit reconstructions, Japanese translations, and explanations of the
mantras.) As there are close to two thousand repetitions to perform, the
dispersed invocations can take upwards of an hour to complete, consti-
tuting one-third to one-half of the duration of the rite.
Given their duration and their placement within the ritual sequence
- situated immediately after the climax of the three mysteries - one
might suppose that the dispersed invocations comprise a particularly
important section of the Shidokegyo. practices. Yet traditional Shingon
commentators have little to say about the meaning and function of this
segment, and what they do say is often vague and equivocal. The mean-
ing of the term sannenju itself is ambiguous (see below), and my use of
"dispersed invocations" is little more than an expedient; "supplemen-
tal" or "scattered invocations" might serve just as well.
54 This list, typical of Chiiin-ryii manuals, is taken from Oyama 1987: 129-133, and
Nakagawa 1986: ]ahachidi5. The mantras will differ depending in part on the identity of
the principal deity used for the rite. On the dispersed invocations, in addition to Oyama
see Toganoo 1982b: 66-67; Takai 1953: 206-208; Tanaka 1962: 147-148; Ueda 1986:
182-187; Miyata 1984: 91-94; Sawa 1975: 271b; Foguang da cidian zongwu weiyuan-
hui 1989: 5.4973c-d; Nakamura 1981: 496d; and Ding 1984: 1145d.
55 The rosary is used along with counting sticks to keep track of mantra recitations.
The confusions are in part due to the absence of authoritative textual
soirrces for the dispersed invocations. The sannenju segment does not appear
in any of the dozens of Chinese texts on which Shidokegy6 liturgies
were based; nor does it appear in the lahachi geiin, which concludes, as
mentioned above, with the offerings at the end of the eighteen
The dispersed invocations are mentioned, however, in many of the manu-
als attributed to KUkai, including the luhachido nenju shidai (KDZ 2.627)57.
And even when the term sannenju does not appear, early Shingon ritual
manuals that conform to the eighteen-methods structure will often pre-
scribe mantra recitations immediately following the three-mysteries rites;
these recitations appear to be the functional equivalent of the sannenju.
The absence of a canonical Chinese precedent meant that Japanese
practitioners enjoyed considerable latitude in their approach to the
dispersed invocations. Manuals and commentaries agree that practitioners
- or at least advanced practitioners (itatsu - are free to add, sub-
tract, or substitute mantras in accord with their own predilections, to aug-
ment or decrease the prescribed number of recitations, or to omit this sec-
tion entirely. Accordingly, the dispersed invocations were also known as
the "discretionary invocations" (zuii nenju and some medieval
commentators alternate freely between the two terms
. Moreover, the
56 I have found only a single reference to the term san niansong in the "Esoteric
teachings section" (mikkyo bu <til!:llil) of the Taish6 canon. This is in the Yaoshi yigui yizhu
(T.924c: 19.32c23), a text of uncertain authorship and provenance, but here it
refers to a segment occurring before the major invocations of the rite, and hence it appears
of limited relevance to the discussion at hand. The lingangdingjing yuqie shibahui zhigui
i!1!IJlfll&",ihu-ti\1tlliiliil translated by Amoghavajra contains a reference to an "esoteric dis-
persed recitation that augments skillful means" (bimi zhucheng fangbian sansong
iiiI<tilliJpx1Ji!miili, T.869: l8.286a22). This locution may have influenced Kiikai's use of the
phrase to sannenju shitchi hob en in the Ninno hannyakyo nenju shidai
(KDZ 4.751); see the discussion in Ueda 1986: 183.
57 The sannenju is also mentioned in the following manuals ascribed to Kiikai: Fudo
myoo nenju shidai (KDZ 2.677), SahO shidai (KDZ 2.507); Issai nyorai
taisho kongo shidai (KDZ 2.609); Taizo bizai shidai (KDZ 4.617); Taizo
bonji shidai Rilili!1t'l";9(jfl (KDZ 2.286); Bizai shidai friil':E;9(jfl (KDZ 4.659); Gumonji shidai
*llIIffl;9(jfl (KDZ 4.701); and Ninno hannyakyo nenju shidai (KDZ 4.751). Usually the texts
simply say, "Next, the dispersed invocations" although occasionally, as in the
Fudo myoo nenju shidai and Issai nyorai taishO kongo shidai, the text will list the names
of the mantras to be used.
58 See, for example, the Taizokai nenju shidai yoshaki by G6h6 (SZ 25.519b). Kiikai's
SaM shidai contains a short gloss under the sannenju saying it is "optional" (nin'i ff:1ii:;
dispersed invocations are not found in the liturgies of the Tendai esoteric
tradition (Tairnitsu trW), marking it as one of the few notable differences
between the Shidokegy6rites of the Tendai and Shingon schools.
The origins and meaning of the term sannenju are unclear
mentators typically begin their discussions of the term by opposing the
sannenju to the shonenju or "formal invocation" of the three-mys-
teries segment
. In its narrow sense, the "formal invocation" refers to
the second of the three mysteries - the "mystery of speech" (gomitsu
mW) - realized through a stylized recitation of the mantra of the prin-
cipal deity accompanied by an elaborate contemplation of the mantra
circulating between the deity and the practitioner. However, the term
"formal invocation" can also denote the entire three-mysteries sequence.
In either case, the shO IE of shOnenju is interpreted as "formal," "solemn,"
and "direct," while san is understood as "scattered," "dispersed,"
and "diffuse" (santa Whereas the shOnenju is a highly stylized
invocation directed toward the principal deity alone, the sannenju is a
less stringent "scattering" of invocations among a variety of supple-
mentary deities. Thus the shOnenju, which is accompanied by a mudrii
as well as an elaborate "visualization," is considered the "primary" recita-
tion, while the sannenju, which is accompanied by a mudrii alone, is
treated as "secondary."
Commentators also suggest that the sho of shonenju has the sense
of shoshin IEJL\ meaning a "focused" or "directed mind," in contrast
to san as sanshin meaning a "diffused" or even "distracted
mind. ,,61 According to this reading, during the three-mysteries segment
KDZ 2.507). And the reference to the dispersed invocations in the Jilhachido sata i-i\jlfJI/t
by Kakuban lists the mantras to be included in this section, namely Buddha-locana,
Mahavairocana, Trailokyavijaya-raja, Vajrasattva, and Acalanatha, following which one can
"continue at one's own discretion" (sonogo zuii :Jt1fIl1l;'@;; T.2517: 79.26c8). Lexical sources
note other names for the sannenju as well, including "supplementary invocations" (kayo
nenju and "miscellaneous invocations" (shozo nenju Foguang da cidian
zongwu weiyuanhui 1989: 5.4973b; Sawa 1975: 271b).
59 See the sources mentioned in note 54 above.
60 The shOnenju is also known as sanmaya nenju. jo nenju and kaji
nenju (Sawa 1975: 388b-389a).
61 See, for example, Ueda 1986: 182, who cites chapter six of the Zuigyo shishO
the practitioner is one with the deity, in a state of meditation (jocha
in which the practitioner enters into the deity's samiidhi. In the
dispersed invocations that follow, the practitioner emerges from
samiidhi in order to fulfill the bodhisattva vows, enlightening others by
"scattering" mantras in all directions. How, one might ask, is the prac-
titioner to practice if the mind is "scattered"? This question is raised
in the luhachido kuketsu by Raiyu, who provides one of the more
detailed discussions of the dispersed invocations
. Citing the Hizoki
WJ1.iilac (thought to be Kilkai's record of Huiguo's teaching), Raiyu says
that the practitioner and the deity have both merged into the single
dharma realm (ichi hokkai during the previous invocations,
and thus the practitioner is able to retain control fr3 even though his or
her mind is scattered. Having just merged with the deity, the practi-
tioner is able to reenter the phenomenal world while remaining iden-
tified with the principal deity63. Raiyu goes on to equate the formal
invocation with meditation and the dispersed invocations with wisdom
Modern commentators pick up this opposition, saying that the formal
invocation is the "practice of inner realization" (jinaisho no homon
while the dispersed invocations effect the liberation of
others (keta {l::At!!)65. The liberation of others is achieved through the invo-
cation of a host of deities (shoson that have a karmic bond (en
with either the principal deity or the practitioner, thereby augmenting the
grace and power of the principal deity (Tanaka 1962: 147). The structural
relationship between the formal invocation and the dispersed invocations
is thus equated with the standard Mahayana moieties of buddha versus
bodhisattva, emptiness versus skillful means, and so on.
62 T.2529: 79.70bll-c4. See also his Usuzoshi kuketsu, T.2535: 79.178b22-179al.
63 Moriya Eishun j:l<:jotJl!if31, a priest at K6fukuji, explained this to me by saying that dur-
ing the sannenju the wandering mind of the practitioner is identical with the wandering
mind of the principal deity, eliminating the need for any prescribed contemplations to
accompany the sannenju recitations.
64 T.2529: 79.70c3. Raiyu also comments that since the dispersed invocations do not
appear in the scriptures and early recitation manuals, it is optional in the Rishoin-ryu
65 Tanaka 1962: 147. Tanaka notes that, according to exegetes such as Kakuban, the
formal invocation is also effective in liberating others.
Ritual Incoherence
One question that arises is how, if at all, the dispersed iIlvocations fit
into the overall narrative program of the Shidokegyo rituals. We have
seen that there is a break between (1) the offerings segment that complete
the traditional eighteen methods, and (2) the enactment of the three mys-
teries that follows. At best, the rites of the three mysteries would seem
to render the preceding guest-host narrative an upiiya; the true relation-
ship between practitioner and deity is not that of host and guest after all
but rather one of identity. Yet the guest-host narrative recommences with
the "latter offerings" that follow the dispersed invocations. If this denoue-
ment to the narrative, in which the host unbinds the sanctuary and bids
farewell to the guest, is taken at face value, then at what point in the nar-
rative does the practitioner "emerge from samadhi" and disengage from
the deity? This question bears directly on the narrative significance and
function of the dispersed invocations.
Commentators have explored, explicitly or implicitly, three possibil-
ities. The first is that the practitioner disengages from the deity and
reverts to his or her former self with the commencement of the dispersed
invocations. The dispersed invocations then represent the activity of a
bodhisattva; the practitioner, having "reentered the marketplace" (to
borrow a popular Zen image), scatters invocations for the liberation of
all sentient beings. The second possibility is that the dispersed invo-
cations are themselves intended to reintegrate the practitioner into'
the world; they facilitate a gradual and controlled emergence from
The third possibility is that the entire sequence of dispersed
invocations is performed while ensconced in the samadhi of the princi-
pal deity.
There is a certain elegance to the last position, according to which
the dispersed invocations are the manifest performance of the princi-
pal deity himself. This renders the dramatic narrative of the Shidokegyo
rituals structurally analogous to the performance of a shaman or spirit
66 See esp. Miyata 1984: 91-94, who views everything following the three mysteries
as a gradual process of "dissociation." According to Miyata, the process continues through
the dispersed invocations to the latter offerings, the unsealing of the realm, and the depar-
ture of the deity.
medium, in which the raison d' etre of the ritual prologue is to efface
the agency of the practitioner and invoke in his place the presence of
the deity. In Shihgon, this is viewed not as possession, of course, but
rather as an extended communion, referred to as "reciprocal resonance
[with the deity]" (kanno doko wherein practitioner and god
act in total accord
This should not be construed merely as an inte-
rior "meditative state"; rather, the physical activity of the performer
is precisely the physical activity of the embodied deity (sokushin
This is how the practitioner is instructed to approach the fourth and final
rite of the Shidokegy6 sequence, namely, the Goma. Like all Shidokegy6
rites, the Goma ritual is built around the eighteen methods of the
juhachido. But there is an important difference: the Kong6kai and
Taiz6kai rituals are constructed as expansions of the liihachid6 rite, with
dozens of additional ritual elements interspersed among those of the
IUhachid6. The Goma, in contrast, is constructed by taking the entire fire
ritual segment and nesting it whole in the midst of the liihachid6 dis-
persed invocations. Thus the Shidokegy6 Goma opens with the liihachid6
sequence, running it all the way through the main offerings, three-
mysteries invocations, and most but not all of the dispersed invocations.
The fire ritual proper commences just before the final three mantras of
the dispersed invocations (Mahavajracakra, and
Buddha-Iocana). When the Goma is complete the practitioner performs
the three mantras that remain from the dispersed invocations and then
continues through the "latter offerings" of the liihachid6 (Takai 1953:
389). The fire ritual is thus framed by the recitations of the dispersed
invocations, and the practitioner is to remain in a state of unity with the
principal deity throughout the fire offerings.
In the end, there is little agreement among traditional or modem Japan-
ese exegetes as to the specific point at which the practitioner emerges
from samadhi - the point at which, according to the logic of the narra-
tive, guest and host are not one but two. This narrative ambiguity mir-
rors an ambiguity in the rites' underlying soteriology.
67 Toganoo 1982b: 66. On the notion of reciprocal accord with the deity see Hiz6ki,
KDZ 2.36; Toganoo 1982b: 151; and Sharf 2002: 77-133.
Mahayana notions of tathagatagarbha and intrinsic l;lUddha-nature gave
rise to a conundrum that captivated generations of scholiasts: if buddha-
nature is innate, why practice? The Zen patriarch Degen (1200-
1253) is often associated with the response that one practices not in order
to attain buddhahood but in order to manifest it. But in viuiousguises this
"solution" to the problem predates Degen by many centuries, and Degen's
own approach may have been influenced by his mikkyo training at
Enryakuji In any case, Shingon ritual is predicated on a view of
the phenomenal universe as the theophany of the dharmakiiya, a view
that confutes, at least in theory, the notion that Shingon ritual is intended
to bring about a fundamental change in the ontological status of either the
practitioner or the world. The point of the rites, in other words, is not the
attainment of buddhahood but rather its expression. This expression takes
the form of an elaborately scripted drama wherein the practitioner com-
pels the presence of a buddha only to reveal that the buddha was never
. Among other things,this notion provides doctrinal justification
for the seemingly obsessive character of mikkyo ritual; since there is no
ultimate "goal" to be achieved, one is left, like SeW practitioners of
zazen, with practice for its own sake. This also provides conceptual
grounds for the ambiguity in the ritual narrative noted above: from the
standpoint of tathagatagarbha theory and the doctrine of intrinsic bud-
dha-nature, it makes little sense to mark a ritual moment at which one
ceases to be a buddha.
Finally, a similar conceptual ambiguity can be discerned in the treat-
ment of the central image (honzon) enshrined on East Asian Buddhist
68 This is made explicit throughout the liturgical recitations and contemplations of the
Shingon ritual manuals. Take, for example, the "interpenetration of self [and deity]," the
fIrst of the three-mysteries rites which brings about the union of the body of the practi-
tioner and the body of the deity. The contemplation associated with this rite reads:
"The principal deity sits on a ma7;uJala. I sit on a mm:uJala. The principal deity enters my
body and my body enters the body of the principal deity. It is like many luminous mirrors
facing each other, their images interpenetrating each other (Miyano
and Mizuhara 1933: Nyoirin 27-28; cf. Ozawa 1962: Jilhachid6 78-79). Traditional
exegetes interpret the use of the mirror image as showing that the body of the principal
deity does not literally "enter" the practitioner; rather, one is to look upon the principal
deity as if gazing at one's own reflection. The body of the principal deity and the body of
the practitioner have always subsumed each other (Takai 1953: 194).
temple altars. In order to be ritually efficacious, such images must be
consecrated in an "eye-opening" (kaigen ~ ~ N ) ceremony when first
installed. Such a consecration transforms an image from a mere physical
likeness into a vivified icon that literally embodies the deity69. At the
same time, if one looks at the structure of the services regularly per-
formed before such images (J: kuy6 ~ ~ , from Sk: pilja, "rites of offer-
ing"), they typically involve a ritual segment, however brief, that invites
the deity to descend into the image
. This raises the question: if the icon
was successfully consecrated at the time of its installation, thereby trans-
forming it into the living body of a deity, what need is there to request
the descent of the deity yet again at the time of worship? Is this merely
a case of ritual anxiety fueling a ritual obsession that betrays a lingering
doubt over the efficacy of the rites?
Phyllis Granoff has argued that this "confusion" can be explained by
reference to the historical evolution of image worship in India
. Accord-
ing to Granoff, the two moments of invocation - one during the initial
consecration of the image and the other during regular "feedings" -
may derive from two different paradigms of worship that became incor-
porated into the later image cult. One is an earlier "Vedic" model, in
which the worshipper must solicit the presence of the deity prior to each
sacrifice. This paradigm was established long before the use of sacred
icons in India; Brahmin priests invoked invisible beings on an altar that
was often a temporary structure built specifically for the occasion.
The spread of the cult of the image is associated with a later "Purfu:!ic"
mode of worship focused around a consecrated icon permanently
enshrined in a temple 72. The image, which some believe was introduced
from Greece, was approached as the animate physical incarnation of the
69 See Sharf 2001 a; for references to the secondary literature on eye-opening cere-
monies see Sharf2001b: 248n. 64.
70 This is true of both "exoteric" (kengyo IIllfl) and "esoteric" (mikkyo) Buddhism, but
as far as I am aware only in the latter case is the deity explicitly sent back to his or her
abode at the close of the rite.
71 See Granoff 2001, n.d.a, and n.d.b. I want to thank Phyllis Granoff for generously
sharing and discussing her unpublished work with me.
n It is worth noting that the gap between these two paradigms was so great that the
orthodox Brahmin priests originally distanced themselves from and castigated the emerging
temple cult.
deity. The icon/deity became a permanent resident j.n the community;
it needed to be bathed, dressed, fed, and entertained on a regular basis.
But the earlier Vedic paradigm was soon superimposed on the treatment
of these images; VediC-style incantations (mantras) were used to impel
the deity's descent at the initial consecration and again during regular
pftjii offerings. The Buddhist treatment of images appears to be based on
this pan-Indian synthesis of Vedic and Purar:uc models. (In Shingon, the
Vedic antecedents are somewhat more pronounced, as a new "tempo-
rary" altar is ritually constructed during the course of each performance.)
If Granoff is correct, then the ambiguities, if not the discursive incoher-
ence, that result from the fusion of Vedic and Purar:uc modes of worship
is analogous in many respects to the narrative ambiguities that result from
the fusion of Vedic ritual and Mahayana bhiivanii found in the EastAsian
Esoteric rites discussed above.
Ritual Meaning
My musings on the history of Buddhist ritual and image worship are
just that: the musings of an outsider based largely on the evidence of rit-
ual texts the provenance and historical development of which are still
poorly understood. Historical questions aside, however, my overview of
the narrative content, structural logic, and doctrinal import of the ritual
procedures is by no means an etic imposition. The guest-host narrative is
made explicit in the sequence of Shidokegyo ritual procedures and is
further amplified in oral and written commentaries from early on.
Moreover, basic Shingon teachings concerning one's identity with the
principal deity, the dependently arisen nature of all phenomena, the bod-
hisattva vows, and so on, are reiterated ad nauseam in the content of Shi-
dokegyo recitations and contemplations

73 For example, the cuhninating moment of the Shidokegyo rites is the ritual identifi-
cation of the mind of the deity with the mind of the practitioner. This takes place in the
"syllable-wheel contemplation," which consists in a Madhyamika-style "deconstruction"
of the principal deity's mantra. According to the discursive logic of this rite, to appreci-
ate the dependently arisen nature of the deity's mantra, and thus the emptiness of the deity
himself, is precisely to become one with the deity's mind. See Sharf 200lb: 184-185.
In a parenthetical comment earlier in this paper I suggested that what
makes tantra "tantra," in any critical cross-cultural sense, lies not in its
"meanings" but in its techniques. Tantra is an applied knowledge per-
taining to the use of a cornucopia of ritual implements, icons, occult ges-
tures and utterances. These techniques were adopted into diverse reli-
gious contexts across Asia and reinterpreted in the light of local tradition.
There is thus no reason to assume that the specifically Shingon under-
standing of the narrative or doctrinal content of the rites examined above
is commensurate with non-Buddhist interpretations of Tantra found in
South or Southeast Asia. Buddhist exegetes would agree with this assess-
ment, since by their own account the Buddha borrowed the outward forms
of Vedic worship and supplied them with new Mahayana meanings.
But by the same measure, any robust account of Shingon Tantra must
acknowledge the discursive content of the rites that was salient in the
Shingon school. Each element in the rite was understood in the context
of this content - its place in the overarching guest-host narrative - and
modifications to the ritual form were made in full awareness of their nar-
rative and doctrinal consequences. As such, Staal's thesis as to the essen-
tial invariance and meaninglessness of ritual cannot stand, for in Shingon
we have a sophisticated ritual tradition of considerable antiquity in which
(1) rituals underwent continual, albeit incremental, change, and (2) seman-
tic content clearly mattered.
This still leaves us with the question as to why anyone would perform
these rites in the first place. Here Staal raises an important point, for the
meanings themselves cannot account for or justify the tremendous com-
mitment of human and institutional resources necessary for the perform-
ance of these rites. Considerable expense is involved in the acquisition
and preparation of the essential ritual paraphernalia, and a monastery must
be willing to offer material support to the priests in cloistered retreat. More
important, the rituals themselves are hard work: the retreats are long, ardu-
ous, and mentally and physically exhausting. Why spend years of one's life
perfecting a surfeit of rites that all end up "saying" much the same thing?
Any full response to this question must take into account a host of
sociological and psychological factors, bearing on everything from insti-
tutional structure, to issues of social status, to questions of identity for-
mation and personal faith - issues that cannot be addressed here. But our
response must also take into account the power and. allure of the rituals
themselves, an allure derived in part from the narrative explored above.
This narrative situates the practitioner as the protagonist in a dramatic
encounter with powerful and mysterious forces. The construted, fictive,
dramatic, and patently playful aspects of the encounter make it no less
74 See Sharf n.d. for an analysis of the element of "play" in the workings of ritual.
Procedural Sequence for the Eighteen-Methods Practice
(Jiihachido nenju shidai
The following outline of the Eighteen Methods sequence is that used by con-
temporary priests in the SanbOin-ryli, in which Nyoirin Kannon functions as the
principal deity (see Ozawa 1962: Jahachid6, and Takai 1953: 109-216). Among
the prototypes for the contemporary rite, the most influential manual is the Sh6-
nyoirin Kanjizai bosatsu nenju shidai by Genge j[;:li':
(914-995). Genge's manual is in tum based on the Jahachi geiin
Jahachid6 nenju shidai and Jahachid6 kubi shidai all
of which are attributed to Klikai.
The symbol" b," indicates a procedure included in the traditional list of eight-
een procedures traced to the Jahachi geiin. The symbol "-:/' indicates a section
included in the "six practices" (rokuhO r;#i).
Various ritual purifications precede the formal entrance to the hall.
3. 3fm
4. m{;!l;
Enter the Sanctuary
Universal Prostration [to all Tathiigatas]
Sit Down
Separate the hnplements
Universal Prostration (as above)
Rub Powdered mcense (powdered incense is rubbed
on the hands and arms, and then across the chest,
anointing the five-part dharma-body li?t#i,!!:t.)
Contemplate the Three Mysteries (This contem-
plation uses the "urn" syllable to purify body,
speech, and mind.)
The following five procedures constitute the
goshin b6 or bodily purification and pro-
Purify the Three Karmic Actions (body, speech,
Buddha Family Assembly
Lotus Family Assembly
Vajra Family Assembly
Don Annor and Protect the. Body



21. t$j11JTJ!l.ll
22. n'ljij



Empower the Perfumed Water (ambrosia)
Empower the Implements
Contemplate the Character "ran"
Purify the Earth
Contemplate the Buddhas
Arouse the Vajra
Universal Prostration
Declaration of Intent
Siltra Offerings to Divine Spirits
Five Repentances (Samantabhadra's vows)
Purify the Three Karmic Actions (as above)
Universal Prostration (as above)
Give rise to the Mind of Awakening (Bodhicitta)
Three Samaya Precepts
Recite the Vows
Five Great Vows (to save all beings, to cultivate
all merits and wisdoms, to awaken to all the
dharma-gates, to serve all tathagatas, and to real-
ize unexcelled awakening)
Universal Offering
::, 29. :tll!*Ii
::, 30.
Great Vajra Wheel
Bind the Earth (also called the Vajra Pillar)
Bind the Perimeter, or the Vajra Wall

::, 31.
::, 32.
Contemplate the Sanctuary
Universal Offerings of the Great Sky-Repository
Small Vajra Wheel
t;, 34.
t;, 35.
Send Forth the Jeweled Vehicle
Invite the Deities to Ascend the Vehicle and Ride
to the Sanctuary
t;, 36. Welcome the Deities
37. IZEIjIj Four Syllable Mantra
38. Clap Hands (in welcome)
t;, 39. }i!iililjlj.:E Invoke the Horse-headed Wisdom King
Sky Net (or Vajra Net) t;, 40.
t;, 41. Vajra Fire
42. *':::,*Jfll Great Samaya [Assembly]
t;, 43. fMHlJo Offer Pure Water
t;, 44.

Offer Lotus Seats
45. JJret,t Offer the [Five Pronged] Vajra and Bell

Five Offerings (powdered incense, garland, burnt
incense, food and drink, light)
47. Eulogy of the Four Wisdoms (accompanied by

Eulogy to the Principal Deity
t;, 49. Offer the Great Wish-fulfilling Gem (or Univer-
sal Offerings)

Worship the Buddhas' [Names]
51. Interpenetration of Self [and Deity]
52. Empowerment of the Principal Deity
53. Formal Invocation
54. Empowerment of the Principal Deity (as above, 52)
55. Syllable Wheel Contemplation
56. Empowerment of the Principal Deity (as above, 52)
57. Dispersed Invocations
58. Five Offerings (as above, 46)

Offer Pure Water (as above, 43)
60. f&Mi Latter Offering of Bell and Vajra (as above, 45)
Eulogy (as above, 47)
62. Universal Offering with the Verses of the Three
Strengths (as above, 27)
63. ijlJ{jil Worship the Buddhas' [Names] (as above, 50)
64. Dedication of Merits
65. Five Repentances and Vows of a Sincere Mind
66. fW-!JI. Release the Realm, consisting of the following
five segments, in the reverse of their order above:
Great Samaya [Assembly] (as above, 42)
Vajra Fire (as above, 41)
Sky Net (as above, 40)
,i!llAAIjIj.:E Horse-headed Wisdom King (as above, 39)
Vajra Wall (as above, 30)
67. Send Off [the Principal Deity and His Assembly]
68. =:1!EIli!<J!ll
Three-fold Samaya
69. Don Armor and Protect the Body (as above, 12)
70. Universal Prostration
71. Leave the Sanctuary
Works Cited
KDZ Kobo Daishi zenshii 6 vols. Edited by Sofil senyokai
Tokyo: Rokudai shinposha, 1910.
SZ ShingonshU zensho 44 vols. Edited by Shingonshil zensho
kankokai Koyasan: Koyasan daigaku, 1933-36.
T TaishO daizokyo j(iE;kiXfil!. 100 vols. Edited by Takakusu Junjiro
and Watanabe Kaigyoku lJj;jjfjj:1i!l.. Tokyo: Taish5 issaikyo
kankokai, 1924-32. Texts are indicated by the text number ("T.") fol-
lowed by the volume and (when appropriate) page, register (a, b, or c),
and line number(s).
Premodern Ritual Manuals and Commentaries
Achu rulai niansong gongyang fa Trans. Amoghavajra. T.921: 19.
Bizai shidai Attr. Kilkai !,gi/ii. KDZ 4.620-663.
Dabao guangbo louge shanzhu bimi tuoluoni jing
Trans. Amoghavajra. T.1005a: 19.
Dapiluzhe'na chengfo jingshu ;kJ!BJJ!J!ltIlB.6Xi?U!J!Ml. Yixing -fl'. T.1796: 39.
Issai nyorai daishO kongo shidai Attr. Kilkai !,g#if.. KDZ
Fudo myoo nenju shidai Attr. Kilkai !,gi/ii. KDZ 2.656-678.
Guanzizai pusa ruyilun yuqie ill J3 Trans. Amoghavajra. T.1086: 20.
Guanzizai pusa ruyilun niansong yigui Trans. Amoghava-
jra. T.1085: 20.
Gumonji shidai Attr. Kilkai !,gi/ii. KDZ 4.698-702.
Jiho kongo nenju shidai Attr. Kilkai !,gi/ii. KDZ 2.556-570.
Jingangding lianhuabu xinniansong yigui Trans. Amoghava-
jra. T.873: 18.
Jingangdingjing dayuqie bimi xindi famen yijue
Amoghavajra. T.1798: 39.
Jingangdingjing yuqie shibahui zhigui 7\ 1rlirJ$. Trans. Amoghavajra.
T.869: 18.
Juhachi geiin Attr. Kilkai !,g#if.. KDZ 2.634-645.
Juhachi geiin gishaku shOki J5jin Ai::i?I1. T.2475: 78.
Jlihachido kubi shidai Attr. Kilkai !,gi/ii. KDZ 2.611-612.
Juhachido kuketsu Raiyu !fjlit. T.2529: 79.
Juhachido nenju shidai Attr. Kilkai !,gi/ii. KDZ 2.616-628.
Juhachido sata Kakuban T.2517: 79.
Kongokai dai giki ii:Jiiltl:l'l-;kiiiIiJL. Attr. Kilkai !,gi/ii. KDZ 4.466-496.
Kongokai dciihO taijuki Annen 'ti:t!.\. T.2391: 75.
Kongokai kue mikki Gengo T.2471: 78.
Mujin shOgonzo shidai Attr. Kiikai KDZ 4.497-53l.
Ninno hannyakyo nenju shidai Attr. Kiikai KDZ 4.720-752.
SahO shidai fti*;>zm. Attr. Kiikai KDZ 2.495-508.
Sanmitsu shOryoken :=:'ifii'.P''-fI'lil. Kakuch6 1!!m. T.2399: 75.
Senju Kannon gyohO shidai =f'FlIJ!i'fl'i*;>zm. Attr. Kiikai KDZ 2.533-554.
Shari hO Attr. Kiikai KDZ 4.770-773.
Sheng huanxitian shifa Trans. Prajfiacakra (Zhihuilun
T.1275: 21.
Shiba qiyin -t- Attr. Huiguo T.900: 18. (This is a reprint, with minor
changes, of the Juhachi geiin; see above.)
Shingon denju sahO Attr. Kiikai KDZ 4.417-422.
Shonyoirin Kanjizai bosatsu nenju shidai Geng6
Shugokyo nenju shidai Attr. Kiikai KDZ 4.755-769.
Taizo bizai shidai JIiliilXiilii';>zm. Attr. Kiikai KDZ 4.559-617.
Taizo bonji shidai nliilXJiI::"i";>zm. Attr. Kiikai KDZ 2.247-290.
Taizokai daihO taijuki Annen *?:ft. T.2390: 75.
Taizokai nenju shidai yoshUki G6M SZ 25.1-519.
Taizokai shOki Kakuch6 T.2404: 75.
Taizokai unji shidai Attr. Kiikai KDZ 4.665-697.
Usuzoshi kuketsu ji1jl:'TIJ#t. Raiyu ili:lm. T.2535: 79.
Wuliangshou rulai guanxing gongyang yigui Trans.
Amoghavajra. T.930: 19.
Modern Sources
Abe, Ryiiichi
1999 The Weaving of Mantra: Kukai and the Construction of Esoteric Buddhist
Discourse. New York: Columbia University Press.
Andersen, Poul
2001 "Concepts of Meaning in Chinese Ritual." Cahiers d'Extreme-Asie 12:
Davis, Edward L.
2001 . Society and the Supernatural in Song China. Honolulu: University of
Hawai'i Press.
Ding Fubao TWii*, ed.
1984 Foxue da cidian 1919. Reprint, Peking: Wenwu Publishing.
Dobbins, James C.
2001 "Portraits of Shinran in Medieval Pure Land Buddhism." In Living
Images: Japanese Buddhist Icons in Context (Asian Religions and Cul-
tures), edited by Robert H. Sharf and Elizabeth Horton Sharf, 19-48.
Stanford: Stanford University Press.
Foguang da cidian zongwu weiyuanhui
1989 Foguang da cidian f1t:l't*iff!ll!. 8 vols. Taipei: Foguang chubanshe.
Gennep, Arnold van
1960 The Rites of Passage. Trans. Morrika B. Vizedom and Gabrielle Caffee.
Chicago: University of Chicago Press. First published in 1909.
Goepper, Roger '
1979 "Some Thoughts on the Icon in Esoteric Buddhism of East Asia." In Stu-
dia Sino-Mongolica, Festschriftfiir Herbert Franke, MiinchenerOstasi-
atische Studien, bd. 25, 245-254. Wiesbaden: Steiner.
Granoff, Phyllis
2001 "Images and Their Ritual Use in Medieval India: Hesitations and
Contradictions." Paper delivered at the conference "Images in Asian
Religions: Texts and Contexts," held at McMaster University and
University of Toronto, May 10-12,2001.
n.d.a "The Absent Artist as an Apology for Image Worship: An Investigation
of Some Medieval Indian Accounts of the Origins of Sacred Images."
Chapter of a Festschrift in honor of Padmanabh S. Jaini being edited by
Olle Qvamstrom.
n.d.b "Reading Between the Lines: Colliding Attitudes Towards Image Wor-
ship in Indian Religious Texts." Chapter of a volume being edited by
Gilles Tarabout and Gerard Colas.
Hakeda Yoshita S.
1972 Kilkai: Major Works. New York: Columbia University Press.
Heesterman, J. C.
1993 The Broken World of Sacrifice: An Essay in Ancient Indian Ritual.
Chicago and London: University of Chicago Press.
Hubert, Henri, and Marcel Mauss
1981 Sacrifice: Its Nature and Functions. Trans. W. D. Halls. 1964. Reprint:
Chicago: The University of Chicago Press. A translation of "Essai sur
la Nature et 1a Fonction du Sacrifice," L'Annee sociologique (Paris,
Ishida Hisatoyo
1987 Esoteric Buddhist Painting. Japanese Arts Library, no. 15. Trans. E. Dale
Saunders. Tokyo: Kodansha International. Originally published under the
title Mikkyoga Nihon no bijutsu vol. 33 (Tokyo: Shibundo, 1969).
Kamata Shigeo et aI., eds.
1998 Daizokyo zenkaisetsu daijiten Tokyo: Yilzankaku.
Kiyota, Minoru
1978 Shingon Buddhism: Theory and Practice. Los Angeles and Tokyo:
Buddhist Books International.
Matsunaga Yilkei fk&1fiJ!
1989 Mikkyo: Indo kara Nihon e no denshO -1 / B:;js:"O){;!Jjjl;. Tokyo:
ChilO koron.
1990 "Esoteric Buddhism: A Defmition." In Mikkyo: Kobo Daishi Kilkai and
Shingon Buddhism (Bulletin of the Research Institute of Esoteric Buddhist
Culture, Special Issue, October 1990), 23-40. K6yasan: Research Insti-
tute of Esoteric Buddhist Culture, K6yasan University.
Mikky6 Jiten Hensankai ed.
1983 Mikkyo daijiten *f;[:;k1i1'!l\l;. 1970. Reprint in 1 vol., Kyoto: H6z6kan.
First published in 6 vols, 1931.
Miyano Yilchi and Mlzuhara Gy6ei eds.
1933 Shidokegyo shidai 6 vols. K6yasan: Matsumoto nisshind6
honten. (This text is based on a set of manuals dated TenshO .xlE 4 [1576].)
Miyata Taisen
1984 A Study of the Ritual Mudras in the Shingon Tradition: A Phenomeno-
logical Study of the Eighteen Ways of Esoteric Recitation (Jilhachido
nenju kubi shidai: Chain) in the Koyasan Tradition - Including trans-
lation of Kilkai's "Jilhachido Nenju Kubi Shidai." Published privately.
Miyata Taisan and Dale Todaro, eds.
1988 Handbook on the Four Stages of Prayoga, Chilin Branch of Shingon
Tradition. K6yasan: Department of K6yasan Shingon Foreign Mission.
Mochizuki Shink6
1933-36 Bukkyo daijiten {!tii&:;k1i1'!l\l;. 10 vols. Tokyo: Sekai seiten kank6 ky6kai.
Nakagawa Zenky6 ed.
1986 Shidokegyo shidai: Chain K6yasan: Shinn6in.
Nakamura Hajime <pH3i;, ed.
1981 Bukkyogo daijiten 1975. Reprint in 1 vol., Tokyo: T6ky6
Ono Gemmy6
1932-36Bussho kaisetsu daijiten f!t.:mm:;k1i1'JJ!\. 12 vols. Tokyo: Dait6 shuppansha.
Oyama K6jun :;kLlJ01f
1987 Chilinryil no kenkyil Osaka: T6h6 shupp an. Originally pub-
lished as Himitsu Bukkyo Koyasan chainryil no
(K6yasan: Oyama K6jun hOin shOshin kinen shuppankai, 1962).
Ozawa ShOki IJ\lWll!ll:ijii
1962 Shido shidai: koshin hi) Kyoto: S6honzan chishakuin.
Payne, Richard Karl
1991 The Tantric Ritual of Japan: Feeding the Gods - The Shingon Fire
Ritual. Sata-Pitaka Series, voL 365. New Delhi: Aditya Prakashan.
Sawa Ryilken ed.
1975 Mikkyo jiten *ii&1i1'!l\l;. Kyoto: H6z6kan.
Sharf, Robert H.
1995 "Buddhist Modernism and the Rhetoric of Meditative Experience."
Numen 42 (3): 228-283.
1998 "Experience." In Critical Terms for Religious Studies, edited by Mark
C. Taylor, 94-116. Chicago: The University of Chicago Press.
200la "Prolegomenon to the Study of Japanese Buddhist Icons." In Living
Images: Japanese Buddhist Icons in Context (Asian Religions and
Cultures), edited by Robert H. Sharf and Elizabeth Horton Sharf, 1-18.
Stanford: Stanford University Press. .
2001 b "Visualization and Mandala in Shingon Buddhism." In Living Images:
Japanese Buddhist !cons in Context (Asian Religions and Cultures),
edited by Robert H. Sharf and Elizabeth Horton Sharf, 151-197. Stan-
ford: Stanford University Press.
2002 Coming to Terms with Chinese Buddhism: A Reading of the Treasure
Store Treatise. Kuroda Institute Studies in East Asian Buddhism, no. 14.
Honolulu: University of Hawai'i Press.
n.d. "Ritual." In Critical Terms for Buddhist Studies, edited by Donald S.
Lopez, Jr. Chicago: University of Chicago Press, forthcoming.
Staal, Frits
1979a "The Meaninglessness of Ritual." Numen 26 (1): 2-22.
1979b "Ritual Syntax." In Sanskrit and Indian Studies: Essays in Honor
of Daniel H. H. Ingalls, edited by M. Nagatomi, B. K. MatilaI and
J. M. Masson. Dordrecht: Reidel.
1983 Agni: The Vedic Ritual of the Fire Altar. 2 vols. Berkeley: Asia Human-
ities Press.
1990 Rules Without Meaning: Ritual, Mantras and the Human Sciences.
Toronto Studies in Religion, vol. 4. New York: Peter Lang.
Strickmann, Michel
1989 "Canons of Giant Art: Ritual of Land and Water." Unpublished paper
presented at the conference "Art and the Emperor: A Public Interdisci-
plinary Symposium on the Arts of China," Ohio State University, 14
April 1989.
2002 Chinese Magical Medicine. Edited by Bernard Faure. Asian Religions and
Cultures. Stanford: Stanford University Press.
Takai Kankai .
1953 Mikkyo jiso taikei Kyoto: Takai zenkeshu chosaku kankokai.
Tanaka Kaio
1962 Mikkyo jiso no kaisetsu Tokyo: Rokuyaon.
Todaro, Dale Allen
1985 "An Annotated Translaton of the 'Tattvasal1lgraha' (Part 1) with an
Explanation of the Role of the 'Tattvasarp.graha' Lineage in the Teach-
ings of Kilkai." Ph.D. dissertation, Columbia University.
1986 "A Study of the Earliest Garbha Vidhi of the Shingon Sect." Journal of
the International Association of Buddhist Studies 9 (2): 109-146.
Toganoo Shoun
1982a Himitsu Bukkyoshi 1933. Reprint, Toganoo ShOun zenshu
voL 1. Koyasan: Koyasan daigaku mikkyo bunka kenkyfijo.
1982b Himitsu jiso no kenkyu jfI;*11lOJliifyj;. 1935. Reprint, Toganoo ShOun zen-
shu vol. 2. Koyasan: Koyasan daigaku mikkyo bunka
Tylor, Edward B.
1920 Primitive Culture: Researches into the Development of Mythology, Phi-
losophy, Religion, Language, Art, and Custom. 2 vols., 6th ed. London:
John Murray.
Veda Reij6
1986 Shingon mikkyo jiso gaisetsu - shidobu - shin'anryu 0 chU"shin toshite
'" .c.' C: G"[. Tokyo: D6b6sha.
Yamasaki Taik6
1988 Shingon: Japanese Esoteric Buddhism. Trans. Richard and Cynthia Peter-
son. Boston: Shambhala.
The SrIghanacarasailgraha is a Sanskrit text in verses on the conduct of
Buddhist novices. It is extant only in the fonn of quotations found in a
commentary on it, the Sphutarilia SrIghanacarasailgrahatIka, written by
J 1. The name of the author of the verses is unknown

