THE BUDDHIST FORUM

VOLUME II
Seminar Papers 198890

Edited by
Tadeusz Skorupski

THE INSTITUTE OF BUDDHIST STUDIES, TRING, UK

THE INSTITUTE OF BUDDHIST STUDIES, BERKELEY, USA
2012

First published by the School of Oriental and African Studies (University of London), 1992
First published in India by Heritage Publishers, 1992

Online copyright 2012 belongs to:

The Institute of Buddhist Studies, Tring, UK &
The Institute of Buddhist Studies, Berkeley, USA

British Library Cataloguing in Publication Data

The Buddhist forum. Vol. II
1. Buddhism
I. University of London, School of Oriental and African Studies
294.3
ISBN 81-7026-179-1

CONTENTS
The online pagination 2012 corresponds to the hard copy pagination 1992

Abbreviations..................................................................................................................................vii
List of Illustrations...........................................................................................................................ix
Introduction......................................................................................................................................xi
T.H. Barrett
Devils Valley ot Omega Point: Reflections on the Emergence
of a Theme from the N....................................................................................................................1
T.H. Barrett
Buddhism, Taoism and the Rise of the City Gods..........................................................................13
L.S. Cousins
The Five Points and the Origins of the Buddhist Schools............................................................27
P.T. Denwood
Some Formative Inf1uences in Mahyna Buddhist Art............................................................61
G. Dorje
The rNying-ma Interpretation of Commitment and Vow...........................................................71
Ch.E. Freeman
Savti, Vyavahra and Paramrtha in
the Akamatinirdea and its Commentary by Vasubandhu........................................................97
D.N. Gellner
Monk, Househo1der and Priest: What the Three Ynas
Mean to Newar Buddhists.............................................................................................................115
C. Hallisey
Councils as Ideas and Events in the Theravda........................................................133
S. Hookham
The Practical Implications of the Doctrine of Buddha-nature............................................149

R. Mayer
Observations on the Tibetan Phur-ba and the Indian Kla .........................................................163
K.R. Norman
Theravda Buddhism and Brahmanical Hinduism:
Brahmanical Terms in a Buddhist Guise..................................................................193
References.....................................................................................................................................201

ABBREVIATIONS

A
AO
AM
As
BEFEO
BHSD
BM
BSOAS
BSR
CIS
CPD
CSSH
CSLCY
D
Dp
EA
EFEO
EJS
EI
ERE
HJAS
HR
IASWR
IBK
IHQ
IIJ
IT
JA
JAS
JHR
JIABS

Aguttara-nikya
Acta Orientalia
Asia Major
Ahaslin
Bulletin de lEcole Franaise dExtrme Orient
F. Edgerton, Buddhist Hybrid Sanskrit Grammar and Dictionary
Burlington Magazine
Bulletin of the School of Oriental and African Studies
Buddhist Studies Review
Contributions to Indian Sociology
Critical Pli Dictionary
Comparative Studies in Society and History
Chin-so liu-chu yin, in TC, no. 1015
Dgha-nikya
Dpavasa
tudes Asiatiques
Ecole Franaise dExtrme Orient
European Journal of Sociology
Epigraphia Indica
Encyclopedia of Religion and Ethics, edited by James Hastings, Edinburgh,
T.&T. Clark, 1911
Harvard Journal of Asiatic Studies
History of Religions
Institute for Advanced Studies of World Religions
Indogaku Bukkygaku kenky
Indian Historical Quarterly
Indo-Iranian Journal
Indologica Taurinensia
Journal Asiatique
Journal of Asian Studies
Journal of the History of Religions
Journal of the International Association of Buddhist Studies

JNCBRAS
JNRC
JPTS
JRAS
JS
Kv
Kv-a
MCB
M
Mhbv
Mhv
Mp
MSMS
Ptis
PTS
RH
RO
S
SBE
Saddhamma-s
SLJBS
Sp
SSAC
T
Th
TMKFTCC
TP
TC
TTD
TTP
UCR
VBA
Vin
Vism
WZKSO
ZDMG

Journal of the North China Branch of the Royal Asiatic Society

Journal of the Nepal Research Centre
Journal of the Pali Texts Society
Journal of the Royal Asiatic Society
Journal des Savants
Kathvatthu
Kathvatthu-ahakath
Mlanges Chinois et Bouddhiques
Majjhima-nikya
Mahbodhivasa
Mahvasa
Manoratha-pran
Monumenta Serica Monograph Series
Paisambhid-magga
Pali Text Society
Revue Historique
Rocznik Orientalistyczny
Sayutta-nikya
Sacred Books of the East
Saddhamma-sagaha
Sri Lanka Journal of Buddhist Studies
Samantapsdik
Studies in South Asian Culture
The Taish edition of the Buddhist Canon in Chinese (vol. no.)
Theragth
Tao-men ko-fa ta-chan-chi, in TC, no. 1215
Toung Pao
The Taoist Canon, text numbered in accordance with the Harvard-Yenching
Index to its titles
Tibetan Tripiaka, sDe-dge Edition
Tibetan Tripiaka, Peking Edition
Univeristy of Ceylon Review, Colombo
Visva-bharati Annals
Vinaya-piaka
Visuddhimagga
Wiener Zeitschrift fr die Kunde Sd- (und Ost) asiens
Zeitschrift der Deutschen Morgenlndischen Gesellschaft

LIST OF ILLUSTRATIONS

Between pages 6465:

Fig. 1. Plan of martyrium of St Babylas, Antioch-Kausiye.
Fig. 2. Plan and reconstructed section of cathedral, Bosra.
Fig. 3. Plan and section of Audience Hall of Al-Mundhir, Resafa.
Fig. 4. Plan of Parthian palace at Assur-Labbana.
Fig. 5. Plan of Palace at Firuzabad with Audience Hall in centre.
Fig. 6. Plan of Palace at Sarvistan.
Fig. 7. Section of the Chahar Qapu at Qasr-i-Shirin.
Fig. 8. Elevation of model maala-palace made by Tibetans at Dharamsala, India.
Fig. 9. Reconstructed elevation of cruciform reliquary in the shape of a building,
Shaikhan Dheri, Pakistan.
Fig. 10. Cruciform Hindu Temple at Patan, Kashmir.
Fig. 11. Section of Char Narayan Temple, Nepal.
Between pages 168169:
Fig. 12. Vajrakla rolls the Mt Meru Phur-ba
Fig. 13. A standard Tibetan phur-ba (kla).
Acknowledgements: Figures 1, 2 & 3 taken from C. Mango, Byzantine Architecture, New York,
1975; Figures 4, 5, 6 & 7 from A.U. Pope, A Survey of Persian Art, Oxford, 193839; Figures 12
& 13 were drawn by Cathy Cantwell.

INTRODUCTION

The present publication, the second volume in the Buddhist Forum series, contains a
selection of the papers presented for discussion at the SOAS Buddhist Forum seminars
during the past two academic years, namely, 198889 and 198990. The seminars which
are held once a month during term time are open to everyone who has an interest in
Buddhism and cover a whole range of topics relating to the various areas of Buddhist
studies such as philosophy, doctrines, philology, history, art and other relevant subjects.
The seminars are in progress and it is hoped that further volumes will appear in future
years.
The rather enigmatic title of the first paper only becomes clear on a full reading,
but its subject matter could more or less be summed up as Buddhahood for the nonsentient: an encounter between Buddhist and Taoist traditions. Professor Barretts main
concerns in this paper are not the doctrinal issues involved in the interpretation of
Buddhahood with regard to inanimate things, but rather certain historical and religious
interactions between the Buddhists and the Taoists in China between the fourth and the
eighth centuries. At the outset the paper establishes that so far as Taoism is concerned,
the nature of the Tao, the ultimate and inherent Way, as being present in all inanimate
things, was already formulated several centuries before the common era, a considerable
time before the advent of Buddhism in China Apart from quoting a conversation
attributed to Chuang-llu, the author also refers to several other Tao philosophers,
including Kuo Pu. It is through Kuo Pu that a connecting link is made to Kuei-ku-tzu,
the Master of Devils Valley, a place which is difficult to locate. So far as Buddhism is
concerned, although there may be some indications in the canonical texts, the Indian
tradition has never taken much interest in the question of Buddhahood for non-sentient
things. But in China this matter was widely discussed and various opinions were
expressed. It seems that a clear formulation of this doctrine developed in China as part of
a much larger discussion which was concerned with the sudden attainment of
Buddhahood. The theme of the attainment of Buddhahood by the non-sentient was given
a most prominent place in Japanese N plays of the fourteenth and fifteenth centuries.
However, as already mentioned, the major part of Professor Barretts paper focuses not
on doctrinal issues but on the mutual influences and

xi

interactions between the Buddhists and the Taoists in connection with this concept. Thus
on the whole this is a paper that falls within the category of comparative religion. Those
who are specifically interested in the question of Buddhahood for the non-sentient or its
treatment in the N plays should refer to the sources provided in Professor Barretts
informative annotations.
The second paper, also by Professor Barrett, is similar in nature to the first, but
this time deals with the origin of city gods. Like the preceding contribution this is an
excellent study that takes into account not only the religious elements but also the sociocultural factors that might have influenced the emergence of the cult of city gods within
the Taoist tradition. Here the concern is to trace the adaptations made within Taoism to
accommodate and justify a new group of minor deities.
A long and learned paper by Lance Cousins discusses the formulation and nature
of the five propositions attributed to Mahdeva. His detailed treatment of the Five
Points is not restricted to doctrinal matters but covers a wide range of issues such as the
historical background, the validity of the available sources and the formulation of the
Kathvattu. Among the historical considerations arc included the dates of the early
Buddhist councils, the origin of the Mahsghika school, the accession of Aoka and the
chronology of events after the Buddhas final demise. The author attaches more
importance and validity to the Theravda sources than to the material preserved in
Chinese and Tibetan. However, in treating his subject he makes use of all the available
sources. An examination of the historical formulation and of the doctrinal content of the
Kathvtthu leads him to suggest a new logical sequence for the Five Pomts. In doing so
Cousins seeks to demonstrate that there were certain complex developments that
influenced their nature and interpretation, something that is reflected in the three
formative phases of the Kathvatthu. Apart from exploring the original sources Cousins
also expresses his opinion on certain important contributions written in Western
languages, such as those by Louis de La Valle Poussin, Etienne Lamotte and others.
While Lance Cousins attaches great importance to the Pli sources and follows
meticulously the classical path of interpreting historical events and doctrinal issues, the
paper by Professor Charles Hallisey of Loyola University, Chicago, offers a new
approach to the study and interpretation of the Buddhist councils. Its specific concern is
more with the functional role of the Buddhist councils in the Theravda tradition and
rather less with establishing historical facts. Professor Hallisey suggests two possibilities
or orientations which taken together can be helpful in enhancing or even rectifying our
understanding of the Buddhist councils and of the Theravda tradition as such. One
orientation is referred to as event and the other as idea, the former being more rooted
in history and the latter more in anthropology. It is argued that past events, whether truly
or allegedly historical, in