mentions three other commentators
who worked on the same
verses, but their commentaries are not extant. The SrIghanacarasailgraha
was therefore an important text for atleast one monastic community.
According to Singh, "the text probably belongs to the Mahasamghika
school"4, because of the following passage
1 The text was first edited in 1968 by Sanghasena. In 1983 the same scholar (as Sang-
hasen Singh) republished the same text (with different pagination), adding a translation and
a reconstruction of the original verses. In this paper I always refer to this edition as Singh
1983. In the same year, Derrett published his own translation, based on the 1968 edition
and on a microfilm (Derrett 1983: 5, n. 1). The two translations are independent from one
2 Derrett (1983: 6) took the term irfghana as the "pen-name" of a "vinaya specialist"
who authored the verses. He knew (ib.: 14, n. 3) that according to lexical sources this
term may be applied to buddhas (add now examples in Handurukande 2000: 6 and in
inscriptions from the eleventh century in Tsukamoto 1996-1998: I 154, 200). Derrett (1983:
14, n. 3) even regarded it "as quite possible that really believed a Buddha
called Srlghana wrote the verses!" However, the term irfghana is used in the text in the
meaning of 'novice' (Singh 1983: 3-4). In 1961 Singh, too (paper re-published in Singh
1983: "Appendix i", p. 241), had taken the term irfghana as the name of the author, but
in the same year he pointed out that this term in the text merely refers to novices (paper in
PaIi of 1961, published as Singh 1974 and again as Singh 1983: "Appendix ii"). Shimoda
(1990: 495) also took irfghana as meaning 'novice'.
3 bhadanta bhadanta Prajiiasirhha, and bhadanta Dharmavalokitamitra
(Singh 1983: 57,63, 119 = ff. 19a, 28b, 93b).
Journal of the International Association of Buddhist Studies
Volume 26 Number 1 2003
yadi kascit agatyagantukaJ:t a[bhivadayati], tadanena
vaktavyam " . katamas te nikayaJ:t, kati ca tasya nikiiyasya bhedaJ:t '" so 'pi
yadi aryamahiisiimghiko bhavati tadiinena vacyam aryamahiisiimghiko 'smi
/ tasya ced bhediiJ:t
viidinas ciirthasiddhiirthiiJ:t sailadvayaniviisinaJ:t /
bhiidrayana haimavataJ:t mulasiimghikiiJ:t / /
If any arriving [ascetic] addresses one settled and seated, then the latter
should say io him: " ... Which is your nikiiya? How many divisions are
there in your nikiiya? ... " And if he is an Arya-Mahasarilghika, then he
should say: "I am an Arya-Mahasamghika". Here are their divisions:
1. Vadins, 2. Mhasiddharthas, 3. 4. Sailadvayanivasins, 5. Bhadrayanas,
and 6. Haimavata; the Mi1lasamghikas are divided into six [nikayas].
Derrett, however, made a distinction between the original verses, which
constitute the SrlghanacarasaiJ.graha, and the prose commentary by
According to Derrett, the passage quoted above is merely
evidence that the commentator is evidently interested in the
Mahasfu:hgbikas but it has been doubted whether he did appertain to that
sect,,6. As for the author of the verses, Derrett's "impression" is that he
"worked for all nikdyas, and deliberately eschewed allegiance to any"
(ib.). Therefore, I assume, Derrett did not think that the verse quoted
above belongs to the mula text, but implied that it is a mnemonic loka
produced by himself, and indeed Singh did not use it in his
reconstruction of the mi1la verses
. Derrett also noticed that the author of
the verses refers to a sequence of the ten precepts unknown to any other
school, including the Mahasfu:hghikas

4. Singh 1983: 7. In a later article (id.: 1986), the work is Mahasamghika in the title,
but only "probably" so in the text (ib.: 6). Singh (1983: 7) also maintained that "the word
'sanvara' [sic] for the conduct of a 'sramru;tera' " implies Mahayana, but it doesn't. The only
Mahayana element is the mention of Mafijusri in the mangalaslokas (ib.: 45).
5 Singh 1983: 119-120. Translation as in Derrett (1983: 80-81; slightly modified).
6 Derrett 1983: 7. He refers to Ejima (1976: 918-919). Following Ejima, Yuyama, in
his survey of Vinaya literature (1979: 39
40), classifies commentary as a
Mahasamghika text, but adds a question mark. However, Shimoda (1990: 492, n. 4) notices
that Ejima does not give any reason for doubting a Mahasamghika affiliation.
7 According to Derrett (1983: 6, 82), the original verses numbered 102. According to
Singh (1983: 219 and 239, n. 128), they numbered 200, although this figure does not need
to indicate the exact number (ib.: 286). Both rely on the expression satadvayeneti (Singh
1983: 121). Singh tries to reconstruct exactly two hundred verses in his "Appendix vi"
(ib.: 289-313). Shimoda (1990: 495-494) agrees with Singh.
More recently Masahiro Shimoda, disagreeing with Derrett, has sought
toprove the Mahasamghika affiliation of the commentator,
By comparing passages from the section on theft (adattadana) in
commentary and in the Mahasamghika Vinaya, Shimoda
was able to show that both texts agree very much in terms of cOntents,
although the agreement "is not word for word"lO. This result, based only
on this section, led him to conclude that work must be a
summary from the Mahasamghika Vinaya
Shimoda's arguments are not conclusive. He did not explain why the
order of the ten precepts in work is different from the one
in the Mahasamghika Vinaya. Also, he only examined one section of the
commentary to the Srighanacarasailgraha, a section that does not contain
any verbatim quotations from the Vinaya. Therefore, he only showed that
knew a Vinaya which is similar to the Mahasamghika Vinaya.
We do not know yet if it was identical to it. For, similarity does not
entail identity, as it is known for example that two texts belonging to the
Lokottaravadins, the and the Abhisamacarika-Dharma,
are similar, but not identical to the corresponding sections of the
Mahasamghika Vinaya. Moreover, Shimoda did not disprove Derrett's
contention that the author of the verses "worked for all nikayas, and delib-
erately eschewed allegiance to any". For, Shimoda only worked on
commentary, not on the verses
8 Derrett 1983: 8, where a comparative table of the ten precepts is given.
sequence of the ten precepts is as follows (Singh 1983: 51-52): 1. killing, 2. stealing,
3. sexual intercourse, 4. lying, 5. drinking liquor, 6. high beds and seats, 7. dancing, singing,
and playing music, 8. perfumes, garlands, and unguents, 9. eating at the wrong time,
lD. taking gold and silver. The greater part of the text is made of ten sections on each of
the ten precepts. If one exchanges item 6 with item 8, the result is the sequence of the
Mahiisamghikas, of the Dharmaguptakas (who add suicide as the eleventh precept), and
of the Abhidharmakosa (see Derrett 1983: 8).
9 Shimada 1990. See also Shimada 1987, where the text is taken to be a Mahiisamghika
commentary on the ten precepts.
10 Shimada 1990: 494.
11 Shimada 1990: 495. Prebish (1994: 60-61) seems to follow Shimada and classifies
J s work as a Mahiisamghika text.
12 Shimoda 1990: 495.
Here I shall present evidence to deternrine whether both the author of the
verses and refer to one and the same Vinaya, and to indicate
how close this Vinaya was to the Mahasamghika Vinaya, extant in Chinese.
J comments on five items that define homicide and that
were mentioned by the author of the verses in the following order:
1. upakrama (taking a weapon, etc.), 2. n[sarhjiiii (the idea that one is a
man), 3. nara (there is man), 4. vadhakacetanii (there is the intention of
killing), and 5. jlvitasya k!jaya (destruction of life)13. We are not told
whether these terms occur in the Vinaya. complains that
some "stupid" (mandadhiya/:t) fellows misread the verse: they read nare
vadhakacetanii instead of the correct naro vadhakacetanii. Since the
verse in question is about a list of five items in the nominative case, the
wrong reading would yield a list of four items only, and the Vinaya
would be "curtailed"14. Therefore, has to justify his reading
by quoting an analogous list, not exactly the same one, from his Vinaya
hi vinaye nirddalJ / 1. prclIJlca bhavati, 2. prllIJisamjfil ca bhavati,
3. vadhakacittafi ca pratyupasthitam bhavati, 4. upakramafi ca karoti, 5. jlv-
itl1d vyavaropito bhavati iyatl1 prl1lJlltiplltl bhavati /
For, this is the explanation in the Vinaya: "TIlere is a breathing being. One is
conscious that it is a breathing being. A thought of killing is present. One starts
to act. One is deprived of life. To this extent is one a killer of a breathing being".
This passage in itself is probably pan-Buddhist. For example, parallel
passages occur in Theravada commentaries
. The style, moreover, has the
13 ebhir aizgair bhtivatfti darsayan params codayann aha / upakrama
ityadi / tatra sastradigrahaJ;am / iti samjfia nrsamjfia / kadacid
upakramam karoti nrsamjfia bhavati na tv asau ity aha / naTa iti / yady asau
bhavati / kadacid etani tril}y aizgani sambhavanti, na vadhakacetaneti / ato
vadhacetanavacanam / kadacid ... na ttl jfvitad vyaparopayatfti / ata aha / jfvitasya
ceti (Singh 1983: 59).
14 nare vadhakacetaneti saptamyantam pa?hanti / paficaizgani na siddhyanti nare
vadhakacetanety asya padasyaikaizgatvtit / vinayas ca tair vilopito bhavati (Singh 1983:
15 Singh 1983: 59.
16 See references in Saddhatissa 1970: 89, n. 1.
flavor of a later scholastic elaboration 17. Of all extant Vinayas, only the
Mahasamghika Vinaya has anything similar. Here, both the section on the
third piiriijika (killing human beings) and the section on the sixty-first
piitayantika (killing animalS)IS include comparable lists. In the section on
the third piiriijika we read
If one fulfills five conditions and kills a man, one commits a piiriijika offence:
1. a man, 2. the idea of a man, 3. starting to find a means [to kill], 4. the
thought of killing, 5. cutting off the life [ofa man]. These are called the five
In the section on the sixty-first piitayantika we read
If a monk is possessed of five dharmas and cuts off the life of an ani-
mal, [he is guilty of a] piitayantika offence. What five? [1.] An animal,
[2.] the idea of an animal, [3.] the thought of killing, [ 4.] arising of
bodily action, [5.] cutting off the faculty of life. These are called the five
The Sanskrit list mentioned by author of the verses agrees with the
Chinese list from the piiriijika section, but the order is different. It is
impossible to know whether the author of the verses took these terms
directly from a Vinaya. who is quoting a Vinaya, mentions
a list that corresponds to the Chinese one from the piitayantika section,
in the same order. For, he uses the word priilJin (lEt, 'animal'), not
nr/nara (A, 'man'). Given that context is 'killing human
beings', it is strange that he does not quote the passage from the perti-
nent section, using the same terms that the author of the verses used.
Perhaps, Vinaya did not list these five items in the section
on killing human beings, and therefore quotes an analogous
list from the section on killing animals. On the basis of this comparison,
17 So Gombrich (1984: 99), referring to the analogous list in the Pili commentaries.
18 I use the Sanskrit piitayantika instead of any middle Indic form for convenience of
reference with the comparative table of the monastic precepts published by Rosen (1959:
19 -'if .A. 0 = 'if =:'ifJ!!1J{f 0 [Il'if4l1:,L.'o 1l'if1!li
frro (T.1425 xxn 257c3-5).
20 Ii:'!:. Ii:'!::'''\!o 4I1:,L.,. frr */lilio :&\:
(T.l425 xxn 378a24-25).
Vinaya was similar to the Mahasamghika Vinaya, but per-
haps it was not identical to it.
refers to his Vinaya also in the context of ukravisHti,
'emission of semen'. This is an offence, unless it occurs in a dream.
pafica svapnal; vinaye uktal; / satyasvapno yathii bodhisattvena dutal;,
allkasvapno yathii dHtal; tathii na bhavdty allkam mHeti k[tva,
acfn:/asvapno yat satatakarm/iyam vastu d[Syate, anantasvapno yal; sakalam
ratrim drsyate na paricchidyate, svapnasvapno yal; svapna evanyal; svapno
d[Syate /
Five [kinds at] dreams are mentioned in the Vinaya:
1. A truthful dream as seen by a bodhisattva.
2. A false dream. [A real occurrence] is not the same as it is seen [in a
dream], taking 'false' as 'deceitful'.
3. An unfulfilled dream. Something is seen that remains to be done.
4. A ceaseless dream, which is seen during the entire night and is not
S. A dream in a dream, i.e. another dream which is seen within the very
same dream.
Only the Mahasamghika Vinaya, in the corresponding section, mentions
five types of dreams. Still, there are some differences
. M.E.-.

. . . .
. .

21 Singh 1983: 92. My translation is mainly based on Derrett (1983: 56), and less on
Singh (1983: 186-187). The Pali Vinaya-atthakatha in the same context mentions
four dreams: annatra supinantii ti ettha supino eva supinanto, tam rhapetvii apanetvii ti
vuttam hoti. tan ca pana supinam passanto catuhi kiiralJehi passati dhiitukkhobhato vii
anubhUtapubbato vii devatopasamht'irato, vii pubbanimittato vii ti (Samantapas1idika III
520 = T.l462 XXN 760a2- ... ).
22 T.1425 XXII 263b8-16.
There are five types of dreams. What five 7 1. A truthful dream, 2. an untruth-
ful dream, 3. an unclear dream, 4. a dream in a dream, and S. dreaming later
what someop.e has thought of earlier. These are the five.
1. What is a truthful dream? The Tathagata, when he was a bodhisattva, saw
five dreams [which were] not different from the truth. This is called a
truthful dream.
2. Untruthful dream: if a man sees a dream and, when he wakes up, it is
not true. This is called an untruthful dream. .
3. An unclear dream: if one does not remember the beginning, the end, and
the middle part of one's dream.
4. A dream in a dream: if a man sees a dream, and then in [that] dream he
tells a dream to [other] men, this is called a dream in a dream.
S. As for dreaming later what one has thought of earlier, if one dreams at
night what has been done and thought during the day, this is called
'dreaming later what one has thought of earlier'.
first and second types correspond to the first and the
second ones of the Mahasamghika Vinaya. But, fifth type
corresponds to the fourth one of the Mahasamghika passage.
third and fourth types do not correspond to any type in
the Mahasamghika Vinaya. Therefore, from this passage it appears
that Vinaya was similar, but not identical to the
Mahiisamghika Vinaya.
teaches how a junior monk should salute a senior one. In this
context, he says23: tasyarh jaflghiiyarh sphotarh na dadyat / kim
ivety aha / avina yathii, "one [a junior monk] should not make a smack-
ing sound with the shaven head against the lower legs [of a senior monk]
( ... ). Like what? As by a sheep "24 The words avina yathii are part of
the original verses, as they are introduced by aha. goes on to
say that bhadanta Avalokitarnitra read ravina yathii (ib.). He is wrong,
says because "in the Vinaya only the example of a ram is
23 Singh 1983: 119:
24 Derrett 1983: 80. Skr. spho{a also means 'boil'. Although it is difficult to see how
this meaning could fit into this sentence, boils are part of the context, as it is clear from
the passages quoted below.
given", vinaye rnerjhakasyaiva d[.Jtdntaddndt25. Therefore, one should find
this example in the Mahasamghika or related Vinayas. .
As Derrett noticed
, this entire section is very close to a passage from
a canonical Vinaya text extant in hybrid Sanskrit, the Abhisamacarika-
Dharma of the Lokottaravadins. The Mahasamghika Vinaya also has a cor-
responding chapter on rules of deportment which often runs
exactly as the Lokottaravadin chapter. I now present two parallel pas-
sages from both Vinayas, where the example of the ram occurs. I divide
both passages into four paragraphs. For each paragraph I quote the Chinese
of the Mahasamghika Vinaya, my translation of it, and the Sanskrit from
the Lokottaravadin Vinaya. I shall then point out any correspondence
with the wording of the Vinaya quoted by the author of SrIghanacarasan-
graha and by
"One may not, covering the head, covering the
right shoulder, wearing leather shoes, make a salutation".
na k$amati I ogulJ.{hitakiiyena na k$amati I ohitahastena na k$amati / upiina-
hiiruq.hena siimfcfkarentenal
"One may not revere the knees, revere the legs,
revere the shins. One should revere by touching the feet".
na k$amati I jiinukena viijanghiihi vii vanditum I atha khalu piidii vanditavyii I
:ll:tl)\l'WtlQ1l52$l'il!o "The man who receives the salutation may
not keep silent like a dumb sheep. He should reply with polite questions".
na diini melJ.q.hena viya iisitavyam piidehi vandayantehi I atha khalu prati-
sammodayitavyam I [in the original text this paragraph is the fourth one]
*ru)\Jl'ilP...t:ff:!fo "If the person in front has a boil on the foot,
one should take care not to hit it with the head".
piidiim vandantena jiinitavyam I yadi kasyaci vralJ.ii bhavati I galJ.q.o vii
pi{ako vii na diini sahasii utpfq.itavyam I athii khalu tathii vanditavyam yathii
25 Derrett 1983: 80. For meejhaka cf. Sanskrit and Piili melJeja.
26 Derrett 1983: 80 nne 4, 5, 9.
27 MahasfuiIghika Vinaya, T.1425 XXII 510b18-21; Lokottaravadin Vinaya,
Abhisamacarika-Dhanna (Abhisamacarika-Dhanna Study Group 1998: 120; I keep the
dalJejas, although they do not make much sense).
na dulJkMpiye padehi vandayantehi / [in the original text this paragraph is
the third one]
As for the fJist paragraph, efr.
idanirh tu yadrgvidhavasthavasthitena navakena yatina na vanditavyarh,
tatM darsayitum aha / avagulJthitetyadi / avagulJthitarh pidhitarh Sfr-!arh
slro yasya tena .. ./ ... tena upanaMrutjhenety /
This passage shows that the author of the verses used the tenn
'by [a cleric] whose head is covered/veiled'. This tenn
is similar, but not identical to the Lokottaravadin tenn ogul}thitakayena, 'by
[a cleric] whose body is covered/veiled'. The reading of the author of the
verses better corresponds to the Chinese Mahasamghika Vinaya phrase "cov-
ering the head". As for the Lokottaravadin tenn upiinahiirurj.hena, "by one
who wears sandals", the author of the verses used the synonym sopiinatkena,
probably metri causa, but clearly refers to the original reading.
The Chinese phrase "wearing leather shoes" may correspond to both tenns.
As for the second paragraph, efr. jiinu ca jiinu ca jiinunr,
tayor jiinuyor yii jaftghii tasyiirh jaftghiiyiirh mUl}rj.e[ na] sphotarh na dadyiit
(ib.). This passage shows that the author of the verses used the tenns jiinu
and jaftghii, 'knee' and 'shin', found in both the Mahiisarhghika and
Lokottaravadin Vinayas.
As for the third paragraph, efr. kim ivety iiha I avinii yathii I
avir I yathii avinii mel}rj.akena dvayor jiinunor jarhghiiyiirh hanyate,
tadvad yatir api v[ddhiintikasya yater jarhghiiyiirh na vandeteti yiivat I ...
vinaye merj.hakasyiiiva dutiintadiiniit (ib.). This passage shows that the
author of the verses used the word avi, 'sheep', but this tenn according
to J is only a substitute for the Vinaya reading merj.haka, 'ram'.
This Vinaya tenn is indeed found in the Lokottaravadin Vinaya, in the
fonn mel}rj.ha
A similar word, if not the same one, was in front of the
Chinese translators 30. Therefore, Vinaya seems to be close
28 Singh 1983: 119.
29 Jinananda (1969: 125), however, reads meriitena, which he emends into
Prasad (1984: 134) follows this emendation in his paraphrasis and takes it as meaning
'banker': "The senior monk whose feet are to be greeted is not to sit like a banker".
30 JQOiil$, 'like a dumb sheep'. According to the Mvy (7684), the similar Chinese phrase
Il52JQ$ translates etjamuka, 'dumb like a sheep' (Tibetan lug ltar lkug pa), which is also
to both the MahasariJ.ghika and Lokottaravadin Vinayas. But, in an impor-
tant point the Srlghanacarasailgraha and its commentarY differ from both
of them. In the Mahasamghika and Lokottaravadin Vinayas, it is the cleric
who receives the salutation who should not be "dumb like sheep", i.e. he
should not keep silent, but should say some polite words. More awkward
is understanding of his own Vinaya, in consonance with the
author of the verses: it is the junior cleric who makes the salutation who
should not hit with his head - like a sheep or a ram - the shins of the
senior cleric. It is difficult to decide whether this understanding is based
on a different wording of Vinaya or on a misinterpretation
of it. Bhadanta Dharmavalokitamitra, as shown above, read ravi, 'sun',
instead of avi, 'sheep', probably because he did not know how to inter-
pret the example of the 'sheep'31. Therefore, the correct understanding of
the example of the sheep was indeed a problem. Perhaps an examination
of the fourth and last paragraph will shed some light on this.
As for the fourth paragraph, cfr. yadii samavasthiiniflat:llJO
yatil:z piidarogel!a gliinal:z syiit ... 32. here is not quoting, but
only referring to a passage close to the last paragraph from both the
MahasariJ.ghika and Lokottaravadin Vinayas, quoted above. This shows
that Vinaya had a passage corresponding, in some form, to
that paragraph. The order of the last two paragraphs in both Vinayas is
the opposite of the one found in In the Lokottaravadin Vinaya
the last two paragraphs follow each other as follows:
padarh vandantena janitavyarh / yadi kasyaci vral:za bhavati / gal:Zo va
pitako va na dani sahasa utpftjitavyarh / atM khalu tatM vanditavyarh
yatM na duJ:tkMpiye padehi vandayantehi / [new paragraph] na dani
me1}tjhena viya asitavyarh padehi vandayantehi / atha khalu pratisarhmo-
dayitavyarh /
attested in the form erjakamuka and corresponds to PaIi e{amuga. See BHSD s.vv. erjaka-
muka and edamuka.
31 Acco;ding to Derrett's emendation and translation, Dharmavalokitamitra's explana-
tion is that "Just as the sun is not to be saluted (?) by one wearing shoes [ ... J so an asce-
tic who has his shoes on should not salute" (Derrett 1983: 80, corresponding to Singh 1983:
32 Singh 1983: 119.
[ ... the feet must be revered]. [The cleric] who reveres the feet must know
if some [cleric in front of him] has a boil, or a pimple, or a blister [on a foot].
He must not inconsiderately sqeeze it, but he must revere [the feet] so as not
to cause pain while the feet are being revered. [new paragraph] [The cleric
who is being revered] must not sit like a ram, while his feet are being revered,
but should make a salutation in return.
It is possible that the author of the verses and J had in mind this
paragraph order, and that they construed na diini melJ4hena viya with the
preceding paragraph: tatM vanditavyarh yathii na du/:lkMpiye piidehi van-
dayantehi I na diini melJ4hena viya, "the salutation must be made so as not
to cause pain while the feet are being revered, not as if [it were made] by
a ram". To be sure, the word diini does not allow such a construction, but
one does not need to assume that had the word diini in his mind
or in his Vinaya. This construction would explain why the author of the
verses and compare the junior cleric to a ram that hits the shins,
as opposed to comparing the senior cleric to a ram that merely keeps silent.
In short, a comparison with these four paragraphs has shown that: 1. the
Vinaya quoted or implied by the author of the SrIghanacarasaIigraha and by
was generically similar to both the Mahasamghika and the Lokot-
taravadin Vinayas; 2. the term used by the author of the
verses, is found in the Mahasamghika Vinaya, not in the Lokottaravadin
Vinaya; 3. the author of the verses and understanding of the
example of the ram/sheep is at variance with both the Mahasamghika and
the Lokottaravadin Vinayas, but could be based on the word order of the
Lokottaraviidin Vinaya, or of a similar one; 4. the author of the verses and
had in mind the same Vinaya, contrary to what Derrett thought.
Another passage shows that not only but also the author
of the verses depends on the Mahasamghika Vinaya or on a very similar
one. It occurs in the section on 'false speech' which corre-
sponds to the fIrst piitayantika in the Vinaya. says that one
becomes a liar because of four elements (aflga): 1. there is
some matter; 2 there is a man who is aware that [something] is false; 3. his
mind is directed toward [lying]; 4. he is aware that it is false speech;
5. he utters words
A few lines below, J adds that the last
four, or the last three, or the last two, or even the very last miga are
enough to define a liar
From quotations of the original
verses, it appears that the author of the SrighanacarasaIigraha himself
referred to the same theory35. Of all extant Vinayas, only the
Mahasamghika Vinaya has anything similar, although it might be based
on a somewhat different terminology36:
... ... ... =$ ... ... !<Qllff:'Ofto
If one, being possessed of five dharmas, consciously lies, one incurs a piitayan-
tika offence. What five? 1. there really is [such and such matter], 2. awareness
that there is [such and such matter], 3. one turns his mind [to it], 4. awareness
of disobeying [a precept], and 5. one utters words different [from the truth].
... If one is possessed of four dharmas [i.e. the last four] ... three dharmas ...
two dharmas ... one dharma ... and consciously lies, it is a patayantika
The Chinese renderings of some of these five items are not completely
clear to me, as my translation shows, but they are clear enough to indi-
cate the similarity with the five items mentioned by the author of the
Srlghanacarasailgraha and by
uses the technical term arthotpatti, which can be translated
as 'particular case>37. This term is peculiar to the Vinaya of the
33 katibhilJ punar angailJ sytid ity tiha I vastu eety tidi slokalJl vastu ea bha-
vati, alfkasmnjiif ea bhavati, vinihitam eittam bhavati, bhavati, vticam ea
(Singh 1983: 99).
34 na kevalam paiieabhir angailJ samprajtinamHtivtido bhavati, kin tu hy ekentipfti
dadayann tiha I eatustrfty tidi I .,. latra caturbhir angailJ bhavati I alfkasamjiif
cety tidi / alfkasamjiiitvena vastunalJ parigrahtin na prthan nirdisyate ... (similarly for the
other angas; Singh 1983: 100)
35 See the reconstruction of slokas 144-145 in Singh (1983: "Appendix vi", p. 306)
and the underlined words in the quotations given in the two preceding notes.
36 T.1425 xxn 325b2-12.
37 See Roth (1970: 109, 142, n. 1) and Nolot (1991: 376, n. 2). The former translates
arthotpatti as the "arising of a particular case" (ib.), the latter as "cas particuliers" (ib.).
Lokottaravadins: after a story that leads to the promulgation of a precept,
other stories follow which represent 'particular cases'. An arthotpatti
does not constitUte a separate precept. This term is conspicuously absent
in the Mahasamghika Vinaya.
J has to comment on a verse that forbids a novice to drink
water containing living beings and then on other verses that forbid a
novice to use. water containing living beings in order, for example, to
water plants. An opponent says that the first verse is redundant, because
it prohibits what is implicitly prohibited in the other verses

defends the author of the verses as follows
: kin cdrthot-
pattivasdn na likhitety arthotpattiprabhdvatvdd vinayasyeti
krtvd, "furthermore, there is no fault that '[the verse] is not [to be]
written because it is a special case' [i.e part of a precept], because the
Vinaya is the source for [what should or should not be accepted as]
special cases". Even though my translation might need improvement
it is clear that uses the term arthotpatti in the same tech-
nical sense as found in the of the Lokottaravadins.
His point is that in the Vinaya itself the rule 'not to drink water con-
taining living beings' is not an arthotpatti contained in the section
about the precept 'not to use water containing living beings'. He and
his readers knew that they are two different precepts41, and therefore
one cannot fault the author of the SrlghanacarasaIigraha for devoting
verses to both of them.
One does not need to assume that Vinaya contained the
term arthotpatti. Even though this term is not represented in the Chinese
version of the Mahasamghika Vinaya, it could have been part of
Mahasamghika exegetical terminology. Therefore, mention
38 atha kimartham iyam kiirikii prthag vyavasthiipyate? ... yatra hi
sutariim tatra (Singh 1983: 62).
39 Singh 1983: 62-63.
40 Derrett (1983: 29; bracketed words are his own) translates: "Moreover there is no
harm if it is written [del. na], so as to bring out the meaning, since the strength of the vinaya
is its meaningfulness". Singh (1983: 149) translates: "That is why, there is no fault (here
in composing separately) taking into account that the Vinaya has got the effect of mean-
ing some thing". Both miss the technical meaning of arthotpatti.
41 Piitayantika 19 (about using water) and 51 (about drinking water) in the
Mahiisarilghika Vinaya.
of the term arthotpatti does not exclude his affiliation to Mahasamghikas
and does not prove his affiliation to the Lokottaravadins.
Finally, we should notice the occurrence of term 'Vinaya
of monks'. The commentator J says that the original verses are
merely "an excerpt from the and maybe the author of
the verses himself had already used this term
It is not a common term
because it does not correspond to any section of the Vinayas of most
The usual arrangement is: vibhaitga, divided into
haitga and and a section made of various chapters
called khandhakas or vastus
The Vinayas of the Mahasamghikas and
of the related Lokottaravadins, however, have a different arrangement: the
is made of a with its own prakfrl)aka (cor-
responding to the khandhakas or vastus), and the is made
of a with its own prakfrl)aka
The Lokottaravadin
is indeed extant in Sanskrit. Therefore the term
in commentary most probably refers to the
appropriate section of the Mahasamghika Vinaya
, or of a similar one.
The evidence so far presented clearly shows that the author of the
SrIghanacarasaIigraha and its commentator J knew the same
42 bhik:fuvinayiit samuddh[tam aciiriintaram na tu svamanf:fikayiinyat "rtam iti daday-
itum iiha / vinaya iti (Singh 1983: 121). According to Singh's reconstruction, the entire
compound bhik:fuvinaya was part of the original verse, although this is not certain (ib.,
"Appendix", p. 313, kiirikii 199).
43 It occurs in the commentary to the Pali Vinaya, but it refers to 'discipline', not to a
. section of the pari Vinaya: vinayaril paccakkhiimf ti na vevacanena paccakkhiinam !
bhikkhuvinayam paccakkhiimi bhikkhunzvinayam ... paccakkhiimf ti evam iidinii vinayave-
vacanena sikkhiipaccakkhiinam hoti (Samantapasadika I 252).
44 Hirakawa 1982: 14-15.
45 Hirakawa 1982: 16-18. The Sarvastivada Vinaya also has a similar structure (ib.).
46 As already suggested by Singh (1983: 238 n. 127).
Vinaya, but it does not definitely solve the problem of their precise nikiiya
affiliation. Their Vinaya appears to have been generally close to the
Mahasamghika and Lokottaravadin Vinayas, but in some details it differed
from both of them. If one dismisses these differences as unimportant and
due to the vagaries of the tradition or the imperfection of the Chinese
translation, one will have to be generic on the nikiiya affiliation of both
the author of the verses and J they belonged either to the
Mahasamghikas or to any related nikiiya. If one stresses the importance
of these differences, one will have to maintain that both the author of the
verses and were neither Mahasamghikas nor Lokottaravadins,
but belonged to a different nikiiya related to the Mahasamghikas.
I am inclined to give importance to these differences, because - as
noted above - both the author of the verses and accept a
sequence of the ten precepts unknown to any other nikiiya. Derrett sug-
gested that the author of the verses used this "curious order" intentionally,
because he wanted to address all nikiiyas
. However, I have shown that
there is no reason to doubt that the author of verses and J referred
to one and the same Vinaya, and it is therefore more simple to suggest that
also their unique sequence of the ten precepts belonged to one Vinaya
school. While the Mahasamghika sequence is known, the sequence adopted
by any sub-schools is not;48. Therefore, the available evidence strongly
suggests that the author of the SrIghanacarasailgraha and were
47 Derrett 1983: 7.
48 In principle, an unknown sequence of the ten precepts can be inferred from a known
sequence of the eight precepts. For, the wording and order of the ten precepts taken by
novices is very similar to the wording and order of the eight precepts taken by laypersons
during fasting days: two precepts taken by novices - abstention from dancing etc. and
abstention from perfumes etc. - correspond to one precept taken by laypersons; novices
also abstain from taking gold and silver (see e.g. Gombrich 1991: 78). A short text on lay
precepts, the (unknown nikiiya affiliation), lists the eight precepts of the weekly
fast in the following order: t\
1lf/l'illr.\'iJl\\':;k:,*. HH!i\t;;&. (T.1486 XXIV
1023c29-1024a3). If one splits the seventh precept into two and adds the precept about gold
and silver, the result is sequence. However, the same text, after the passage
just quoted, adds some verses where the eight lay precepts are listed in a different order:
(ib. a5-6). From this order one can derive the
list of the ten precepts of the Mahasarnghikas, of the Dharmaguptakas, and of the Abhid-
harmakosa (see table in Derrett 1983: 8).
not Mahasamgikas, but belonged to a nikiiya that was related to the
Mahasamghikas. '
Abhisarnacarika-Dhanna Study Group 1998. f:icff!tmlrul.1li!:tml
'if; . A Guide to the Facsimile Edition
of the Abhisamacarika-Dharma of the Mahiisa1?1ghika-Lokottaravadin
(Tokyo: The Institute for Comprehensive Study of Buddhism, Taisho Uni-
BHSD = Franklin Edgerton, Buddhist Hybrid Sanskrit Grammar and Dictionary.
Volume II: Dictionary (New Haven: Yale University Press, 1953). Reprints,
Delhi: Motilal Banarsidass, 1970 etc.
Derrett, J. Duncan M. 1983. A Textbook for Novices: layaraksita's Perspicu-
ous Commentary on the Compendium of Conduct by Srlgh"ana (Torino:
Indologica Taurinensia) (Pubblicazioni di Indologica Taurinensia XV).
Ejima, Yasunori 1976. "Daishubukei riUen bonpon ni tsuite"
(;: -:> c' "( [Some Observations on the Sanskrit Vinaya Texts
of the Mahasamghikas], in Okuda lio Sensei Kiju Kinen "Bukkyo Shiso
Ronshi"t" [Studies in Buddhist Thought
Dedicated to Professor Iio Okuda in Commemoration of His Seventy-
Seventh Birthday] (Kyoto: Heirakuji Shoten pp. 911-922.
Gombrich, Richard. 1984. "Notes on the Brahminical Background to Buddhist
Ethics", in Gatare Dhammapala, Richard Gombrich, and K.R. Norman, Bud-
dhist Studies in Honour of Hammalava Saddhatissa (Nugegoda, Sri Lanka:
University of Sri Jayewardenepura), pp. 91-101.
Gombrich, Richard. 1991. Buddhist Precept and Practice. Traditional Buddhism
in the Rural Highlands of Ceylon (Delhi: Motilal Banarsidass).
Handurukande, Ratna. 2000. Three Sanskrit Texts on Caitya Worship. In Rela-
tion to the Ahoratravrata, an edition and synopses in English (with an intro-
duction) (Tokyo: The International Institute for Buddhist Studies of the
International College for Advanced Buddhist Studies) (Studia Philologica
Buddhica. Monograph Series XVI).
Hirakawa, Akira. 1982. Monastic Discipline for the Buddhist Nuns: An English
translation of the Chinese text of the Mahiisamghika-Bhik:fulJl-Vinaya (patna:
K. P. Jayaswal Research Institute).
Jinananda, B. 1969. Abhisamacarikii {Bhik:fupraklrlJakaJ (Patna: K. P. Jayaswal
Research Institute) (Tibetan Sanskrit works Series volume IX).
Mvy = Sakaki, R. (ed.). 1916-1925. Mahiivyutpatti, part 1 & 2 (Kyoto) (Kyoto
Imperial University Series 3).
Nolot, Edith. 1991. Regles de discipline des nonnes bouddhistes. Le
Bhik:fulJlvinaya de l'ecole Mahiisa1?7ghika-Lokottaravadin. Traduction
annotee, commentaire, collation du manuscrit (Paris: Edition-Diffusion de
Boccard) (College de France. Publications de l'Institut de Civilisation Indi-
enne. Fascic\.!le 60). .
Prasad, Maulichand. 1984. A Comparative Study of Abhisamiiciirikii.
Abhisamiiciirikii-Dharma-Vinaya of the Arya Mahiisiirhghika-Lokottar-
aviidins and the Pali Vinaya of the Theraviidins (Patna: K. P. Jayaswal
Research Institute) (Tibetan Sanskrit Works Series; No XXVI).
Prebish, Charles S. 1994. A Survey of Vi nay a Literature (Taipei: Jill Luen Pub-
lishing House) (The Dharma Lamp Series, Volume 1).
Rosen, Valentina. 1959. Der Vinayavibhanga zum der Sar-
viistiviidins: Sanskritfragmente nebst einer Analyse der chinesischen Uber-
setzung (Berlin: Akademie Verlag) (Deutsche Akademie der Wissenschaften
zu Berlin. Institut fur Orientforschung. Nr. 27) (Sanskrit Texte aus den Tur-
fanfunden IT).
Roth, Gustav. 1970. Including and a Sum-
mary of the of the Arya-MahiisiiY(lghika-Lokottaraviidin.
Edited and annotated for the first time with Introduction and two Indexes (patna:
K. P. J ayaswaI Research Institute) (Tibetan Sanskrit Works Series. Volume Xll).
Saddhatissa, Hammalawa. 1970. Buddhist Ethics: Essence of Buddhism (New
York: George Braziller).
Samantapasadika = J. Takakusu and M. Nagai, Samanta-Piisiidikii: Bud-
dhaghosa's Commentary on the Vinaya Pi[aka, 6 vols. (London: Pali Text
Society, 1938-1966).
Sanghasena. 1968. Sphu{iirthii Srlghaniiciirasangrahatikii (patna: K. P. Jayaswal
Research Institute) (Tibetan Sanskrit texts series, v. Xl). For a second edi-
tion by the same editor see Singh 1983.
Shimoda, Masahiro rB3lE'L. 1987. "Daishubukei sharnikai no tekisuto ni tsuite"
17) 7" ;r, r to: '? ,. '( [A Note on the Sphutartha
SnghanacaraSarilgrahatIka] Indogaku Bukkyogaku Kenkyu
[Journal of Indian and Buddhist Studies] 35, 2: 941(80)-939(82).
Shimoda, Masahiro rB3lE5L.. 1990. "The Sphu{iirthii Srighaniicarasangraha{lkii
and the Chinese Mahiisiinghika Vinaya", Indogaku Bukkyogaku Kenkyu
(Journal of Indian and Buddhist Studies) 39, 1: 495(11)-
Singh, Sanghasen. 1974. "Sirighana-sadda-vImarilsa", Buddhist Studies The Jour-
nal of the Department of Buddhist Studies, University of Delhi 1: 101-103.
Singh, Sanghasen. 1983. A Study of the Sphutiirthii Srlghaniiciirasangraha (patna:
Kashi Prasad Jayasawal Research Institute) (Tibetan Sanskrit Works Series;
No. XXIV). For a previous edition by the same editor see Sanghasena 1968.
Singh, Sanghasen. 1986. "A Study of the Scriptural References and Comments
in the Sphutartha SnghanacarasailgrahatIka of the Mahasanghika Sect", Bud-
dhist Studies (The Journal of the Department of Buddhist Studies, Univer-
sity of Delhi) 10: 1-6.
T. = Takakusu, Junjiro and Kaigyoku Watanabe (eds.), TaishO
shinshU Daizokyo, 55 vols. (Tokyo: 1924-1929). .
Tsukamoto, KeishO J:%.:;$:mt$. 1996-1998. Indo bukkyo himei. no kenkyu
.; :> 0) liJf9i: [A comprehensive study of Indian Buddhist Inscrip-
tions]. Part I: Text, note, wayaku [Text, notes and Japanes'ltranslation]
(Kyoto: Heirakuji shoten, 1996). II: Sakuin, zuhan [Indices, maps and illus-
trations] (Kyoto: Heitakuji shoten, 1998).
Yuyama, Akira. 1979. Systematische Ubersicht iiber die buddhistische Sanskrit-
Literatur - A systematic Survey of Buddhist Sanskrit Literature. Erster Teil:
Vinaya (Wiesbaden: Franz Steiner Verlag).
The Mahayana has sometimes been associated with the doctrine that all
sentient beings will attain complete awakening, a doctrine which is often
linked to some conception of the "embryo of the Tathagata" (tathiigata-
garbha)l. However, according to an alternate Mahayana doctrine, only
some sentient beings will attain the complete awakening of a buddha - and
some may even be excluded from attaining any form of awakening at all.
In this paper, I will examine just such a doctrine, as it is found in ah Indian
Yogacara treatise, the Mahiiyiinasutriilaf!lkiira ("Ornament to the Mahayana
Siitras"; abbr., MSA), a Sanskrit verse-text, and its prose commentary, the
Mahiiyiinasutriilaf!lkiira-bhii:rya (MSABh)2. Particular Tibetan and Chinese
sources attribute the composition of the MSA to the bodhisattva Maitreya
which gives us some indication of the importance this text was understood
to have within certain traditions. Nevertheless, the authorship and date of
1 A concise introduction to this doctrine, and the Mahayana sutras to which it is related,
may be found in Williams 1989, Chapter 5.
2 When referring to both the verse-text and the prose commentary together, I will use the
abbreviation MSA/Bh. By the term "the text" I mean the MSA and the MSABh taken
together, by "the verse-text" I mean the MSA, and by "the commentary" I mean the MSABh.
Throughout this paper, for the Sanskrit I will quote from Levi's edition of the MSA/Bh
(1907); I have also consulted the editions of Bagchi (1970; based on Levi's edition) and
Funahashi (1985; select chapters based on mss. from Nepal). All translations are my own.
The Tibetan canon contains the following relevant works: the MSA (verse-text): DT 4020;
the MSA/Bh (verse-text along with prose commentary): DT 4026; the MSA VBh (Sthiramati's
subcommentary to the text): DT 4034; and the MSAT (Asvabhava's subcommentary to the
text): DT 4029. The MSA/Bh also appears in the Chinese canon (Taish6 1604), although with
some differences from the Sanskrit version; on this, see Nagao 1961: vi.
3 The colophon of the Derge edition of the MSA states that the verse-text was com-
posed by Maitreya. Bu ston (1290-1364) includes the MSA as one of the five Maitreya-
texts; see Obermiller 1987: 53-54. Vi (1928: 221) identifies a Chinese tradition of the "five
treatises of Maitreya," which differs from the Tibetan list of texts, but which also includes
the MSA. Xuanzang (seventh century CE) writes that Asailga received the MSA and other
texts from Maitreya; see Bea11906, voL 1: 226.
Journal of the International Association of Buddhist Studies
Volume 26 Number 1 2003
the verse-text and its commentary are not certain; I hypothesize that the
MSAlBh may be dated to the fourth century CE (perhaps c. 350 CE)4. It is
my hope that an examination of such a source may contribute to the study
of the various ways in which the contours of the Mahayana have been
drawn from a doctrinal perspective. In the MSAlBh, one way in which the
limits of the Mahayana are defIned is through the employment of the gotra-
theory, a theory which identifIes the soteriological potentialities of indi-
viduals through reference to their spiritual "family" or "lineage." So in
order to understand this text's discursive construction of the category
"Mahayana," we must understand its concept of gotra.
In the context of discussions of Buddhism, the term gotra has been
variously translated as "family" (Edgerton 1970, vol. 2: 216), "basis,
source, cause, seed" (ibid.), "kind, class, category" (ibid.), "species"
(Wayman 1961; 58), or "spiritual lineage" (Ruegg 1968; 303, Griffiths
1990b: 49)5. Again, in the MSAlBh, gotra represents the soteriological
category to which a particular sentient being belongs: an individual's
gotra is taken to be indicative ofthat individual's soteriological possibil-
ities, i.e., what type of - or even whether - awakening can be attained.
So if a particular being is said to belong to the bodhisattva-gotra, then that
being has the potentiality for the awakening of a buddha, and if a partic-
ular being is classified as "without gotra" (agotraka), this indicates that,
at least for the present, that being does not have the "seed" to attain any
form of awakening at all.
4 My working hypothesis is that earlier strata of the MSA were compiled, redacted,
added to, and commented upon by one person, and I take the result of this process to be
the received text of the MSA/Bh. An extended introduction to the MSA/Bh - its editions
and translations, structure and contents, authorship, date, and relation to a larger corpus of
texts - may be found in Chapter 2 of my PhD dissertation (D' Amato 2000).
5 Ruegg (1976: 354) offers the following meanings of gotra in Buddhist usage: i. "mine,
matrix"; ii. "family, clan, lineage"; iii. "germ, seed"; and (iv.) "class, category." He also
offers a preliminary discussion of the relation of the term gotra to other possible cognate
terms in Iranian languages; his provisional hypothesis is that it might be possible to derive
the various meanings of the cognate terms - including the Vedic meaning "cattle-pen"
- from a root meaning "origin, source" (ibid., 354-356). In the context of the MSA/Bh,
the translation "spiritual lineage" is perhaps most appropriate. "Spiritual lineage," how-
ever, should not be confused with the notion of a lineage of transmission or tradition
(paramparti). In any case, in this paper when using the term gotra, I will leave it untrans-
lated and unitalicized.
Given. that for the MSA/Bh gotra is indicative of a sentient being's
soteriological potentiality, an important issue is whether gotra is able to
be acquired by every sentient being or whether there are some beings
who are excluded from ever acquiring a gotra; and furthermore if some
beings are excluded, in what sense they are excluded
. It will be neces-
sary to understand these issues in order to address the question of whether,
from the perspective of the MSA/Bh, all sentient beings can potentially
attain awakening? And so I will begin by discussing the meanings of the
term gotra in the MSA/Bh-including its relations to other important
terms and concepts. Then I will tum to a presentation of the various cat-
egories or subdivisions of gotra according to the text. Following this,
I will consider the text's gotra-theory in relation to some related doc-
trines in the MSA/Bh. I will then conclude with a response to the ques-
tion of whether all sentient beings can potentially attain aWakening.
Gotra defined
As I stated above, a number of translations have been offered for the
term gotra. The MSA/Bh itself offers an interesting interpretation of the
term. In explaining the use of the phrase gUlJottiiralJatii ("having the char-
acteristic of increasing virtues") in 3.4, the commentary states:
guI?ottaraI).arthena gotrarp. veditavyarp. guI?:l uttaranty asmad udbhavantiti lqtv1i/
Gotra should be known as that which increases virtues, since virtues arise
and increase because of it
6 For example, according to the Buddhabhumyupade.sa certain beings are excluded in
a predetermined sense: "From the beginningless beginning all sentient beings are divided
into five kinds of lineages (gotra) .... the first four of the above will ultimately attain final
cessation ... But the fifth lineage ... will never reach a time of final cessation" (Keenan
1980: 494). For other references to gotra as a predetermined category in Buddhist litera-
ture, see Lamotte 1976: 304.
7 In my general formulation of this question, I have benefited from Ruegg's work on
the gotra-theory, especially 1968, 1969: 73-107, and 1976.
8 Mano (1967: 970) points out that this etymological interpretation (nirukta) is also
given by Haribhadra (end of the eighth c. CE) in the Abhisamayiilaf!lkiiriilokii (the
Abhisamayiilaf!lkiira [AA] is considered by Bu ston to be another of the five Maitreya-texts),
as well as by both Arya-Vimuktisena and Bhadanta-Vimuktisena in their earlier com-
mentaries to the AA.
From this we may see that for the MSAlBh, acqui:ripg a gotra means
acquiring the ability to increase virtues. And this ability is of no small.
importance according to the text, because the development of a number
of virtues is understood as necessary for the attainment of particular sote-
riological goals. More specifically, gotra is posited as the cause of dif-
ferences in inclination towards a particular vehicle (adhimukti) - that is
to say, which soteriological vehicle one will be inclined to follow; reli-
gious practice (pratipatti); and awakening itself (bodhi) (MSA 3.2). It is
said to be the basis of knowledge (jfiiina), purification from the afflic-
tions (kleSa-nairmalya), and the supernormal powers, such as the higher
knowledges (abhijiiii) (MSABh ad 3.9). It is said to be the cause of great
awakening, great knowledge, concentration (samiidhi), and the matura-
tion of sentient beings (MSA 3.10). Gotra is also identified as one cause
for the production of the thought of awakening (cittotpiida) (MSABh ad
4.7), compassion (ad 17.34), the practice of the perfections (ad 16.16),
and the specific perfection of patience (ad 8.6). Indeed at MSA 19.61
gotra is identified as the [lIst of the ten aspects of the Mahayana, thus it
is understood to be the foundation upon which the practice of the
Mahayana is based
. According to the MSA/Bh, then, having a gotra is
foundational to attaining any specific Buddhist soteriological goal.
Having considered the MSA/Bh's explanation of the term gotra, we now
tum to terms that are used as equivalent to it in the text. One such equiva-
lent term used by the text is dhiitu. In fact, there are two places in the text
in which gotra and dhiitu are used interchangeably 10. The [lIst is at 11.8,
where the verse-text uses the term tridhiituka. Here the commentary states:
tatra dhatu-niyato yaJ:t sravakadi-gotra-niyataJ:t1
There [in the previous line of the text] a defInite dhiitu is a definite gotra,
such as sravaka.
Although the term dhiitu has a number of meanings in this and other
texts, in this instance, "stage" or "level of attainment" seems to be the
9 Chapter 19 of the MSA;Bh is specifically devoted to the topic of the gU(las; further-
more, at MSA;Bh 19.59-61, all of the virtues referred to in this paragraph are either explic-
itly mentioned or implicitly contained in the lists that occur there.
10 Furthermore, as Ruegg (1969: 85) points out, in a number of places the Tibetan text
of the MSA;Bh has rigs (normally gotra) for khams (normally dhiitu).
most appropriate. In another context (at 11.43) the verse-text uses the
term arya-gotra, the "noble gotra," which the commentary glosses with
anasrava-dhatu, the "undefiled realm" - i.e., the realm or level of attain-
ment in which there is no longer the influx of afflictions which bind one
to smp.sara. Here, then, we may infer that gotra is understood in terms of
a spiritual stage or level of attainment
The term gotra is also used twice in the commentary as a gloss for
praJerti, or "nature." The commentary to 8.5 glosses svaprakrtya - "by
one's nature" or "according to one's nature" - with gotrel}a - an instru-
mental form of the term meaning "by gotra" or "according to gotra."
Also, the commentary to 18.19-21 glosses the term prakrtya with gotrata/:t,
an ablative form of the term meaning "due to one's gotra." So here the
term gotra refers to one's "nature."12
While the term gotra has been equated with spiritual stage or level of
attainment (dhiitu) and nature (prakrti)13, it is so far unclear whether this
stage or nature represents something predetermined or something acquired.
That is to say, if the development of certain capacities or virtues is due
to one's nature - or due to one's gotra - then does gotra represent a
predetermined and predetermining category, or does it represent an
acquirable and alterable category of spiritual potentiality? And if it is
acquirable, is it acquirable by all? These are questions that I will return
to below. In any case, belonging to a certain gotra means having the
potentiality for reaching specific soteriological goals. The next step in
understanding gotra, then, is understanding the different categories or
subdivisions of gotra according to the text.
Categories of gotra
. In Indian Buddhist literature, a list of the following five gotras may be
found: sravaka-gotra, pratyekabuddha-gotra, bodhisattva-gotra, indefinite
11 In this connection, according to Ruegg (1974: 204), the Visuddhimagga equates
ariya-gotta with ariya-bhilmi, which Ruegg translates as "spiritual stage of the saint."
12 However, note that at neither of these locations does the text posit that one's nature
(prakrti) is to be understood as beginningless or unalterable.
13 These synonyms for gotra are also found in the Bodhisattvabhiimi (Sanskrit edition,
Dutt 1966: 2); a third synonym given in that text is bfja, "seed."
gotra (aniyata-gotra), and without gotra (agotraka)14. Spmetimes this list
of five is shortened to just the fIrst three members
, which the MSA/Bh
then aligns with the three. vehicles: those of the sravaka-gotra go by the
sravakayana, etc. But such a shortening of the list need not reflect any
serious philosophical differences. Only the first three gotras result in par-
ticular Buddhist soteriological goals, since only the fIrst three gotras cor-
respond to particular Buddhist soteriological vehicles. Furthermore, being
classified under the category of indefinite gotra may be understood as a
liminal state: when one's gotra becomes definite, it will be in terms of
one of the three standard gotras of sravaka, pratyekabuddha, or bod-
hisattva. And the category "without gotra," after ali, is not properly an
element of the list of categories of gotra. Therefore, the lists of five and
three gotras may be seen as having different conceptual emphases, rather
than different theoretical bases.
While the MSA/Bh offers no specific list of gotras that corresponds to
those given above, each member of the list of five is either explicitly
mentioned or implicitly referred to
. Furthermore, it is clear that the
14 For example, the Mahiivyutpatti contains the following fivefold list of gotras (1261-
1265): sriivaka-yiiniibhisamaya-gotraJ:t, pratyekabuddha-yiiniibhisamaya-gotraJ:t, tathii-
gata-yiiniibhisamaya-gotraJ:t (gotra for the realization of the vehicle of sravakas, pratyek-
abuddhas, and tathagatas, respectively), aniyata-gotraJ:t (indefinite gotra), agotrakaJ:t
(without gotra); see Sakaki 1926. The Mahiivyutpatti is a compilation of lists of Sanskrit
Buddhist terms, along with their Tibetan translations, that dates to the early ninth century
CE, centuries after the time of the MSA/Bh. Nevertheless, when a particular list is found
in the Mahiivyutpatti, this suggests that the list was important to Indian Buddhist thought
at the time. The fivefold list of gotras is also significant for our purposes since Sthiramati
offers the same one in his subcommentary to the MSA/Bh (DT sems tsam MI 48a3-4).
15 The three gotras are referred to in the Sa7[ldhinirmocana-sutra (SNS); Lamotte 1935:
73-74 and 198-199 offers the Tibetan text and French translation, respectively; cf. Pow-
ers 1995: 110-115. The SNS is considered to be one of the earliest Yogacara siitras (along
with the Mahiiyiiniibhidharma-sutra, which is no longer extant). Although the MSABh
does not explicitly refer to or quote the SNS, Schmithausen (1976: 240, note 2) makes the
convincing point that MSA 19.44ab presupposes SNS 8.20.2 in its discussion of the seven
types of thusness (tathatii). The Abhidharmakosa-bhii$ya ad 6.23cd also identifies the three
gotras (Sanskrit edition, Pradhan 1967: 348). Other lists of gotras appear in Indian Buddhist
literature. For example, the Mahiivibhii$iiSiistra mentions six different gotras; but even in
this case the principal gotras of the are the standard three identified here; see
Davidson 1985: 94-95.
16 The bodhisattva-gotra is mentioned in various places, e.g., MSABh ad 3.5, 3.7, 3.8,
etc.; the verse-text refers to it as the "foremost-gotra" (MSA 3.13: agra-gotra) or "noble-
gotra" (MSA 11.43: iirya-gotra). The sravaka-gotra is mentioned at MSABh ad 11.8,
MSA/Bh posits the superiority of the bodhisattva-gotra: in linking gotra
to the roots of virtue (kusala-mula), the commentary to 3.3 states that the
roots of virtue of the bodhisattva-gotra are far superior to those of the
sravaka-gotra - those of the sravaka-gotra, for example, lack the special
powers of a buddha. And in 11.43 the commentary states that the noble
gotra of buddhas - i.e., the bodhisattva-gotra
- is distinct from those
of the sravaka and pratyekabuddha for five reasons: (1) it is purified from
the impregnating afflictions (sawisana-kleSa), (2) it purifies a buddha-
field, and (3-5) it attains the three buddha-bodies.
The MSA/Bh explicitly aligns each of the first three gotras with one of
the soteriological vehicles (yanas). The commentary to 3.2 states that there
is a difference in gotra with respect to the three vehicles - the three vehi-
cles here being the sravakayana, pratyekabuddhayana, and bodhisattvayana
The commentary also goes on to distinguish three types of awakening (infe-
rior, middling, or superior), stating that each corresponds to a particular
gotra in the way that a fruit corresponds to its seed
And at 5.4-5 the com-
mentary states that there are three gotras: again, inferior, middling, or supe-
rior. Thus the MSA/Bh posits the following threefold structure:
spiritual category
soteriological vehicle
form of awakening
It is clear, then, that the category of gotra is of importance to the
MSA/Bh's soteriological scheme because different gotras lead to different
11.53, etc. The pratyekabuddha-gotra is implicitly referred to in the phrase sravakadi-
gotra, "the gotra of sravakas, etc." [i.e., pratyekabuddhas and bodhisattvas] at MSABh ad
11.8. The indefinite gotra is mentioned at MSA 3.6, etc., and the category "without gotra"
is mentioned at MSA 1.14, etc. Again, Sthiramati's subcornrnentary does offer the specific
list of five gotras; see note 14.
17 It is quite clear that buddha-gotra is another term for bodhisattva-gotra, and not a
separate category. MSABh ad 3.2 states there are three forms of awakening, and that each
form corresponds to a gotra as a fruit corresponds to its seed: thus the awakening of a bud-
dha corresponds to the bodhisattva-gotra. Then MSABh ad 3.4 states that gotra does not
exist along with its fruit; so when the awakening of a buddha is attained, no gotra exists.
18 While the first two vehicles are specifically mentioned in the text (see, e.g., MSABh
ad 19.44), the third is not; the term mahayana is used rather than bodhisattvayana.
19 Here we see gotra understood in terms of a seed, as in definition iii. offered by
Ruegg; see note 5.
soteriological goals. To trace this out a bit further, we mention the
following points: (1) the first two gotras belong to the lITnayana, while
the third gotra belongs to the Mahayana
; (2) the Hfuayana is said to bring
about the termination of the afflictive obstructions (kldavaral}a), while
the Mahayana is said to bring about the termination of both the atflictive
and the cognitive obstructions (jfieyavaral}a)21; (3) the lITnayana leads to
lesser forms of awakening, and ultimately to nirvfu:1a without residual
conditioning - which the text interprets as a
form of extinction
; (4) the Mahayana leads to a superior form of awak-
ening - the complete awakening of a buddha, an awareness of all objects
of knowledge and all modes of appearance (sarva-jfieya-sarvakara-jnana),
viz., omniscience - which is a state of being coextensive with reality
(thusness, tathata) itself, since the text posits that ultimately there is no
distinction between subject and object23; furthermore, the Mahayana does
not lead to the extinction of nirvaI;la without residual conditioning, but
rather to non-abiding nirvaI;la - an attainment
which allows for continued manifestations in the world in order to aid sen-
tient beings24. So the MSA/Bh's gotra-theory is of central importance to
the text's soteriological theory, since the first two gotras lead to lesser
forms of awakening which ultimately terminate in extinction (non-
existence), while the third gotra leads to complete awakening which is
nothing less than omniscience (coextension with reality itself).
While the preceding gives us a sense of the MSA/Bh's presentation of the
gotras of sravaka, pratyekabuddha, and bodhisattva, in order to understand
20 MSA 15.4 refers to the "twofold Hlnayana" (nihfnayiina-dvividha), viz., the
sriivakayiina and pratyekabuddha-yiina. There is further textual evidence for identifying
the sriivaka- andpratyekabuddhayiinas with the hfnayiina; on this, see D'Amato 2000: 177-
21 See, e.g., MSABh ad 17.4-5 and MSA/Bh 20-21.44.
22 See MSABh ad 3.3, 16.50, etc. In every place where the term or
is used in the text, some form of the term ("loss, destruction,
end, termination") is also used.
23 While this is by no means the standard account of omniscience offered in Buddhist
traditions, it is the one that I believe is most defensible as a reading of the MSA/Bh; see
MSA/Bh 20-21.58, Griffiths 1990a: 106-108, and D'Amato 2000: 130-131, 141-146, and
24 See MSA/Bh 17.32, 19.61-62, etc. On both forms of nirvfu.1a in the text, see D' Amato
2000, Chapter 5.
the text's gotra-theory more fully, we must also consider the latter two
gotras in our initial fivefold list: indefinite gotra and without gotra. Regard-
ing the indefinite-gotra, we may begin with a consideration of MSA 3.6,
where the verse-text introduces a fourfold classification of gotras:
niyatamyatarp. gotram aharyarp. haryam eva cal
pratyayair gotra-bhedo 'yarp. samasena catur-vidhalJII
Gotra can be definite or indefInite, incapable of being lost or able to be lost.
In sum, according to conditions, the distinctions of gotra are fourfold.
While the verse seems to set up a fourfold classification system (def-
inite-incapable of being lost, definite-able to be lost, etc.), the commen-
tary reduces this to a twofold system: "definite" corresponding only to
"incapable of being lost," and "indefinite" corresponding only to "able
to be lost. "25 The commentary's move here is supported by the fact that
while the term "indefinite" (aniyata) is again used by the text in con-
nection with the term gotra (at MSA/Bh 11.54), neither term of the pair
"incapable of being lost/able to be lost" (ahiirya/hiirya) is again linked
to it. Furthermore, Sthiramati's subcommentary agrees with the MSABh
in specifying that "definite" corresponds to "incapable of being lost"
and "indefinite" corresponds to "able to be lost. "26
Regarding the classification definite/indefinite, it is significant that for
the MSA/Bh indefinite gotra does not represent a fourth gotra alongside
the three standard gotras. Rather, any sentient being belonging to one of
the three specific gotras may be further classified according to whether
that specific gotra is definite or indefinite. If the gotra is definite that
means it is fixed and will not be lost, but if the gotra is indefinite that
means it is not fixed and there is the possibility that it can be lost or
changed after it has been acquired.
The text further discusses the indefinite gotra in a section devoted to
the analysis of the ekayana (a doctrine which the MSA/Bh does not
25 MSABh ad 3.6: samasena caturvidharp. gotrarp. niyataruyatarp. tad eva yatha-kramarp.
pratyayair ahiiryarp. hiiryarp. ceti/.
26 MSA VBh ad 3.6 (DT sems tsam MI 45b4-5): de bas na rigs nes pa mams ni rkyen
gyis mi 'phrogs pa :les bya ste/ ... rigs rna nes pa roams ni rkyen gyis 'phrog pa :les bya
consider to be definitive, but rather in need of interpretation). At 11.54,
the verse-text states:
anya-sarpdh1iraIJ.aya cal
desitaniyataniiIp. hi sarpbuddhair ekayanatall
For the purpose of attracting some, and for supporting others, the fully
awakened ones taught the fact of one vehicle for those who are indefInite.
The commentary goes on to specify that those who are attracted are
those with an indefinite sravaka-gotra and those who are supported are
those with an indefinite bodhisattva-gotra. Also, in discussing the func-
tion of buddhahood as a refuge, the commentary to 9.8 states:
[Buddhahood] protects those of an indefInite gotra from the Hfuayana by con-
structing the uniform path of the Mahayana.
Although this comment refers to the indefinite gotra without linking it
to one of the three specific gotras, it seems reasonable to read it in terms
of 11.54: buddhas teach the unity of vehicles in order to lure those of an
indefinite sravaka-gotra away from the Hinayana, and in order to keep
those of an indefinite bodhisattva-gotra from entering the Hlnayana.
Furthermore, in discussing the ten types of sentient beings towards whom
bodhisattvas are compassionate (MSA/Bh 17.29-30), the verse-text refers
to one type as those who have gone astray, which the commentary spec-
ifies as those who are indefinite in adhering to the Hinayana - a refer-
ence to the indefinite sravaka- and pratyekabuddha-gotras

In the commentary to verse 11.53 the MSABh mentions the indefinite
sravaka-gotra, stating that those of this gotra may attain final liberation
through the Mahayana
Thus there is the possibility for one of an indef-
inite sravaka-gotra to acquire the bodhisattva-gotra. Furthermore, if as
27 MSABh ad 17.29-30: utpatha-prasthita hInayana-prayukta aniyatlQI/. Again, hfnayiina
refers to the sriivakayiina and pratyekabuddhayiina. While the indefinite pratyekabuddha-
gotra is not specifically mentioned in the MSA/Bh, it is mentioned in Sthiramati's sub-
commentary. In fact, Sthiramati specifies that each of the three specific gotras may be def-
inite or indefmite; see MSAVBh ad 3.6 (DT sems tsam MI 45a6-7).
28 MSABh ad 11.53: aniyata-sravaka-gotrfu,lfup. mahayanena niryfu,lad yanti tena yanam
iti lqtva/.
11.54 states some bodhisattvas are in need of support, this implies that
one of an indefinite bodhisattva-gotra has the possibility of losing that
gotra. So for one of an indefinite gotra there is the possibility of losing
one's gotra and transferring to another