xii

this case the Buddhist councils, provide inspiration for certain conceptual formulations
that derive their authority from those events and in turn inspire or sanction new events; or
to put it differently, the idealized past validates present actions. When applied to the
Theravda, this general model can provide new insights into the pattern in which the
Theravda developed, justified and validated its own internal evolution. The concept of
events and ideas as such is not introduced here as something entirely new. However,
Professor Hallisey makes excellent use of it to demonstrate how the history of
phenomena with a dual character as events and ideas can help us to understand the
Theravda as it actually evolved.
Roy Normans paper, like those by Cousins and Hallisey, is also related to the
Theravda tradition and is concerned with Buddhist terminology. It has been argued by
some scholars in the past that Buddhism was merely an offshoot of Hinduism. There
were also scholars who tended to assume that there were no traces of Hinduism in
Buddhism. However, these two extremely opposed approaches to the understanding of
Buddhism arc hardly tenable in the light of the vast number of studies produced by
modem scholarship. It is difficult to see how Buddhism could exist in isolation and
produce a whole body of complex doctrines without employing or making adaptations of
the already existing religious and philosophical terminology. Modem comparative studies
demonstrate beyond any reasonable doubt that Buddhism has always interacted with
other contemporary religious traditions of India. Some scholars refer to this interaction
between Buddhism and the other traditions as drawing from the common repository of
Indias religious and philosophical wealth, while others tend to stress borrowing on the
part of Buddhism. Norman does not accept that Buddhism can be seen as an offshoot of
Hinduism nor as being without any trace of Hinduism. His paper demonstrates that
Buddhism remains indebted to Brahmanical Hinduism, in particular insofar as
terminology is concerned. Taking into account the vast body of Buddhist terminology,
this is rather a succinct paper. Nevertheless it offers a full enough coverage of a number
of important Brahmanical terms to demonstrate how they were adapted, reinterpreted or
modified within the context of the Buddhas teachings. There is no doubt that more
studies of this kind are needed to unravel the complexities of Buddhist terminology; they
would also enhance our understanding of Buddhist thought as a whole.
The papers by Gyurme Dorje and Shenpen Hookham relate to Tibetan
Buddhism. Gyurme Dorjes paper is based on two of his substantial works, namely, his
doctoral thesis The Guhyagarbha-tantra and its Commentary, and his forthcoming
annotated translation of bDud-joms Rin-po-ches history of the rNying-ma school. His
doctoral thesis, over one thousand pages long, consists of an English translation and a
critical study of the Guhyagarbha-tantra and its fourteenth-century commentary, entitled
Phyogs-bcu mun-sel. This work,

xiii

unfortunately still unpublished, and the forthcoming translation of bDud-joms

Rin-po-ches text constitute an important contribution to the study of Tibetan
Buddhism. The seminar paper itself represents a scholarly assessment and interpretation
of vows and commitments as found in the rNying-ma tantras. Since the notion of vows
and commitments is considered within the context of the Nine Vehicles, one learns about
their interpretation across the whole spectrum of Buddhist traditions; but of course
through the rNying-ma approach.
The fierce and at times violent controversy which evolved in Tibet between the
Buddhist thinkers who upheld the view of self-emptiness and those who defended the
view of other-emptiness is well known to scholars of Tibetan Buddhism. What is
interesting about Dr Hookhams papers is the connection she makes between these two
controversial interpretations of the notion of emptiness and two distinct trends within
Buddhism as a whole with regard to the nature of man and the path to spiritual perfection.
It is possible to argue that within Buddhism, right from the outset, or at least from the
time of the early schools, there emerged two very broadly defined orientations with
regard to doctrinal formulations and spiritual practices: one stressing the weaknesses of
human nature and the importance of spiritual purification and elimination of impediments
on the path to liberation; and one placing more emphasis on the positive aspects of man,
such as the view of mind as being pure by nature although temporarily obscured by
adventitious impurities. The doctrine of Buddha-nature as being present in all living
beings can be comfortably accommodated into the second category. It is interesting to
follow Hookhams argument and see how she divides Buddhism into these two doctrinal
orientations and relates them to particular schools of Tibetan Buddhism. However, she
docs not limit herself to doctrinal considerations alone but takes a step further to apply
her findings to the structure and functioning of Tibetan society.
Charlotte Freeman, a SOAS Ph D student, has for the past five or more years been
working under the supervision of Dr Piatigorsky on the Akyamatinirdea-stra and its
commentary by Vasubandhu. The material contained in her paper represents a small part
of her extensive researches. Although savti and other concepts discussed in the paper
have been fairly well explored in Indian sources and in the writings of Western scholars,
Freemans particular style and her persuasive presentation of Vasubandhus interpretation
add a fresh and further dimension to our knowledge of those concepts.
The papers by David Gellner and Rob Mayer, while dealing with different topics,
have in common the challenge they offer to earlier scholarship in their specific fields.
David Gellners main concern is to correct certain approaches to the treatment of
Newar Buddhism and to provide a new and positive method for its interpretation.
He makes an opening assertion to the effect that some scholars in the past and

xiv

contemporary Westerners educated by modem Western means have failed to understand

Newar Buddhism because of certain preconceptions which they had or have about
Buddhism as such. In order to remedy this situation, he proposes a different approach
which would justify and validate the nature of Newar Buddhism. He makes use of certain
anthropological insights which he relates to historical developments within Buddhism in
order to explain the pattern on which Newar Buddhism evolved and crystalized into a
unique religious and socio-cultural entity. The main aim of the paper is to disprove the
views of Newar Buddhism as anything like a remnant of Indian Buddhism or its decadent
descendant, and to demonstrate that when understood on its own terms and within its own
context it represents a perfectly valid form of Buddhism.
Rob Mayers paper deals with the origin, history and symbolism of the ritual
implement known as kla in Sanskrit and phur-ba in Tibetan. This is the second longest
paper in this volume and it covers a wide range of issues connected with the nature and
use of the kla. He surveys and assesses not only the material produced in Western
languages but also refers to sources in Sanskrit, Tibetan, and even Chinese; his survey of
Indian literary material includes both Buddhist and Hindu sources. A number of
secondary issues apart, the main aim of the paper is to demonstrate that the kla had a
long and complex history in India and that it was not a Tibetan invention.
Philip Denwood, apart from being specialist teacher of Tibetan language and
culture, takes a keen and professional interest in art and architecture. In this particular
paper, he aims to demonstrate that certain architectural elements present in Iran and
developed in connection with the concept of universal sovereignty had some inspirational
influence on the design of the Buddhist maala. He also makes similar observations
with regard to the position of images in architectural constructions such as the maala palace.
Finally, I would like to express my gratitude to those who helped me in bringing
out this volume. Apart from the people who gave me their moral support and
encouragement, I wish to thank Susan Madigan for typing the majority of the papers,
Martin Daly for his administrative help and advice, Diana Matias for her professional
editorial help, and to Anne Glazier for reading the proofs. Words of gratitude are also due
to the School of Oriental and African Studies for accepting responsibility for the cost and
distribution of this second volume of the Buddhist Forum.

xv

Devils Valley to Omega Point: Reflections on the

Emergence of a Theme from the N
T.H. Barrett

Despite the mention of Japanese literature in my subtitle, and the naming of a destination
on a specifically Christian horizon, the itinerary I propose to lay before the reader lies
strictly within the confines of China. The Japanese section of the road has already been
well covered by other scholars dealing with the ultimate salvation of plants and trees as a
literary theme in mediaeval Japan.1 As for the twentieth-century conception of an Omega
Point towards which creation yet proceeds, although this, too, was arguably the product
of a sojourn in China, my reading so far suggests that no more than coincidence is
involved here: the singular, lonely intelligence responsible for fusing Christian doctrine
and evolution theory into this splendid, dazzling vision seems to have been if anything
completely at odds with the religious sensibilities of its Chinese environment,2 and those
who have made it their business to compare Buddhist and Catholic notions of spirituality
have (sometimes quite explicitly) kept this particular thinkers work out of
consideration.3
By bringing in the term Omega Point I wish to indicate merely that more than one
civilization has come up with the ultimate in what I would call cosmic optimism , the
notion that the entire phenomenal world can look forward to a glorious future of religious
fulfilment: no more detailed comparison is intended. I do not even wish to tackle the
question as to why such cosmic optimism should

By Donald H. Shively, Buddhahood for the Nonsentient: a theme in No plays , HJAS, 20, 1957,
135161, whence (p. 150) my epigraph; W. LaFleur, Saigy and the Buddhist Value of Nature , HR,
13.2, 1973, 93128, and 13.3, 1974, 227248. LaFleur, 94, note 1, lists the main Japanese studies of
the development of the doctrine.
2
See Claude Rivire, En Chine avec Teilhard, 19381944, Paris, 1968, 1306, 1435: her summary
on page 132 (le Pre aimait peu la Chine et les Chinois) alas says it all.
3
One recent comparative work does consider the grass and trees becoming Buddhas as understood
in the Tien-tai tradition of Buddhism, but simply as a prologue to some inter-faith reflections on
environmentalism: see Nanzan shky bunka kenkyjo, ed., Tendai Bukky to Kirisuto-ky, Tokyo,
1988, 183206. For an explicit exclusion of Teilhard from consideration in a comparative discussion
of Zen and Christian mysticism, see the following note.
1

emergethough I note that some would argue that a sentiment of hope constitutes one of
the primary forms of religious life4my only interest is in how the idea came to the fore
in China; who was the Chinese equivalent of that remarkable Jesuit thinker?
Our journey of discovery must, however, start a long way from Omega Point, a
long way even from Devils Valley, in a dilapidated barnyard (or so it would seem)
somewhere in the state of Sung, some four centuries before the Christian era. Here the
iconoclastic philosopher Chuang-tzu is represented as discoursing on the Tao, the
ultimate unseen Way, with a neighbour. In the translation of Angus Graham:
Tung-kuo-tzu inquired of Chuang-tzu
Where is it, that which we call the Way?
There is nowhere it is not.
Unallowable unless you specify.
It is in molecrickets and ants.
What, so low?
It is in the weeds of the ricefields.
What, still lower?
It is in the tiles and shards.
What, worse than ever!
It is in the shit and piss.
Tung-kuo-tzu did not reply.5
Now as a statement of the omnipresence of the underlying Tao, these remarks
attributed to Chuang-tzu could hardly be improved upon, but what they do not tell us
about this Tao, this Way, is where it was supposed to lead. Indeed, throughout all the
writings ascribed to Chuang-tzu, or to his alleged predecessor Lao-tzu, this does not seem
to be an issue: the unchanging nature of the hidden Way is affirmed; any suggestion that
it is going anywhere is completely avoided.6 So we must pass on, and stop off again in
the early fifth century of the Christian era, in the southern Chinese capital of Nanking, or
Chien-kang as it was then known. Here in this vast metropolis the monasteries of an
originally alien way of thought,