The final classification to consider is that of being without gotra; a cat-
egory that is the topic of MSA/Bh 3.11. Here, the verse-text and com-
mentary state:
aikantiko duscarite 'sti kascit kascit samudghatita-sukla-dharma/
'sti kascin nihIna-suklo 'sty api hetu-hlnal}/'po
aparinirvaI).a-dharmaka etasminn agotrastho 'bhipretal}/ sa ca samasato
dvividhal}l tat-kalaparinirvaI).a-dharma atyantarp. cal tat-kalaparinirviiI).a-
dharma caturvidhaJ:!/ ... atyantaparinirviiI).a-dharma tu hetu-hlno yasya
parinirviiI).a-gotram eva nasti/
Some have solely ill conduct, some have pure qualities that have been
destroyed, some have purity that is not associated with liberation, or an infe-
rior purity, and some also lack the cause.
This [verse] refers to those who are without gotra, those who lack the qual-
ities associated with parinirviiI).a. And this is concisely in two ways: lack-
ing the qualities associated with parinirviiI).a at the present time and for ever
(or "absolutely"; atyantam). Lacking the qualities associated with
parinirviiI).a at the present time can be in four ways .... But those who for ever
(or "absolutely"; atyanta) lack the qualities associated with parinirvaI).a
- those who lack the cause - simply do not have the parinirviiI).a gotra.
So according to the text, being without gotra means lacking the qual-
ities associated with parinirvfu).a (aparinirviilJa-dharmaka)31. And there
are two ways in which this might occur: lacking the qualities associ-
ated with parinirvaJ:?a at the present time (tat-kala) and lacking them
for ever or absolutely (atyantam). The first option -lacking the qual-
ities at the present time (tat-kiila) - is explained with reference to the
29 For a discussion of this issue in the Buddhabhumyupadda, see Keenan 1980: 678-
684; briefly, those of an indefinite gotra may attain nirvfu.la either through the Mahayana
or through one of the other vehicles.
30 Following the commentary, I do not interpret this verse in terms of the standard four
padas; I read the last pada as identifying two elements in a list, rather than one.
31 The term aparinirvaIJa-dharmaIJal; is used at MSABh ad 17.29-30: it refers to those
do not have the qualities associated with parinirvfu.la because they have never put an end
to sarpsara (sarrzsara-vartmatyantanupacchedat). The context here is a discussion of the
types of beings towards whom a bodhisattva should be compassionate.
first four reasons stated in the verse: having solely ill conduct, having
cut-off roots of virtue (samucchinna-kusala-mula)3Z; having roots of
virtue unrelated to liberation and hav-
ing inferior roots of virtue (hlna-kusala-mula). So sentient beings belong-
ing to this category are without gotra because of some deficiency in
roots of virtue. However, they are understood to be without gotra only
for the present, with the implication that they can acquire a gotra at
some point in time through accumulating an adequate store of the appro-
priate roots of virtue.
The second option - lacking the qualities associated with parinirval).a
for ever or absolutely (atyantam) - makes reference to the fifth reason
stated in the verse: lacking the cause, which ostensibly means lacking
any roots of virtue whatsoever. So here we see that there is a certain cat-
egory of sentient beings who are excluded from acquiring a gotra.
But there is some difficulty in determining in precisely what sense they
are excluded, a difficulty which hinges on the way in which the term
atyantam is translated in this context, a term which has a semantic range
which includes both "for ever" and "absolutely."33
Translators have dealt with the term atyantam in different ways in this
context. While Ruegg (1969: 80ff.) translates it as absolument, Levi
(1911: 30) suggests indefiniment (although this is not included as one of
the meanings ofthe term in the standard dictionaries). The Tibetan trans-
lation (DT sems tsam PHl, 138b3) gives gtan [du], which Das's diction-
ary defines as "always, continually, for ever." The Chinese translation
(Taish6 vol. 31, no. 1604, p. 595a25) renders it as bijing, which accord-
ing to the Digital Dictionary of Buddhism has the basic meaning of
"absolute[ly]," but also has the senses of "finally, in the long run.,,34
Hence there has not been a consensus on the meaning of atyantam in this
32 See Davidson 1985: 98-99 regarding samucchinna-kusala-mUla in the Vaibhfuiika tra-
33 The standard Sanskrit-English dictionaries include both of these senses of the term
atyantam. Monier-Williams has "in perpetuity" and "absolutely, completely"; Apte lists
"for ever" and "absolutely"; and the Poona dictionary (edited by Ghatage) also gives
"for ever" and "absolutely."
34 My thanks to Peter Gregory for providing me with the reference to the Chinese trans-
lation of the MSAlBh. The term bijing, according to Muller's Digital Dictionary of
Buddhism, also includes the senses of "positively, decidedly" and "necessarily."
context. However, since the term is used in MSABh ad 3.11 in opposi-
tion to tat-kala (lit., "that time"; "immediately, the present time"), I think
there is some reason to understand atyantam here in a temporal sense
(viz., as "for ever"). Furthermore, there is evidence that the term atyan-
tam is used in its temporal sense in other locations in the text. For exam-
ple, at 8.22 the verse-text states that the bodhisattva instructs beings for
as long as the world exists, which the commentary explains as follows:
yavallokasya bhavas tat-samanaya gatya atyantam ity arthaJ:!/
For as long as the world exists - with that same duration - means "for
ever" (atyantam).
And the commentary to 18.44 states that the practice of bodhisattvas
goes on "for ever" (atyantam) because it does not end in nirvfu).a with-
out residual conditioning (nirupadhise.ya-nirvalJe 'pi tad-ak,yayat). Fur-
thermore, in the commentary to the Madhyantavibhiiga (a text referred to
in the MSABh), the term atyanta is used to gloss sada ("always")35. Thus
it can be seen that the term atyantam does mean "for ever" in certain con-
texts. In any case, it is clear that sentient beings belonging to this cate-
gory - those.who are atyantaparinirvalJa-dharma - are without roots
of virtue, without gotra, and hence excluded from attaining any form of
It should be emphasized that according to 3.11 roots of virtue (kusala-
mula) are understood to be the cause of gotra: gotra is acquired through
amassing an adequate store of the appropriate roots of virtue. This is not
so unusual a claim for the MSAlBh to make, since obtaining roots of
virtue has traditionally been understood as a necessary aspect of the path
to awakening
. Furthermore, Sthiramati's subcommentary to the Mad-
hyantavibhiiga states that one standard definition of gotra is "roots of
MA V 1.18b states: sada sattva-hitaya cal "And for the benefit of sentient beings
always." MA VBh ad 1.18b then glosses with: atyanta-sattva-hitarthaJPI "For the sake of
the benefit of sentient beings for ever (atyanta)." See Sanskrit ed., Nagao 1964: 25.
36 On the importance of the kusala-mula to Buddhist conceptions of the path, see
Buswell 1992; he identifies one basic classification of the kusala-mula as nongreed
(alobha), nonhatred (adveSa), and nonignorance (amoha). The MSA/Bh is not the only
text that links the roots of virtue to gotra; for a discussion of the relation between these
two in the context of the and the Abhidharmakosa, see Davidson 1985:
virtue. "37 But the MSAlBh also posits that when the bodhisattva-gotra is
acquired, it in turn becomes a source of further roots of virtue. At 3.3 the
verse-text states that the pre-eminence of the [bodhisattva-]gotra is indi-
cated by the vastness, totality, greatness of purpose, and imperishability
of its purity (subha) - and here the commentary identifies purity with
the roots of virtue. And in 3.9, when the verse-text compares gotra to a
mine of gold
, the commentary states that the bodhisattva-gotra is like a
source of abundant gold since it is the basis of unlimited roots of virtue.
So for the MSAlBh, acquiring roots of virtue is a necessary condition for
acquiring a gotra (3.11), but when the bodhisattva-gotra is acquired, it
becomes a source of unlimited roots of virtue (3.9).
To conclude this section, it is necessary to consider one further char-
acterization of gotra in the text. At 3.4, the verse-text states:
pralqtya ca asrayas casritarp. ca tatl
sad asac caiva vijiieyarp. gUl.J.0ttaraI).atiirthataI:!/ I
By nature, nourished, support and supported, existing and not existing,
it should truly be known as that which increases virtues.
First we should note that this verse does not offer an addition to the
fivefold list of gotras discussed above; rather, it is introduced by the
commentary as a verse onthe characteristics (lak:;ar:tas) of gotras in gen-
eral. The commentary identifies the following four characteristics of
gotra: (1) gotra is natural (prkrtistha) insofar as it has the nature (svab-
hiiva) of a support, (2) it is attained (samudiinUa, a gloss for paripu:;taY[l
- "nourished" - in the verse) insofar as it has the nature of being
supported, (3) it exists along with its cause (hem), (4) but it does not exist
along with its fruit. From this we can see that gotra can be attained
- at least by some sentient beings - and that it has a cause. And again,
according to MSAlBh 3.11, the cause of gotra - the condition for its
acquisition - is the roots of virtue: gotra is acquired when adequate roots
of virtue of the appropriate kind are accumulated. But we must also
remember that, according to MSA/Bh 3.9, when the bodhisattva-gotra is
37 Sanskrit ed., Yamaguchi 1934: 188.
38 As stated in note 5, one of the meanings of the tenn gotra is "mine" (as in "a mine
of gems or ores"; see Edgerton 1970, vol. 2: 216, def. 2), thus here the MSA/Bh is play-
ing off that defmition of the tenn.
acquired, it serves as a basis for further roots of virtue. It is in this
sense that we should read characteristic (1) according to the verse above:
gotra is natural'insofar as it is a support or basis for further roots of

From all that has been said above, we may summarize the MSA/Bh's
gotra-theory as follows: (1) having a gotra is a prerequisite for attaining
any form of awakening; (2) there are three specific gotras: the sravaka,
pratyekabuddha-, and bodhisattva-gotras; (3) the first two gotras lead to
lesser forms of awakening, and ultimately to the HJnayana goal of nirvfu).a
without residual conditioning - understood as extinction - while the
bodhisattva-gotra leads to complete awakening, the Mahayana goal of
buddhahood - understood as omniscience; (4) some sentient beings are
of an indefinite gotra: they are able to lose their gotra and acquire a dif-
ferent one; (5) some beings are presently without gotra, but can acquire
one through amassing roots of virtue; and (6) some beings are excluded
from acquiring any gotra. Considering these points we are now in a bet-
ter position to attempt to address the question of whether all sentient
beings can potentially attain awakening.
39 The Bodhisattvabhiimi ([BBh] the fifteenth section of the first division of the
Yogiicarabhiimi, which contains in parts some of the oldest Yogaciira materials
[Schmithausen 1969]) also makes use of the classification of prakrtistha-gotra and samu-
diinfta-gotra. Yamabe (1997: 195ff.) offers a discussion of the BBh's interpretation of
these categories; according to that text, the prakrtistha-gotra is beginningless (aniidikii-
lika), while the samudiinfta-gotra is acquired through the accumulation of roots of virtue.
While the MSA/Bh is closely related to the BBh in the selection and order of topics that
it addresses, the two texts do not always address those topics in the same way. For exam-
ple, the MSA/Bh does not defme gotra in terms of the ifaeJiiyatana-viseifa (" distinct state
of the six sense bases"). Furthermore, the MSA/Bh does not use the term aniidikiilika
("beginningless") in connection with the topic of gotra at all. Thus the two texts differ in
their treatments of gotra. For the BBh, a gotra is prakrtistha if it is beginningless (viz., pre-
determined in some way), whereas for the MSA/Bh a gotra is prakrtistha only in the sense
that it serves as a causal basis for the further accumulation of roots of virtue. [Here note
that one meaning of the termprkrti is "cause," so prkrtistha may be interpreted as "exist-
ing/operating as a cause."] The MSA/Bh's interpretation of prkrtistha may also be seen
at 3.12, where the text states that the bodhisattva-gotra possesses virtues both naturally
(prakrtyii) and by nourishment (paripuiftasya); thus a bodhisattva-gotra is in one sense
natural (or causal), and in another sense attained (or caused). Finally, both Ruegg (1969:
476-477) and Davidson (1985: 100) state that the prakrtistha/samudiinfta distinction seems
to be strikingly similar to two types of roots of virtue in the V a i b h ~ i k a tradition - those
that are congenital (upapattiliibhika) and those due to application (priiyogika).
Can all beings potentially attain awakening?
Before responding to this question, we must fIrst consider a few relevant
doctrines in the MSA/Bh that have not been addressed thus far. The fIrst,
and most significant for our purposes, is that of the tathagatagarbha, the
"embryo of the Tathagata. " The MSA contains one reference to the tatha-
gatagarbha, which is found at 9.37. This verse and its commentary state:
tathata suddhirrl agata/
tathagatatvarp. tasmac ca tad-garbhaI). sarva-dehinaI:t11
tathata tad-visuddhi-svabhavas ca tathagataI:t1 atal;t sarve
sattvas tathagata-garbha ity ucyatel
Although thusness is in all [living beings] without distinction, when it is
pure it is the nature of the Tathagata; thus all living beings have its embryo.
Thusness is in all [living beings] without distinction, and the Tathagata has
the nature of the purity of that. Hence it is said that all sentient beings have
the embryo of the Tathagata.
This verse states that all sentient beings have the embryo of the Tatha-
gata (tathagatagarbha) since all sentient beings have the nature of thus-
ness (tathata). It should be noted that it is possible to interpret this verse
as stating that all beings "are" the embryo of the Tathagata, rather than
all beings "have" the embry040. The former would imply, however, that
all sentient beings will attain buddhahood, a claim that the MSA/Bh
does not seem inclined to make. In fact, as we have already seen in
the material on gotra, the text states quite clearly that different beings
belong to different gotras and that different gotras lead to different forms
of awakening41; thus not all sentient beings attain buddhahood. Never-
theless, the text does here claim that all sentient beings have the poten-
tiality for attaining buddhahood, even if this potentiality is not actually
The claim that all sentient beings have the potentiality for attaining bud-
dhahood is not such a strange one for the MSA/Bh to make given its
40 For more on this see Griffiths 1990b: 62-63. Here I am following Griffiths' trans-
41 Compare this to Haribhadra's theory of gotra as found in the AbhisamayaiaJ?1kara/okii,
where from the ultimate point of view gotra is seen as non-distinct in all sentient beings;
see Ruegg 1968, especially: 316-317, and Mana 1967.
affirmation of mind as fundamentally pure in nature. The first half of
13..19 states:
matarp. ca chtarp. pralqti-prabhiisvararp. sada tad
Mind should properly be thought of as always luminous by nature; it is
impure due to adventitious defilements.
The commentary further states that mind is like space, or like water that
is pure in itself but made impure by pollutants; and like water, mind can be
purified through removing the defIlements. Thus from the perspective of the
text, insofar as sentient beings have (or just are) minds, they may attain the
state of fundamental purity through the removal of adventitious defIlements

The theme of purity is also discussed at 11.13-14, although here it is
the nature of reality that is fundamentally pure. MSA 11.13 states:
tattvarp. yat satatarp. dvayena rahita:qJ. bhriintes ca sa:qJ.nisraya}:t
sakya:qJ. naiva ca sarvathabhilapitu:qJ. yac caprapaiicatmaka:qJ./
jfieya:qJ. heyam atho visodhyam amala:qJ. yac ca pralqiya matarp.
yasyakasa-suvan;ta-vari-sadrsI klesad visuddhir matal/
Reality - which is always without duality, is the basis of error, and is entirely
inexpressible - does not have the nature of discursivity. It is to be known,
abandoned, and purified. It should properly be thought of as naturally immac-
ulate, since it is purified from defilements, as are space, gold, and water.
As in the commentary to 13.19, the nature of reality - like the nature
of mind - is said to be similar to that of space and water: they are nat-
urally pure and defIled only adventitiously. The next verse goes on to state
that there is nothing else in the world besides this fundamentally pure real-
ity. Thus at an ontological level the MSA/Bh posits that, even though it
serves as the basis of error, reality is fundamentally pure.
This brief excursus into the domain of the MSA/Bh' s ontological dis-
course is to be understood in relation to our original question. The moves
towards understanding the MSA/Bh's position on the tathrigatagarbha-
theory and the MSA/Bh's doctrine of the fundamental purity of mind and
reality were, I think, necessary in order to more fully consider a response
42 The theme of the fundamental purity of mind and the adventitious nature of defile-
ments may also be found in certain passages in the nikayasjagamas; see Keenan (1980:
21-22) on passages from the Anguttara-niktiya and Majjhima-nikaya that posit the funda-
mental purity of mind.
to the question of whether all sentient beings can potentially attain awak-
ening. According to the passages discussed here we see that
(1) All sentient beingshave the potentiality for attaining complete awak-
ening (i.e., all beings have the embryo of the Tathagata; MSA/Bh 9.37),
and all beings have the potentiality for purifying their minds, smce mind
- like reality itself - is fundamentally pure by nature (MSA/Bh 13.19
and 11.13).
However, we must also consider this claim in relation to the MSA/Bh's
discourse on gotra, according to which
(2) Having a gotra is a prerequisite for attaining any form of awakening, but
some beings are excluded from acquiring a gotra (MSA/Bh 3.11).
Considering these claims together, we may note a degree of tension
between (1) and (2). More specifically, according to (1) all sentient beings
have the potentiality for complete awakening, while the implication of
(2) is that some beings are excluded from the attainment of any form of
awakening at all, in that they are excluded from acquiring the "seed"
(gotra) necessary for awakening. The issue here is in what sense we
should understand the state of being excluded - and, more specifically,
in what sense we should understand the term atyantam in the commen-
tary to 3.11. Are sentient beings of this category - those who are atyan-
tiiparinirviilJa-dharmii - excluded "absolutely"? "For ever"? Does any-
thing hinge on deciding one way or the other?
I would argue that something does indeed hinge on such a decision, that
it is not philosophically insignificant whether atyantam is translated as
"absolutely" or "for ever" in this context. To say that some beings are
"absolutely" without the qualities associated with parinirvaJ).a -
absolutely without gotra - implies that some beings are "uncondition-
ally" in this state
: it implies that these sentient beings unconditionally
lack gotra - hence they simply do not attain any form of awakening,
without reference to any other conditions or qualifications. This would
pose a problem in interpreting the text consistently, in that we have already
seen that gotra is not unconditional: the condition for its acquisition is the
43 The O:iford English Dictionary offers one defInition of "absolutely" as "without
conditiou or limitation; unconditionally."
roots of virtue. Furthermore, understanding atyantam here as "absolutely"
intensifies the tension between (1) and (2): it would imply the problem-
atic conclusion that although all beings have the potentiality for complete
awakening, some beings are absolutely unable to attain any form of awak-
ening at all. On the other hand, saying that some beings are "for ever"
without the qualities associated with parinirvfu;a - for ever without
gotra - does not imply that any beings are unconditionally in this cate-
gory. Rather, it implies that some beings simply always remain in this cat-
egory due to a conditional lack in roots of virtue. And translating atyan-
tam here as "for ever" would significantly reduce the tension between (1)
and (2): it would allow that while all beings have the potentiality for
complete awakening, some beings simply never actualize this potential-
ity. In fact, according to the text there is always a surplus or remainder
of sentient beings who have not been ripened to awakening, since the
world is infinite

I would propose that a more perspicuous means of clarifying and
addressing the tension between (1) and (2) - between a doctrine of uni-
versal potentiality for buddhahood and the exclusion of certain sentient
beings from attaining awakening - may be found through introducing the
modal concepts of necessity, possibility, and contingency. It should first
be noted that the MSA/Bh does not employ these concepts in this or any
other context; in fact, to my knowledge, the concepts of modal logic are
not fully articulated anywhere in the history of Indian Buddhist thought.
What I propose then is of the nature of a rational reconstruction. Briefly,
Haack specifies the distinction between necessary and contingent truths
as follows: "a necessary truth is one which could not be otherwise, a con-
tingent truth one which could; or, the negation of a necessary truth is
impossible or contradictory, the negation of a contingent truth possible or
consistent; or, a necessary truth is true in all possible worlds, a contingent
truth is true in the actual but not in all possible worlds" (1978: 170).
To this we may add that a possible truth is one whose negation is not nec-
essary. I would argue that a reconstruction of (1) and (2) employing modal
concepts allows for a clarification of the claims at MSA/Bh 9.37 and 3.11,
and the elimination of the tension between them. Through introducing the
44 MSABh ad 9.49: na ca lokasyanantatvat/.
concepts of possibility and contingency, we might restate propositions (1)
and (2) in the following terms: '
(1') For all x such that x is a sentient being, it is possible that x will attain
complete awakening,
(2') For some x such that x is a sentient being, it is contingent that x is
without gotra, and hence without any form of awakening.
These two propositions are consistent with one another, since there is
no contradiction in stating that awakening is possible for all beings even
though some beings do not in fact attain it. Note, however, that if we
were to translate atyantam as "absolutely" at MSABh ad 3.11, then (2')
would be restated as follows:
(2") For some x such that x is a sentient being, it is necessary that x is
without gotra, and hence without any form of awakening.
Such a proposition would be inconsistent with (1 '), since it is contra-
dictory to state that it is possible for all beings to attain awakening but
necessary that some do not. Thus my reconstruction of (2) entails inter-
preting atyantam as a term implying contingency rather than necessity45.
And so I propose that the claims at MSA/Bh 9.37 and 3.11 be read in
terms of propositions (1 ') and (2'), respectively.
The tension which I raise between (1) and (2) is one which has been
noticed by Tibetan and Indian Buddhist traditions. In his study of the the-
ories of tathiigatagarbha and gotra, Ruegg (1969: 82) states that the appar-
ent contradiction between MSAJBh 9.37 and 3.11 has divided Tibetan com-
mentators, and that certain Tibetan commentators have argued that the
tathiigatagarbha-verse (9.37) is to be understood as having a sens inten-
tionnel in this context - that its claim is not definitive for the MSA/Bh.
Also notable is the fact that in the subcommentary to the Madhyiintavibhiiga
(again, a text cited in the MSABh), Sthiramati offers two rather conflicting
interpretations of gotra
: according to the first interpretation, different
gotras are "inherent" (sviibhiivikam) and "beginningless" (aniidikiilikam)
in different individuals - for example, some have the sravaka-gotra and
45 My thanks to Jay Garlield for suggesting this way of stating the point to me in
46 See Sanskrit ed., Yamaguchi 1934: 55-56.
others the buddha-gotra - a view that implies a theory of predetermined
and distinct "seeds" of aWakening. According to the second interpretation,
however, all beihgs have the tathagata-gotra - a view that implies a the-
ory of universal potentiality for buddhahood. And Sthiramati does not indi-
cate which interpretation is to be understood as definitive. Thus we can see
that even in the Indian context there was some debate over whether all
beings have the tathiigatagarbha or whether different beings just have dif-
ferent gotras, with some beings excluded from the attainment of complete
awakening, and others - those who are inherently without gotra -
excluded from the attainment of any form of awakening at all. We might
speculate that had the modal concepts of necessity, possibility, and contin-
gency been developed in a rigorous fashion and employed in the context of
a controversy between the theories of tathiigatagarbha and gotra, any incon-
sistency between the two theories - at least as they occur in the MSAlBh
- could have been resolved. Again, the reconstruction which I propose
involves the two steps of interpreting atyantam at MSABh ad 3.11 as "for
ever" - a step supported by both internal and external evidence - and
interpreting (1) as a statement of possibility and (2) as a statement of con-
To conclude, in response to our initial question of whether all sentient
beings can potentially attain awakening, we may state the following: in
the terms of the MSAlBh itself, while all beings have the embryo of the
Tathagata, some beings are simply for ever without the "seed" (gotra) of
awakening. And in the terms of my proposed reconstruction: while all sen-
tient beings can potentially attaining awakening, it is contingently the
case that some beings will never actually do S047.
Derge Tanjur (Sde dge bstan 'gyur)
47 It may be interesting to consider this interpretation in relation to Anguttara-nikiiya
V: 193-195, where, after a discussion of the fourteen restricted points, the Buddha remains
silent in response to the question of whether the whole world will attain deliverance.
MSA: Mahiiyanasatralal[lkiira (verse-text)
MSABh: (commentary)
MSA/Bh: Mahiiyanasatrala"r[zkara and
MSAT: Mahiiyanasatriilal[lkiira-tfka (Asvabhava's subcommentary)
MSA VBh: (Sthiramati's subcommentary)
SNS: Sal[ldhinirmocana-satra
Works Cited
Bagchi, Sitansusekhar, ed. 1970. Mahiiyana-Satralmikiira of AsaJiga. Darbhanga,
Bihar: Mithila Institute.
Beal, Samuel, trans. 1906. Si-Yu-Ki: Buddhist Records of the Western World,
2 vols. London: Kegan Paul, Trench, Triibner and Co.
Buswell, Robert E., Jf. 1992. "The path to perdition: the wholesome roots and
their eradication." Paths to Liberation: The Marga and Its Transformations
in Buddhist Thought. Edited by Robert E. Buswell, Jr. and Robert M.
Gimello. Honolulu: University of Hawaii Press: 107-134.
D'Amato, Mario. 2000. "The Mahayana-ffinayana Distinction in the Mahiiyana-
satriilal[lkara: A Terminological Analysis." PhD diss., University of
Davidson, Ronald M. 1985. "Buddhist Systems of Transformation: Asraya-
parivrttii-paravrtti among the Yogacara." PhD diss., University of Califor-
nia, Berkeley.
Dutt, Nalinaksha, ed. 1966. Bodhisattvabhiimi. Patna: K. P. Jayaswal Research
Edgerton, Franklin. 1970. Buddhist Hybrid Sanskrit Grammar and Dictionary,
2 vols. Delhi: Motilal Banarsidass. Originally published in 1953.
Funahashi Naoya, ed. 1985. Mahayanasatralarpkiira (Chapters I, II, III, IX, X):
Revised on the Basis of Nepalese Manuscripts. Tokyo: Kokushokankokai,
Griffiths, Paul J. 1990a. "Omniscience in the Mahayanasutralailkara and its com-
mentaries." Indo-Iranian Journal vol. 33.2: 85-120.
Griffiths, Paul J. 1990b. "Painting space with colors: Tathagatagarbha in the
Mahayana-satralaJikiira-corpus IX.22-37." Buddha Nature: A F estschrijt for
Minoru Kiyota. Edited by Paul J. Griffiths and John P. Keenan. Reno: Bud-
dhist Books Int.: 41-63.
Haack, Susan. 1978. Philosophy of Logics. Cambridge: Cambridge University
Keenan, John P. 1980. "A Study of the Buddhabhiimyupadda." PhD diss.,
University of Wisconsin, Madison.
Lamotte, Etienne, ed. and trans. 1935. Sarpdhinirmocana Satra: L 'Explication des
Mysteres. Louvain: Universite de Louvain.
Lamotte, Etienne, trans. 1976. The Teaching of Vimalakfrti. Translated by
Sara Boin. London: Pali Text Society. Originally published as L'Enseigne-
ment de Vimalakfrti in 1962.
Levi, Sylvain, ed. and trans. 1907-1911. Mahiiytina-Siltrtilarpktira, 2 vols. Paris:
Librairie Honore Champion.
Mano Ryukai. 1967. "Gotra in Haribhadra's theory." Journal of Indian and Bud-
dhist Studies (Indogaku Bukkyogaku Kenkyil) vol. 15.2: 972-964.
Nagao Gadjin. 1958-1961. Index to the Mahiiytina-Siltrtilarp.ktira, 2 vols. Tokyo:
Nippon Gakujutsu Shinko-kai.
Nagao Gadjin, ed. 1964. Tokyo: Suzuki Research
Obermiller, E., trans. 1987. The Jewelry of Scripture by Bu ston. Delhi: Sri Sat-
guru. Originally published in 1931.
Powers, John, trans. 1995. Wisdom of Buddha: The Sarp..dhinirmocana Siltra.
Berkeley: Dharma Publishing.
Pradhan, P., ed. 1967. of Vasubandhu. Patna:
K. P. Jayaswal Research Institute.
Ruegg, David Seyfort. 1968. "Arya and Bhadanta Vimuktisena on the gotra-the-
ory of the Prajfiaparamita." Wiener Zeitschrift fur die Kunde Sud- und
Ostasiens vol. 12-13: 303-317.
Ruegg, David Seyfort. 1969. La Theorie du Tathiigatagarbha et du Gotra. Paris:
Ecole Fran<;aise d'Extreme-Orient.
Ruegg, David Seyfort. 1974. "Pali gotta/gotra and the term gotrabhil in
Pali and Buddhist Sanskrit." Buddhist Studies in Honour of I. B. Horner.
Edited by L. Cousins, A. Kunst, and K. R. Norman. Dordrecht: D. Rei-
del: 199-210.
Ruegg, David Seyfort. 1976. "The meanings of the term gotra and the textual
history of the Ratnagotravibhiiga." Bulletin of the School of Oriental and
African Studies vol. 39: 341-363.
Sakaki R., ed. 1926. Mahiivyutpatti, 2 vols. Kyoto: Suzuki Gakujutsu Zaidan.
Schrnithausen, Lambert. 1969. "Zur Literaturgeschichte der aIteren Yogacara-
Schule." ZeitschriJt der Deutschen Morgenliindischen Gesellschaft supple-
menta I: 811-823.
Schrnithausen, Lambert. 1976. "On the problem of the relation between spiritual
practice and philosophical theory in Buddhism." German Scholars on India,
vol. 2. Edited by the Cultural Department of the Embassy of the Federal
Republic of Germany. Delhi: Motilal Banarsidass: 235-250.
Ui Hakuju. 1928. "On the author of the Mahayana-sutralarp.kara." ZeitschriJt flir
Indologie und Iranistik vol. 6: 215-225.
Wayman, Alex. 1961. Analysis of the Srtivakabhilmi Manuscript. Berkeley and
Los Angeles: University of California Press.
Williams, Paul. 1989. Mahiiytina Buddhism: The Doctrinal Foundations. London:
Yamabe Nobuyoshi. 1997. "The idea of dhiitu-vada in Y o g a c ~ a and Tathiigata-
garbha texts." Pruning the Bodhi Tree: The Storm over Critical Buddhism.
Edited by Jamie Hubbard and Paul L. Swanson. Honolulu: University of
Hawaii Press: 193-204, 208-219.
Yamaguchi Susumu, ed. 1934. MadhyantavibhiigatJkii. Nagoya: Librairie Hajinkaku.
I would like to thank the members of the Five College Buddhist Studies Semi-
nar (Jay Garfield, Peter Gregory, Jamie Hubbard, Marylin M. Rhie, Young H.
Rhie, Andrew Rotman, and Taitetsu Unno) for reading an earlier draft of this
paper and participating in a helpful discussion on it. While I have greatly bene-
fited from their comments, I alone am responsible for any errors or deficiencies
that remain.
l. Introduction
Since the publication of Th. Stcherbatsky's Buddhist Logic in 1932,
many scholars have followed the great Russian Buddhologist's lead in
looking to the works of DharmakIrti for help in understanding the works
of Dignaga. Among other things, this has meant taking Dignaga to have
understood in terms of what Stcherbatsky characterized as
"point-instants," a translation which perhaps plausibly conveys a sense
of DharmakIrti's understanding of this concept.
It is not surprising that scholars should thus have relied on DharmakIrti
in interpreting Dignaga, since understanding Dignaga's works is a task that
presents significant interpretive difficulties. Unlike the case of Dhar-
makIrti (several of whose works survive in the original Sanskrit), Dig-
naga's works come down to us only in Tibetan translation!. Moreover, in
the case of the Pramar;asamuccaya, what we have are in fact two often
quite divergent Tibetan translations, a state of affairs that reflects Richard
Hayes's contention that the available translations "show signs of having
been done by translators who were themselves not certain of the mean-
ings of many passages in the original texts ... "2. The available texts of
Dignaga's works are thus more than usually underdetermined. Even more
than is typically the case with respect to the characteristically elliptical
works of Indian philosophers, then, a full understanding of Dignaga
requires recourse to his commentators. In this regard, it is not surprising
that a great many modem scholars have tread in Stcherbatsky's footsteps
1 Randle (1926) has compiled such Sanskrit fragments of Dignaga as can be gleaned
from the quotations of him in other extant works of Indian philosophy.
2 Hayes (1988). p.6. Note that Hattori's edition and translation of the first chapter of
the PramiilJasamuccaya (1968) gives editions of both Tibetan translations (i.e., the one
supervised by the Indian palJrjita and the one supervised by Kanakavar-
man). Both Hayes and Hattori take the translation of Kanakavarman as their basic text.
Journal of the International Association of Buddhist Studies
Volume 26 Number 1 2003
and read Dignaga primarily through the lens of DharmakIrti, who is, after
all, traditionally taken to have been Dignaga's grand-disciple, and whose
Prarnii1J.aviirttika represents itself as what would thus be the earliest sur-
viving "commentary" on Dignaga's magnum opus.
In recent decades, however, several scholars have urged that Dharmaldrti
is a commentator in name only, and that his works in fact represent inno-
vative departures from Dignaga's works. Indeed, Radhika Herzberger has
gone so far as to urge that "Diimaga's thought is not encompassed by the
greater depth of Dharmaklrti' s, rather it is washed away by it"3. The ques-
tion of Dignaga's understanding of is one of the chief issues
with respect to which recent scholars have particularly questioned the value
of Dharmaklrti. Thus, for example, Hayes (1988, p.15) says that among the
views erroneously "imputed" to Dignaga by Stcherbatsky is " ... the view
of particulars as point-instants, which amounts to a commitment to a doc-
trine of radical momentariness Similarly, Shoryu Katsura
(1991, p.l44) has urged that "[Dharmaklrti's view ot] reality is charac-
terized by momentariness, an idea which has no place in Dignaga."
While such cautions may be appropriate, it nevertheless remains diffi-
cult to be sure exactly what Dignaga does understand to be
like, since about the only thing he ever says about them is that they are
"ineffable," "unspecifiable," or "indefmable" - with all of these being
plausible renderings of avyapadeSya (Tib., bstan par bya ba rna yin pa),
the word that Dignaga uses. Thus, there is still light to be shed on the issue.
I propose, then, to weigh in on the question of how Dignaga ought to be
understood with regard to I propose to do so, however, not
by recourse to any of Dignaga' s commentators4, but by looking at an
3 Herzberger (1986), p.241; quoted by Hayes (1988), p.30. Hayes concurs with
Herzberger's assessment, adding that Dharmaldrti "also washed away much of the accom-
plishment of the Buddha as well." (p.310) Among other things, this reflects a significant
tendentiousness in Hayes's lucid presentation of Dignaga, but that is a subject for another
4 Hayes has urged that Iinendrabuddhi's VisalamalavatinamapramaIJasamuccayaflka
represents a more helpful commentary on the Pramti7JaSamuccaya than DharmakIrti's
PramaIJavarttika, despite the latter's being much earlier. However, not only is it the case
that Iinendrabuddhi's commentary (like Dignaga's text) survives only in Tibetan transla-
tion (as the Yans-pa dan dri-ma med-pa ldan-pa ses-bya-ba tshad-ma kun-las-btus-pa'i
'grel-bsad, Tohoku 4268); moreover, Iinendrabuddhi (800-850) significantly post-dates
Dharmaldrti, and Iinendrabuddhi himself thus tends to read Dignaga through the lens of
early critic of him: specifically, Candrakirti, the first chapter of whose
Pfasannapadii comprises a highly under-appreciated engagement with an
unnamed interlocutor whose thought looks very much like that of Dignaga.
In the standard edition of the Prasannapadii, this section spans some
twenty pages
. Typical of the neglect of this section is the fact that, while
it thus constitutes more than a fifth of Candrakirti's opening chapter,
Cesare Rizzi's 36-page summary of the chapter devotes a scant two pages
to this "controversy with the Buddhist Logicians"6. This neglect perhaps
owes something to the fact that some influential Tibetan discussions of
at least parts of this section take Candraldrti to have been continuing his
attack on Bhiivaviveka, so that what is almost certainly an engagement with
Dignaga's epistemology gets subsumed in the sviitantrika-priisangika dis-
cussion that has instead preoccupied most scholars.?
DharmakIrti. Cf., Hayes (1988), pp.224-6, for comments on Jinendrabuddhi's nevertheless
being preferable to DharmakIrti as a commentator on Dignaga.
5 The standard edition is that of Louis de La Vallee :poussin (1903-1913), which was
printed as volume IV in the Bibliotheca Buddhica. Based on additional manuscripts from
Nepal, J. W. de Jong (1978) suggested extensive revisions to this edition. All translations
in the present essay are my own, and are from the edition of La Vallee Poussin as revised
by de Jong (with de long'S changes noted). (The edition ofVaidya [1960], which provides
the pagination from the Biliotheca Buddhica edition, can be used, but is effectively just a
reprinting of La Vallee Poussin's edition without La Vallee Poussin's extraordinarily
erudite and helpful footnotes. Vaidya can, though, occasionally prove useful for his
judgments regarding which of La Vallee Poussin' s variants to adopt.) In La Vallee Poussin' s
edition, the engagement with Dignaga runs from 55.1l to 75.13 (with references thus being
to page and line numbers), with the entire first chapter spanning 91 pages.
6 Rizzi (1987). The bulk of Rizzi's short book (pp.23-59) consists in what is usually a
detailed paraphrase of the first chapter of the Prasannapadt'i; Rizzi's account of the sec-
tion in question (i.e., that spanning pages 55.1l to 75.13 of La Vallee Poussin's edition)
occupies a page and a half at pp.47-49.
7 For the view that CandrakIrti is still occupied with Bhavaviveka in at least part of the
section I will consider, d., Thurman (1991), pp.292-295 (and especially p.293, n.13); this
translates a section of Tsoil-kha-pa' s Legs bsad snyiiz po that addresses (and quotes exten-
sively from) a discussion occurring at pp.66.1-68.4 of the Prasannapadt'i. Cf., also, Eckel
(1978), who similarly follows Tsoil-kha-pa's lead in taking this section to be addressed to
Bhavaviveka. Stcherbatsky (1927) understood this whole section of the Pras(mnapadt'i
(i.e., pp.55.11-75.13) to have been addressing Dignaga, introducing his translation of it
(p.142) as a "Controversy about the Validity of Logic," and characterizing CandrakIrti's
as "The Logician." (Stcherbatsky's translation of this section is at pp.142-174.)
Hattori (1968) also understands CandrakIrti to have been addressing Dignaga, and his
annotations to his translation from Dignaga's Pramt'il}asamuccaya frequently provide use-
ful cross-references to CandrakIrti. The only other significant treatments of this section that
Whatever the reason, this neglect is regrettable, and not least because,
in my view, an understanding of Candraldrti's engagement particularly
with Dignaga affords us an unusually good opportunity for appreciating
the logically distinct character of Candraldrti's Madhyamaka
For the
present, however, I am chiefly interested in what CandrakIrti can tell us
about how to understand Dignaga - and specifically, how to understand
Dignaga's concept of I would like to show this by following
Candraldrti in elaborating, in effect, on what seems to me an apt obser-
vation from Shoryu Katsura (which will be noted in due course). As we
will see, Candraldrti elaborates a similar insight in such a way as to
make clear precisely how Dignaga had (on Candraldrti's reading, anyway)
transformed the A.bhidharmika notion of - specifically, how
Dignaga had used a word which A.bhidharmikas understood to denote a
species of universal to refer instead to what can plausibly be character-
ized as bare particulars. We can, then, flesh out Dignaga's spare account
of by appreciating what Candraldrti thought Dignaga would
find to be an unwanted consequence of his own view.
II. in the Abhidharmika context
Insofar as it is characteristically A.bhidharmika usage that Candraldrti
will press against Dignaga, our account should begin with the discourse
I have been able to fmd are those of Mookerjee (1957, pp.42-58; this is basically a para-
phrase of Candraldrti's text) and Siderits (1981), who also takes Candraldrti's target here
to be Dignaga.
Recently, an interesting bit of evidence regarding Candraldrti's purvapakein has come
to light: Yoshiyasu Yonezawa (1999, 2001) has been preparing a critical edition of the
* Lakea/}atfka, from a Sanskrit manuscript in Tibetan dbu-med script, recovered at Zha lu
monastery by Rahula Sai1lq1:yayana (1937, p.35). This very brief commentary on
the Prasannapada was, Y onezawa speculates, written under the supervision of
Abhayakaragupta (2001, p.27), which would place it roughly in the 12th century. Among
the things which this concise commentary does is identify the various unnamed inter-
locutors, and with regard to the section that will concern us, the anonymous author of this
commentary specifically identifies Dignaga; cf., n.43, below. (I would like to thank
Prof. Yonezawa for sharing this information with me.) Be that as it may, I would argue
(and indeed, have done so in Arnold, 2002) that Candraldrti has good reasons for fmding
the epistemological project of Dignaga in principle problematic, and that Candraldrti's
unnamed interlocuto"r throughout this section is in fact Dignaga.
8 I have argued this at length elsewhere; see Arnold 2002.
of Abhidharma. According to standard Abhidharmika accounts of the
Buddhist reductionist project, dharmas are the really (though fleetingly)
existent elements that survive characteristically Buddhist reductionist
analysis; dharmas are, in other words, the elements to which existents (and
paradigmatic ally, of course, persons) can be reduced, and a great deal of
Abhidharmika literature is devoted to the enumeration of the" dharmas"
which should thus be permitted into a final ontology (where "ontology"
thus has something like its standard meaning of a catalogue of ontologi-
cally primitive categories). A standard such enumeration, for example,
lists 75 dharmas that constitute the ontological primitives upon which all
other, derivative existents are supervenient
Note, though, that the idea
of there being 75 dharmas is not the idea that there exist only 75 unique
particulars in the world; rather, these clearly delineate 75 ontologically
primitive categories - types of which there can be, presumably, innu-
merable tokens. The Abhidharmika notion of dharmas is closely related
to what are, in this literature, the conceptually cognate notions of
(in this context, "defining characteristic") and svabhiiva
("essence" or "intrinsic nature"). Thus, for example, Vasubandhu's
explains that dharmas (literally, "bearers") are
so called "because they bear (...Jdhr) That is, what dis-
tinguishes something as exemplifying one of the 75 categories of onto-
logical primitives (one of the dharmas) is the fact of its sharing the same
defining characteristic that is common to all instantiations of that dharma.
Thus, to bear such a "defining characteristic" or "essence" is, in effect,
to qualify for inclusion in this final ontology.
Among the significant points about this understanding of
is that each of these amounts to a sort of property belonging to a
9 Cf., Cox (1995), p.12; Chaudhuri (1976), p.14(a). The language of "supervenience"
is borrowed from Kapstein (1987), pp.90, ff.
10 Pradhan (1975), p.2.1O: On the connection between
(as "defining characteristic") and svabhava (as "essence"), cf., inter alia, Cox
(1995), p.12, as well as ad 6.14cd (Pradhan 1975, p.341.11-12):
KayafTl svasarnanyalak(!alJabhyafTl VedanafTl cittafTl dharrnas ca. Svabhava
"The body is investigated in terms of its defining and general
characteristics. Feeling and thought are dharmas; the essence of these is their defining
characteristic. "
dharma. That is, this discourse speaks of dharmas as, the irreducible
remainder of reductionist analysis, and speaks of these, in turn, as indi-
viduated or characterized by the defIning properties that belong to them
- as, for example, perceptual awareness (vijfulna) is definitively char-
acterized in terms of some "conception regarding an object"
tivijfiapti), or as earth (Prthivf) is definitively characterized by "hard-
ness" or "resistance" (khara or kiithtnya)Y There is thus an important
sense in which the in virtue of which dharmas qualify as
such are, in fact, universals or abstractions; for, say, a "conception
regarding an object" is something that belongs to (and definitively
characterizes) every instance of perceptual awareness - characterizes
each, that is, as a token of the type of thing that belongs in a final
ontology. The abstract nature of such "defining characteristics" fig-
ures particularly prominently in Sarvastivadin arguments for the exis-
tential status of past and future moments of time. Thus, as Collett Cox
The tenn "intrinsic nature" [svaZak,l'a('la] does not indicate a factor's
[i.e., dharma's] temporal status, but rather refers to its atemporal underly-
ing and defining nature. Intrinsic nature thus determines the atemporal, exis-
tential status of a factor as a real entity (dravya). Nevertheless, it is pre-
cisely in this sense of intrinsic nature that factors can be said to exist at all
times (svabhiivaly. sarvadil cast!); intrinsic nature, as the particular inherent
characteristic, pertains to or defmes a factor in the past, present, and future,
regardless of its temporal status. (1995, p.139)
But even for Sautrantikas who, following Vasubandhu's Abhidhar-
reject this specifIcally temporal application of the point,
it is nevertheless the case that the that individuate existents
as belonging to one or another dharmic category are fundamentally
abstract. This is, I will suggest, among the salient points that will be trans-
formed by Dignaga's use of the term, and it will be rendered clear by
CandrakIrti's urging of .Abhidharmika usage against Dignaga.
11 The adducing of as the of vijniina occurs at Abhid-
harmakosa 1.16a (Pradhan, p.ll), and khara as that of prthivi at Abhidharmakosa 1.12
(Pradhan, p.8). For kiifhinya as synonymous with khara, cf., Pradhan, pp.24.3, 78.7-8.
These are the examples of which, as we will see, CandrakIrti adduces contra
Dignaga; cf.,nA8, below.
While macro-objects such as (paradigmatically) persons can thus be
reduced to their basic parts, such reductionist analysis is thought by Abhid-
harmikas to be capable of reaching bedrock, in the form of the dharmas
that are individuated by uniquely defining characteristics. In Abhidharmika
literature, this intuition that reductionist analysis can yield ontological
primitives is also advanced in terms of a debate regarding what is
dravyasat and what is prajfiaptisat - that is, regarding, respectively,
what "exists as a substance," and "what exists as a prajfiapti"12.
Paul Williams, borrowing from Brentano, aptly renders these as (respec-
tively) primary and secondary existence
, and emphasizes that what is
at stake here is not so much what exists, as how it exists. Thus, things that
exist as prajfiapti (prajfiaptitab) are invariably reducible to things that
exist as ontological primitives (dravysat); the latter, in turn, exist irre-
ducibly. In Vasubandhu's massive Abhidharmakosa and his
thereon, the most prominently recurrent debate concerns the question of
precisely which things are to be admitted as being dravyasat. Thus, if we
follow the traditional doxographic view (according to which Vasubandhu's
commentary reflects a Sautrantika critique of the perspective
reflected in the kiirikiis), we might characterize the as onto-
logically promiscuous, and the Sautrantikas as ontologically parsimo-
nious; for throughout the course of Vasubandhu' s massive work, various
Buddhist categories are introduced and considered, with the
characteristically asserting that they exist dravyatas ("substantially"),
and the Sautrantikas invariably rejoining that, in fact, they only exist
prajfiaptitas ("derivatively" or "superveniently," we might say).14
12 I leave prajfiapti untranslated since I am dissatisfied with the customary rendering
of this as "concept" (cf., e.g., Warder 1971). In the Madhyamika context, much depends
on the rendering, since the notion of upadaya prajfiapti is pivotal for Nagarjuna and
CandrakIrti. I have argued elsewhere (Arnold 2002) that, particularly as deployed by Bur-
ton (1999), the translation of this as "concept" is highly misleading.
13 Williams (1981). This is one of the best discussions of the conceptual motivation
behind Abhidharmika discussions of dravyasat and prajfiaptisat. See also the discussion
by Kapstein (1987), pp.90, ff.
14 So, for example, the famous debate, in the fifth chapter of the Abhidharmakosa,
regarding the existential status of past, present, and future moments. The characteristically
claim is that all three "really" exist, and that this reflects the proper interpre-
tation of the Buddhist scriptural passage (sarvam asti, "everything exists") that gives
adherents of this school the name "Sarvastivarla" (the '''everything exists'-affmners").
The reason it matters so much how these terms are ,allocated is that,
given the intuitions that motivate the Abhidharmika project, what is
dravyasat ("substantially" or "primarily existent") is, ipso jacto, admit-
ted as being paramarthasat ("ultimately existent," "real," or "true").
On this view, in other words, the characteristically Buddhist contention
that there are two levels of "truth" (" conventional truth," saY(lvrtisatya,
and "ultimate truth," paramiirthasatya) has a specifically ontological
correlate: what is conventionally true is what is reducible, by way of
critical analysis, to what is "ultimately real"; the latter category, in
tum, thus consists in an enumerable set of ontological primitives. In an
often-cited passage, V asubandhu' s Abhidharmakosa makes this point
There are also two truths, conventional truth and ultimate truth. What are
the characteristics of these two? ... The conventionally true is that with
respect to which the concept does not arise when it is broken into parts, as
for example a jar; for with respect to that, when it is broken into pieces
(kapalaso bhinne), the idea of a jar does not arise. And that with respect to
which, having excluded other dharmas by way of the intellect, the idea does
not arise - that, too, should be known as conventionally true, as for exam-
ple water; for with respect to that, having excluded, through the intellect,
other dharmas such as form, the idea of water does not arise. Everything else
is ultimately true; with respect to this, even when broken, the idea still
arises, even when other dharmas are excluded by way of the intellect - that
is ultimately true, as for example, form. IS
Vasubandhu the Sautrantika rejoins that he does not deny that these exist; he simply rejects
the claim regarding how they exist. Thus, "We, too, say the past exists; but
the past is what existed previously, and the future will exist with respect to [presently] exis-
tent causes. And in this sense they exist, but not substantially." (Pradhan [1975], p.299.1ff:
vayam api briimo 'sty atftanagatam iti; atftaT(l tu yad bhiitapiirvam, anagataT(l yat sati
hetau EvaT(l ca krtva-astzty ucyate na tu punar dravyatal:z.) On this whole
debate, see Williams (1981) and Cox (1995, passim), who both provide very illuminating
15 Pradhan (1975), pp.333-4: dve api satye saT(lvrtisatyaT(l parmarthasatyaT(l ca. Tayol:z
... Yasminn avayavaso bhinne na tad buddhir bhavati tat saT(lvrtisat, tady-
atka ghatal:z; tatra hi kapaiaso bhinne ghatabuddhir na bhavati. Taira ca-anyan apohya
dharman buddhya tad buddhir na bhavati taccapi saT!lvrtisad veditavyam, tadyatka-ambu;
tatra hi buddhya riipadfn dharman apohya-ambubuddhir na bhavati .... Atonyatha para-
marthasatyam; talra bhinne 'pi tad buddhir bhavaty eva; anyadharmapohe 'pi buddhya
tat paramarthasat, tadyatha riipam.
Collett Cox explains: "If the notion ofa particular entity disappears
when that entity is broken (e.g., a pot) or can be resolved by cognition
into its components (e.g., water), that entity exists only conventionally.
Entities that are not subject either to this further material or mentalanaly-
sis exist absolutely. Thus, actual existence as a real entity (dravyasat) is
attributed only to the ultimate constituent factors, which are not subject
to further analysis"16. As an example of the latter, Vasubandhu has here
adduced the case of "form" (rupa) - presumably in the sense of the
first of the five skandhas