See Peter L. Berger, A Rumor of Angels, Garden City, New York, 1970, 545. William Johnston,
The Still Point: Reflections on Zen and Christian Mysticism, New York, Evanston, San Francisco,
London, 1971, contains the best part of a chapter on Teilhard, but finds him wanting (pages 1678) in
some important respects, and less similar to Zen (page 172) than might be assumed.
5
A.C. Graham, Chuang-tzu, London, 1981, 161; note (page 158) that this passage is seen as part of a
cycle of stories later than Chuang-tzus authentic work and distinguished by themes adumbrating the
concerns which emerged centuries afterwards in Chinese Buddhism.
6
I have referred to G. Finazzo, The Notion of Tao in Lao Tzu and Chuang Tzu, Taipei, 1968, to check
this.
2

Buddhism, are very much in evidence, and here learned Chinese monks may be found
debating the key concepts of the new religion in the light of their own cultural heritage.
The greatest of these controversialists is Tao-sheng (c. 360434), who is particularly
exercised by a question of latency, of hiddennessnot quite the same one as Chuang-tzu,
though; he seems to have in mind the Confucian thinker Mencius and his theory of the
beginnings of goodness hidden within the individual, which allow every man to
become a sage like the paragons Yao and Shun.
But the question is now posed in a Buddhist form; does everyone possess the
Buddha-nature, fo-hsing, the potential within them for achieving Buddhahood? Taoshengs answer is yes, even the most depraved have this capacity. Nor did he stop there:
such a capacity could, he declared, be realised very rapidly, in one go, as it were. This
proved to be a particularly controversial thesis, generating a protracted argument over
sudden versus gradual enlightenment, even though (to judge by the recent research of
Whalen Lai) Tao-sheng seems to have wandered almost nonchalantly into this
controversy: maybe it was just that there was an argument there already waiting to
happen.7
One point that was not in dispute for Tao-sheng or for his Buddhist successors in
the following two centuries was that Buddhahood, fast or slow, was the prerogative of
sentient beings only: the external world remained, as in Indian Buddhist thought, no more
than a vessel, no more than a backdrop against which the great drama of enlightenment
was played out.8
But, meanwhile, what of Chuang-tzus spiritual descendantsor at least those
who claimed him as their forefather, the Taoists? Their monasteries, too, were in the first
half of the fifth century on the verge of spreading across the land also, in emulation of the
Buddhists. Originally Taoist clergy had not lived apart from the faithful, it seems, except
in small hermitagesthough some of these were by this time well established. A source
contemporary with Tao-sheng, for instance, notes that a Taoist hermitage already existed
in Ching-chou (modern Hubei province) on a mountain by a stream known as the Clear
Brook or (through a minor orthographic alteration) the Blue Brook, Ching-hsi.9 The
author of this record would have known that a poet of the early fourth century, Kuo Pu
(276324), had already written of a mysterious Taoist of the Blue Brook glimpsed in

. See Whalen Lai, Tao-shengs Theory of Sudden Enlightenment Re-examined , in Peter N.

Gregory, ed., Sudden and Gradual: Approaches to Enlightenment in Chinese Thought, Honolulu,
1987, 169200: note that the historical development uncovered by Lai actually suggests a process the
reverse of that given in my summary: the notion of a universal Buddha-nature was a by-product of
polemics over sudden enlightenment.
8
The bhjana-loka of standard Buddhist texts.
9
See the Ching-chou chi, as preserved in Wang Mo, ed., Han-Tang ti-li shu chao, Peking, 1961,
414.
3

the vicinity, and that in his poetic imagination this lone figure had been identified with
Kuei-ku-tzu, the Master of Devils Valley, a shadowy political thinker and strategist
almost as ancient as Chuang-tzu.10 Other places claimed the Master for themselves; other
places claimed the name Blue or Clear Brook.11 But on the basis of Kuos poetic flight of
fancy this scenic valley in Ching-chou seems to have become a favourite haunt of
religious recluses and scholars.
The Buddhists arrived about a century later, in the person of the redoubtable
monk Tao-hsien, who in 517 started a twenty-eight year occupancy of Blue Brook
mountain in the face of sustained hostility from the Taoists already present.12 By 594,
when the Buddhist polemical essayist Fa-lin (572640) arrived, the mountain was
evidently a major centre of learning, for we read that he was able to acquire an education
in both Confucianism and Taoism there, besides a training in his own faith:13 the wealth
of citations in his works of the literature of his opponents traditions gives ample
testimony to the accuracy of this account.14 A brief picture of the communal life of study
and meditation practised by Buddhists on Blue Brook Mountain in the late sixth century
may also be found in the biography of Fa-hsi (572632), who appears to have had some
forty students to his name.15
Another older Buddhist contemporary, Fa-lun (528c. 605), on the other hand,
seems to have supplemented his studies (evidently, again, encompassing Taoism and
Confucianism)16 at the Blue Brook by recourse to a separate Buddhist centre some fifty
kilometers away, the Jade Spring Monastery, Y-chan ssu.17 There are perhaps some
grounds for wondering whether two separate monks, both called

10

See Hsiao Tung, comp., Wen-hsan 21, Hong Kong, 1960, 461: commentary on this passage
provided Wang Mo (see preceding note) with the fragment of the Ching-chou chi just cited.
11
The spot most usually favoured for the location of Devils Valley was quite elsewhere, at Yangcheng in Henan, to judge from the materials collected in Chin En-fu, ed., Kuei-ku-tzu, Beijing, 1985,
reprint of 1805, though the valley of the Blue Brook did at one time lay claim to a cave allegedly
inhabited long ago by the Master: see Wang Hsiang-chih, Y-ti chi-sheng (comp. 1227), Taipei, 1962
reprint of 1860 ed., 73.7b. Cf. also note 33 below. Identifying the Blue Brook has been a matter of
some concern to Japanese scholars: cf. Yoshioka Yoshitoyo, Dky to Bukky, Vol. One, Tokyo,
1959, 31114; Kamata Shigeo, Chgoku Bukky shisshi kenky, Tokyo, 1969, 173218; the matter
has been resolved by the discovery of the new source cited in n. 32 below.
12
See Tao-hsan (596667), comp., Hs Kao-seng chuan 25, 651ab, in T, vol. 50.
13
See Yen-tsung, Tang hu-fa shamen Fa-lin pieh-chuan 1, 198b, in T, vol. 50.
14
See Ishii Masako, fuchi Ninji, comp., Rikuch, T, S no kobunken shoin Dky tenseki
mokuroku, sakuin, Tokyo, 1988, 312321, for a complete listing of the Taoist works cited in his
writings.
15
Hs Kao-seng chuan 19, 587a-b.
16
Hs Kao-seng chuan 9, 500a.
17
To judge by Chih-pan, comp. (c.1270), Fo-tsu tung chi 9, 199c, in T, vol. 49.

Fa-lun, may not have studied consecutively at the two centres,18 yet the Jade Spring and
the Blue Brook were certainly in communication during the seventh century, in the
person of another monk named Tao-yeh, who started his career at the former site, but
for most of his life lived on Blue Brook Mountain.19 The Jade Spring Monastery was a
much newer factor in the religious life of Ching-chou: it had first been constructed and
served as a base during the years 592 to 595 for Chih-i (538597), the great systematizer
of the Tien-tai school of Chinese Buddhism.20
Here later followers of the school maintained throughout the seventh and eighth
centuries a centre of doctrinal study independent of the schools main base in the Tientai mountains of Zhejiang.21 Though the Tien-tai tradition as a whole seems to have
possessed a coherence spanning such local centres unusual for Buddhist groups in China
at this period,22 continuing support for the school in the Ching-chou area was to be
expected, since Chih-i himself hailed from the region,23 and many Ching-chou monks
like Fa-lin and Fa-lun could claim some connection with him.24
But a word is now necessary once more concerning the Taoists, who by the
seventh century were in a difficult position, despite strong imperial support. Taoism
during this period has been described as crypto-Buddhist,25 which is an unjustified slur,
since the religion had ancient and purely indigenous roots, whether it went back to
Chuang-tzu or not. But certainly the Chinese had always been staggered by the
overwhelming prolixity of Buddhist literature,26 and it is quite

18

It is clear that Chih-pan, as cited in the preceding note, does not explicitly identify his Fa-lun with
the famous Fa-lun who died early in the seventh century; there are also grounds for believing that the
famous Fa-lun was already too prominent and too busy by the time of the founding of the Jade Spring
Monastery to have had any opportunity to study there: cf. L. Hurvitz, Chih-i, 538597: An
introduction to the life and ideas of a Chinese Buddhist monk, MCB, 12, 1962, 154.
19
Hs Kao-seng chuan 25, 661c662a.
20
Hurvitz, Chih-i, 154-7.
21
Sekiguchi Shindai, Tendai Shikan no kenky, Tokyo, 1969, 185205.
22
There are signs that Tien-tai saw itself not simply as a teaching tradition but as a much more
organized system of both theory and practice: see Muranaka Yush, Tendai shoki gyh no shsei ni
tsuite , IBK, 23.2, March, 1975, 5616.
23
See Hurvitz, Chih-i, 106.
24
For Fa-lin, note the point raised by Muranaka on the last page of the article cited in note 22 above.
For Fa-lun, see Hurvitz, Chih-i, 154, note 1, and cf. Fo-tsu tung-chi 7, 186c. For a general study of
Chih-is disciples, which reveals a strong Ching-chou connection, see Ikeda Rosan, Chigi metsugo
no Tendai kydan to dk , IBK, 21.1, December, 1972, 338343.
25
Thus A.F. Wright, in The Cambridge History of China, vol. 3, Cambridge, 1979, 77. In fairness to
Professor Wrights memory, he did do much to promote the scholarship which has led since his death
to a revision of this view.
26
See the remarks of Mou-tzu (early fourth century?) preserved in Seng-yu, comp., Hung-ming chi 1,
2b, in T, vol. 52.
5

clear that, faced with such fecund opposition, the Taoists succumbed to the temptation to
plagiarize Buddhist materials more or less wholesalea practice in which they were, of
course, found out and which drew from the Buddhists quite predictable ridicule. For the
Taoists had been at times quite mechanical in their rewriting, replacing Buddha with
Tao27and hence, it seems, Buddha-nature with a new coinage, Tao-nature, taohsing, which first appears in the late sixth or early seventh century.28
But, of course, the Tao-nature is in everything, and so even in inanimate objects.
The explicit assertion that it is in tiles and shards, in trees and stones, is made by the
Taoist Meng An-pai in his Tao-chiao i-shu.29 We do not know exactly when this text
was written, but one thing that is patently obvious is that although it only quotes Taoist
sources,30 it is organized in such a way as to mark a strong debt to Tien-tai Buddhism.31
As for Meng, he is described as a Taoist of the Blue Brook and just by chance a
literary anthology has preserved a record by Chen Tzu-ang (661702), a well-known
poet, of Mengs success in securing imperial support for the renovation of a Taoist
foundation in Ching-chou in 699.32 So Mengs Tao-chiao i-shu is a late seventh or early
eighth century product of the Blue Brook religious environmentan environment which,
incidentally, seems to have continued to attract admirers of the Master of Devils Valley
well into the ninth century.33
What is most striking is that Mengs assertion of the religious potential of
inanimate objects must antedate by several decades the first explicit statement that trees
and stones possess the Buddha-natureand that that statement may be found in the
writings of Chan-jan (711782), next to Chih-i the most famous

27

See the remarks in K. Chen, Buddhism in China, Princeton, 1964, 474, for some occasional lapses
in this process; Kamatas study of Buddhism and Taoism, cited above, note 11, provides ample
examples of borrowing, too.
28
According to Kamata, 53ff.
29
Kamata, 6774.
30
See Ishii and fuchi, 150160, for a complete listing of its copious citations of Taoist texts.
31
See the analysis in Kamata, 185198.
32
Li Fang et al., comp. (987), Wen-yan ying-hua, Beijing, 1966, 822.1a-2b. For Chen, see (pending
the appearance of a SOAS publication on his work) the discussion in Stephen Owen, The Poetry of the
Early Tang, Yale U.P., 1977, 157183: the Wen-yan ying-hua 848.5a-b, 734.8a-b, provides ample
testimony to his interest in Taoism. Meng, like Shen-hsiu (see McRaes remarks cited below, note 39),
seems to have taken advantage of the Empress Wus family connections with the Ching-chou area.
33
See the poem by Li She in Peng Ting-chiu et al., comp., Chan Tang shih 477, Peking, 1960,
54234; Li styled himself Master of the Clear Brook, apparently in imitation of the Master of
Devils Valley.

thinker in the Tien-tai tradition.34 Now it must be conceded that any link between Meng
and Chan-jan can only have been indirect: Chan-jan is most unlikely to have ever visited
Ching-chou, and represents a strand within the Tien-tai tradition apparently
unconnected with the Jade Spring Monastery.35 By the time that Meng was active,
moreover, there is some evidence of Taoist familiarity with Tien-tai thought which may
be traced not to Ching-chou but to a Taoist presence on the Tien-tai mountains
themselves.36
It must also be admitted that there is also a considerable consensus amongst
students of Chinese Buddhism that Buddhahood for the nonsentient was already on the
agenda during the seventh century. No one seems to have felt obliged to make explicit
claims in the way that Chan-jan later did, but such thinkers as Chi-tsang (549623) and
Fa-tsang (643712), it has been argued, already concede the possibility in their
discussions of the Buddha-nature.37
Some evidence has also been adduced which would make the Northern Chan
leader Shen-hsiu (606?706) a very clear (if tacit) supporter of the religious prospects of
the non-sentient.38 This is of particular interest, since Shen-hsiu was
34

See note 31 above. It will have become obvious that this essay owes a very great deal to the
remarkable pioneering research of Kamata Shigeo, which was, however, carried out prior to the
realization by Japanese scholarship of the true date of Meng An-pai. The source listed in note 32
invalidates Kamatas belief that Meng was influenced by Chan-jan, and raises the strong possibility
that the direction of the influence was the other way round; it does not invalidate Kamatas research
into the similarities between Meng and Chiao-jan. Chan-jans writings have long served as the locus
classicus for the notion that vegetation too has a spiritual destiny: they are even treated briefly in
English in Fung Yu-lans well-known History of Chinese Philosophy.
35
Hibi Sensh, Tdai Tendaigaku josetsu, Tokyo, 1966, gives a full account of the historical
background to Chan-jans writings, including a detailed study of his travels, which took him nowhere
near Ching-chou.
36
This is the view taken by Livia Khn; see her Seven Steps to the Tao, Monumenta Serica
Monograph Series XX, Nettetal, 1987, 27, 67, though the matter is somewhat more complex than this
study allows, as is shown by the research of Meng Wen-tung, e.g. as republished in Chung-kuo chehseh 4, 1980, 318.
37
For some discussion in English, see La Fleur, 95, and also T. Unn, The Buddhat theory of Fatsang , Transactions of the International Conference of Orientalists in Japan 8, 1963, 3441.
Japanese scholarship continues to interest itself in the part played by the writings of these two men in
the emergence of the doctrine of Buddhahood for the non-sentient; the most recent study that I have
been able to consult has been Akao Eikei, Hz ni mieru smoku jbutsu ni tsuite , IBK, 32.2,
March, 1984, 404411. All the work done in this vein tends to view the process of development as
taking place solely within the Chinese Buddhist tradition itself: there is, of course, much to be said for
this, but Kamatas work has (to my mind) shown that Taoism cannot simply be left out of the picture.
For one Japanese survey that does adopt a broader perspective, see the following note.
38
Kamata, 2034. Fukunaga Mitsuji, Issai shuj to smoku doseki , Bukky shigaku, 23.2, March,
1981, 103118, gives a masterly diachronic survey of the religious prospects of non-sentient objects in
China, including some reference to Neo-Confucianism: note the problems raised for example by
Wang Yang-ming (14721529) in Julia Ching, To Acquire Wisdom: The Way of Wang Yang-ming,
Columbia U.P., 1976, 127, 142. Fukunaga does not, however, indulge in any excursions into local
history like the present essay.

a resident of the Jade Spring Monastery at the very time we know that Meng An-pai was
active in Ching-chou.39 The interrelationship between the later Tien-tai leader Chan-jan
and the various Chan factions of the eighth century is a complex topic, and one which
may well have a bearing on Chan-jans open advocacy of the potential Buddhahood of
plants and trees.40
But the very fact that eighth-century discussion of their issue in Buddhist circles
uses phrases such as tiles and shards, and even molecrickets and ants suggests that,
whatever their internal disagreements, these Buddhists were conscious of the Taoist
background to the question. And not simply the ultimate source of such phrases in
Chuang-tzu: the Pao-tsang lun, which mentioned molecrickets and ants as well as
vegetation, demonstrates quite irrefutably a familiarity with Taoist doctrinal writings
similar to Mengs Tao-chiao i-shu.41
I set out at the start of this trip to seek a Buddhist Teilhard, but that search must
be abandoned in the face of the evidence presented above suggesting not the flash of
insight of one solitary thinker but the slow emergence of an idea within a protracted
process of interaction encompassing two rival religions and several rival schools. But at
least it seems possible to identify the type of milieu, and maybe even the very locality,
where a Chinese view of the cosmos finally fused with an Indian conception of spiritual
development so as to produce in time an important theme in Japanese literature.42 That
process surely could not have taken place without the existence of communities of
religious scholars such as those on Blue Brook Mountain who looked back to the Master
of Devils Valleywhat better title for the principal of a mediaeval college!as their
founder, and who exchanged ideas (and no doubt traded insults) with other centres such
as the Jade Spring Monastery in a wider world of learning. And if I am right, this raises a
shocking possibility, one undreamed of even in Chuang-tzus scatological imagination:
that the profound truth of the Way may be found even in so vile a place as the Devils

39

See John McRae, The Northern School and the Formation of Early Chan Buddhism, Honolulu,
1986, 5051, and cf. 2656.
40
One attempt to sort out these interrelationships may be found in Yagi Nobuyoshi, Tai-Zen rysh
no ksh ni tsuite , IBK, 18.2, March, 1970, 61112, though both he and the other scholars who have
touched upon this area would appear to have been faced with problems in Chan-jans biography
requiring careful historical research.
41
On this, see yet again Kamata, op. cit., 2045 and 237242.
42
This is not to deny that in Japan purely Japanese attitudes towards nature further played an
important part in the acceptance of the theme in literature, as is noted, for example, by the contributors
to Tendai Bukky to Kirisuto-ky (see note 3 above), 191, 193.
8

Valley of academic debate; that the road to Omega Point may even pass through precincts
of modern universities.
The Postscript
The preceding pages reproduce almost without alteration a paper presented to the
Buddhist Forum in March, 1989; this paper, however, represented a reworking of
unpublished materials first drafted ten years earlier as a result of my discovery of the
source concerning the Taoist Meng An-pai given in n. 32 above. These materials were
laid aside as the result of the almost simultaneous publication in English of a lengthy
footnote discussing the same source by fuchi Ninji,43 and only taken up again because
subsequent discussion in English of the Buddhahood of plants and trees had not taken
Meng An-pais writings, used much earlier by Japanese scholars, into account. During
the intervening period, however, my own concerns had shifted away from the debate over
the Buddhahood of plants and trees as such towards an interest in the institutional
environment in which contacts with Taoist thought were possible. As a result rewriting
produced a certain imbalance between text and notes.
Professor Seyfort Ruegg, who attended the Forum when the paper was presented,
quite rightly drew attention to its consequent tendency to minimize internal development
within the Buddhist tradition as an important cause for the eventual prominence of the
notion of Buddhahood for the non-sentient. Indeed, to the studies mentioned in n. 36
above it would be possible to add at least one further article in English exploring this
theme in the thought of Chi-tsang from a purely Buddhist perspective.44 Of particular
interest was Professor Seyfort Rueggs suggestion that even the most undeveloped
Buddhist materials show a markedly positive attitude to vegetable life. Certainly the
Chinese did notice some of these materials, at least by the tenth century, when they were
excerpted in the Buddhist encyclopedia I-chu liu-tieh:45 earlier encyclopedias may do
the same, but a rapid check has not turned up any exactly parallel examples.46 But these
particular materials simply suggest that all the vegetable kingdom, as well as the animal
kingdom, is provided with tutelary deities; there is nothing here that resonates with
Chinese ideas of spiritual progress, so it is unlikely that they played a part in

43

See note 5, 2256 of Ninji Ofuchi, The Formation of the Taoist Canon, in H. Welch and A.
Seidel, eds., Facets of Taoism, New Haven & London, 1979, 253267.
44
Namely, Aaron K. Koseki, Prajpramit and the Buddhahood of the Non-sentient World: The
San-lun assimilation of Buddha-nature and Middle Path Doctrine, JIABS, 3, 1, 1980, 1633.
45
See Makita Tairy, ed., I-chu, Giso Rokuj/I-chu liu-tieh, Kyoto, 1979, 16.34a, page 364. I-chu
completed this work in A.D. 954.
46
Pao-chang, ed., Ching-l i-hsiang 46, page 240b in T, vol. 53, notes that trees over seven foot tall
and a foot in girth harbour spirits, but this would seem to indicate an acceptance of hamadryads in
decent size trees only (hardly a startling notion), not the sanctity of all vegetation suggested by Ichus citation.
9

nurturing Chinese speculation on Buddhahood for the non-sentient. Buddhist attitudes

towards the natural environment, however, do still deserve to be explored in yet more
depth than hitherto.
But so, too, does the institutional environment which promoted Buddho-Taoist
contacts in seventh-century China, and which in my view must at least have added a
further stimulus to Buddhist thinking. Most accounts of these contacts have concentrated
on events at court;47 indeed, our knowledge of Tang intellectual life as a whole is largely
dominated by the perspective afforded by central government.48 But not all higher
learning was in the hands of, nor at the service of, the state: many mountains other than
the Blue Brook Mountain attracted individuals and communities of different religious
affiliations to pursue contemplation and study in what must have amounted virtually to a
university setting. Some research on this aspect of Tang intellectual life has already
been undertaken in East Asia,49 but in Western languages only accounts of individual
study, such as the poet Po Ch-is retreat to Lu-shan, have been published.50
What is intriguing about the Blue Brook is its rise to major prominence in the
seventh century, followed by a subsequent almost total eclipse: most other centres, like
Lu-shan or Tien-tai, remains important into much later times, even into the twentieth
century. As suggested above, the start of the story of the Blue Brook cannot be pushed
back any further than the fifth century, though in subsequent reading I have come across
a tale, perhaps ultimately of that period, alleging a Buddhist presence by the Blue Brook
in the last quarter of the fourth century A.D.51 This concerns a young Buddhist monk who
dreams that the "Lady of the Blue Brook Temple" appears to him and demands that he
should take over as the god of her shrine: he tells his companions as a result that he must
be fated to die to take up this charge, and asks them to visit him when he assumes his new