III. Digniiga and the culmination of Abhidharmika commitments
That the foregoing represents the basic set of intuitions inherited by Dig-
nag a is perhaps most clear in his Alambanaparzk$ii ("examination of
intentional objects"). This very short text - which consists in only eight
kiirikiis together with a brief auto-commentary - represents Dignaga's
attempt to argue that awareness can satisfactorily be explained provided
only that we posit some mental phenomena as the "objects" intended by
awareness; and indeed, that we cannot coherently posit any non-mental,
external objects as what is directly intended by awareness. The latter is
true insofar as any account of external objects necessarily presupposes
some version of minimal part atomism, which, Dignaga argues, cannot
coherently be adduced to explain our awareness of macro-objects. Clearly,
16 Cox (1995), pp.138-9. Cf., Williams (1981), p.237: "Saq1ghabhadra [i.e., the
whose Nyayanusara - now extant only in Chinese translation - is tradi-
tionally held to represent a rejoinder to Vasubandhu' s Sautrantika criticisms 1 adds that the
distinction between primary and secondary existence corresponds to that between ultimate
and conventional truth (paramartha and saT[lvrtisatya). This point is extremely important
for it shows that in the Sarvastivada the distinction between satyas was not soteriological
but primarily philosophical, in this case ontological."
17 With the represented as admitting the skandhas to be dravyasat.
The Sautrantikas, in contrast, deny that the five skandhas exist as dravyasat, instead favor-
ing the view that what is dravyasat are the 75 dharmas into which, inter alia, the skand-
has can be reduced. Thus, for Sautrantikas the category of rupa-skandha exists only sec-
ondarily (prajfiaptisat) insofar as it comprises the first 11 dharmas in the standard list of
75 (specifically, the five bodily senses, together with their respective objects, plus the cat-
egory of avijfiaptirupa).
Dignaga's argument here owes something to Vasubandhu's Vi1!lsatikti

As with the latter work, there is some scholarly disagreement about
whether Dignaga is best understood as arguing here for an idealist meta-
physics, or simply for something like a representationalist epistemology
involving sense-data (which allows the possibility of remaining neutral
with respect to what might finally exist in the world).19
Be that as it may, what is of greatest interest to me here is Dignaga's
clear allusion to the passage from Vasubandhu (considered above) on the
"two truths." Thus, arguing that there is an unbridgeable gap between
atoms as the putative cause of awareness, and medium-sized dry goods
as the content thereof, Dignaga says: "Things like jars are [merely] con-
ventionally existent, because if the atoms are removed, the awareness that
appears with respect to them is destroyed [k.5c-d]. In the case of what is
substantially existent, such as color, even when one has taken away what
is connected with it, there is no removal of the awareness of the color
itself"2o. Like Vasubandhu, Dignaga thus suggests that what qualifies
medium-sized dry goods (of which jars are a stock Indian example) as
merely "conventionally existent" (kun rdzob tu yod pa; Skt., sa1!lvrtisat)
is the fact of their being reducible, while the constituent parts to which they
can be reduced (such as "color," kha dog, which is shorthand for n1pa and
the other skandhas) in tum exist "substantially" (rdzas su, dravyataJ:t).
18 On Vasubandhu's arguments against atomism in the Vil]lsatika, see, inter alia,
Kapstein (1988).
19 Hayes (1988), for one, opts for the latter characterization (with respect both to
Dignaga and to Vasubandhu), and calls the view "phenomenalism"; see pp.96-104
(on Vasubandhu) and pp.173-178 (on the
20 As with Dignaga's other works, the survives only in its Tibetan
translation. I have used the edition of Tola and Dragonetti (1982), which gives: bum pa
la sogs pa ni kun rdzob tu yod pa Hid do I urdul phran yons su bsal na ni I der snan ses
pa nyams 'gyur phyir II" rdzas su yod pa rnams la ni 'brei pa can bsal du zin !cyan kha
dog la sogs pa biin du ran gi blo 'dor pa med do. (The part italicized in the text repre-
sents kariM. 5c-d, which is, in the edition of Tala and Dragonetti, marked off with quota-
tion marks. I will follow the convention of italicizing portions from the kiirikas in all sub-
sequent citations from the PramiiIJasamuccaya, as well.) Note that Hayes (1988, p.I77)
translates, "In the case of what is rigorously real ... " (my emphasis) - which suggests
that the text reads paramartha;at, for which Hayes has adopted the translation equivalent
"rigorously real." But the text in fact reads rdzas su yod, which suggests instead dravyasat
- though as I have been arguing, the two terms are, in the Abhidharmika context which
presently concerns us, conceptually co-extensive.
In the PramalJasamuccaya, Dignaga alludes to the same discussion, this
time explicitly putting the issue in terms of what is "ultimately existent"
(paramarthasat). Thus, arguing that a cognition cannot properly be named
after the object that produces it, Dignaga says: "These individual [atoms],
when aggregated, are the cause [of cognition], but it is not the aggregate
[itself that is causally efficacious], since this exists only conventionally ....
if [a cognition be produced] from an object, that [object] must be [a real
entity, and what is real is] ultimately unnamable ... "21,
Clearly, then, Dignaga's understanding of the reductionist project (and
correspondingly, his understanding of the two truths as consisting in the
enumerable sets of those existents that are reducible and those that are not)
is substantially the same as Vasubandhu's, and we can safely say that
Dignaga's notion of svalak$alJa thus represents one of several signifi-
cantly correlated terms: svabhava, svalak$alJa, dharma, dravyasat, para-
marthasat. That is, Dignaga's notion of svalak$alJa represents the cul-
mination of the .Abhidharrnika intuition that there exist basic (dravyasat)
and irreducible entities - ontological primitives which are the sole
remainder of critical analysis, and which are defined vis-a-vis svalak$ar;as;
and that the "ultimately real" or "ultimately true" (paramarthasat) con-
sists in an enumerable set of such things
, The same point is particularly
clearly put by DharrnakIrti, who thus elaborates Dignaga's ideas vis-a-vis
the category of pragmatic efficacy (arthakriya): "Whatever has the capac-
ity for pragmatic efficacy is said in this context to be ultimately true;
21 PramiilJasarnuccayavrtti ad 1.15: de dag bsags pa na yan so so ba rgyu yin gyi de
bsags pa ni rna yin te tha sfiad du yod pa'i phyir ro .... "gan las de ni don dam par / de
fa tha sfiad du rna byas /" [1.15c-d]. The Tibetan (per the translation of Kanakavarrnan)
is at Hattori (1968), p.189, with Hattori's translation at pp.34-5. I have followed Hattori's
translation partiCUlarly of the kiirikii, retaining his insertions; cf., Hattori's nn.2.24-25
(p.120) for an elaboration, together with relevant Sanskrit fragments. .
22 In this regard, the thumb-nail doxographical sketch provided by the 18th-century
dGe-lugs-pa dKon-mchog 'jigs-med dbaIi.-po is interesting, and quite accurately states
what I have here taken to be the most significant aspect of this approach: "A phenome-
non that is established as bearing critical analysis with regard to its own way of being, inde-
pendent of the imputation of terms or conceptions: that is the definition of ultimate truth.
'Existent,' 'ultimate truth,' 'svalak:falJa,' 'impermanent,' 'constructed,' and 'truly estab-
lished' are synonyms." (Mimaki, p.84: sgra dan rtog pas btags pa la rna [tos par ran gi
sdod lugs kyi nos nas rigs pas dpyad bzod du grub pa'i chos de don dam bden pa'i rntshan
fiid / dnos po dan / don dam bden pa dan / ran mtshan dan / mi rtag pa dan / 'dus byas
dan / bden grub rnarns don gcig/ )
everything else is conventionally true. These two [sets consist, respec-
tively, in] unique particulars (svalak!fwta) and abstractions (siimiin-
yalak!fa1}a). "23
In what sense, though, does Dignaga's understanding of svalak!fa1}a
represent, as I have put it, the "culmination" of A.bhidharrnika intuitions?
That is, just what are the svalak!fa1}as which, for Dignaga, thus constitute
the set of really existent things? It seems to me that it is in his concep-
tion of this notion that Dignaga perhaps most significantly parts com-
pany from Vasubandhu
I have indicated that a salient point about the
A.bhidharrnika usage of svalak!fa1}a is that it denotes some property
- specifically, the "defining characteristics" of which dharmas are the
"bearers." Indeed, Vasubandhu's etymology of the word dharma
(svalak!fa1}adhiiralJiid dharma, "it is a 'bearer' because of 'bearing' a
defining characteristic") turns on precisely this notion. I have also sug-
gested that the svalak!falJas ("defining characteristics") thus "borne" by
dharmas are abstract or universal, in that any instance of some dharma
qualifies as such by virtue of its sharing with every other instance of that
dharma the property which defines it as belonging in a [mal ontology.25
We can highlight the contrast with this A.bhidharmika usage by noting
that Dharmakfrti understands svalak!falJas as unique, objective particulars
of some sort - specifically, as the kind of vanishingly small bare par-
ticulars that fit with Dhannaklrti's metaphysics of "momentariness"
23 PrarnalJavarttika 2.3 (Miyasaka [1971/72], pA2): arthakriyasarnarthal'Jl yat tad atra
pararnarthasat / anyat sal?lvrtisat proktal'J1 te svasarnalJyalak:jalJe 1/. It is important to note
(as Hayes and Katsura have) that the notion of "pragmatic efficacy" (arthakriya) as the
criterion of the ultimately real is among Dha=akIrti's innovations.
24 Katsura (1991, p.l36) agrees, saying with respect to svalak/ialJas that "Dignaga
accepted the Abhidharrnika's concepts of them at least in general. Nonetheless, he appears
to have attached to them new significances." In characterizing this as Dignaga's most sig-
nificant departure, I am only speaking, of course, in te=s of the issues relevant to the pres-
ent discussion. A more comprehensive account of Dignaga' s innovations would of course
have to assess the significance of his apoha doctrine, and of his fo=ulation of rules for
valid inferences.
25 Again, cf. Katsura (1991, p.l37): " .. .it is clear that svalak/ialJas of Abhidharrna,
viz. dharrnas which are actually named as rupa, vedana, etc., should be regarded by Dig-
naga not as svalak:jal}as but as sarnanyalak:jalJas. Consequently, Dignaga's sarnanyalak/ia(w
corresponds to both sva- and sarnanyalak/ialJa of the Abhidharrna, which cannot be regarded
as real in Dignaga' s system."
Thus, for example, John DUIDle has urged that, on Dhar-
maklrti's understanding, it must be the case that have no spa-
tial extension.
At least in Dharmaklrti's thought, then, the Abhidharmika
tradition reaches its culmination in the insight that the irreducible. onto-
logical prinlitives in the system Camlot be said themselves to have any
properties; for if they did, they would be reducible (i.e., into dharma and
dharmin, "property" and "property-possessor"). Thus, it is no longer the
case that are the "defining characteristics" possessed by dhar-
mas; rather, just are the ontological primitives on this view,
and they are simply "self-characterizing." Dunne (1995: 195) nicely
expresses the upshot: "This is best illustrated by a genitive construction
such as, 'The nature of the infinitesimal particle.' Dharmaldrti maintains
that in such expressions the dharma is actually identical to the dharmin
itself. The apparent separation of the dharma from the dharmin is simply
part of the exclusion process, and is hence conceptual." This reflects an
extension or "culmination" of the Abhidharmika project, then, insofar as
the idea of irreducibility has here been taken to its logical extreme, such
that what is irreducible cannot be said even to have aI!-y properties -
here, in other words, we have the idea that for ontological prinlitives even
to be simply logically reducible is to compromise the basic idea.
As I have noted, though, several scholars have recently challenged the
idea that Dignaga understood in the way that Dharmaklrti thus
understood them. That may be the case. In fact, though, it seems to me
that it is not altogether clear what Dignaga means by since
he never formally defines the concept. Indeed, about the only thing Dig-
naga says about is that they are "indefinable" or "unspeci-
fiable" (avyapaddya; "ineffable" is a frequently met translation for this).
Thus, Dignaga begins his Pramii7J.asamuccaya by arguing:
Perception and inference are reliable warrants. There are only two, since
there are [only] two [kinds of] warrantable objects; there is nothing war-
rantable other than svalak,ymJas and abstractions. It is perception that has
svalak,ymJas as its objects, and inference that has abstractions as its objects
26 Cf., Dunne (1999), p.13l.
27 Pramiil:zasamuccayavrtti ad 1.2: .. mnon sum dan ni rjes su dpag / tshad ma dag niH
gfiis kho na ste, gan gi phyir .. mtshan fiid gfiis / gial byaH ran dan spyi'i mtshan fiid dag
As for the "sphere of operation" (gocara) of the perceptual senses
(indriya): it is the "indefinable (anirdeya) form which is to be known
in itself. "28 Later on, in. contesting the Naiyayika account of percep-
tion (which has it that perceptual awareness is "ineffable"),29 Dignaga
urges that this qualification is unnecessary, because redundant.
He explains: "It is not possible that a definable (bstan par bya ba)
object be the object of a sense-cognition (dbail po 'j blo, =Skt.
indriyabuddhi) , since what is definable is [always] the object of infer-
ence. [Therefore,] there is no [possibility of a sense-cognition's] vari-
ance in regard to indefinability. "30
Just what does it mean, though, for svalak$WLas thus to be "indefin-
able"? Here again, Dignaga's own account is frustratingly underdeter-
mined. What nevertheless seems clear, though, is that this idea has some-
thing to do with Dignaga's taking them to be the objects (Tib., yul, Skt.,
vi$aya) of perception - and, in tum, with his characteristic insistence on
the fact that perception (pratyak$a) is definitively characterized by its
being "free of conceptual elaboration" (kalpanapocJha). Apropos of this,
Dignaga says: "Perception is free from conceptual elaboration; that
awareness which is without conceptual elaboration is perception. And
what is this which is called 'conceptual elaboration'? Association with
name, genus, etc"31. The basic idea here is that a bare perceptual event
is constitutively non-linguistic, with the subsequent addition of linguis-
tic interpretation representing, among other things, the point at which
las gian pa'i gial bar bya ba med do. ran gi mtshan Rid kyi yut can ni milon sum yin la
spyi' mtshan Rid kyi yul can ni rjes su dpag pa'o .... Tibetan per Hattori (1968), p.l77.
28 Pram{lI;asamuccaya 1.5c-d: svasaf!lvedyam anirdesyaf!l rupam indriyagocaraJ:t (San-
skrit fragment in Hattori 1968, p.91, n.1.43, which also provides some useful elaboration;
among other things, Hattori reports an alternative reading from another source: svalaksanam
anirdeSyam .... ).
29 Cf., Nyiiyasutra 1.1.4, given by Hattori (1968) at p.121, n.3.1.
30 PramiiIJasamuccayavrtti ad 1.17: dbail po'i blo la bstan par bya ba'i yul Rid srid
pa rna yin te, bstan par bya ba ni rjes su dpag pa'i yul yin pa'i phyir yo. bstan par bya
ba rna yin pa Rid la yan 'khrul ba yod pa rna yin te .... (Hattori, p.191) Cf., also,
PramiiIJasamuccya 2.2a: ran gi mtshan Rid bstan bya min ("the svalak$alJa is indefin-
31 PramiiIJasamuccaya 1.3, with vrtti: "mnon sum rtog pa dan bral ba." ses pa gan
la rtog pa med pa de ni mnon sum mo. rtog pa ses bya ba 'di ji Ita bu sig ce na, "min
dan rigs sogs bsres pa'o."
cognitive error can creep in. To be sure, it is not necessarily the case that
any subsequent linguistic elaboration introduces error, as some such is nec-
essary merely to yield the kind of propositional knowledge which alone
could make the initial perception useful. Thus, for example, Dignaga
exemplifies the steps of the cognitive process by saying: "One [initially]
apprehends the non-conventional [i.e., because ultimately real]
(rail. .. rntshan fiid dag tha sfiad du bya ba rna yin; Skt., *avyavahiir-
and the abstraction 'being colored.' Then, by means
of the operation of the mind, one relates [being colored] to [the univer-
sal] impermanence, and expresses [the resulting cognition in the judg-
ment] 'colored things and so forth are impermanent. "'32
While discursive elaboration in terms of universals is thus held to be
indispensable to the development of propositional knowledge, it is nev-
ertheless the case that a part of Buddhism's "deep grammar," as it were,
is the idea that our cognitive and soteriological defIlements are adventi-
tious to our basic epistemic faculties, such that the removal of these defIle-
ments would leave untrammeled perception free to register things as they
really are. If discursive elaboration of our basic percepts is thus neces-
sary to yield propositional knowledge, then, it is nevertheless the case
that such, in one form or another, is also precisely the problem to be over-
come by Buddhist practice. That intuitions such as these are in play is
made more clear by DharmakIrti, who revises Dignaga's account by
adding that perception is not only "free of conceptual elaboration," but
also "non-mistaken."33 In this way, "conceptual elaboration" (kalpana)
32 PrarniilJasarnuccayavrtti ad 1.2c-d (Hattori 1968, p.177): rail dail spyi'i rntshan iiid
dag tha siiad du bya ba rna yin pa dan kha dog iiid dag las kha dog la sogs pa bzuil nas,
kha dog la sogs pa rni rtag go ses rni rtag pa iiid fa sogs par yid kyis rab tu sbyor bar
byed do. Here, I have basically followed Hattori's translation (p.24), with some adjustments;
cf., Hattori's n.1.19, p.81, for extensive Sanskrit fragments from commentaries on Dhar-
makIrti. In these, avyapadeya is again the word used to characterize svalaklfalJas. For the
Sanskrit underlying the Tibetan translation of Dignaga' s tha siiad du bya ba rna yin, I have
taken avyavahiirtavya from Chandra (1959-1961, p.lOlO), whose usage is from the
Nyiiyabindu of DharmakIrti.
33 Nyiiyabindu 1.4 (Shastri 1985, p.20): tatra pratyaklfaJ?! kalpaniiporjharn abhriintarn;
cf., PrarniilJaviirttika 2.123, ff. (Miyasaka 1971n2, pp.56, ff.). While the introduction of
this as a definitive feature perhaps represents an innovation by Dharmaklrti,
cf., PrarniilJasarnuccayavrtti ad 1.17: Having said that the Nyaya definition of perception
involves a redundant reference to avyapadeyatva (cf., n.29, above), Dignaga adds: "Nor is
is implicated as the point in the cognitive process at 'Yhich error comes
in. Moreover, DharmakIrti also expands Dignaga's contention that con-
ceptualization involves with name, genus, etc.," with the
significant adjustment that conceptualization involves simply any idea
that is suitable for association with discourse
With this emphasis, Dhar-
makIrti means to allow that conceptual activity is the sort of thing which
may be (and is in fact) found even in such pre- or non-linguistic creatures
as infants and animals - which must be the case if one is to avoid the
unwanted consequence that the main soteriological defilement does not
exist for infants or animals