47

Such is the emphasis of standard accounts like Kubota Ryon, Shina Ju-D-Butsu kshshi, Tokyo,
1943, and Lo Hsiang-lin, Tang-tai san-chiao chiang-lun kao, Tung-fang wen-hua, 1,1, 1954, 8597, following the emphasis in their main sources.
48
This is the standpoint taken, for example, in D.L. McMullens excellent State and Scholars in Tang
China, Cambridge, 1988.
49
The classic study in this vein is Yen Keng-wang, Tang-jen tu-shu shan-lin ssu-yan chih fengshang, Chung-yang yen-chiu-yan Li-shih y-yen yen-chiu-so chi-kan, 30.2, 1959, 689728, which
does not, however, collect materials on Blue Brook Mountain. Earlier works on the eremitic tradition
in Chinese history, such as Chiang Hsing-y, Chung-kuo yin-shih y Chung-kuo wen-hua, Shanghai,
1947, also make clear the role of mountain centres in higher edcuation.
50
See Arthur Waley, The life and times of Po Ch-i, London, 1949, 120.
51
See Wang Shao-ying, ed., Sou-shen hou-chi, 5, Beijing, 1981), 312: this text, attributed to the poet
Tao Chien (365-427), is already cited under his name in the early sixth century, as noted by Y
Chia-hsi, who is quoted on page 147 of Wangs edition. The passage on the Lady of the Blue Brook
is cited in a seventh-century Buddhist encyclopedia, Tao-shih, ed., Fa-yan chu-lin 90, page 953a in
T, vol. 53.
10

role. When, after his subsequent sudden death, they do visit the temple, his voice
converses with them, requesting them to sing psalms and regretting his separation from
their company. This tale is clearly to do with a problem of previous spiritual occupancy,
a topic to which I revert in the next paper, though it provides an unusual twist when
compared with better known accounts of pioneer monks vanquishing chthonic deities,
often in ophidian form.52 One may presume that it dates to a time when the Buddhists
were relatively insecure newcomers, but it is difficult to pin down the period of
composition precisely, and a late Tang reworking of the story (moralizing ponderously
over the young monks fate as the karmic result of earlier persistent gross immorality)
sets it in the fifth century, under the Liu-Sung dynasty.53
During the mid-sixth century the Blue Brook Mountain became (though
apparently not for long) the scene of a genuine historical episode almost as bizarre, when
it was chosen as a base by the remarkable Lu Fa-ho, a charismatic leader who combined
the roles of warlord and magus, and who reconciled a declared commitment to Buddhism
with a penchant for Taoist thaumaturgy.54 Lus regime, one presumes, must have featured
in the monograph on Blue Brook Mountain said to have been compiled by Fa-lin:
unfortunately, not a single word of this survives, so far as I am aware.55 Fa-lin, however,
very probably did not live long enough to include in his work any account of the Blue
Brooks most illustrious Buddhist alumnus, the monk Tao-lin, who succeeded in the midseventh century in reaching India by the sea route and studying at the famous Buddhist
university of Nland, where he is said to have much impressed the great Buddhist
logician Dignga.56
Tao-lins biography reveals that he was also associated in his early days with the
Jade Spring Monastery; one of the advantages of the Blue Brook Mountain seems to have
been its ease of communication with Chiang-ling, administrative centre of Ching-chou,
nearer to which the Jade Spring Monastery lay. The biography of the late sixth century
monk Fa-hsing implies that he frequently made the journey from mountain to town,
passing the future site of the Jade Spring Monastery on the way.57 Chiang-ling itself was
at this point no mean city; it had briefly been capital of the Liang dynasty, and then
became the chief city of the Later Liang dynasty

52

For a classic example of such an account, see pages 946 of Hisayuki Miyakawa, Local cults
around Mount Lu at the time of Sun Ens rebellion, in Welch and Seidel, Facets of Taoism, pp. 83
101, concerning the second century A.D. Buddhist missionary An Shih-kao.
53
See Huai-hsin, Shih-men tzu-ching lu 1, 808a in T, vol. 51.
54
The chief account of Lus extraordinary career may be found in Li Pai-yao, Pei-Chi shu, 32,
Beijing, 1972, 427431; Lu is said to have started his active military career from Blue Brook
Mountain at the time of the rebellion of Hou Ching (548552) on page 427.
55
See the preface to Fa-lin, Po-hsieh lun, 475a, in T, vol. 52.
56
See I-ching, Ta-Tang Hsi-y chiu-fa kao-seng chuan, 2, 6c7a in T, vol. 51.
57
Tao-hsan, Hs Kao-seng chuan 25, 658a.
11

not one of the periods most powerful political units, admittedly, but one which continued
to maintain its nominal independence for about a generation, from 555 to 588.58 But from
the seventh century onwards the region quite manifestly declined in importance and lost
populationprobably to the south bank and to the delta of the Yangtseand it also
suffered increasing administrative fragmentation into the tenth century, as it came within
the orbit of several more successful centres.59 This may explain the paucity of later
references to Blue Brook Mountain: a medical text by the Master of the Clear Brook
which is said to have circulated in the Tang must presumably be attributed to the ninthcentury poet mentioned in n. 33 above,60 but it is difficult to find any eminent Buddhists
associated with the site in later times.61 The connection with the Master of Devils
Valleyremarked on already in the preface to one of Fa-lins works62was reaffirmed
constantly in Taoist hagiography,63 but the area does not seem to have supported any later
Taoist communities of note, either. Thus while monographs on other mountains like Lushan and Tien-tai were compiled and recompiled as the communities on them continued
to flourish, Fa-lins work on the Blue Brook was lost, and its role as an ancient seat of
ecumenical learning became obscured.
In sum, then, I would stress not simply the need to see the development of
Buddhist ideas within their full Chinese intellectual context. What is also necessary is
some appreciation of the institutional arrangements which made interaction between
different religious traditions possible, and here a study of local history can be of value
also: if one can find ones way through Devils Valley, the road to Omega Point becomes
far clearer.

58

See Hurvitz, Chih-i, 105, 107 note 2, etc., for some of this political background as it affected
Buddhism.
59
See Wang Gungwu, The Middle Yangtse in Tang politics, 203 (and page 199 for the tenthcentury outcome), in A.F. Wright and D.C. Twitchett, eds., Perspectives on the Tang, New Haven
and London, 1973, 193235.
60
See Lo Shih-lin, chen Li, Liu Wen-chi and Liu Y-sung, comps., Tang-shu ching-chi i-went hochih, New Haven & London, 1973, 193235.
61
The tenth-century Chan monk Hung-chin may have resided on our Clear Brook Mountain, though
it is described as lying in the territory of Hsiang-yang, further north, rather than Ching-chou: see Taoyan, Ching-te chuan-teng lu, 24, 400a in Taish Canon vol. 51. An inscription, apparently
describing events dated in the late eleventh century, preserved in Wang, Y-ti chi-sheng, 73-8a,
records that a Dharma-master of the Severed Arm had connections with both the Jade Spring and the
Clear Brook. I suspect that the date is a mistake for 617: the author of the inscription, Li Chou, would
appear to be the same man who in the late eighth century wrote a biography of the Sixth Patriarch of
Chan: see Yanagida Seizan, Shoki Zenshu shisho no kenky, Tokyo, 1967, 99, note 16, for the
significance of this lost biography.
62
In the preface to his Po-hsieh lun, 475a.
63
The standard form for the Masters biography from the early tenth century onward, which
reconciles his sojourn at the Blue Brook with his residency of Devils Valley, appears on page 9 of Tu
Kuang-ting, Hsien-chuan shih-i, in Yen I-ping, ed., Tao-chiao yen-chiu tzu-liao, vol. 1, Taipei, 1974.
12

Buddhism, Taoism and the Rise of the City Gods

T.H. Barrett*

If one can learn much about a religious tradition from observing its interaction with its
rivals, and particularly from the encounter of a great, well-established tradition like
Buddhism with new forms of religiosity, then China is perhaps the best place of all to
observe this process. This is not simply because of the outstanding Chinese
historiographic traditions which allow us to put so much of its religious history into
contextin fact for the type of religion examined here historiographic problems are far
from absent, as we shall see. Rather, the simultaneous existence in mediaeval times of
another great and established tradition, Taoism, allows us to treat the question of
reactions to novelty from a comparative perspective by tracing through the same period
of time the evolving reactions of both old religions to the newcomer. In this study the
role of newcomer is played by the so-called cheng-huang cults.
The class of Chinese deities known as cheng-huang or city gods has long excited
the interest of scholars. I have only glanced at what I take to be the earliest serious studies
of these gods, which appeared in 1910 and 1924;1 as with David Johnsons recent and
very stimulating historical examination of these cults,2 the point of departure for this
study goes back no further than Teng Ssu-ys research of 1935.3 One of Tengs aims at
this time was to explore the connections between Buddhism and Taoism and the city gods
in order to assess the possible influences of these religions on the appearance of the city
god cults. The evidence which he uncovered, to which we shall revert, led him to believe
that the origins of the city