Clearly, then, the idea of the "indefinability" of can
serve important intuitions about the non-linguistic character of perception
- intuitions according to which perception is thought to yield access to
uniquely uninterpreted data, which, being "knowable in themselves"
(svasarrzvedyam), amount to something that is simply "given" to aware-
ness as the uniquely certain foundation for all other knOWledge. Such has
been the contention of Tom Tillemans, who aptly appeals to Wilfred
Sellars's characterization of the "myth of the given" :
One of the forms taken by the Myth of the Given is the idea that there is,
indeed must be, a structure of particular matter of fact such that (a) each fact
can not only be noninferentially known to be the case, but presupposes no
other knowledge either of particular matter of fact, or general truths; and
(b) such that the noninferential knowledge of facts belonging to this struc-
ture constitutes the ultimate court of appeals for all factual claims - par-
ticular and general- about the world. (Sellars 1963, p.164)
there a possibility of [perception's] having an erroneous object, since an erroneous cogni-
tion has as its object an illusion produced by the mind" (Tibetan at Hattori, p.193: 'khrul
ba'i yul nyid kymi srid pa rna yin te, 'khrul ba ni yid kyi 'khrul ba'i yul Rid yin pa'i phyir
TO; cf., Hattori's n.3.7, p.122).
34 Nyayabindu 1.5 (Shastri 1985, p.25): abhilapasarrzsargayogyapratibhasapratftil;
kalpana ("Kalpana is a conception which has an appearance suitable for association with
35 It is thus important to note just how much is excluded, by DharrnakIrti, with the
characterization of perception as "kalpanapof/,ha"; for to the extent that contemporary
scholars hope to explicate DharmakIrti's thought vis-a.-vis developments in contemporary
philosophy, it becomes quite significant that DharmakIrti' s idea of what it would mean for
perceptual cognitions to be (in one contemporary idiom) uninterpreted is thus meant to
include instances (to retain the idiom) of "interpretation" even on the part of infants and
animals. The kind of "conceptualization" thus ruled out must be very general indeed.
Following this lead, Tillemans explains that for the Buddhist
Epistemologists, represent "the purely particular, known without prior
reliance on concepts or any general truths," such that "particulars
be they accepted as external or as only mental, are the sort
of thing naturally suited to be present to non-inferential awareness, and
hence can be considered as a type of given - this is what is involved in
Buddhists saying that particulars are the exclusive objects of perception"36.
While the A.bhidharmika usage had it that were "defining
characteristics," it thus seems clear that Dignaga's characterization of
as avyapaddya is part of a project in which these are now
understood as unique particulars of some sort. Surely, then, the avya-
padefyatva of thus goes hand in hand with the idea of per-
ception as definitively free of conceptual elaboration, and as having
as its object. In attempting to understand just what kind of
"unique particulars" we are talking about, though, I would like to con-
sider the possibility that Dignaga' s characterization of as
"indefinable" is meant to advance a stronger claim - one such that Dig-
naga's version of might resemble Dharmaldrti's, after all (or
at least, such that it is not clearly incompatible with Dharmaldrti's). In this
connection, it is interesting to start by noting Masaaki Hattori's transla-
tion of part of PramiilJasamuccaya 1.2 and the vrtti thereon - specifi-
cally, the passage that reads [tshad ma] dag ni gfiis kho na ste, gail gi phyir
mtshan fiid gfiis / gzal bya ..
. Hattori translates: "They [i.e.,pramiilJas]
are only two, because the object to be cognized has [only] two aspects,"
reading (with my emphasis) as though mtshan fiid gfiis
were a bahuvrzhi compound standing for gzal bya (*prameya)38. Against
such a reading, Shoryu Katsura makes what seems to me exactly the right
point about Hattori's translation: viz., that it "may suggest that the object
to be cognized is a possessor of the two and [is] something dif-
ferent from them .... [But] I do not think that Dignaga admitted any bearer
of the two And it is at this point, finally, that rwould like
36 Tillemans (2003), p. 98. See also Tillemans (1990), pp.41, ff.
37 Cf., n.27, above. Hattori's translation (1968) is at p.24.
38 Cf., the Sanskrit reconstrnction given in Katsura (1991), p.136, n.29: .,' pratyak$am
anumanarrz ca prama[le dve eva, yasmad lak$a[ladvayarrz prameyarJ1.
39 Ibid,
to introduce Candraldrti; for one reason I am inclined to opt for Kat-
sura's more straightforward reading (according to which the compound
is not a bahuvrfhi, and prameyam is more literally gerundive) is that Can-
drakfrti has seized on precisely the same conceptual issue that Katsura
here notes. Let us, then, see how Candraldrti develops this point.
Iv. Candrakfrti on Digniiga on svalalqaIfas
Having devoted a considerable part of the first chapter of the Prasan-
napadii to refuting Bhavaviveka, Candraldrti turns to address the objec-
tion of an epistemologist, who wants to know what pramiil}as warrant
the claims made in MUlamadhyamakakiirikii 1.1. Specifically, Can-
draldrti's interlocutor here wants to know whether Candrakirti's "cer-
tainty" or "conviction" (niscaya) is or is not produced by an accredited
pramiil}a ("reliable warrant")40. Candraldrti's initial rejoinder is remi-
niscent of Nagarjuna's well-known disavowal, in the Vigrahavyiivartanf,
of any "thesis" (pratijfzii), with Candraldrti here disavowing any claim
to the conceptually cognate category of "certainty" (niscaya)41.
After ringing the changes on this theme, CandrakIrti turns to consider
commitments such as are specific to Dignaga. He sets up this considera-
tion by anticipating the claim that Dignaga is merely thematizing our con-
ventional epistemic practices, and so cannot be charged with striving for the
sort of ultimacy that is only the purview of a fully realized Buddha
. Thus:
[Objection:] Or perhaps [the Epistemologist will suggest:] "It is [simply]
worldly convention (vyavahara) regarding warrants and warrantable objects
which has been explained by us through [our system's] treatise."
40 Prasannapada 55.11-15: Atra kedt paricadayanti: Anutpanna bhava iti kim ayaTJ1
prama[laja niscaya Tatra, yadi prama[laja i:jyate, tada-idaTJ1 vaktavyaTJ1:
kati prama[lani, kiTJ1lak:ja[lani, kiTJ1vi:jaya[li, kiTJ1 svata utpannani, kiTJ1 parata ubhayata
'hetuta va-iti? sa na prama[ladhfnatvat prameyadhigamasya.
Anadhigata hy artha na vina prama[lair adhigantuTJ1 sakyata iti, prama[labhiivad arthiid-
higamabhiive sati, kuta 'yam samyagniscaya iti?
41 56.4-5: Ucyate: Yadi kascinniscaya nama-asmakaTJ1 syat, sa prama[laja va syad
aprama[laja va. Na tv asti.
42 For Candraldrti had urged that the demand for putatively probative arguments (upa-
patti) makes no sense insofar as "ultimate truth is a matter of venerable silence" (57.7-8:
KiTJ1 khalv arya[lam upapattir na-asti? Kena-etad uktam asti va nasti va-iti? Paramartha
hy aryas [per de Jong]).
[Response:] Then it should be stated what the fruit of [your] explanation of
. this [i.e., of worldly usage] is. .
[The Epistemologist continues:] It [i.e., worldly usage] has been destroyed
by bad logicians (kutiJrkikaif:z), through their predication (abhidhiJna) of false
characteristics. Its correct characteristics have been explained by us.
[Reply:] If [this is said, we rejoin:] This doesn't make sense, either. For if,
based on the composition of a false definition by bad logicians, everyone
made a mistake regarding what's under definition (krtal'(!lak.tyavaiparftyal'(!
lokasya syiJt), [then] the point of this [i.e., of your proposed alternative to
Nyaya epistemology] would be one whose effort was fruitful. But it's not
so, and this effort is pointless.
It is significant that CandrakIrti thus introduces his consideration of
commitments such as are specific to Dignaga with this exchange about
whether or not Dignaga can credibly claim to be offering an account of
conventional epistemic practices; for from this point on, CandrakIrti's
governing concern will simply be that of rejecting this claim. CandrakIrti
has, in other words, introduced his survey of commitments specific to
Dignaga by setting it up in such a way that he will only need to show that
Dignaga's categories are not only not used conventionally, but cannot
even account for conventional usage. To show this, on CandrakIrti's view,
is to show that, notwithstanding his likely protests to the contrary, Dig-
naga is really trying to explain conventions by getting behind them to
something that is more "real" than they are (specifically, really existent
43 58.14-59.3: Atha syad qa eva pramaIJaprameyavyavahiiro laukiko 'smabhi1.1 sas-
treIJanuvarIJita iti. TadanuvarIJasya tarhi phalal]l vacyal]1. KutarkikaiJ:t sa nasito viparfta-
Tasya asmabhiJ:t samyaglakijalJam uktam iti cet. Etad apy ayuktal]1. Yadi
hi kutarkikair [Tib., praIJayanat ... ; adopted by Vaidya] lq-tal]1
syat. Tadarthal]1 prayatnasaphalyal]1 syat. Na ca etad evam iti
vyartha evayal]1 prayatna iti. It is with respect to this passage that the anonymous author
of the (cf.,n.7, above) specifically identifies Dignaga as Candraldrti's inter-
locutor: "He says that on this view, it makes sense only [to speak of] the worldly convention
regarding warrants and warrantable objects, not [what is] ultimate[ly the case]. [This is what
is said in the passage] beginning 'Atha ... .' ['Its correct characteristics have been explained]
by us' means by Dignaga, et al. It's the master [i.e., Candraldrti] who says, at this point,
'the fruit of this intention should be explained,' and it's Dignaga who rejoins, '[It has been
destroyed] by bad logicians.' 'It' [here] means convention." 2b4: laukika eva
pramaIJaprameyavyavahiiro yukto na paramarthika ity asmin pakije aha / athetyadi /
asmabhkr> DignagadibhiJ:t / tadanubandhanasya phalal]1 vacyam ityatraryaJ:t,
kutarkkikair iti DignagaJ:t, sa iti vyavahiiraf:t). Thanks to Y oshiyasu Y onezawa for sharing
this fragment with me.
And having shown only this much, CandrakIrti will have
reduced Dignaga's project to absurdity, insofar as CaridrakIrti's project,
as contra Dignaga's, consists (as Jay Garfield aptly says of Nagarjuna) in
"taking conventions as the foundation of ontology, hence rejecting the
very enterprise of a philosophical search for the ontological foundations
of convention"44.
With this set-up in place, then, CandrakIrti devotes the remainder of his
lengthy engagement with this interlocutor to showing that the categories
of and - specifically as they are understood and
correlated by Buddhist philosophers like Dignaga - cannot accommo-
date familiar examples of what are, after all, perfectly ordinary words in
the Indian context. Our concern here, of course, is particularly with Can-
drakIrti's reading of Dignaga' s notion of CandrakIrti begins:
Moreover, if you say there are [only] two reliable warrants, corresponding
respectively to the two [kinds of warrantable objects, Le.,] unique particu-
and abstractions, [then we are entitled to ask,] does the subject (lak.rya)
which has these two characteristics exist?46 Or does it not exist? !fit exists,
then there is an additional warrantable object; how, then, are there [only]
two reliable warrants? Or perhaps [you will say] the subject [which is
characterized by these characteristics] does not exist. In that case, even
the characteristic, being without a locus, doesn't exist, [and] how, [in that
case,] are there [as many as] two reliable warrants? As [Nagarjuna] will
say: "When a characteristic is not in play, a subject to be characterized
doesn't stand to reason; and given the unreasonableness of a subject to be
characterized, there is no possibility of a characteristic, either."
Or this could be said [by Dignaga]: It is not that' lak:jafla' means "that by
which [something] is characterized." Rather, [following Pal!ini's rule at
IIL3.113, Le.,] "krtyalyuto bahulam" ["the gerundive affix is variously
applicable"], taking the affix in the sense of an object (karmm)i), 'lak:jar;a'
means "what is characterized."
44 Garfield (1995), p.l22.
45 This is how I render svalak:ralJa when it is Dignaga's usage that is in play; one of
the difficulties in this whole section of the Prasannapadii is that of keeping clear on whose
usage is in play, with CandrakIrti urging that the word conventionally means "defIning
characteristic. "
46 Alternatively: " ... is that which has these two characteristics a 1ak:rya, or not?"; or,
to take lak:rya more literally as a gerundive, "is that which has these to be characterized,
or not?"
[Reply:] Even so, this same fault [still obtains], since that instrument by
. which something [i.e., some object] is characterized has the quality of being
a thing other $an an object (yena tal lak.yyate tasya karmJasya karmmJo
'rthiintaratviit), owing to the impossibility of something's being character-
ized by itself (tenaiva tasya lak.yyamiilJatviisarrtbhaviid). 47
In this way, Candraldrti's opening salvo trades on the notion that the
idea of a "characteristic" is by definition the idea of a rela-
tionship - specifically, a relationship between a "characteristic"
and the "thing characterized" thereby Thus, CandrakIrti urges
that Dignaga's sva- and precisely insofar as they are
(etymologically) types of "characteristics," must be instantiated in some
subject of characterization - which Dignaga cannot admit with-
out compromising his commitment to the view that there are only two
types of existents, since the subject in which these were instantiated would
seem to represent an additional existent. On the other hand, it is incoherent
to suppose that these are not the "characteristics" of anything, since the
conventional understanding of the term definitionally involves the char-
acteristic / characterized relationship.
What is particularly interesting for our present purposes is that Can-
drakIrti presses this point - viz., that the idea of a "characteristic" defin-
itively includes the idea of a "subject of characterization" - as an
unwanted consequence for Dignaga. That he does so clearly reflects his
having read Dignaga as wanting to claim, to the contrary, that
are irreducible, being neither the characteristics of nor characterized by
anything else. Thus, on Candraldrti's reading (as on Katsura's), Dignaga
did not wish to admit any separate "bearer" (dharma!) of his
which is precisely why CandrakIrti can (as he does) urge that Dignaga
must think of as simply "self-characterizing" - against
47 59.7-60.3: Ki1?1 ca yadi svasamanyalak$a1Jadvayanurodhena pramal}ldvayam ukta1?1,
yasya tallak$a1Jadvaya1?1 ki1?1 tallak$yam[per de Jong; so, too, Vaidya] asti? Atha nasti?
Yadyasti, tada tadapara1?1 prameyam astfti, katha1?1 pramaradvaya1?1? Atha nasti lak$ya1?1,
tada lak$aram api nirasraya1?1 niistfti katha1?1 pramiiradvaya1?1? Vak$yati hi:
" lak$ariisa1?1pravrttau ca na lak$yam upapadyate, lak$yasya anupapattau ca lak$arasyapy
asa1?1bhaval;," iti. (This quotes MMK 5.4.) Atha syiin na lak$yate 'neneti lak$ara1?1. Ki1?1
tarhi "krtyalyuro bahulam" iti karmari lyufa1?1 krtvii lak$yate tad iti lak$ara1?1. Evam api
tenaiva tasva [per de Jong]lak$yamiiratvasa1?1bhavad - vena tallaksvate [per de Jong]
tasya kararasya karmaro 'rthiintaratviit - sa eva dO$al;.
which view, CandrakIrti here exploits standard grammatical analyses of
the "characterizing" relationship as necessarily involvIng both a charac-
teristic and a thing characterized
To complete CandraIdrti's basic characterization of Dignaga's posi-
tion, we need only attend to the passage that immediately follows what
we have just seen. Having thus charged that his interlocutor's account
incoherently posits something essentially self-characterizing, CandrakIrti
now anticipates moves intended to salvage the possibility of such a thing.
Ultimately, this will lead to a consideration of Dignaga's account of
svasarrzvitti (reflexive awareness), which will be adduced as the unique
example of something that is at the same time both an object and an
instrument. We can, however, appreciate CandrakIrti's basic point with-
out entering that thicket; for the main point in what follows is CandrakIrti's
clarification of the sense in which Dignaga's understanding of
differs from what CandrakIrti takes to be the conventional sense. Thus:
[The Epistemologist rejoins:] Well, perhaps this could be said: Because
awareness (jfllina) is an instrument, and because this [i.e.,jiiiina] is included
in [our concept of] the unique particular, there is not the fault [with which
you charge us].
[Reply:] Here [in the world], that which is the nature (svariipa) of existents,
[i.e.,] their own, not shared with anything else, that is their defining charac-
teristic For example, the earth's [defining characteristic] is
resistance, [the defining characteristic] offeeling is experience, [and the defin-
ing characteristic] of awareness is a conception regarding any object; for [in
each of these cases,] by that [quality the thing in question] is characterized.
By one who, disregarding (avadhaya) the usage which follows the familiar
sense based on this (iti krtvii), [instead] accepts the definition [of
as an object, and positing [at the same time] the instrumental nature of per-
ceptual awareness, it is said [in effect] that one unique particular has the
quality of being an object, and another unique particular has the quality of
being an instrument. In this case, if the unique particular which is perceptual
awareness is an instrument, then it must have a separate object (tasya vyatirik-
tena karmalJii bhavitavyam). This is the fault [in your position].48
48 60.4-61.2: Atha syiit: Jniinasya karaIJatviit, tasya ca svalak:falJiintarbhiiviid, ayam
ado:fa iti. Ucyate: Iha bhiiviiniim anyiisiidhiiraIJam iitmfya/?1 yat svariipal'{l, tat svalaqalJal'{l.
Tadyathii prthivyiif:t kii!inyal'{l, vedaniiyii anubhavo [per de Jong, Tibetan], vijniinasya
vi:fayaprativijnaptif:t. Tena hi tal lak:fyata iti /qtvii, prasiddhyanugatiim [per de Jong,
Tibetan] ca vyutpattim avadhiiya karmasiidhanam abhyupagacchati. Vijniinasya ca
In this particular case, the examples of the conventional usage - which,
we note, are taken from the Abhidharmakosabhiiryam
- are particularly
contrasted with a"usage which takes the word as "denoting an
object" (karmasadhanam)50. It is not only with reference to Candraldrti's
chosen examples, then, but also with this contrast in mind that I render
when CandrakIrti uses the word as he thinks it is conven-
tionally used, as "defining characteristic"; and what seems to character-
ize the relationship between, say, earth and its "defining characteristic"
(viz., "hardness" or "resistance") is the fact of their being inseparable,
such that "hardness" is not an object that could be perceived apart from
karaJ}abhavafJ1 pratipadyamanena-idam uktafJ1 [per de Jong, Tibetan] bhavati,
svalak$Gf}asyaiva karmatii, svalak$alJantarasya karaIJabhavasceti. Tatra yadi vijRanas-
valak$alJafJ1 karaIJafJ1, tasya vyatiriktena karmaIJii bhavitavyam iti sa eva
Though conceptually fairly straightforward, particularly the latter part of this section
is grammatically tricky, and I have found it helpful to consult the Tibetan here, with par-
ticularly those passages underlined in the following having proved useful: ci ste ses pa
byed ba yin pa'i phyir la I de yan ran gi mtshan Rid kyi khons su 'du ba'i phyir Res pa
'di med do sRam na Mad par bya ste I re iig 'dir ji Itar des de mtshon par byed 'di sa'i
sra ba dan I tshor ba'i myons ba dan I rnam par ses pa'i yul so sor rnam par rig pa Itar
bdag Rid kyi ran gi flO bo gian dan thun mon ma yin pa gan yin pa de ni ran gi mtshan
Rid yin na I rab tu grags pa dan ries su 'brei pa'i bye brag tu Mad pa bor nas I las su
sgrub pa khas len iin rnam par ses pa byed pa'i no bor rtogs pas ni I ran gi mtshan Rid
kho na las Rid yin iin ran gi mtshan Rid gian ni byed pa'i no bo yin no ies bya ba 'di
smras par 'gyur ro I.
49 Cf., n.ll, above. Cf., also, Madhyamakavatara 6.202-3 (La Vallee Poussin's edition,
p.316), where CandrakIrti trots out a similarly Abhidharmika list of "defming character-
istics" (svalak$alJas) of all" of the skandhas: "Form has the defining property (svalak$alJa)
of color and shape; vedanii has the nature of experience; safJ1jfzii grasps characteristics;
safJ1skaras fashion [things]; the defining property of perceptual awareness is a conception
regarding any object" (gzugs ni gzugs run [sic; read ran] mtshan Rid can I tshor ba myon
ba'i bdag fzid can I 'du ses mtshan mar 'dzin pa ste I 'du byed mnon par 'du byed pa'o II
yulla so sor rnam rig pa I rnam ses ran gi mtshan Rid do I). Cf., inter alia (and in addi-
tion to n.10, above), Abhidharmakosa 1.14 (Pradhan, p.lO: safJ1jfzii
nimittodgrahalJiitmika), which gives some more of the defming characteristics repeated
here by CandrakIrti.
50 As is evident in nA8, above, the Tibetan translation renders this as las su sgrub pa,
"established as an object." For the sense of -sadhana as "denoting" or "expressive of,"
I follow Apte, p.1666, meaning #4. On the compound karmasiidhana, see also Renou
(1942), p.125, who gives: "qui a l'objet-transitif (i.e. une notion passive) pour mode de
realisation." We could easily follow this lead and transpose this discussion into the key
of grammatical terms (hence, e.g., "denoting an accusative"), with little change in sig-
"earth." That is, when one encounters an instance of "earth," one just is
encountering an instance of "hardness"; indeed, this is Just what it means
for the latter to be a defining characteristic of the former. Moreover, as
I have suggested, this understanding of qualifies it as what
Dignaga would consider to be a universal; for a "defining charaCteristic"
is what is common to any and all instances of the thing defined thereby.
Thus, to be sure, we can separate a thing and its defining characteristic
analytically, as we do when we specify which is the thing being defined
("earth," the and which is the thing adduced as its definition
("hardness," its What we cannot do, though, is encounter
these separately as ontologically "given" entities. And this is precisely
what is required, according to CandrakIrti, on Dignaga's usage of the
term, which is such that we can render as "unique particular"
when it is Dignaga who is using the tenn. For Dignaga, are
the unique objects of the cognitive act which is perception, they are what
(following Tillemans) perception encounters as "naturally suited to be
present to non-inferential awareness."
Taken together, the two passages we have considered from CandrakIrti
provide a clear sense of how he considers Dignaga' s doctrine of
to differ from the examples of the A.bhidharmika usage of the
tenn adduced by CandrakIrti. Thus, CandrakIrti rightly understands the
A.bhidharmika usage as not denoting any kind of object; rather, it denotes
the sort of "defining characteristics" which are, in fact, abstractions.
Candraldrti would concur, then, with Shoryu Katsura, who notes that
of Abhidharma ... [must] be regarded by Dignaga not as
but as This point is underscored by Can-
drakIrti's characterization of Dignaga's usage as karmasadhanam, "denot-
ing an object." Moreover, as CandrakIrti stresses in the first of the two
passages we have just considered, "defining characteristics" are the kinds
of abstractions that are definitively inextricable from the existents they
define - they are not only not objects, but it makes no sense to think of
encountering them in the way we encounter objects, since they are defin-
itively relational abstractions. Among other things, this means they are
necessarily instantiated in some some "bearer" of the defining
51 Cf. n. 25, above.
property in question - and on Candraldrti's reading, Dignaga cannot
coherently concede this, since his position requires that there be no addi-
tional kind of eXIstent to which could belong.
V. Are Digniiga' s svalalqaI).as "bare particulars" ?
As I have suggested, what is most significant for Candraldrti is that Dig-
naga's usage cannot accommodate what Candraldrti takes to be attested
usage, with demonstration of such failure of adequacy to conventions
being sufficient for Candraldrti's purposes. It is interesting, though, that
Candraldrti reduces this failure to absurdity particularly by way of an
argument to the effect that any attempt by Dignaga to accommodate con-
ventional usage will issue in infinite regress. This fact makes clear how
Candraldrti reads Dignaga's doctrine of for all of the fore-
going centrally has to do with CandrakIrti's basic rejection of
understood as self-characterizing. Candraldrti's argument here can plau-
sibly be characterized, I think, as fundamentally similar to some con-
temporary arguments against the sort of "bare particulars" presupposed
by "substratum theories" similar to the Abhidharmika version of reduc-
tionism. Thus, the view that medium-sized dry goods are reducible to
more fundamental constituents is often expressed in terms of a "bare sub-
stratum" in which various properties are instantiated, but which is itself
without any properties. Such an account is intended to bring the exercise
of reductionism to rest, explaining the numerical diversity of ontological
primitives without presupposing that the reducible properties are them-
selves such primitives. It has been persuasively argued, however, that the
idea of bare particulars as the "ultimate" (i.e., because themselves irre-
ducible) exemplifiers of the properties of a whole is incoherent, insofar
as putatively bare particulars can always be essentially characterized
- that is, characterized by such "essential" properties as being a sub-
stratum or a human being. Michael Loux succinctly summarizes this line
of argument:
The difficulty is that once we concede this fact, we fmd that the very prob-
lem substrata were introduced to resolve arises in their case. Substrata tum
out to be complexes or wholes themselves, complexes or wholes constituted
by the attributes essential to them. Unfortunately, the attributes essential to
anyone substratum seem to be precisely those esseptial to any other.
They are all essentially subjects for attributes, all essentially diversifiers, all
essentially different from the number seven, all essentially colored if green,
all essentially red or nbt red. But, then, while being numerically different
from each other, they begin to look like qualitatively indiscernible entities.
And so we need an account of their numerical diversity; and the only account
that will do is one that posits a lower-level substratum in each of our original
substrata, a lower-level substratum that makes each of our original substrata
different from each other. But since nothing can be bare, the same problem
arises for these new, lower-level substrata; and we seem once again to be
off on an infInite regress

We can express Loux's argument more perspicuously simply by point-
ing out that any particular must at least have the "property" of being a
unique particular - with the latter being an abstract state of affairs that
can be said to be a universal or abstraction (i.e., since the property "being
a unique particular" is one that is shared by all unique particulars! )53.
But in that case, the basic problem of how particulars are characterized
(which is essentially the problem of how particulars are related to their
defining properties) is not avoided by claiming that particulars are defined
as such simply by their having only themselves as "characteristics";
for this move opens an infinite regress insofar as there remains a sense
in which this characterization itself necessarily involves a relationship
between characteristic and thing characterized (lak:;alJa and lak:;ya). I sug-
gest that Candraklrti's opening argument against Dignaga's svalak:;alJas
trades on a fundamentally similar point. Thus, the point that Candraklrti
makes in terms of the "characterizing" relationship is that it is incoher-
ent to think that anything without characteristics (any "bare particular")
could in the end be all that really exists, insofar as any object (karman)
we encounter as possessing characteristics must be in relation to what
52 Loux (1998), pp.116-17. Among the thinkers whom Loux here follows is Sellars;
cf., Sellars (1963), pp.282-3n.
53 Thus to suggest that "bare particulars" must at least be capable of being "essentially"
characterized is, it seems to me, basically to make David Armstrong's point that a truly
bare particular "would have no nature, be of no kind or sort" (Armstrong 1989, p.94); and
the argument is that this is self-referentially incoherent insofar as saying something is a
"bare particular" just is to say that it is of some kind or sort. As Armstrong puts it: "Per-
haps a particular need not have any relations to any other particular - perhaps it could
be quite isolated. But it must instantiate at least one property." (Ibid.)
characterizes it (karal}a) - with the force of necessity here coming from
the unavoidability of talk about what Loux has called "essential charac-
teristics"54. And just as with the line of argument summarized by Loux,
the logic of Candraldrti's argument against Dignaga similarly trades on
the charge that Dignaga's account involves an infinite regress - with
such an argument gaining its power insofar as it is precisely the point of
Dignaga's project to bring the reductionist project to rest in something not
further reducible.
It can of course be questioned, though, whether abstract properties (like
the property being a unique particular) should be admitted as in any sense
"real." Indeed, to the extent that Dignaga's whole project centrally
involves the denial even of first-order property-universals, it might be
thought that the adducing of what Loux has called "essential character-
istics" (which are basically second-order properties: the property of being
something with such-and-such properties) will have little purchase against
. But there is a non-trivial point at stake here, and we would do
well to take seriously the problem raised by these cases. Thus, Candraldrti
has argued that svalak:fal}a (in the sense of "defining characteristic") nec-
essarily involves a relationship between two things; and I have proposed
reconstructing this as an argument to the effect that even an irreducibly
unique particular necessarily has (hence, stands in relation to) the prop-
erty of "being a unique particular." Such a reconstruction helps to make
54 It is particularly this part of my exegesis of CandrakIrti that involves an effort at
rational reconstruction; for CandrakIrti, of course, argues simply on the basis of standard
Sanskritic grammatical analyses of the various parts of speech necessarily involved in any
. instance of l a k ~ a T : z a ("characterization"). (That CandrakIrti's procedure here is standard in
Sanskritic philosophical discourse is suggested, I think, by the perceptive remarks of Ingalls
1954.) What I have suggested is that the necessarily relational quality of any instance of
"characterization" can be argued by appeal to the unavoidability of saying at least that par-
ticulars can be "essentially characterized."
55 Note, though, that these seemingly second-order constructions ("being X or Y") in
fact neatly reflect one of the main ways of discussing universals in Indian philosophy.
Thus, one of the points at issue between apohaviidins such as Dignaga and, say,
MIrnfupsakas, is whether or not a word such as go ("cow") gains its usefulness by refer-
ring to some universal abstraction (gotva) that is common to all cows. In much of the sec-
ondary literature on such debates, words like gotva are often rendered as "cow-ness." But
in fact, the traditional commentarial gloss on the -tva suffix involves bhiiva ("being"),
and we are probably better off thinking of gotva as expressing the property" being a cow."
clear how CandrakIrti can plausibly argue that on Dignaga's account of
svalak:fm:za (i.e., as neither being nor having any characteristic), it becomes
impossible to say of any svalak:falJ.a even that it is one!
And indeed, might not this radical reading make sense of Dignaga's
own claim that svalak:falJ.as are characterized only by their unspecifiability
(their avyapadeiyatva)? But here we are on the verge of a very dense
thicket, and any full accounting of what Dignaga may have been up to
would surely require significant attention to his anyiipoha theory of mean-
ing (which CandrakIrti seems not to have engaged). For I have tried to
give CandrakIrti's argument against Dignaga greater purchase by intro-
ducing the case of "essential characteristics"; but I have also indicated
that thus adducing a problem involving universals would count for little
in Dignaga's view, insofar as his whole project centrally involves deny-
ing the reality of universals. And in the context of that project, it is the
anyiipoha ("exclusion") theory of meaning that is meant to explain how
language is possible - how, e.g., it is possible to predicate of some par-
ticular a certain characteristic (such as "being a unique particular") -
without reference to any really existent universals. The apoha doctrine,
then, is meant precisely to explain how one can do away with really exis-
tent universals (how one can deny, that is, that an abstraction like being
a unique particular should be allowed to count as a really existent state
of affairs), while yet retaining the ability at least usefully to say that a
unique particular is "of a certain kind or sort." Perhaps, then, an answer
to the difficulties with bare particulars that Loux and Armstrong have
identified would be forthcoming from an attempt to link this discussion
with Dignaga's account of apoha. If the attribution to Dignaga of a the-
ory "bare particulars" (in abstraction, at least, from other crucial parts of
his program) does not, then, finally turn out to represent the most
hermeneutically charitable reading of Dignaga's project, it is nevertheless
a plausible reading of his contention that svalak:falJ.as are "characterized"
only by their avyapadeSyatva; and it is, I am suggesting, the reading that
is recommended by CandrakIrti's engagement with Dignaga.
It is clear, in any case, that CandrakIrti (like Katsura) read Dignaga as
wanting to affirm that svalak:falJ.as do not themselves have any charac-
teristics (and that they are not, in turn, the characteristics of anything else)
- which is precisely why Candraldrti can (as he does) take it as an
unwanted consequence for Dignaga that must be the char-
acteristics of something else. With this in mind, we might put the differ-
ence between the Abhidharmika usage of (which is what Can-
drakIrti favors) and Dignaga's Sanskritically, in terms of two different
analyses of the compound on the Abhidharmika usage, the
compound is a karmadhiiraya, such that a denotes simply
whatever "property" or "characteristic" is defInitively "proper"
or "specific" (sva-) to something (i.e., something's "own characteris-
tic"); Dignaga, on the other hand, can be said to read the compound as
a bahuvrzhi, such that denotes what "has itself (sva) as [its
only] characteristic." But recall Dunne's characterization of DharmakIrti's
notion of the irreducibility of "is best illustrated
by a genitive construction such as, 'The nature of the infinitesimal parti-
cle.' DharmakIrti maintains that in such expressions the dharma is actu-
ally identical to the dharmin itself. The apparent separation of the dharma
from the dharmin is simply part of the exclusion process, and is hence
conceptual." If CandrakIrti and Katsura are right, it seems to me that
Dignaga is after essentially the same idea: in order to be consistent, the
Abhidharmika version of Buddhist reductionism cannot come to rest with
the idea of dharmas, if such are thought to "have" some defining char-
acteristic; rather, it must be pressed to the point where the only ontolog-
ical primitives in the system are not even logically resolvable even into
"properties" and "property-possessors." on such a view,
thus become not only unique, objective particulars, but bare particulars.
Such a reading has the advantage, at least, that it might tell us some-
thing about what Dignaga meant in characterizing as "inde-
finable." That is, perhaps Dignaga's point is that cannot be
"defined" or "specified" (vyapadisyate) specifically as having any prop-
erties; rather, the only irreducible worth the name must be
"indefinable" in that they admit of no logical reduction into dharma and
dharmin, "property" and "property-possessor. "56 It is no longer the case,
56 Cf., Dignaga's recurrent point that the distinguishing of separate vise.ra and vise.rya
("characteristic" and "thing characterized") is a constitutively conceptual operation - in
which case, perception can never itself register such a distinction. Thus, e.g.,
PramiilJ.asamuccaya 1.23, where Dignaga adduces the case of perception's perceiving such
a distinction as a counterfactual entailing problematic consequences: "If it were admitted
that is, that svalak:;Gl;as are the "defining characteristics" possessed by
dharmas; rather, svalak:;Gl;as just are the ontological primitives on
this view, and they are. simply self-characterizing. Having followed
Candraldrti, then, we see that the upshot of Katsura's apt point regarding
Hattori's translation is thus to emphasize that, on Dignag;i's view,
svalak:;Gl;as are no longer "borne" by anything; they are simply them-
selves the direct objects of perception, and that only insofar as percep-
tion is uniquely devoid of the sort of conceptual activity that is concerned
with discerning distinguishing properties.
VI. Conclusion
This stronger claim about what it takes for something to qualify as
irreducible seems to me to cut against several of those interpretations of
Dignaga's svalak:;ar,tas that emphasize his differences from Dharmaldrti.
Consider, for example, the interpretation of Dignaga put forward by Jonar-
don Ganeri, according to whose trope-theoretical reconstruction, Dignaga's
svalak:;ar,ta seems to denote simply any "object" of perception - i.e.,
such as the garden variety macro-objects we typically take ourselves to
? Thus, Ganeri would seem to agree with CandrakIrti that Dig-
naga takes svalak:;ar,ta as "denoting an object" (karmasadhanam)58, but
that both [viseea(la and vise.)'ya] were objects of the same [sense,] unaccepted consequences
would follow" (Tibetan at Hattori, p.207: yul mtshuns fiid du 'dod ce na / mi 'dod pa yan
thai bar 'gyur /I).
57 This reading would seem to be recommended by, inter alia, dKon-mchog 'jigs-med
dbang-po, who adduces a pot as an example (Ita bu) of a svalak.)'a(la as the "Sautrantikas"
understand the latter. See Mimaki (1977: 85): don dam par don byed nus pa'i chos de /
ran mtshan gyi mtshan iiid / mtshan gii ni / bum pa Zta bu .... (Of course, dKon-mchog
has DharrnakIrti in mind here.) While it is not my task here to elaborate a complete inter-
pretation of Dignaga (much less of DharrnakIrti) in relation to later Tibetan interpretation,
it seems worth at least noting that this reading of svalakealJ.as (i.e., as exemplified by
macro-objects like pots) could in principle be reconciled with Dunne's contention that, for
DharrnakIrti, svalak.)'a(las have no spatial extension (n.26, above); for the latter claim could
have chiefly to do with a representationalist epistemology, according to which dKon-
mchog 'jigs-med dbang-po's "pot" could simply be understood as a sense-datum.
58 Which would, to be sure, be sufficient to distinguish Dignaga' s usage from that of
the Abhidharrnikas. In making this point, Ganeri follows Katsura (1991, p.138), who says:
"I would like to assume that in Diirnaga's system svaZak.)'a(la is the object itself which is
to be grasped directly by perception, which is neither expressible nor identifiable at that
moment.. .. "
would disagree with the further claim that it denotes something like "bare
particulars." On Ganeri's reading, then, the "indefinability" (avya-
pade.yatva) of these consists simply in their being unavailable to any
comprehensive intuition. As Ganeri says,
Properties are conceptual constructs. They are potential contents of con-
ception because it is possible, in principle, to know everything about them ....
Objects, on the other hand, are not potential constructs of conception because
it is not possible, even in principle, to know everything about them. Again,
on the trope-theoretic analysis, what this means is that one cannot know
every member of a class of concurrent tropes - all the trope-constituents
of this vase, for example
On this reading, the point is that "objects" are inde-
finable simply as given to perceptual awareness, and that insofar as
perceptual awareness can never comprehensively apprehend all facets
of an object.
Note, though, that Ganeri's interpretation seems to be licensed by a
reading particularly of Hattori's translation of Dignaga - and specifically,
of PramiilJ.asamuccaya 1.5a-b, which Hattori renders: "a thing possess-
ing many properties cannot be cognized in all its aspects by the sense"60.
Richard Hayes (1988, p.138; my emphasis) instead translates: "no knowl-
edge at all of a possessor of properties that has many characteristics is
derived from a sense faculty." Explaining the difference from Hattori, he
ventures an interesting point:
Please note that the Tibetan translation construes the modifier 'sarvatha' as
governing the negative 'na' and so renders the core of the sentence modally:
'rtogs srid rna yin' or 'knowledge is impossible.' The point is that knowl-
edge of a multi-propertied whole is impossible through the senses. Hattori's
translation ... implies [the] weaker claim ... that while sensation can capture
some of the aspects of a multi-propertied whole, it cannot know the whole
exhaustively. But I think the point is clearly that the whole cannot be known
59 Ganeri (2001), p.106. On "trope" theories, cf., Armstrong 1989: 113-133.
60 Hattori (1968), p.2?; my emphasis. Ganeri (2001, p.lOl) follows Hattori, modify-
ing slightly: "A thing possessing many forms (rupa) cannot be cognised in all its aspects
by a sense-faculty." Kanakavannan's Tibetan is at p.181: du rna'i no bo'i chos can ni /
dban po las rtogs srid rna yin. Hattori (p.91, n.1.43) gives the Sanskrit as quoted by
Prajiiakaragupta: dharrnil:.zo 'nekarupasya nendriyat sarvatha gatib. Cf., n.28, above, for
PrarnalJasarnuccaya l.5c-d.
at all by the senses, because the notion of a whole is superimposed upon a
multiplicity of discrete data of sense

Thus, while Hayes is (as I have noted) critical of those who follow
Stcherbatsky in seeing Dignaga's as the "point-instants" of
DhannakIrti, he nevertheless reads Dignaga's point about the "indefin-
ability" of as a strong claim that they are radically different
from what is present to propositional awareness; and, in keeping with his
emphasis on Dignaga's as a "phenomenalist" epistemology62, he never-
theless reads Dignaga's not as (macro-) objects themselves,
but as the component sense-data out of which such are constructed:
" ... individuals, which are the referents of singular terms, are regarded
by Diimaga to be the synthesis of a multiplicity of cognitions and hence
are treated as classes rather than as particulars"63.
Hayes's point seems to me to be generally correct, and not obviously
incompatible with the reading I have developed following DhannakIrti.
As I indicated in beginning this essay, though, it is hard to be sure pre-
cisely what Dignaga means by his use of the term since he
says so little explicitly about it, and that in texts that come down to us
mainly in divergent Tibetan translations. What is nonetheless clear, in
any event, is that Dignaga has transformed the Abhidharmika sense of the
word, and that CandrakIrti can help us to understand the nature of this
transformation. For the present, I am not concerned with whether or not
CandrakIrti's arguments against Dignaga's notion of have any
purchase (though such is, of course, an interesting question). Rather,
I wish only to have suggested that CandrakIrti's engagement with Dig-
naga represents, inter alia, one traditional reading of what Dignaga claimed
regarding - a reading that, having been developed perhaps
only a generation or two removed from Dignaga, might well be helpful
to our understanding of what is, in the texts of Dignaga, a frustratingly
underdetermined concept.
And what CandrakIrti clearly tells us is that Dignaga's notion of
represents a transformation of the Abhidharmika notion.
61 Hayes (1988), p.l70, n.20.
62 Cf., n.19, above.
63 Hayes (1988), p.189.
The latter notion is that of uniquely defIning "properties" or "character-
istics" borne by dharmas, with Vasubandhu having invoked the latter's
"bearing" (....Jdhr) of these to explain their name (dharma, "bearer").
Clearly, on the spartan epistemology espoused by Dignaga, such "defin-
ing characteristics" would not be the sort of thing that could be encoun-
tered in perception, and would instead have to be counted as among the
things Dignaga considers to be "abstractions" (samanyalak:ja1}as). On Dig-
naga's usage, in contrast, svalak:ja1}as are the unique particulars encoun-
tered by perception, and are "characterized" only by their "indefmability"
(avyapadeSyatva) - which is, perhaps, simply to emphasize the irreducible
uniqueness of particulars, as opposed to the eminently categoreal notion
at play in the idea of dharmas. Whether or not we understand the latter point
as intended to delimit the kinds of vanishingly small "point-instants" that
DharmakIrti will have in mind, it is clear that this characterization advances
the intuition that our epistemic faculties yield some sort of access to a sim-
ply given, uninterpreted sort of data. And whether or not Dignaga can
rightly be thought to have upheld a Dharmaldrtian doctrine of momen-
tariness, it is at least not obviously the case that his doctrine of svalak:ja1}as
is incompatible with one. In any case, what is clear is that, at least as Can-
drakmi reads him, Dignaga has eschewed the conventional usage and
clearly posited something very much like bare particulars.
Apte, Varnan Shivararn, The Practical Sanskrit-English Dictionary. Revised and
enlarged edition, edited by P. K. Gode and C. G. Karve. Kyoto: Rinsen,
1992. (Reprint; revised and enlarged edition first published Poona, 1957.)
Armstrong, D. M., Universals: An Opinionated Introduction. London: Westview
Press, 1989.
Arnold, Daniel, "Mima.rp.sakas and Madhyarnikas Against the Buddhist Episte-
mologists: A Comparative Study of Two Indian Answers to the Question
of Justification." Ph.D: dissertation, University of Chicago, 2002.
Burton, David F., Emptiness Appraised: A Critical Study of Niigiirjuna's Phi-
losophy. London: Curzon, 1999.
Chandra, Lokesh, Tibetan-Sanskrit Dictionary. Kyoto: Rinsen, 1998. (Reprint of
Satapitaka series, vol. 3. New Delhi: International Academy of Indian Cul-
ture, 1959-1961.)
Chaudhuri, Sukomal, Analytical Study of the Abhidharmakosa. (Calcutta San-
skrit College Research Series No. CXIV.) Calcutta: Sanskrit College, 1976.
Cox, Collett, Disputed Dharmas: Early Buddhist Theories on Existence: An Anno-
tated Translation of the Section on Factors Dissociatedfr6m Thoughtfrom
Saizghabhadra's Nyayanusara. Tokyo: The International Institute for
Buddhist Studies, 1995.
De Jong, J. W., "Text-critical Notes on the Prasannapadii," Indo-Iranian lour-
nal20 (1978), nos.1/2: pp.25-59; nos.3/4: pp.217-252. .
Dunne, John D., "Foundations of Dharrnaldrti's Philosophy: A Study of the Cen-
tral Issues in His Ontology, Logic and Epistemology with Particular Atten-
tion to the Svopajfiavrtti." Ph.D. dissertation, Harvard University, 1999.
Eckel, Malcolm D., "Bhavaviveka and the early Madhyamika theories of lan-
guage," Philosophy East & West 28/3 (1978): 323-337.
Ganeri, Jonardon, Philosophy in Classical India: The Proper Work of Reason.
London: Routledge, 200l.
Garfield, Jay, The Fundamental Wisdom of the Middle Way: Niigiirjuna's Milla-
madhyamakakarika. New York: Oxford University Press, 1995.
Hattori Masaaki, Digniiga, On Perception, being the of
Digniiga's Pramfu).asamuccayafrom the Sanskrit fragments and the Tibetan
versions. Cambridge: Harvard University Press, 1968.
Hayes, Richard P., Digniiga on the Interpretation of Signs. Dordrecht: Kluwer, 1988.
Herzberger, Radhika, Bhartrhari and the Buddhists: An Essay in the Develop-
ment of Fifth and Sixth Century Indian Thought. Studies in Classical India,
8. Dordrecht: Kluwer, 1986.
Ingalls, Daniel H. H., "The Comparison of Indian and Western Philosophy,"
lournal of Oriental Research, Madras 22 (1954): 1-11.
Kapstein, Matthew, "Self and Personal Identity in Indian Buddhist Scholasti-
cism: A Philosophical Investigation." Ph.D. dissertation, Brown Univer-
sity, 1987.
Kapstein, Matthew, "Mereological Considerations in Vasubandhu's Proof oflde-
alism," Idealistic Studies 18 (1988): 32-54.
Katsura Shoryu, "Dignaga and Dharrnaldrti on apoha," in Ernst Steinkellner,
ed., Studies in the Buddhist Epist.emological Tradition: Proceedings of the
Second International Dharmakfrti Conference, Vienna, lunell-16, 1989,
pp.129-146. Osterreichische Akademie der Wissenschaften, Philosophisch-
Historische Klasse Denkschriften, 222. Band. Wien: Verlag der Osterre-
ichischen Akademie der Wissenschaften, 1991.
La Vallee Poussin, Louis de, ed., Mulamadhyamakakiirikiis (Miidhyamikasutras)
de Niigiirjuna, avec la Prasannapadii Commentaire de Candrakfrti. Biblio-
theca Buddhica, Vol. IV. Osnabrock: Biblio Verlag, 1970. (Reprint; origi-
nally St.-Petersbourg: Academie Imperiale des Sciences, 1903-1913.)
La Vallee Poussin, Louis de, ed., Madhyamakiivatiira par Candrakfrti: Traduc-
tion tibhaine. Bibliotheca Buddhica, Vol. IX. Osnabrock: Biblio Verlag,
1970. (Reprint; originally St.-Petersbourg: Academie Imperiale des Sci-
ences, 1907-1912.)
Loux, Michael J., Metaphysics: A Contemporary Introduction. London: Rout-
ledge, 1998.
Mimaki Katsumi, ed" "Le Grub mtha' rnam bZag rin chen phren ba de dKon
mchog 'jigs med dbaiJ. po (1728-1791): Texte tiMtain edite, avec une intro-
duction," Zinbun: Memoirs of the Research Institute for Humanistic Stud-
ies, Kyoto University 14 (1977): 55-112. .
Miyasaka Yusho, ed., "Pramfu.1avarttika-karika (Sanskrit and Tibetan)," Acta
Indologica II (1971/2): 1-206.
Mookerjee, Satkari, "The Absolutist's Standpoint in Logic," in Satkari Mook-
erjee, ed., The Nava-Nalanda-Mahavihara Research Publication (Patna),
vol. 1 (1957): 3-175. (Mookerjee's summary of CandrakIrti's critique of
the Buddhist Epistemologists is at pp.42-58.)
Pradhan, Prahlad, ed., of Vasubandhu. Patna:
K. P. Jayaswal Research Institute, 1975.
Randle, H. N., Fragments from Dinniiga. Delhi: Motilal Banarsidass, 1981.
(Reprint; originally published in London, 1926.)
Renou, Louis, Terminologie Grammaticale du Sanskrit. Paris: Librairie Anci-
enne Honore Champion, 1942.
Rizzi, Cesare, Candraklrti. Delhi: Motilal Banarsidass, 1987.
Sanlqtyayana, Riliula, "Second Research of Sanskrit Palm-leaf Mss. in Tibet,"
Journal of the Bihar and Orissa Research Society 23/1 (1937): 11-57.
Sellars, Wilfrid, Science, Perception and Reality. Atascadero, CA: Ridgeview, 1963.
Shastri, Swami Dwarikadas, ed., Nyiiyabindu of Acharya Dharmakirtti [sic], with
the commentaries by Arya Vinitadeva and Dharmottara. Varanasi: Bauddha
Bharati, 1985.
Siderits, Mark, "The Madhyamaka Critique of Epistemology II," Journal of
Indian Philosophy 9 (1981): 121-160.
Stcherbatsky, Th., Buddhist Logic. Two volumes. 'S-Gravenhage: Mouton &
Co., 1958. (Indo-Iranian Reprints, voL IV. First published in Leningrad,
1932, as volume XXVI of the Bibliotheca Buddhica.)
-, The Conception of Buddhist NirviilJa. Delhi: Motilal Banarsidass, 1989.
(Reprint; originally published in Leningrad by the Publishing Office of the
Academy of Sciences of the USSR, 1927.
Thurman, Robert, The Central Philosophy of Tibet: A Study and Translation of
Jey Tsong Khapa's Essence of True Eloquence. Princeton: Princeton Uni-
versity Press, 1991. (Paperback reprint of Tsong Khapa's Speech of Gold in
the "Essence of True Eloquence," 1984.)
Tillemans, Tom J. F., Materials for the Study of Aryadeva, Dharmapiila and
Candraklrti: The CatuJ:zsataka of Aryadeva, Chapters XII and XIII, with the
Commentaries of Dharmapiila and Candraklrti: Introduction, Translation,
Sanskrit, Tibetan and Chinese Texts, Notes. Wiener Studien zur Tibetolo-
gie und Buddhismuskunde, Heft 24, 1-2. Wien: Arbeitskreis fUr tibetische
und buddhistische Studien, 1990.
Tillemans, Tom J. F., "Metaphysics for Madhyamikas," in G. Dreyfus and
S. McClintock eds. The Svatantrika-PrasaJigika Distinction. What Difference
does a Difference make? Studies in Indian and Tibetan Buddhism. Boston:
Wisdom Publications, 2Q03: 93-123.
Tola, Fernando, and Carmen Dragonetti, "Dignaga's
Journal of Indian Philosophy 10 (1982): 105-134.
Warder, A. K., "The Concept of a Concept," Journal of Indian Philosophy 1
(1971): 181-196.
Williams, Paul, "On the Abhidharma Ontology," Journal of Indian Philosophy 9
(1981): 227-257.
Yonezawa Yoshiyasu, A Sanskrit Manuscript of an Anonymous
Commentary on the Prasannapada," Indogaku Bukkyogaku Kenkyu 47/2
(1999): 1024-1022.
-, et aI, "Introduction to the Facsimile Edition of a Collection of Sanskrit Palm-
leaf Manuscripts in Tibetan dBu med Script." TaishO University: The Insti-
tute for Comprehensive Studies of Buddhism (Study Group of Sanskrit
Manuscripts in Tibetan dBu med Script), 2001.
1. Object of research
The 9
century treatise bSam gtan mig sgron [Torch of the Eye of Med-
itation], composed by gNub chen Sangs rgyas ye shes, is the only known
work which discusses in detail the four Buddhist approaches prevalent dur-
ing the early spread of Buddhism in Tibet, namely the discussion of (1) the
gradual path, known as Rim gyis pa, (2) the sudden approach of Chinese
Meditation Buddhism known as Cig car ba, (3) the Mahayoga tradition,
and (4) the rDzogs chen teachings
. The author aims at distinguishing
these four schools in the four main chapters of the bSam gtan mig sgron
1 An earlier Chinese draft of this paper was fIrst presented at the Conference of Tibetan
Studies held by the Centre for Tibetan Studies in Beijing in the summer of 2001 (that Chi-
nese draft is to be published in the proceedings of the conference). The idea of this paper
results from discussions with Master Tam Shek-wing (a disciple of Dudjom Rinpoche)
and Henry C. H. Shiu in the summer 2000 in Toronto on the bSam gtan mig sgron and
the ArylivikalpapraveanlimadhliralJl. I am thankful to Mr. Tam's pointing out the con-
nection in structure and in content between these two texts. Moreover, I would like to
express my thanks to Prof. Schmithausen for recent inspiring comments on the present ver-
sion of this paper during a colloquium in Hamburg in winter 2002.
2 gNubs chen Sangs rgyas ye shes, gNubs chen sangs rgyas ye she rin po ches mdzad
pa'i sgom gyi gnang gsal bar phye ba bsam gtan mig sgron [Torch of the Eye of Medita-
tion Elucidating the Very Heart of Meditation, Composed by gNubs chen Sangs rgyas ye
shes Rin po che], short title: bSam gtan mig sgron [Torch of the Eye of Meditation], repro-
duced from a manuscript made presumably from an Eastern Tibetan print by 'Khor gdon
gter sprul 'Chi med rig dzin, Leh: Smarts is shesrig spendzod, Vol. 74, 1974 (hereafter in
the footnotes abbreviated as SM). According to different historical data the dates of gNubs
range from 772 as the earliest date of his birth (cf the sources listed in: Herbert V. Guen-
ther, '''Meditation' Trends in Early Tibet", in: Lewis Lancaster/Whalen Lai (ed.), Early
Ch'an in China and Tibet, Berkeley: Berkeley Buddhist Studies Series, 1983,352) up to
the late 10
century (cf the discussion in: Samten Gyaltsen Karmay, The Great Peifection
Journal of the International Association of Buddhist Studies
Volume 26 Number I 2003
in order to clarify the misunderstandings about some of their apparent
His analysis is undertaken in the light of t h ~ soteriological
idea of "non-conceptuality" (rnam par mi rtog pa). Therefore, gNubs
chen Sangs rgyas ye shes quotes extensively from canonical scriptures and
also from texts which are now only preserved in Dunhuang manuscripts
in order to exemplify "non-conceptuality" according to the understand-
ing of each particular school.
The Japanese scholar Ueyama Daishun already pointed out that in the
eighth and ninth centuries in the Sino-Tibetan border regions, e.g. in
Dunhuang as one geographical junction in the encounter between Chinese
and Tibetan Buddhism, the AryavikalpapraveSaniimadharalJl [The Supreme
dharalJl of Entering into Non-Conceptuality] has been widely known and
was of particular regional importance in the spread of Buddhism from
Central Asia to Tibet
. This short siUra is a teaching attributed to the
historical Buddha on how to give up clinging to discursive thoughts in
order to enter into the "non-conceptual sphere" (rnam par mi rtog pa'i
The AryiivikalpapraveSaniimadhiiralJl and gNubs chen Sangs rgyas ye
shes likewise emphasise the importance of the understanding of "non-con-
ceptuality". In this initial approach the present paper shall open a win-
dow to develop our insight into the important but difficult bSam gtan mig
sgron through a structural analysis which, however, does neither claim to
be fmal nor complete. Thus, in order to analyse the structure of its' four
main chapters as a possible soteriological path in itself, that is from Rim
gyis pa to Cig car ba, Mahayoga, and finally to rDzogs chen
, this paper
(rDzogs chen). A Philosophical and Meditative Teaching in Tibetan Buddhism, Leiden:
E. J. Brill, 1988,101t). If we follow the argumentation of H. Guenther (1983: 352), that
is placing the birth of gNubs in the late eighth century, gNubs presumably composed the
bSam gtan mig sgron in an old age in the late 9
3 The four main chapters that the SM covers: chapter four on Rim gyis pa SM: 65-118,
chapter five on Cig car ba SM: 118-186, chapter six on Mahayoga SM: 186-290 and chap-
ter seven on rDzogs chen SM: 290-494.
4 Ueyama, Daishun/Kenneth W. Eastman/Ieffrey L. Broughton, "The Avikalpapravda-
dhiiralJz: The Dharani of Entering Non-Discrimination", in: BBK (1983), 35.
5 In his unusual doxographical description ranging from Rim gyis pa to rDzogs chen
gNubs seems to omit Anuyoga as a link between Mahayoga and rDzogs chen intention-
ally even though he briefly distinguishes the essence of Anuyoga in regard to the other
traditions in his concluding remarks (cf. SM: 490.6-491.3,492.4-6,493.5-6).
proposes to understand it through the fourfold correct practice as it is also
taught in the AryavikalpapravesanamadhiiralJl. There, this fourfold prac-
tice is described in terms of: (1) "perception" (dmigs pa), (2) "non-
perception" (mi dmigs pa), (3) "non-perception of perception" (dmigs
pa mi dmigs pa) and (4) "perception of non-perception" (mf dri1igs pa
dmigs pa). Even though the AryavikalpapraveanamadhiiralJl is not the
first and only place where a fourfold correct practice is discussed
, this
paper claims that gNubs chen Sangs rgyas ye shes was familiar with such
a fourfold structure in arranging his treatise. Therefore, in the present
paper the fourfold correct practice as it is also exemplified in the
AryavikalpapravesanamadharalJl is used as such an example of that prac-
tice and as a bridge to demonstrate how to enter into the "non-concep-
tual sphere". Having analysed the insight into "non-conceptuality" accord-
ing to this fourfold practice, we shall compare it to the understanding of
"non-conceptuality" in the four different schools as described in the bSam
gtan mig sgron. This structural analysis of the bSam gtan mig sgron may
shed new light on distinguishing the traditions of Rim gyis pa, Cig car
ba, Mahayoga and rDzogs chen as distinct approaches on the Buddhist
path. It will furthermore highlight some philosophical differences between
the apparently similar foundations of Chinese Meditation Buddhism (Cig
car ba) and Tibetan rDzogs chen.
2. The Aryavikalpapravesanamadhara
2.1 Dunhuang manuscripts and canonical versions
The AryavikalpapravesanamadhiiralJl played an important role in the
spread of Buddhism from Central Asia to Tibet. D. Ueyama even argues
6 -A detailed and comprehensive analysis of the fourfold correct practice in the Bud-
dhist literature awaits further research. Yael Bentor is the fIrst to have investigated different
fourfold meditation systems in Tibet and proposes in her closing remarks of her article even
a kind of prototype for most of the systems she discussed (cf. Yeal Bentor, "Fourfold
Mediation: Outer, Inner, Secret, and Suchness", in: Religion and Secular Culture in Tibet.
Tibetan Studies II, Leiden: Brill, PIATS 2000, vol. 2, 41-58). On this topic cf. also Tam
Shek-wing (ed./transl.), Sianfa fax:ing lun. Shiqin shi lun l!IitiHtOi [Dhar-
madharmatavibhtiga. The Commentary of Vasubandhu] , Hongkong: Vajrayana Buddhism
Association Limited, 1999, 159-160.
that the knowledge of Sanskrit originals and the philosophical discus-
sions on how to enter into a "non-conceptual" state may have even moti-
vated further translations of the text in the Sino-Tibetan border regions
such as in Dunhuang
. WithIn the corpus of Duriliuang manuscripts we
find one Chinese and two Tibetan translations
The Dunhuang Chinese
translation is entitled the Ru wu fenbie zongchi jing A ~ 'TJ' 3u * ~ t ~ *1
[The dharal}l of Entering into Non-Conceptuality] (jiang :Ii: 23) preserved
in the Beijing National Library, and S. tib. 51 and S. tib. 52 preserved in
the British Library in London
Moreover, the AryavikalpapraveSana-
madharal}l was translated around the same time, that is the 9
into Tibetan by the prominent translators Jinamitra, Danaslla, and dKa'
ba dpal brtsegs. Their translation 'Phags pa rnam par mi rtog par 'jug
pa zhes bya ba'i gzungs was then included in the tripitakalO. In Central
China, however, the AryavikalpapraveSanamadharal}1 did not have the
same lasting impact. The text was only translated in the eleventh century
by Danapala ( n t ! ! ~ , active in Kaifeng from 982 to roughly 1017) under
the title F oshuo ru wu fenbie famen jing f?f) ~ A ~ 'if 3u it; ~ ~ * ~ [The sutra
of Entering into the Dharma Gate of Non-Conceptuality taught by the
7 Ueyama/Eastman/Broughton 1983: 35.
8 Kazunobu Matsuda t.llB3;j!]j" has also published two Sanskrit fragments of the
AryiivikalpapraveSaniimadhiirQ/JI, one found among the Gilgit manuscripts (cf "Nirvikalpa-
pravesa-dharrup: ni tsuite [On the Nirvikalpa-pravda-dharal)I]", in: Buddhist Seminar 34
(1981, 40-49) and one in the St. Petersburg manuscript collection of the Institute of Oriental
Studies at the Academy of Sciences of Russia (cf "Nirvikalpapravesadharrup:. Sanskrit
Text and Japanese Translation", reprint from Bulletin of the Research Institute of Bukkyo
University (3/1996), 89-113).
9 Ueyama/Eastman/Broughton 1983: 32-33. D. Ueyama (loc. cit.: 38-40) first pub-
lished jiang 23; I am preparing an English translation of jiang 23 and a structural analy-
sis of the text. The Catalogue of the Tibetan Manuscripts from Tun-Huang in the India
Office Library by Louis de la Vallee Poussin (Oxford: Oxford University Press, 1962,
24-25) gives for the Tibetan versions S. tib 51 and 52 the title rNam par mi rtog pa 'jug
pa zhes bya ba'i gzungs. D. Ueyama noted that he did not yet identify S. tib. 52. I was not
able to look at the Tibetan originals myself so far.
10 "'Phags pa mam par mi rtog par 'jug pa zhes bya ba'i gzungs [The Supreme dhiiral}l
of Entering into Non-Conceptuality]", translated by Jinamitra, DiinasUa and dKa' ba dpal
brtsegs, in: IT. 32, no. 810, la-6b. Furthermore, the ninth century catalogue !Dan dkar
rna lists under no. 196 also the title 'Phags pa rnam par mi rtog par 'jug pa'i gzungs
[The Supreme dhiiral}! of Entering into Non-Conceptuality] (cf Yoshimura, Shyuki,
The Denkar rna. An Oldest Catalogue of the Tibetan Buddhist Canons with Introductory
Notes, Kyoto: Ryiikoku University, 1950).
Buddha]l1. Apparently, the Buddhist traditions in Central China were not
aware of the regional importance of this text in the Sino-Tibetan border
areas and Tibet. Furthermore, the Chinese translation in the Dunhuang
manuscript jiang 23 is very close to the Tibetan canonical version, whereas
in the later Chinese translation Danapala either provided a rather free trans-
lation or used a different Sanskrit original. In our discussion we shall there-
fore pay attention to jiang 23 and the Tibetan canonical version likewise.
2.2 Influence of the Text in Tibet
In the literary history of Buddhism in Tibet the contents of the
Aryavikaipapraveanamadharmy,Z is a reoccurring theme. Shortly after the
translation of the text into Tibetan, none other than the Indian scholar
Kamalasila wrote a commentary to the Aryavikalpapraveanamadhiirm}l,
namely the 'Phags pa rnam par mi rtog par 'jug pa zhes bya ba'i gzungs
kyi rgya cher 'grel pa [Extensive Commentary to The Supreme dhiirar;z
of Entering into Non-ConceptualityF2. Kamalasila is said to have been the
advocate of a gradual path towards awakening in the great debate of bSam
yas that apparently took place in the late eighth century13. Unlike his
II "Foshuo ru wu fenbie famen jing [The siltra of Entering into the
Dharma Gate of Non-Conceptuality taught by the Buddha]", translated by Danapala
in: T. 15, no. 654, 805-806.
12 "'Phags pa mam par mi rtog par 'jug pa zhes bya ba'i gzungs kyi rgya cher 'grel
pa [Extensive Commentary to The Supreme dhiirGl}1 o/Entering into Non-Conceptual-
ity]", by Kamalasna, in: IT. 105, no. 5501, f. 146b.6-174b.1.
13 A lot of recent research has been done concerning the reliability of historical mate-
rial about the great debate of bSam yas and thus concerning the question whether the
debate can be regarded as an actual historical event at all. D. Seyfort Ruegg also provides
a comprehensive bibliography in this field in his footnotes (cf Buddha-nature, Mind and
the Problem o/Gradualism in a Comparative Perspective, London: School of Oriental and
African Studies, 1989). For earlier research concerning the debate cf also Paul Demieville,
Le concile de Lhasa, reprint, 1
edition 1952, Paris: College de France Institnt des Hautes
Etudes Chinoises, 1987; Giuseppe Tucci, Minor Buddhist Texts II, Rome: Is. M. E. 0.,
1958; Ueyama, Daishun, "The Study of Tibetan Ch'an Manuscripts Recovered from Tun-
huang: A Review of the Field and its Prospects", in: L. Lancaster/W. Lai (ed.), Early
Ch'an in China and Tibet, Berkeley: Berkeley Buddhist Studies Series, 1983, 327-350 and
L. Gomez, "The Direct and Gradual Approaches of Zen Master Mahayana: Fragments of
the Teachings of Mo-ho-yen", in: M. Gimello/p. N. Gregory (ed.), Studies in Ch'an and
Hua-yen, Studies in East Asian Buddhism No. 1, Honolulu: University of Hawaii
Press,1983[a], 69-168. Despite the uncertainties surrounding the question of whether the
opponents in the Chinese School of Meditation Buddhisll). who are said
to have quoted the text to emphasise their subitist outlook
, KamalasIla
interpreted the text to support his view of a gradual path.
Apart from this immediate historical connection, the contents of the
Aryiivikalpaprave.saniimadhiirm;l is a central theme in a scholastic treatise
of the yogaciira school, namely in the Dharmadharmatiivibhiiga [Dis-
crimination of dharma and dharmatii] 15. A later commentary of Mi pham
Rin po che to this yogaciira text explicitly states that the passage in the
Dharmadharmatiivibhiiga on "non-conceptual wisdom" (rnam par mi rtog
pa'i ye shes) summarises systematically the essence of the Aryiivikalpa-
prave.saniimadhiiraf}116. The Dharmadharmatiivibhiiga as a scholastic work
debate ever actually took place, a symbolic meaning was attached to it in the course of
Tibetan history that gave rise to discussions up to the present. Concerning the arguments
of Heshang Moheyan and of Kama1asila cf. Dunwu dacheng zhenglijue 1lf:kJlHl:i!!lili [Rat-
ification of the True Principle of the Mahayana Teachings of Sudden Awakening], P. chin.
4646 (copy edited by Rao Zongyi (Jao Tsung-I) 1!iii*1!li in: "Wangxi Dunwu dacheng zhengli
jue xushuo bing jiaoji [Preface and Notes to Wang Xi's Dunwu
dacheng zhenglijue (J?atification of the True principle of the Mahayana Teachings ofSud-
den Awakening)]", in Chongji xuebao iJii'MHifl Chung Chi Journal] 9/2 (1970), 127-148)
and "sGom pa'i rim pa [Stages of Meditation (Third Bhiivanakrama)]", by Karnalasila,
translated by Prajfiavarrna and Ye shes sde, in: IT. 102, no; 5312; 60b.8-74bA.
14 Cf e.g. the Tibetan manuscript on Chinese Meditation Buddhism P. tib. 116: VIa,
153.2-3 and Vimalarnitra's Cig car 'jug pa rnam par mi rtog pa'i bsgom don [The Mean-
ing of 'Non-Conceptual' Meditation in the School of Simultaneous Entry] (in: IT. 102,
no. 5306, llb.3).
15 "Chos dang chos nyid mam par 'byed pa'i gzhung [Discrimination of dharma and
dharmata]", attributed to Maitreya, in: IT. 108, no. 5523, f. 48b.1-51b.6. Cf also Henry
C. H. Shiu llIIUiUi, Bianfafaxing lu. Bubai shi lun !IOliiiilfffff' 1'lIHHif [Dharmadharmatavib-
hiiga. The Commentary of Mi pham Rin po che], ed. by Tarn Shek-wing Hongkong:
Vajrayana Buddhism Association Limited, 2000, 67-77 and preface by Tarn Shek-wing,
16 Cf. Klaus-Dieter Mathes, Unterscheidung der Gegebenheiten von ihrem wahren
Wesen (Dharmadharmatavibhiiga), Swisttal-Odendorf: Indica et Tibetica Verlag, 1996,
215. The commentary of Mi pham Rin po che consulted by K. Mathes was prepared on
the basis of the block prints from Kathmandu and Rumtek. K. Mathes provides the fac-
simile in the annex. For the passage in question compare this facsimile: "Chos dang chos
nyid marns par 'byed pa'i tshig le'ur byas pa'i 'grel pa ye shes snang ba mam 'byed [Com-
mentary to the Verses of the Discrimination of dharma and dharmata. Discrimination of
Primordial Wisdom and Appearances]", by Mi pham Rin po che (1846-1912), in: Mathes
1996, annex, f. l6a.1. Moreover, K. Mathes (1996: 83-84 and 138-139) also provides a
transliteration and translation of Vasubandhu's commentary to the Dharmadharmatavib-
haga that in some cases even uses the same vocabulary on the two main topics of the
AryavikalpapraveSanamadhiiraJJI, that is on "abandoning marks" (mtshan rna spong ba)
is more systematic and also offers a more affording vocabulary than the
sutric text. Therefore, we shall also pay attention to the relevant passages
of the Dharmadharmatiivibhiiga in our discussion of the fourfold correct
practice as it is taught in the AryiivikalpapraveSaniimadhiiralY-z.
2.3 Fourfold correct practice
In demonstrating how to enter into the "non-conceptual sphere,"
the AryiivikalpapraveSaniimadhiira'!l focuses on two corresponding meth-
ods, namely the method of "abandoning marks" (mtshan ma yongs su
spong ba) of the conceptual framework and the method of "correct prac-
tice" (yangdag par sbyor ba)17. The approach of "abandoning marks" is
described in a fourfold way, namely as the abandoning of marks of
(1) "own nature" (rang bzhin), (2) "antidotes" (gnyen po), (3) "thusness"
(de kho na nyid), and (4) of "realisation" (thob pa)18. However, in the con-
text of the present research we shall only focus on the method of "correct
and on "correct practice" (yang dag pa'i sbyor ba). These equivalents will also be proved
in my English translation of the Tibetan text of the AryiivikalpapraveSaniimadht'iral;l. It will
soon be published together with my translations of KamalaSlla's commentary and Vimalami-
tra's Cig car 'jug pa mam par mi rtog pa'i bsgom don in the anthology Studies on the
AryiivikalpapraveSaniimadht'iral!! in the Sino-Tibetan Series of Wisdom Publications in
cooperation with Tam Chek-wing, Henry Shiu and Shen Weirong.
I7 For the Tibetan cf "'Phags pa mam par mi rtog par 'jug pa zhes bya ba'i gzungs",
in: IT. 32, f. 2b.3-3b.l (abandoning marks) and f. 5a.3-6b.2 (correct practice); for the
Chinese cf Jiang 23, I. 17-25 (abandoning marks) and I. 83-125 (correct practice) (here-
after the Tibetan version is referred to as IT. 32, p. x and the Chinese version is referred
to as jiang 23. I. x). In Kamalaslla's commentary to the AryiivikalpapraveSaniimadhiiraf}!
the discussion on "abandoning marks" and on "correct practice" (mtshan ma yongs su
spong bar yang dag par sbyor ba) is the most relevant (Luis O. Gomez, "Indian Materi-
als on the Doctrine of Sudden Enlightenment", in: Lewis Lancaster/Whalen Lai (ed.),
Early Ch 'an in China and Tibet, Berkeley: Berkeley Buddhist Studies Series, 1983[b], 408).
Cf "'Phags pa mam par mi rtog par 'jug pa zhes bya ba'i gzungs kyi rgya cher 'grel pa",
in: IT. 105, f. 156a.5-l63b.8. My above mentioned translation of the Tibetan text will also
provide a structural analysis of the AryiivikalpapraveSaniimadht'iraf}l.
18 The abandoning of these four marks in the Aryiivikalpapraveaniimadhiiraf}! corre-
sponds to a similar passage in the Dharmadharmatiivibhiiga on abandoning the marks
of (1) "non-conducive" (mi thun pa'i phyogs), (2) "antidote" (gnyen po), (3) "suchness"
(de bzhin nyid), and (4) "realisation" (rtogs pa). Cf "Chos dang chos nyid mam par 'byed
pa'i gzhung", in: IT. 108, f. 49b. Thanks to Henry C. H. Shiu for the discussions on this
practice". If we apply the vocabulary of the Dharmadha1:matavibhiiga it
is described as the following fourfold practice: (1) "perception" (dmigs
pa), (2) "non-perception" (mi dmigs pa), (3) "non-perception ofpercep-
tion" (dmigs pa mi dmigs pa), and (4) "perception of non-perception" (mi
dmigs pa dmigs pa)19. The Aryiivikalpapravesaniimadhiira/Jl a p p l i ~ s such
a fourfold correct practice to the practice about form and to the practice
about omniscience, which respectively pertain to the "marks of own
nature" in the case of form and to the "marks of realisation" in the case
of omniscience
. In order to demonstrate this gradual path of cognition
we shall, however, merely look at the practice about form as one exam-
ple of this structural process.
Generally speaking, conceptual thinking - which is itself cause for the
appearance of duality - arises when thusness is not cognised. Thus, the
manifestations of cause and effect appear, yet they are not inherently exis-
tent. Only when those manifestations do not appear anymore as seemly
inherently existent then the nature of all dharmas comes to light and
"non-conceptual wisdom" may be cognised. Concerning the above-
mentioned fourfold correct practice, the four steps in this process of enter-
ing into the "non-conceptual sphere" are described as follows:
(1) In the first step of "perception" one cognises all dharmas as the
manifestation of "mere cognition" (rig pa tsam), that is, all dharmas
are an expression of one's own mind. Even though conceptual
thinking still arises, one does not mistake it for existent, but rather
takes it as "mere cognition"21. The Aryiivikalpapraveaniimadhiira/Jl
19 Cf "Chos dang chos nyid mam par 'byed pa'i gzhung", in: IT. 108, f. 50a.6.
The equivalent passages in the Tibetan and Chinese version of the Aryiivikalpa-
praveaniimadhiiraJ;zi are in IT. 32, f. 5a.3-6b.2 and injiang 23,1. 83-125. However, in
these translations the AryiivikalpapraveaniimadhiiraJ;zi does not use those pithy designa-
tions of the fourfold practice, as the AryiivikalpapraveaniimadhiiraJ;ti itself is of rather deno-
tative character and instead gives lengthy explanations of how to give up different kinds
of concepts - which, nonetheless, correspond to the essence of the four correct practices.
20 Cf IT. 32, f 5a.3-Sb.4 (practice about form) and f. 5b.6-6b.2 (practice about omniscience)
andjiang 23, 1. 83-101 (practice about form) and 1. 106-121 (practice about omniscience).
21 Cf IT. 32, f. 5a.3-5 and jiang 23, 1. 83-88. For the term "practice of perception"
(dmigs pa'i sbyor ba) in the Dharmadharmatiivibhiiga cf. "Chos dang chos nyid mam par
'byed pa'i gzhung", f. 50a.6.
describes this first step as follows: If one takes form as existent, then
. one is still practising in the conceptual sphere
(2) In the second step one cognises the "non-perception" of objects, to
which the ordinary apprehension generally adheres. External dharmas
are non-existent because in the first step of "perception" the "mere
cognition" already emerged as an object. Therefore, to speak with the
A.ryavikalpapraveanamadharGl}f, form is also non-existent. And if
one takes non-form for true, one is again practising in the concep-
tual sphere