An earlier version of this paper was read at UCLA, October, 1989; my thanks to all those who
offered comments at that time.
1
P.A. Volpert, Tschng huang, Anthropos, 5,5/6 (Sept.Dec., 1910), 9911026; F. Ayscough,
The Chinese cult of Cheng-huang lao-yeh, JNCBRAS, 55, 1924, 131155. To judge by the last
couple of pages of each study, the line of inquiry pursued here was not contemplated by these
pioneers.
2
David Johnson, The City-god cults of Tang and Sung China, HJAS, 45,2, Dec. 1985, 363457. As
an examination of the rise of these cults from the point of view of social history, this article leaves
little room for additional comment; the approach adopted here, however, concentrates on the history
of religion.
3
Teng Ssu-y, Cheng-huang kao, Shih-hseh nien-pao, 2.2, 1935, 24976.
13

gods might not be found within either traditions, and this conclusion is further underlined
by Johnsons article.
Our present concern is somewhat different. I take it as more or less axiomatic that
neither Buddhists nor Taoists were responsible as such for the emergence of this type of
cult. I do not know who precisely was; Johnson points to the role of groups described as
elders, whom he suspects of having had merchant connections.4 The terminology as
such does not carry such connotations; intriguingly enough we find it used at a very early
date to describe the religious leadership of non-Chinese peoples of the South.5 But
merchant connections cannot be ruled out: the research of Kanai Noriyuki has uncovered
direct evidence for merchant involvement in local religion during the Sung,6 if not in
support of the cheng-huang cults themselves. The only instance I have noticed in my
own reading where the source of financial support for refurbishing a cheng-huang
temple is specified (in the eleventh century) uses the unrevealing term hao to describe the
man who stepped in to take over the cost from the community as a wholein other
words, a boss, a man of local power who was not considered by the writer (a Buddhist
monk) part of the elite.7
This, of course, fits in well with the recent findings of Robert P. Hymes
concerning Sung religion,8 but the whole question lies rather beyond the scope of the
present essay. My own question is simple: confronted with what was indubitably a novel
phenomenon in the history of Chinese religion, how did the established traditions react?
Convention requires that some words should be said about Confucianism also, and I have
certainly not excluded Confucianism from my title because this tradition had nothing to
say in response to the appearance of the cheng-huang. For though Confucianism tended
to keep a low profile as a religion, it did possess a clear-cut canon, and the rise of a noncanonical cult could be treated in but two ways: either it was denounced (as those NeoConfucian zealots, the Cheng brothers, preferred),9 or it was explained away as a
development implicitly present in the Confucian Classics, by means of the philological
sleight of hand at which Chinese scholars have always excelled; examples of the latter
can be found going back to the ninth century.10 But during the period when the chenghuang

Johnson, The City-god cults of Tang and Sung China, 419.

Hs Chia-jui, Ta-li ku-tai wen-hua-shih kao, Hong Kong, 1979 reprint of second ed., 266.
6
See pp. 456 of Kanai Noriyuki, Sdai no sonsha to Bukky, Bukky shigaku kenky, 18.2, March
1976, 3156: here elders actually appear to be distinguished from merchants, to include other local
notables, patrons.
7
Wen-ying, Hsiang-shan yeh-lu 2, Peking, 1984, 26.
8
See Robert P. Hymes, Statesmen and Gentlemen, Cambridge, 1986, especially ch. 7.
9
Cheng I and Cheng Hao, Ho-nan Cheng-shih i-shu 22A, in Erh-Cheng chi, Peking, 1981, 2956.
10
Tengs article (n. 3 above) opens with some good examples of this. An inscription dated 840
preserved (via a Sung source) in Lu Hsin-yan, ed., Tang-wen shih-i, Taipei, 1962 reprint, 29.3a4b,
provides an excellent early instance of Confucian casuistry: this is no. 51 in Johnsons list (see next
note).
5

14

spread across the land (roughly, the Tang and early Sung), Confucianism itself
underwent a transformation affecting its own self-image far more deeply than the
concurrent changes in Buddhism and Taoism, thus rendering it far less suitable for
purposes of comparison, albeit historiographically far more convenient.
For when one forsakes Johnsons master list of datable inscriptions,11 in which
typically a representative of the state, usually in some sense a Confucian, records his
interaction with a specific community, in order to trawl through literature of an explicitly
Buddhist or Taoist inspiration, precise details of time and place are often harder to come
by. I would, nonetheless, like to think that I have pinned down the earliest Taoist source
to mention the city gods, the Chin-so liu-chu yin, to the century following the An Lu-shan
rebellion of 755: still a little vague, but better than can be achieved for many Taoist
texts.12 It provides, at all events, an intriguing picture of change and continuity, quite
consistent with what we know of other aspects of Taoism at this time.
The Chin-so liu-chu yin is nominally divided between text and commentary, the
former ascribed to late Han revelation, the latter to the early Tang. The first of its
references to city-gods is in the text itself, in the middle of a long list of gods of the
powers of nature (wind, rain) and of localities (mountains, rivers) who will respond to a
Taoist possessed of the right magic; a note from the commentator observes that the
granting of city-godships is in the gift of the Lord of the Latter Days (hou-sheng chn),
the supreme deity presiding over the text as a whole, and quotes the Su-Wu chi (evidently
a topographical text, of which no trace survives) as listing Jui, King of Wu, as an
exampleof him, more shortly.13
Later on in the commentary, in directions for how to counteract drought, we are
told that anciently there were no cities, but now their spirit officers (cheng-huang shenkuan) should head the list of local deities prayed to,14 in dealing with epidemics and
hauntings, too, another note advises adding them in;15 and in supplications for
journeying mercies the note suggesting their addition remarks that such deities had
already appeared in the Chin dynasty.16 These reminders are

11

Johnson, The City-god cults of Tang and Sung China, 4517.

See T.H. Barrett, Towards a date for the Chin-so liu-chu yin, BSOAS, 52,2, July 1990. Further
consideration of this text, especially in relation to its references to Buddhism, has not caused me to
alter my opinion, but I should point out that this dating as yet awaits acceptance.
13
CSLCY 16.5b. For the Lord of the Latter Days, see Isabelle Robinet, La Rvlation du Shangqing
dans lhistoire du Taoisme, Tome II, Paris, 1984, 107.
14
CSLCY 18.5a.
15
CSLCY 23.7b.
16
CSLCY 24.7a.
12

15

followed later by a rite of exorcism in which temple-shrines (miao-she) and cities

(cheng-huang) are, according to the main text itself, to be notified: no comment at all is
offered on this.17 The next main text reference, a method of counteracting black magic
calling, inter alia, upon the spirit officers of cities, provokes a lengthy burst of
aetiological commentary, as follows, in summary:
In high antiquity they only spoke of she-miao, there were no cheng-huang shen.
These started from King Jui of Wu, whose tomb is in Wu. In Chin times the
prefect Liu Wen-ching was building a city-wall in Wu and reached the top of the
tomb. At night the walls were all broken down and the earth carted off by spirit
soldiers so that it was nowhere to be seen.
After this happened seven or eight times Liu posted a watch, which arrested Jui
when he came out of the ground at the head of a ghost army several tens of thousands
strong. Jui complained about the disturbance to his tomb, but offered to protect the city if
the wall was resited one hundred paces away. Nowadays, our text says, anywhere that has
a prior burial (presumably of someone numinous) in its underworld, may, as the result of
consultation between the local underworld and God (tai-shang tien-ti) be granted a
landlord (ti-chu, sic!), who is appointed chief of the spirit-officials of the city.18
Here is a story worthy of comparison with any discussed in Johnsons article.
Note in particular the attempted historical reference. I cannot at this stage comment on
the historicity of Liu Wen-ching, but there never was a King Jui of Wu. Wu Jui, King of
Chang-sha, is however well known to history. He was a leader of the Yeh tribes of the
mid-Yangtse who distinguished himself through his constant loyalty to the founder of the
Han; Chang-sha was the kingdom he was granted as a reward for his services.19 Though
it is no more than a guess, I would suggest that one of the underlying reasons for the rise
of the city gods was the expansion of Chinese settlement into non-Chinese areas: walls
and moats (the literal meaning of cheng-huang) were needed to give a sense of security,
and even more than these material and visible symbols of reassurance, in a territory thick
with alien ghosts a spiritual bulwark against the psychic resentment of the land was an
even greater necessity. What better champion than a local leader who in his life had
demonstrated himself to be a true friend of the Chinese, especially one

17

CSLCY 25.4b.
CSLCY 25.7b8a. This use of the term ti-chu as a type of god rather than a specific god is not listed
in dictionaries (except in a Japanese context in the Engi-shiki of 927), but cf. its mention in the
passage dealt with below, at n. 43, in T. 21, p. 375b, and in the same volume, p. 466c in no. 1315
another one of the problematic group of texts mentioned below, n. 47. See also Yn-chi chi-chien
79.19a, TC, n. 1026, the precise date of which passage is also difficult to determine.
19
Pan Ku, Han Shu 34, Peking, 1962, 1894.
18

16

whose family line (according to the histories) had died out entirely, thus making him, as
it were a supernatural freelance, suitable for deification?20
To return once more to the Chin-so liu-chu yin, the next note on the cheng-huang
gods in the text deals with history on a grand scale. A passage on exorcism which
mentions shrines to gods of the earth is interrupted with the following information: in
high antiquity there were only gods of the heavens and of the earth (tien-shen, ti-chih,
Japanese jingi), and no she-miao; in middle antiquity there was the altar of the soil on the
left, of grain on the right, and still no she-miao; since the Han, Wei, Chin and Sung (note
the Southern outlook) there had been spirit officials of the altars of soil and grain; since
the Chi, Liang, Chen and Sui there had also been spirit officials of cities; spirit officials
of she-miao and cheng-huang now cooperated in maintaining otherworldly law and
order just like chou and hsien officials.21 Two references further on in the commentary
also support the view that the cheng-huang should be seen as a new force collaborating
with the she-miao.22 Now periodization of religious history is basic to Taoism: the Taoist
religion, after all, first appeared in the late Han as a self-declared new dispensation,23 and
far from believing in a single historical revelation, the notion of a series of revelations
seems, too, to have been present from the start.24 But what reminds one (as so often in
Taoism) of secular, bureaucratic parallels is the way here in which these changes are
treated largely as changes in nomenclature in our text: we are presented with a fragment
of an otherworldly Li-tai chih-kuan piao, a diachronic table of the bureaucracy in which
titles of posts change, but the overall structure is alleged to be more or less constant, even
if to some extent there is (as there was in the Tang bureaucracy) a cumulation of new
posts upon old.25 It is