(3) In the following step of "non-perception of perception" one trains
oneself in the non-perception of the perception that "mere cogni-
tion" is non-existent. Since cognition is not possible without an
object, cognition itself is also impossible
. In the A.ryavikalpa-
praveanamadharGl}fit is explained: If the bodhisattva engages in the
notion that form is mere-cognition, he engages in conceptualization.
If he just as he engages in the non-existence of form similarly engages
in the non-existence of cognition manifesting in from, he engages in

(4) In the final step of "perception of non-perception" one perceives
neither an apprehending subject nor an apprehensible object. As sub-
ject and object are not of separate natures, non-duality may be
This is said to be non-dual thusness, the nature of reality
beyond any designations. Again, in the A.ryavikalpapraveanama-
dharGl}f this final step in this perceptual process is expressed as fol-
lows: In regard to not perceiving any dharma apart from cognition,
the bodhisattva neither completely sees the absence of phenomena in
22 Cf. IT. 32, f. 5a.3-4 andjiang 23, I. 85-86.
23 Cf. IT. 32, f. 5a.5-6 and Jiang 23, I. 88. For the term "practice of non-perception"
(mi dmigs pa'i sbyor ba) in the Dharmadharmatavibhaga cf. "Chos dang chos nyid mam
par 'byed pa'i gzhung", f. 50a.6.
24 For the term "practice of non-perception of perception" (dmigs pa mi dmigs pa'i
sbyor ba) in the Dharmadharmatavibhiiga cf. "Chos dang chos nyid mam par 'byed pa'i
gzhung", f. 50a.6.
25 Cf. IT. 32, f. 5a.6-7 andjiang 23,1. 91-92.
26 For the term "practice of perception of non-perception" (mi dmigs pa dmigs pa'i
sbyor ba) in the Dharmadharmatavibhiiga cf. "Chos dang chos nyid mam par 'byed pa'i
gzhung", f. 50a.6.
regard to that cognition nor apart from cognition. In regard to the non-
existence of cognition manifesting in form and to that cognition he
neither completely sees them as same nor as different
Just this very
non-perception of an apprehending subject and an apprehensible
object, or of mere cognition and form, is "non-conceptual wisdom".
In this fourfold investigative practice phenomena are simply a feature
of this very perceptual process itself. And to summarise again, this process
to non-conceptual wisdom leads through the fourfold cognition that (1) all
dharmas are manifestation of one's own mind, (2) that the external world
is inherently non-existent, (3) that "mere cognition" is non-existent and
(4) that cognisable objects and cognition are non-dual. Now, we shall
look at the understanding of "non-conceptuality" according to the sys-
tem of Rim gyis pa, Cig car ba, Mahayoga and rDzogs chen in the four
main chapters of the bSam gtan mig sgron in order to compare it to the
four steps of entering into the "non-conceptual sphere" as they were just
described according to the Aryiivikalpapraveaniimadhiira1J,1.
3. The bSam gtan mig sgron
3.1 The Topic of 'Non-conceptuality'
The composition of the bSam gtan mig sgron by gNubs chen Sangs
rgyas ye shes is to be understood in the broader historical context of the
eighth and ninth centuries. The debate of bSam yas that is said to have
taken place in the late eighth century between KamalasTIa, the Indian
advocate of a gradual path, and Heshang Moheyan, a Chinese Meditation
master rather favouring the subitist approach, is according to the histor-
ical data one of the major events giving evidence for the development of
Chinese Meditation Buddhism in Tibet during those early times
over, from other Dunhuang manuscripts we also know about the spread
of the Mahayoga tradition
gNubs chen Sangs rgyas ye shes' treatise
27 Cf jiang 23, 1. 96-97 and IT. 32, f. 5b.1-2.
28 Cf footnote 13 above and the two text Dunwu dacheng zhenglijue and sGom pa'i
29 Cf K. W. Eastman, "Mahayoga Texts at Tun-huang", in: Bulletin of Institute of
Buddhist Cultural Studies 22, Kyoto: Ryiikoku University, 1983,42-60.
bSam gtan mig sgron is the only known work which discusses in detail
the differences between the four prevalent traditions in the eighth and
ninth centuries, namely Rim gyis pa, Cig car ba, Mahayoga and his own
tradition of rDzogs chen. He clearly saw the potential that the teachings
of these different schools may be intermingled, and thus states:
In [writing] the bSam gtan mig sgron, I gave a detailed description [of the
Cig car ba tradition], because I fear that one mistakes [the meaning of the]
Cig car ba to be similar to rDzogs chen

Furthermore, he may have had an actual syncretistic movement in mind
which fused elements of both traditions alike, namely those of Cig car ba
and rDzogs chen
. In the bSam gtan mig sgron, he clearly refers to con-
temporaries who neither understood the meaning of Cig car ba nor of
rDzogs chen, yet simply mistook their own erroneous view to be rDzogs
Therefore, gNubs chen Sangs rgyas ye shes was concerned to dis-
tinguish the doctrirIal differences of the four above-mentioned schools.
He undertook his analysis in the light of the soteriological idea of "non-
conceptuality" (rnam par mi rtog pa) and said:
[ ... ] in regard to the benefit of myself and others to thoroughly comprehend
the authoritative scriptures about 'non-conceptuality' in each vehicle (of Rim
gyis pa, Cig car ba, Mahayoga and Atiyoga) [ ... ]33
However, how does gNubs chen Sangs rgyas ye shes analyse the view
of "non-conceptuality" according to each of the four schools? The title
of this treatise, Torch of the Eye of Meditation (bSam gtan mig sgron),
30 SM: 186.1-3: Irnal 'byor mig gi bsam gtan gyi skabs 'dirl ston mun dang rdzogs chen
cha 'dra bas gol du dogs pa'i phyir rgyas par bkod dol. Cf Samten Gyaltsen Kannay,
The Great Peifection (rDzogs chen). A Philosophical and Meditative Teaching in Tibetan
Buddhism, Leiden: E. J. Brill, 1988, 105.
31 For a discussion of such a syncretistic outlook in P. tib. 699, a commentary to a
manuscript on Chinese Meditation Buddhism (S. tib 689) cf my article: "Chinese Chan
and Tibetan Rdzogs Chen: Prelinlinary Remarks on Two Tibetan Dunhuang Manuscripts",
in: Religion and Secular Culture in Tibet. Tibetan Studies II, Leiden: Brill, PIATS 2000,
vol. 2, 2002, 289-307 and my revised and extended article on this topic "Conjunction of
Chinese Chan and Tibetan rDzogs chen Thought: Reflections on the Tibetan Dunhuang
Manuscripts S. tib. 689-1 and P. tib. 699", forthcoming, in: SCEAR.
32 SM: 311.1-6 and my article (2002: 304). Cf also Ka=ay 1988: 112.
33 SM: 12.5-6: I de nas bdag gzhan gyi don du gnas der las brtsam pa 'ang theg pa so
so'i mi rtog pa'i gzhung gzhi [= bzhiJ legs par khong du chud pasl [ ... ]. A later interpolation
(SM: 12.1-2) lists the four vehicles (theg pa bzhi) as [s]tonl tsenl mahiil a ni [= ti].
may give a hint as it is also understood in the broader c ~ n t e x t of "non-
conceptuality." At the very beginning of the book, gNubs chen Sangs
rgyas ye shes makes reference to this title as follows:
Even though nature did not vacillate from the condition
Of primordial spontaneous presence,
If it is not seen, nature manifests [as if] dual;
I pay homage to what has become this very condition [of primordial spon-
taneous presence].
Thence, this meditation called the "eye of the yogi", and which is the king of
direct transmission making one definitively understand the spontaneously Great
Perfection (rdzogs chen), [namely] the ground-of-all, the awakened mind
The "eye" (mig) of the "yogi" (mal 'byor pa) who practices "medi-
tation" (bsam gtan) recognizes the direct transmission of rDzogs chen.
Its very essence is the "condition of primordial spontaneous presence"
(gdod nas lhun gyis grub pa'i ngang). This condition is twofold: it is
empty in nature, nonetheless it is also luminous in nature and is the ground
of all which has the potential to allow phenomenal world arise effort-
lessly. Therefore, within this process of cognition "meditation" (bsam
gtan) is of primary importance. However, bsam gtan is here not to be
misunderstood in its ordinary meaning of being a specific meditation
about something, but rather is to be understood in a much broader sense:
bsam gtan is here - as Herbert Guenther has already put it - part of
the process of "spiritual maturity". In this process, the human being is
cured of the feeling of being separated from the world - a feeling that
originates in believing in the "conceptual aspect" (rtog pa) of experi-
Therefore, in this process of convalescence the understanding of
the "conceptual aspect" is especially important as it leads to the under-
. standing of "non-conceptuality" (mi rtog pa).
Moreover, H. Guenther also brought attention to the term bsam gtan
in a similar context dealing with rDzogs chen teachings and gave the fol-
lowing definition:
34 SM: 2.1-3: I gdod nas lhun gyis grub pa'i ngangl rang bzhin ngang las ma g.yos
kyangl ma mthong rang bzhin gnyis su snangl de nyid ngang gyur bdag phyag 'tshall de
fa kun gzhi byang chub kyi sems lhun gyis rdzogs pa chen po gtan la dbab pa'i man ngag
gi rgyaf pol rnal 'byor pa'i mig zhes bya ba'i bsam gtan 'di [ ... J.
35 Cf Guenther 1983: 353.
The tenn bsam-gtan applies to this 'setting' of an as yet preconscious intend-
ing, which gradually becomes frozen into the customary subject-object
division, on any.level where the noetic-noematic ['mind' (sems)] correlation
is in its fonnation. [ ... ]
To be more precise, bsam-gtan characterizes the moment of transition when
the latent discriminating determinations, that become an explicating and
concentrating attention, begin stirring and are going to move freely in the
context of explicit themes such as subject and object, whereby they harden
into 'mind' (sems). This particular transitional moment within experience
is tenned the 'spontaneous' [lhun grub] or 'self-present' [rang snang] or
'natural' [rang bzhin] setting
According to this definition, bsam gtan is the crucial moment when
the "conceptual aspect" of reality arises and thus can also be cut
through. This "conceptual aspect" of experience already means a lim-
itation of the openness of being. In regard to perception, this openness
means "intrinsic awareness" (rig pa) and is identical with the aspect
of "non-conceptuality". Thus in "meditation" (bsam gtan) one is able
to see through the limiting factor of perception of reality. Or in
the words of the bSam gtan mig sgron itself: in meditation "the con-
dition of primordial spontaneous presence" is illuminated. Finally,
gNubs chen Sangs rgyas ye shes distinguishes this meaning of "non-
conceptuality" in rDzogs chen meditation in comparing it to the under-
standing of "non-conceptuality" according to the other three schools.
We shall now look at the contents of the four main chapters of the
bSam gtan mig sgron.
3.2 "Non-Conceptuality" According To The Four Schools
The four main chapters of the bSam gtan mig sgron, that is chapter four
to seven, demonstrate in great detail the respective understanding of "non-
conceptuality" of the four above-mentioned schools. In the context of the
present research, however, which is merely interested in the fundamen-
tal differences and not in particular details, we shall look at the summary
in the third chapter of the bSam gtan mig sgron which provides a general
36 Herbert Guenther, Kindly Bent to Ease Us, Emeryville: Dharma Publishing, vol. 2,
idea of the basic differences in the understanding of "non,-conceptuality"
of these schools

gNubs chen Sangs rgyas ye shes states that there are different degrees
of insight into "non-conceptuality." He makes the following comparison:
The differences [of insight into non-conceptuality] are like the steps of a
ladder. Just as there are high and low steps of the ladder, there are differ-
ences [according to] these four [schools of Rim gyis pa, Cig car ba,
Mahayoga, and rDzogs chen in regard to their respective understanding of]
Then, how does gNubs chen Sangs rgyas ye shes grade these different
steps of the ladder? We shall look at his explanations one by one.
(1) Concerning Rim gyis pa, it is said that "self-nature" (rang bzhin) is
recognized in a step by step meditation. The four kinds of conceptual
thinking, namely about signs of "own nature", "antidotes", "thus-
ness" and "realisation" are abandoned successively. Therefore, it is
said to be a gradual meditation on the "three gates of liberation,"
namely on those of "emptiness" (stong pa nyid), "marklessness"
(mtshan ma med pa), and "aspirationlessness" (smon pa med pa?9.
(2) The Cig car ba teaches from the very beginning "instantaneously"
(Gig car) the "unborn absolute" (don dam pa ma sykes pa) - beyond
any expectation and striving. This means that one shall learn from the
beginning that all phenomena are "without a fixed frame of refer-
ence" (dmigs su med pa)40.
(3) gNubs chen Sangs rgyas ye shes describes the insight gained through
the method of Mahayoga as "non-dual non-conceptuality" (gnyis su
med pa'i mi rtog pa)41. According to Mahayoga texts, "non-dual
thusness" (gnyis su med pa'i de bzhin nyid) means that "sphere"
(dbyings) and "primordial wisdom" (ye shes) are non-dual. Therefore,
37 H. Guenther already translated this important passage of the SM into English.
Cf Guenther 1983: 351-366.
38 SM: 60.6-61.1: Ide dag gi khyad par skad [= skasl kyi gdang bu bzhin tel dper na
skad [= skasl gdang la mtho dman yod par dang 'dra ste/ mi rtog pa 'di bzhi yang khyad
par yodl. Cf Guenther 1983: 360.
39 SM: 55.6-56.1,56.6-57.1. Cf Guenther 1983: 354-355.
40 SM: 57.1-4. Cf Guenther 1983: 357.
41 SM: 55.5. Cf for a translation of this passage cf also Guenther 1983: 355.
"primordial wisdom" does not even take "sphere" as a "referential
object" (dmigs par mi byed pa)42.
(4) Finally, the result of rDzogs chen meditation is described as "spon-
taneously present supreme non-conceptuality" (lhun gyis grub pa'i
mi rtog pa chen pO)43. In "spontaneously present thusness" the whole
phenomenal world is inherently and, perfectly from primordial times,
naturally luminous in the completely pure expanse of "intrinsic
primordial wisdom" (rang byung gi ye shes). It is the "supreme pri-
mordial non-conceptuality" (ye mi rtog pa chen po) in which mani-
festations are not blocked
Therefore, we may label it as the insight
into 'dynamic emptiness,' which is in its empty aspect "non-
existence" (med pa) beyond duality and at the same time in its lumi-
nous aspect "intrinsic awareness" (rang rig pa) allowing the kalei-
doscope of manifestations arise. Therefore, in rDzogs chen meditation
the real issue is not simply a non-referential (mi dmigs pa) situation,
but innate and luminous awareness itself.
4. Comparison of the Four Practices in the
Aryavikalpapravesanamadhararp: and the Contents of the Four Main
Chapters in the bSam gtan mig sgron
In the fourfold perceptual process, as it is described above according
to the AryavikalpapraveSanamadharal}l, insight into "non-conceptuality"
is gained gradually through the steps of (1) "perception," (2) "non-
perception," (3) "non-perception of perception'," and (4) "perception of
non-perception." In order to apply these four stages of understanding of
"non-conceptuality" to the above illustrated account of four schools Rim
gyis pa, Cig car ba, Mahayoga and rDzogs chen and to clearly demon-
strate how they correspond to each other, we shall look again at gNubs
chen Sangs rgyas ye shes' analysis.
gNubs chen Sangs rgyas ye shes criticises each single school respectively
from one step above on the ladder, that is, he disapprove Rim gyis pa from
the perspective of Cig car ba, disapprove Cig car ba from the view of
42 SM: 59.4,6. Cf Guenther 1983: 359.
43 SM: 55.6. Cf Guenther 1983: 355.
44 SM: 60.2-3,5. For a translation of this passage cf also Guenther 1983: 360.
Mahayoga, and fmally disapprove Mahayoga from the angle of rDzogs chen.
Accordingly, (1) the fault of Rim gyis pa is that it is merely occupied with
"perception" (dmigspa) in order to let the expelience of "non-perception"
(ma dmigs pa) arise
(2) Cig car ba seeks for the unborn absolute, yet it
simply corrupts mind
as it still has a concept of the unborn. As already men-
tioned above, the Cig car ba adept learns from the beginning the "non-
perception" (dmigs su med pa) in regard to all phenomena
(3) Even though
Mahayoga talks about "non-duality" (gyis su med pa), it does not realise
"spontaneously present supreme non-conceptuality" of rDzogs chen. This is
so because Mahayoga gets accustomed to thusness by virtue of "examining
reality" (dngos po gzhal ba) and "different means" (thabs mang po). There-
fore, even though Mahayoga comes close to the rDzogs chen realisation of
"spontaneous presence", it still takes it as an "object of perception" (dmigs
pa yod pa )48 and thus regards it as something supreme orreal. (4) Accord-
ing to gNubs chen Sangs rgyas ye shes, the supreme realisation of non-
duality is only accomplished in rDzogs chen meditation. It refers to the
supreme equality of all manifestations of both sarrzsara and nirval}a. At the
end of his treatise he summarises it again as follows:
Since the Great Perfection (rdzogs chen) is spontaneously perfected and
ultimate thusness, supreme non-duality is without divisions; thus [the Great
Perfection] is the stage of unexcelled primordial wisdom. [ ... ]49
Now, we shall investigate step by step how gNubs chen Sangs rgyas ye
shes' classification corresponds to the fourfold correct practice of the
Aryavikalpapravesanamadhiiral}l. (1) Regarding Rim gyis pa, gNubs chen
Sangs rgyas ye shes himself describes it in terms of "perception" (dmigs
45 SM: 61.2: tsen man rim gyis 'jug pa nil sngar bshad pa ltar dmigs pa la sha thang
bar 'bad nas rna dmigs pa skye ba dang! [ ... J.
46 SM: 61.3. Cf Guenther 1983: 358.
47 SM: 57.3-4: yang de nyid las! las dang po pas sems dang po bskyed pa nas nye bar
brtams tel chos thams cad dmigs su med pa la! bslab par bya' 01. Cf Guenther 1983: 357.
48 SM: 63.2-4; 64.4-5: mal 'byor chen po nang pas mtshan ma'i ting nge 'dzin las Sit
rung nas rtags than yang! lhun grub la dmigs pa yod pa'i phyir! rna mthong ba nil dper
na nyi ma'i snying po bltas na! slar mi mthong gi mig ljir 'gyur ba bzhin no!.
49 SM: 491.4-5: !rdzogs chen ni lhun rdzogs de bzhin nyid mthar thug nyid pasl !gnyis
med chen po dbye ba med pasl lye shes bla ma'i sa yin pas [ ... ]1.
pa). This corresponds to the first stage of insight as it is explained in the
Aryavikalpaprave.sanamadhiiralJI. (2) For Cig car ba gNubs chen Sangs
rgyas ye shes demonstrates that this tradition is occupied with the "unborn
absolute," the inconceivable or - to speak in terms of the Aryavikalpa-
praveSanamadharalJl - with "non-perception" (mi dmigs pa). It IS the
realisation of the non-existence of the external world. (3) According to the
bSam gtan mig sgron, Mahayoga talks about "spontaneous presence" (lhun
grub), the supreme realisation of "non-conceptuality" in rDzogs chen,
however, still takes it as an "object of perception" (dmigs pa yod pa). It
comes close to the realm of "spontaneous presence", and thus frees from
the attachment to the subject-object dichotomy. However, it still regards
it as real or supreme. In terms of the AryavikalpapraveSanamadharalJl this
would match the third step of "non-perception of perception," that is a
non-perception of the perception of the SUbject-object dichotomy. (4)
Finally, gNubs chen Sangs rgyas ye shes argues that only in rDzogs
chen "spontaneous presence" is perfected which is "supreme non-
conceptuality." It is the realisation of the empty but luminous nature, or
in other words the union of "intrinsic awareness and emptiness" (rig stong).
It refers to the equality of all manifestations of both saf(lsara and nirvalJa
- not regarding anything as supreme. In regard to the Aryavikalpa-
praveSanamadharalJl this kind of insight is explained as "perception of
non-perception," non-duality of cognisable objects and cognition. Only in
realising this final stage one is able to enter into "non-conceptual sphere."
When Mahayoga talks about "spontaneous presence", it is the coming into
contact of such a realm, whereas rDzogs chen is the realization of such.
By analysing the contents of the four main chapters of the bSam gtan
mig sgron - that is the understanding of "non-conceptuality" in Rim gyis
pa, Cig car ba, Mahayoga and rDzogs chen - in the light of the fourfold
correct practice as it is explained in the AryavikalpapravesanamadharalJl,
the structure of the bSam gtan mig sgron becomes transparent. In the eyes
of gNubs chen Sangs rgyas ye shes these four traditions clearly describe
a soteriological path in itself, that is a path on which one gradually
increases insight into "non -conceptuality." In this kind of interpretation
the four traditions may be seen as separate parts of a broader picture, or
to use the analogy of gNubs chen Sangs rgyas ye shes: they are differ-
ent steps of one ladder.
5. Critique of Cig car ba according to the bSam gtan mig sgron
In conclusion, we shall summarise again gNubs chen Sangs rgyas ye
shes' view of the differences between Cig car ba (i.e., Chinese Medita-
tion Buddhism) and rDzogs chen as gNubs chen Sangs rgyas. ye shes
himself explained in the beginning that one reason for writing the bSam
gtan mig sgron was the apparent similarity of both traditions. In this own
concluding remarks he says:
Regarding the Cig car ba [tradition] its terminology is similar to rDzogs chen.
Although it teaches non-activity and non-practice, it speaks of ultimate truth
as the ground which is unborn and empty - having in mind the ground which
is not arising and is the perfect reality. However, if one investigates this [view],
there is [still] effort getting accustomed to the condition of emptiness; it [has
the notion of dealing with the two] truths alternately5o. [The Cig car bas] never
practically engage in the non-duality of [the two] truths. Veiled by the own
[erroneous] view, the [Cig car bas] need yet to have to enter into non-duality51.
According to this criticism Cig car ba does not understand the absolute
truth, thusness, but perceives it as the object "empty nature." gNubs chen
Sangs rgyas ye shes therefore classifies it as the understanding of the
absolute according to "three essential categories" (ngo bo nyid gsum) in
the Yogadira school as "perfect reality" (yongs su grub pa, parini;;panna).
In an earlier classification of the different understanding of "non-con-
ceptuality" according to the various philosophical schools he described
that "non-conceptuality which manifests as empty nature is the medita-
tion on the perfect [reality] of the Yogacara school"s2. It is not clear how
gNubs chen Sangs rgyas ye shes comes to the conclusion to bring the
understanding of the absolute in the Cig car ba tradition together with
50 I am not sure about my translation of this sentence. s. Karmay (1988: 105) choose
the following translation which seems to avoid the problem ( ... bden pa re mos pa ... ): "[ ... ]
If we examine this view, it still hankers after the "truth" and works on becoming accus-
tomed to the state of voidness. [ ... ]".
51 SM: 490.3-5: I ston mun ni rdzogs chen dang skad mthunl bya ba med bsgrub pa
med par ston yangl Igzhi mi 'byung ba yongs su grub pa la dgongs nasI don dam pa'i bden
pa rna skyes stong pa'i gzhi la smra stel de la ni brtags na da dung bden pa re mos pa
dangl stong pa'i ngang la 'dris par byed pa dangl rtsol ba yod del bden pa gnyis med pa
la spyod kyang rna myong stel rang gi Ita bas bsgribs te gnyis med la bzod 'jug dgos sol.
ej. also Karrnay 1988: 105.
52 SM: 55.3-4: Istong pa'i ngo bor snang la mi rtog pa nil rnal 'byor spyod pa'i yongs
su grub pa bsgom pa'ol. Cf. Karmay 1988: 105.
Yogacara philosophy. He does not give further evidence to confirm his
view. However, as we have seen in the above structural analysis of the
contents of the four main chapters of the bSam gtan mig sgron, it makes
perfect sense to place Cig car ba within the broader picture of Buddhist
soteriology. Yet, in my view, an assertion like the final one of gNubs
chen Sangs rgyas ye shes about Cig car ba would call for a more elabo-
rate verification.
BBK Bukkyo bunka kenkyilsho kiyo [Bulletin of the
Research in Buddhist Culture],
Studies in Central and East Asian Religion
bSam gtan mig sgron; cf. gNubs chen Sangs rgyas ye shes
TaishO shinshO daizokyo t:lEfliiU:1U1 [[Chinese] Buddhist Canon
from the Taisho Era], Taibei Reprint
Tibetan Tripitaka
Bentor, Yeal, "Fourfold Mediation: Outer, Inner, Secret, and Suchness", in:
Religion and Secular Culture in Tibet. Tibetan Studies II, Leiden: Brill,
PIATS 2000, vol. 2, 41-58.
"Chos dang chos nyid mam par 'byed pa'i gzhung [Discrimination of dharma
and dharmatii]", attributed to Maitreya, in: IT. 108 no. 5523, f. 48b.1-51b.6.
"Chos dang chos nyid mams par 'byed pa'i tshig le'ur byas pa'i 'grel pa ye shes
snang ba mam 'byed [Commentary to the verses of the Discrimination of
dharma and dharmatii. Discrimination of Primordial Wisdom and Appear-
ances]", by Mi pham Rin po che (1846-1912), in: Mathes 1996, annex.
Demieville, Paul, Le concile de Lhasa, reprint, 1't edition 1952, Paris: College
de France Institut des Hautes Etudes Chinoises, 1987.
"Foshuo ru wu fenbie famen jing [sutra of Entering into the
Dharma Gate of 'Non-Conceptuality' Taught by the Buddha]", translated
by Danaprua :Iitliiii, in: T. 15, no. 654, 8D5-806.
Gomez, Luis 0., "The Direct and Gradual Approaches of Zen Master Mahayana:
Fragments of the Teachings of Mo-ho-yen", in: M. Gimello/p. N. Gregory
(ed.), Studies in Ch'an and Hua-yen, Studies in East Asian Buddhism No.1,
Honolulu: University of Hawaii Press,1983[a], 69-168.
-, "Indian Materials on the Doctrine of Sudden Enlightenment", in: Lewis
Lancaster/Whalen Lai (ed.), Early Ch'an in China and Tibet, Berkeley:
Berkeley Buddhist Studies Series, 1983[b], 393-434.
Guenther, Herbert, Kindly Bent to Ease Us, Emeryville: Dharma Publishing,
vol. 2, 1976. '
-, '''Meditation' Trends in Early Tibet", in: Lewis Lancaster/Whalen Lai (ed.),
Early Ch'an in China and Tibet, Berkeley: Berkeley Buddhist Studies Series,
1983,351-366. .
Karmay, Samten Gyaltsen, The Great Pelfection (rDzogs chen). A Philosophi-
cal and Meditative Teaching in Tibetan Buddhism, Leiden: E. J. Brill, 1988.
Kazunobu Matsuda "Nirvikalpa-pravesa-dharaI)I ni tsuite [On the
Nirvikalpa-pravesa-dharar.n]", in: Buddhist Seminar 34 (1981), 40-49.
-, "Nirvikalpapravesadharar.n. Sanskrit Text and Japanese Translation", reprint
from: Bulletin of the Research Institute ofBukkyo University (3/1996),89-113.
Mathes, Klaus-Dieter, Unterscheidung der Gegebenheiten von ihrem wahren Wesen
(Dharmadharmatavibhiiga), Swisttal-Odendorf: Jndica et Tibetica Verlag, 1996.
Meinert, Carmen, "Chinese Chan and Tibetan Rdzogs Chen: Preliminary Remarks
on Two Tibetan Dunhuang Manuscripts", in: Religion and Secular Culture
in Tibet. Tibetan Studies II, Leiden: Brill, PIATS 2000, voL 2, 289-307.
-, "Conjunction of Chinese Chan and Tibetan rDzogs chen Thought: Reflec-
tions on the Tibetan Dunhuang Manuscripts S. tib. 689-1 and P. tib. 699",
forthcoming, in: SCEAR.
gNubs chen Sangs rgyas ye shes (9
century), gNubs chen sangs rgyas ye she rin
po ches mdzad pa 'i sgom gyi gnang gsal bar phye ba bsam gtan mig sgron
[Torch of the Eye of Meditation Elucidating the Very Heart of Meditation,
Composed by gNubs chen Sangs rgyas ye shes Rin po che], reproduced from
a manuscript made presumably from an Eastern Tibetan print by 'Khor gdon
gter sprul 'Chi med rig dzin, Leh: Smartsis shesrig spendzod, Vol. 74, 1974.
"'Phags pa rnam par rill rtog par 'jug pa zhes bya ba'i gzungs [The Supreme
dharalJ! of Entering into Non-Conceptuality]", translated by Jinamitra,
DanasTIa, and dKa' ba dpal brtsegs, in: IT. 32, no. 810, la-6b.
"'Phags pa rnam par rill rtog par 'jug pa zhes bya ba'i gzungs kyi rgya cher 'grel
pa [Extensive Commentary to The Supreme dharalJI of Entering into Non-
Conceptuality]", by KamalasTIa, in: IT. 105, no. 5501, f. 146b.6-174b.1.
Rao Zongyi (Jao Tsung-I) "Wangxi Dunwu dacheng zhengli jue xushuo
bing jiaoji [preface and Notes to W angXi' s Dunwu
dacheng zhengli jue (Ratification of the True Principle of the Mahayana
Teachings of Sudden Awakening)]", in: Chongji xuebao iMti!iOi'!l [The Chung
Chi Journal] 9/2 (1970), 127-148.
Ru wu fenbie zongchi jing Jd!!! 5t M' [The dharal}-l of Entering into N on-Con-
ceptuality], Jiang it 23, preserved in the Beijing National Library.
Seyfort Ruegg, D., Buddha-nature, Mind and the Problem of Gradualism in a Com-
parative Perspective, London: School of Oriental and African Studies, 1989.
Shiu, Henry C. H. ?lIlUjIllf, Bianfafaxing lun Bubai shi lun 0 [Dhar-
madharmatavibhiiga. The Commentary of Mi pham Rin po che], ed. by Tam
Shek-wing Hongkong: Vajrayana Buddhism Association Limited, 2000.
Tam Shek-wing (ed./transl.), Bianfa faxing lun. Shiqin shi lun 0
ftfJlfHit [Dharmadharmatavibhiiga. The Commentary of Vasubandhu],
Hongkong: Vajrayana Buddhism Association Limited, 1999.
TaishO shinshO daizokyo ;kiEllifiJl;kilW [[Chinese] Buddhist Canon from the TaishO-
Era], 85 vols., ed. by Dazang jing kanxing hui, Taibei reprint of the edition
of 1927/28, Taibei, Xinwenfeng chubanshe, 1973.
Tibetan Tripitaka, Peking edition of the Kanjur and Tanjur, Tokyo reprint, ed.
by Daisetz T. Suzuki, Tokyo/K yoto: Tibetan Tripitaka Research Institute,
168 vols., 1957.
Tucci Guiseppe, Minor Buddhist Texts II, Rome: Is. M. E. 0., 1958.
Ueyama, Daishun, "The Study of Tibetan Ch'an Manuscripts Recovered from
Tun-huang: A Review of the Field and its Prospects", in: L. Lancaster/
W. Lai (ed.), Early Ch'an in China and Tibet, Berkeley: Berkeley Buddhist
Studies Series,1983, 327-350.
Ueyama, Daishun/Kenneth W. Eastman/Jeffrey L. Broughton, "The Avikalpa-
praveSa-dhiiralJi: The Dharani of Entering Non-Discrimination", in: BBK
(1983), 32-42.
de la Vallee Poussin, Louis, Catalogue of the Tibetan Manuscripts from Tun-
Huang in the India Office Library, Oxford: Oxford University Press, 1962.
Yoshimura, Shyuki, The Denkar mao An Oldest Catalogue of the Tibetan Bud-
dhist Canons with Introductory Notes, Kyoto: Ryilkoku University, 1950.
Giulio AGOSTINI earned his PhD in Buddhist Studies at the University of
California, Berkeley, with a thesis on the Buddhist laity in ancient India. He is
currently working as an independent scholar on Abhidhanna and Vinaya mate-
rials concerning lay people.
Dan ARNOLD did his doctorate at the University of Chicago. He is presently
Assistant Professor in the Faculty of Religious Studies at the University of McGill
in Montreal.
Colette CAILLAT, professor emeritus at the University of Paris ill, is a mem-
ber of the Academie des Inscriptions et Belles-Lettres de l'Institut de France.
She has focused on Indo-Aryan grammar and linguistics, collaborating on the
Critical Piili Dictionary II, and has done extensive research on Jainism, especially
on the canonical texts of monastic discipline.
Mario D'AMATO is currently visiting Assistant Professor in Asian Religions
and the Philosophy of Religion at Hampshire College in Amherst, Massachu- .
Paul HARRISON is Professor of Religious Studies at the University of Canter-
bury, Christchurch, New Zealand, where has he taught Buddhism and Asian Reli-
gions in the Religious Studies program since 1983. His main research interests
lie in the fields of Mahayana Buddhist history and literature, the history of the
Tibetan canon, and Buddhist manuscripts.
Cannen MEINERT currently works as a research fellow at the Institute for the
History and Culture of India and Tibet, Tibetan section, at Hamburg University,
Germany. Her research in progress is dealing with the issue of violence in Tantric
Buddhism, considering Tibetan and Chinese materials alike. She obtained here
PhD from Bonn University in 2001 with a comparative study on Chinese Chan
Buddhism and Tibetan rDzogs chen. Her research mainly focuses on Buddhism
between Tibet and China.
Robert SHARF recently took a position as Professor of Buddhist Studies in the
Department of East Asian Languages and Cultures at the University of Califor-
nia, Berkeley, after having taught at McMaster University and the University of
Journal of the International Association of Buddhist Studies
Volume 26 Number 1 2003
Michigan. He is author of Coming to Terms with Chinese Buddhism: A Read-
ing a/the Treasure Store Treatise (Hawai'i, 2002) and coeditor of Living Images:
Japanese Buddhist Icons in Context (Stanford, 2001).
The International Association of Buddhist Studies
Jikido Takasaki
Ernst Steinkellner
Tom Tillemans
General Secretary
Jerome Ducor
Regional Representatives:
Janet Gyatso (Cambridge, MA, USA)
Kazunobu Matsuda (Kyoto)
David Seyfort Ruegg (London)
Peter Skilling (Bangkok)
Directors at large:
Nalini Balbir (Paris), Colette Caillat (Paris),
Georges Dreyfus (Williamstown, MA), Robert M. Gimello (Cambridge, MA),
Oskar von Hinuber (Freiburg), ShOryu Katsura (Hiroshima),
Liying Kuo (Paris), Richard Salomon (Seattle),
Lambert Schmithausen (Hamburg)
The International Association of Buddhist Studies, founded in 1976, is devoted to
promoting and supporting scholarship in Buddhist Studies in all its aspects, past and
present, around the world. Membership is open to scholars of all academic disciplines.
Membership dues are USD 40 for full members, USD 20 for student members,
USD 1000 for life members. Subscriptions to the JIABS for libraries and other institu-
tions are USD 70. Dues and subscriptions may be paid in the following ways:
1. by direct bank to bank transfers in US dollars or in the equivalent amount in
Euros or Swiss francs. The transfers should be to the lABS accounts at the Banque
Cantonale Vaudoise, PI. St-Frans;ois, Ch-1001 Lausanne, Switzerland (SWIFT CODE:
BCVLCH2L, clearing no. 7677), account numbers, 986.51.04 for US dollars,
5042.09.82 for EUR and 983.51.02 for Swiss francs.
2. by Visa or Mastercard
3. by cheque payable to the "Association Internationale d'Etudes Bouddhiques
(lABS)." There will be no supplementary charges for cheques drawn in Swiss
currency on Swiss banks. Otherwise, please add 10 dollars or 15 Swiss francs to cover
our processing charges.
4. by Eurocheque in Swiss francs.
Prospective members from developing countries may contact the Treasurer concerning
subsidized membership rates. Dues and subscriptions are payable per calendar year
by December 31 of the previous year. Payments other than direct bank transfers should
be sent to Dr. Jerome Ducor, Section de langues et civilisations orientales, Universite
de Lausanne, BFSH 2, CH -1015 Lausanne, Switzerland.