20

See the preceding note, and for the significance of this information, pp. 25960 of Alvin P. Cohen,
Coercing the Rain Deities in Ancient China, HR, 17, 3, 4, 1980, 22465.
21
CSLCY 25.13ab. Cf. Johnson, The City-god cults of Tang and Sung China, 436, n. 219, though
given the untenably early date Johnson accepts for our text, he is forced to indulge in some special
pleading at this point.
22
CSLCY 25.14b, 15a.
23
The Yellow Turbans certainly proclaimed a new era at the time; the evidence for the Way of the
Celestial Masters is lacking, but they also looked back to their founding as a new start: see D.C.
Twitchett and M.A.N. Loewe, eds., The Cambridge History of China, vol. 1, Cambridge, 1986, 815
820, 8756; Chen Kuo-fu, Tao-tsang yan-liu kao, Peking, 1963, 311314.
24
See Anna Seidels study, La divinisation de Lao tseu dans le Taoisme des Han, Paris, 1969, for the
early notions of a series of revelations. Though later Taoism chose to be cautious over the appearance
of new Messiahs, the appearance of new texts (sometimes specifically labelled newly appeared, as
with those bestowed on Kou Chien-chih in the early fifth century) continued unabated.
25
For a brief characterization of this work (the product of a long historiographic tradition stressing the
continuities behind institutional change), see S.Y. Teng and K. Biggerstaff, An Annotated
Bibliography of Selected Chinese Reference Works, 3rd ed., Cambridge Mass., 1971, 200. For an
overview of institutional change during the early imperial period, see M.A.N. Loewe, Imperial China,
London, 1966, 150166. The relationship between imperial bureaucracies, mundane and supernatural,
in China has excited much comment; the latest treatment of the theme is Jean Levi, Les
Fonctionnaires Divins, Paris, 1989.
17

this plus a change response to religious change in the Chin-so liu-chu yin which led to
its early quotation in a much more conventional work,26 and which contrasts sharply with
the either/or tendency notes earlier in Confucianism.
At the same time our text is clearly not simply handing out a stock response:
throughout the work one gets a strong sense of an author struggling, as if for the first
time, with a problem of explanation. Subsequent early evidence for Taoist attitudes
towards the city gods becomes steadily more anticlimactic, though the next earliest I have
noticed, in the Hs Hsien Chuan of Shen Fen, a late ninth century or early tenth century
work, is a trifle startling, since in it a Taoist priest flicks nails into the eyes of a city god
statue to alleviate drought.27 This, however, cannot be taken as a sign of religious
antipathy, both because of what we know of the coercing of rain deities,28 and because
the hero of this episode is a holy man of the well-known Chinese outrageous variety,
much addicted to wine and womenthe sort of Taoist, in other words, from whom bad
behaviour is expected. Shen Fens contemporary, the great Taoist courtier and author Tu
Kuang-ting (850933), also counteracts any impression of early Taoist antipathy
towards city gods by including a brief nod in the direction of the lords of the she-miao
and cheng-huang in a mass for the dead preserved in his best-known collection of
rituals.29
Similar passing references to the city gods may be found in materials ascribed to
Tu preserved in another large collection, evidently of Southern Sung date, the Tao-men
ko-fa ta-chan-chi,30 in rituals devoted to combatting baleful stars,31 and securing
rainfall.32 His Sung editor, for his part, includes them in rituals for seeking heirs,33
invoking the Warrior God Chen-wu,34 expressing repentance,35

26

See Roman Malek, Das Chai-chieh-lu, Frankfurt am Main, 1985, 47, 86.
Shen Fen, Hs Hsien chuan, p. 1.13a, TC. no. 295; for the date of this work, cf. Y Chia-hsi, Ssuku ti-yao pien-cheng, Peking, 1958, 19XX, 1214, where the author is identified as having served
under the Tang.
28
See the study of Cohen cited in n. 20 above.
29
Tai-shang huang-lu chai-i 35.6b, TC. no. 507. On Tu, see F. Verellen, Du Guangting (850933):
Taoste de cour la fin de la Chine mdivale, Paris, 1989, and especially page 212 for an entry on
this ritual compilation.
30
Listed in Verellen, Du Guangting, 215. Note the reference, 63.5a, to the period 112730.
31
TMKFTCC 9.3a.
32
TMKFTCC 12.6a, 13.3a, 5b, 14.3a, 15.3a, 17.3a.
33
TMKFTCC 25.4b, 26.3a, 27.3a, 28.3a.
34
TMKFTCC 63.1b.
35
TMKFTCC 70.2b, 71.2b, 72.2b, 73.2b.
27

18

and achieving immortality.36 From other rituals, though, they are excluded. It is not,
however, possible to make much of this pattern of inclusion and exclusion, since it covers
not only the city gods but also other lesser local divinities: what one can say is that within
this source the city gods are firmly ensconced among the ranks of such lower-order
spiritual powers. There is indeed ample evidence from the Sung for the incorporation of
the city gods cults into the Taoist liturgy: one can point to pro forma petitions to
them37precursors of a whole Ming text devoted to their supplication38and even a
lengthy note on why the cheng-huang god in the capital could not (contrary to common
opinion) be held to have empire-wide responsibilities.39
This last discussion seems to have convinced Teng Ssu-y that the city gods could
not be Taoist in origin: if the Taoists had invented them, they would not have been so
divided over their beliefs concerning them.40 This argument appears not entirely sound
it could be used, for example, to prove that the eucharist was not in origin a Christian
ritualbut the conclusion reached is beyond dispute. Equally admirable are his
conclusions concerning the non-Buddhist origins of the city gods, though here he does
linger over the legend of the Buddhist warrior king Vairavana, Pi-sha-men, who is
depicted in China as a defender of Chinese city walls.41 This legend has intrigued others,
too, since it could be taken to indicate some connection with the city gods, but a recent
study by Yoritomi Motohiro reveals that it draws upon a number of separate sources, and
emerges too late to have itself prompted the cheng-huang cults.42 For reasons which will
become clear shortly, I am not even sure that it could have been promoted by Buddhist
monks as a deliberate alternative to the new gods, either.
For in the case of Buddhism again it is not possible to find any signs of religious
antipathy to the cheng-huang deities: in ritual, and in anecdote, the new gods simply turn
up and make themselves at homeor are made at home. The only problem is when. The
text which in all likelihood marks the earliest Buddhist occurrence of the new cult, the
Yen-lo wang kung-hsing fa tzu-ti, ascribed to

36

TMKFTCC 75.3b, 76.2b, 77.2b, 78.2bthe first reference is actually under Tus name, but the rest
are not attributed to him.
37
Tao-men ting-chih 2.19b etc., TC. no. 1214. On this text, see Judith Boltz, A Survey of Taoist
Literature: Tenth to Seventeenth Centuries, Berkeley, 1987, 5051.
38
Tai-shang Lao-chn cheng-huang kan-ying hsiao-tsai chi-fu miao-ching, TC. no. 1435. My
judgement on the date of this text is based on the administrative geography which it mentions.
39
Shang-ching ling-pao ta-fa, in 44 ch., TC. no. 1213, 27.25b26b. For this text, see Boltz, op. cit.,
4546.
40
Teng, Cheng-huang kao, 272: I may be characterising his remarks a little unfairly. Teng
provides a number of references to Sung ritual compilations of the type I have touched on more
briefly.
41
Teng, Cheng-huang kao, 269.
42
Yoritomi Motohiro, Chgoku Mikky no kenky, Tokyo, 1979, 147159.
19

Amoghavajra (Pu-kung, 705774),43 has defeated the best efforts of the most erudite
Japanese scholar Osabe Kazuo at achieving any precise dating for it.44 Its subject is the
abode of the dead, here organized according to a scheme already attested in the seventh
century in which the Buddhist King Yama ranks higher than the Lord of Mount Tai, who
ranks higher than the generals of the Five Ways in their government of the underworld.45
Yet the text, presumably, is later than Amoghavajra, since though that master himself
may conceivably have sponsored such a highly syncretic form of religion,46 it forms one
of a number of highly sinified works surveyed by Osabe which, far from having been part
of the canon in China, are never mentioned, and only show up in Japan, usually in fairly
late copies.47 But in many cases they cannot be Japanese forgeries: their terminology (as
here) relates distinctly to Chinese forms of religion the Japanese never practised. Despite
the lack of bibliographical supporting evidence, this text is perhaps late Tang, since (as
Osabe points out) we must provisionally assume that it antedates the development of a
new vision of hell associated with the cult of the Ten Kings, which rose to prominence in
the tenth century.48 Admittedly, such a judgement ignores the possibility of regional
variations in the speed of development of the Ten Kings system, and assumes that their
domination became total, though conversely the failure of Tsung-mi (780841), our chief
authority on Tang Buddhist ritual, to mention the city gods in the list of minor deities he
uses

43

Yen-lo wang kung-hsing fa tzu-ti, p. 375b, T.21, no. 1290.

Osabe Kazuo, T-S Mikkyshi Ronk, Kobe, 1982, 41-48.
45
See S.F. Teiser, The Medieval Chinese Ghost Festival, Princeton, 1988, 188.
46
As has been argued most recently by Charles Z. Orbach, Seeing Chen-yen Buddhism: Traditional
Scholarship and the Vajrayana in China, HR, 29,2, Nov. 1989, 87114; see also the earlier remarks
on page 642 of Iyanaga Nobumi, Rcits de la soumission de Mahevara par Trailokavijaya, daprs
les sources chinoises et japonaises, in M. Strickmann, ed., Tantric and Taoist Studies in Honour of
R.A. Stein, vol. 3, MCB, Brussels, 1985, 633745. The monograph of Yoritomi, cited above, n. 42,
also reinforces the point.
47
See Osabes listings on page 214 of T-S Mikky, and quotations, pp. 228231; note for
comparison his conclusion on no. 28 in his list, pp. 229230.
48
Osabe, T-S Mikky, page 50. Cf. Teiser, Ghost Festival, 1825, from whom a more detailed study
is forthcoming.
44

20

in his work may equally be due to his own regional or sectarian biases.49 There is room,
as they say, for further research on this point.50
The first appearance of a city god playing a Buddhist role in anecdotal literature,
on the other hand, coincides fairly exactly with Shen Fen and Tu Kuang-ting, for as
David Johnson notes, the collection including the tale, the Pao-ying lu, dates to the late
ninth century,51 or at the most pessimistic to the early tenth.52 Another tale in which a city
god appears as part of the otherworldly mechanism of the karmic system turns up in the
Sung Kao-seng chuan of 988, but appear to be set in about the same period as the Paoying lu.53 And even if the text discussed earlier attributed to Amoghavajra is actually the
product of some conservative out-group of an unusually late date, the city gods were
definitely incorporated into Buddhist ritual by round about the year 1000 of the Western
calendar, since they are included in a text by Tsun-shih (9461032).54
As in the case of Taoism, acceptance thereafter appears to have been total: they
have kept their place in Buddhist ritual into modern times,55 and Buddhist clergy seem
actively to have supported their cults. I have already referred to one Buddhist author of
the eleventh century who gives some information on the source of financial support for
one temple restoration, and this cleric, Wen-ying, tells us that he tried to secure support
for an inscription recording a miracle associated

49

Kamata Shigeo, Shmitsu kygaku no shisshiteki kenky, Tokyo, 1975, 512.

The questions raised by T. 21, no. 1290 and similar texts are at the moment virtually insoluble,
unless perhaps art historical evidence is used to break the impasse. We do find at Tun-huang false
attributions to Amoghavajra, dating perhaps to the early tenth century: see Tanaka Rysh, Tonk
Zensh bunken no kenky, Tokyo, 1983, 64448. But how late might such attributions occur?
Orbachs article (note 46 above) strongly contests the idea that interest in Amoghavajra waned after
the Tang; even such standard works as Kamio Katsuharu, Kittan bunkashi k, Dairen, 1937, 102106,
attest that an interest in Tantric Buddhism was still live and well under the Liao dynasty in the
eleventh century. Again, if this and other doubtful works attributed to Amoghavajra were from the
start (perhaps even from their genuine composition by Amoghavajra) only transmitted secretly, this
would explain their late appearance. By the same token, however, works transmitted because they
were in constant use for ritual purposes would have been more susceptible to interpolation designed to
update the religious practices involved to accord with shifts in the surrounding religious environment
than works enshrined in the imperially-sponsored canon, which was in theory tamper-proof. By far the
best discussion of the problem of authenticating late Tantric works is in Osabe, T-S Mikky, 150
181 (note especially his conclusions concerning Amoghavajra, pages 1778), but even this is only a
beginning.
51
Johnson, The City-god cults of Tang and Sung China, 446.
52
The author of the collection, Wang Ku, was an examination graduate (chin-shih) of 898, according
to the Tang tsai-tzu chuan, as quoted in Hs Sung, Teng-ko chi kao 24.20b, Taipei, 1972 reprint;
other sources agree in putting his career at this time.
53
Sung Kao-seng chuan 11, 771c772a in T. 50: the tale is dated circa 920.
54
Chih-sheng-kuang tao-chang nien-sung, page 980b, in T. 46.
55
Kamata Shigeo, Chgoku no Bukky girei, Tokyo, 1986, 700.
50

21

with this 1067 rebuilding.56 Kanai Noriyuki has noted in an early-fourteenth-century

gazetteer a reference to a monk actually organizing the construction of a cheng-huang
temple in the early twelfth century.57 Though I have not searched systematically in Sung
sources, I have not so far found like examples of Taoist support.
Nor, looking back, is it possible to see in any of the early references I have found any of
the sense of innovation detectable in the Chin-so liu-chu yin. That such a phase never
existed for Buddhism may in one sense be an argumentum ex silentio, given the
historiographical thinness of my evidence. But there is material to support the claim that
the absence of a Buddhist source similar to the Chin-so liu-chu yin is no accident. For a
reading of the Buddhist canon, as translated for Indian sources, reveals that the cult of the
cheng-huang gods can only be termed a Buddhist cult from the startnot, it should be
stressed again, a Buddhist cult in origin, but a Chinese cult for which there was a readymade place in Chinese Buddhismbecause Indians, too, worshipped city gods. Look into
any type of Buddhist source and you will find them: vinaya,58 Mahayana stra,59 Tantric
text,60 biography of the Buddha.61
Mostly they are just faces in a crowd of lesser gods, in the same sort of minor
league as the dryads and hamadryads of our Classical World, but in the last-named type
of materials, the biographies, they are sometimes assigned brief speaking parts. Typical is
a story later included in the Tang Buddhist encyclopaedia Fa-yan chu-lin, in which the
Buddha-to-be, having decided to forsake his fathers palace for a religious career by
leaving the city at the dead of night, is accosted by the god of the city wall (in another
version of this incident, the god of the gate in the wall),62 and is assured by him that over
this thirteen kalpa career as a cheng-shen he has seen all the past Buddhas behave in just
the same wayand here is a present from one of them.63
In the massive, and yet in China highly popular Hua-yen ching (that is, the
Avatasaka-stra), the role of the city-god, as Teng Ssu-y noticed (up to a

56

See n. 7, above.
Kanai Noriyuki, on page 402 of his Sdai no gsha to tochishin, Nakajima Satoshi sensei koki
kinen ronsh, vol. 1, Hamamatsu-shi, 1980, 385407.
58
Ssu-fen l 46, 911a, in T. 22: this vinaya became standard in China.
59
Lo-mo-chieh ching 3, 871c in T. 10: this is actually an early translation of part of the Gaavyha,
dealt with below.
60
Pu-ti-chang so-shuo i-tzu ting lun-wang ching, 193c, in T.19 (translated by Amoghavajra); in the
same volume, 225a, is the same passage as translated by Bodhiruci.
61
Ta chuang-yen lun ching 15, 344c in T. 4.
62
Hsiu-hsing pen-chi ching 2, 468a in T. 3; cf. also Te-kuang tai-tzu ching, 417c418a, in the same
volume.
63
Fa-yan chu-lin 35, 562 in T. 53.
57

22

point),64 is unusually prominent. Here the Buddha himself declares that in the distant
beginnings of his spiritual career in earlier incarnations, he too had spent some time as a
cheng-shen.65 Under the distinctive name chu-cheng-shen, gods in charge of cities, this
type of divinity is mentioned in many other places in the text, and in the section known as
the Gaavyha, usually dubbed a Buddhist equivalent to Pilgrims Progress, a specific
chu-cheng-shen named Ratnanetra delivers a brief homily to the young seeker Sudhana
on guarding the city walls of ones mind,66 a passage which in due course attracted a
considerable amount of Chinese commentary. The existence of the Gaavyha in
Sanskrit even enables us to establish the original terminology upon which the Chinese
translations are based: city of the mind is revealed (not very surprisingly) as
cittanagara;67 gods in charge of cities renders nagaradevat.68
Yet not all these are cheng-huang gods: throughout such translated Buddhist
sources the native term is either studiously avoided, or was perhaps simply unknown. The
closest one gets is in a text translated circa 400 A.D., which speaks of cheng-chih shen,
gods of city wall and pond (meaning moat?).69 But the existence of these miscellaneous
Indian deities, even if functionally different from cheng-huang gods (they do not in India
seem to connect with the world of the dead, for example) must nonetheless have made it
easier for Buddhists in China to accept the new cults as compatible with their own
religion. Or is this the right way to look at it?
If we review the admittedly somewhat slender evidence discussed so far once
more from our initial standpoint of reactions to the cheng-huang gods, we may now be
able to make one or two provisional and not particularly exciting generalizations. In
Buddhism, for example, the new gods take their place in the organization of the afterlife,
in line with Buddhisms primary religious concern

64

Teng, Cheng-huang kao, 26970: he seems to have been aware of some Hua-yen materials, but
does not systematically exploit the resources contained in the Buddhist canon.
65
Ta-fang-kuang Hua-yen ching (Tang version) 73, 400b in T. 10; cf. also the same text, 1, 2c; 4,
19b; 61, 330c; 75, 408b. For the popularity of this scripture, see Kamata Shigeo, Chgoku Kegon
shisshi no kenky, Tokyo, 1965, 1750.
66
Hua-yen ching 76, 413c414a. Cf. the English translation by Thomas Cleary, Entry into the Realm
of Reality: the Text, Boston and Shaftesbury, 1987, 3067. The name of the goddess is found in P.L.
Vaidya, ed., Gaavyhastra, Darbhanga, 1960, 339, 1.14. Here and below I am indebted to Dr. T.
Skorupski for checking the Sanskrit reference on my behalf.
67
Vaidya, op. cit.., 36, 1.12; 239, 1.20; 339, 1.16; etc. Passages such as these, too, excited
considerable comment within the Chinese Hua-yen school and among Japanese commentators also.
68
Vaidya, op. cit.., 36, 1.13; 278, 1.26; 339, 1.14: cf. Cleary, Entry into the Realm of Reality, 45-6;
254; 306, respectively, for a sample of such passages.
69
Mi-lo ta cheng-fo ching, 431c, in T. 14.
23

with the fate of the individual (or non-individual); Taoist reactions encompass a
somewhat broader spread of roles, including rain-maker, in line with Taoisms more
general aim of providing expertise on the occult. But is there, after all, still something we
can say about influence on the cults, at any rate in the case of Buddhism?
I would like to conclude by raising the possibility that Buddhism may have played
a part in promoting a sense of unity between diverse phenomena. I am aware that I touch
here upon one of the most sensitive topics in the study of Chinese religion.70 I am also
consciously recycling here an argument of Japanese origin, which I accept, concerning
the importance of Buddhism in providing a model of unity for Taoism, an argument that
not everyone would accept.71 But the notion that listings in well-known Buddhist texts of
a class of city gods, which (in the Hua-yen ching, at least) are given individual names and
(minimally) personalities but are usually subsumed into one lump, might have provided
the means whereby a diversity of local cults concerned with the protection of walled
communities came to be treated as expressions of a single ideathat does have its
appeal.72 Such classifications of gods are not unknown in earlier Chinese history: perhaps
mountain gods and river gods were dealt with both as individual cults and also as a
group even before the arrival of Buddhism; the rise of the cheng-huang cults as a group
would, I think, be quite conceivable without Buddhism.73

70

The problem is confronted e.g. by Robert Weller, Unities and Diversities in Chinese Religion,
Seattle, 1987, one of a number of recent studies reviewed by Catherine Bell, Religion and Chinese
Culture: Toward an Assessment of Popular Religion, HR, 29.1, Aug. 1989, 3557. From the
historians standpoint, I would suggest that what has been missing from discussion so far is an
awareness that as Chinese religion grew in complexity, it presented conceptual problems for the
Chinese themselves, and that a study of these historical attempts to cope with unity and diversity can
yield important clues to guide our own contemporary analytical efforts.
71
My argument is presented (in severely compressed form) in an essay, Religious Traditions in
Chinese Civilization: Buddhism and Taoism, in Paul S. Ropp, ed., The Heritage of China: Essays on
Chinese Civilization in Comparative Perspective, Berkeley, 1990.
72
There are certain hints that the cheng-huang cults became over time far more standardized than
they were at the start. Thus one late Ming author, commenting on inscription no. 1 on Johnsons list
(see n. 11 above), notes that it implies a temple outside the city walls, which was only a highly
localized phenomenon in some large villages to the west of Sian in his day: see Chao Han, Shih-mo
chien-hua (Chih-pu-tsu-chai tsung-shu ed.) 4.8a. Wen-yings temple (see n. 7 above) also appears to
have been some way from the community it protected.
73
In favour of this supposition, it must be pointed out that in the earliest texts used above, Buddhist
and Taoist, cheng-huang appear primarily in long lists of such class divinities. But (to put the
problem one stage further back) long lists of class divinities may themselves be an innovation
brought in by Buddhist translationsthey are not conspicuously present in early Chinese literature,
though perhaps examples could be found.
24

But cults associated with mountains and rivers go back to the dawn of time in China; as
far as we are aware on current evidence cults centred on cities as such do not.74 I still
believe that the cheng-huang gods were an East Asian development, and one probably
intimately linked to the tensions between an expanding China and the non-Chinese south.
In coming to terms with this novelty, however, the Chinese would have found in the
Indian experience a ready, if imprecise analogy for what was before them.
For, after all, Buddhism had once been in India a new religion itself, a religion
in opposition to existing beliefs. In cutting down to size its opponents and assigning
them to the other ranks of figures worshipped it had evolved a technique which was
suitable in theory to any time and any place. True, observers of the cheng-huang cults in
China such as those cited at the start of this study suggest that the city gods obstinately
refused to play the extremely minor role assigned to them by Buddhist textsTaoism
seems to have been prepared to be more realistic in that respectbut as a normative
statement of religious priorities the Buddhist writings we have examined cannot be
faulted. Some further cross-cultural comparisons of Buddhisms relations with other
religions might, however, prove illuminating; one suspects, for example, that Buddhism
and Shint in Japan interrelated in a somewhat different fashion. And on this point the
forthcoming 1990 Jordan lectures will doubtless provide much food for thought.

74

Art-historical evidence has not been used in this essay: as noted above (n. 50) it remains our best
hope of locating problematic textual materials in their correct context, but the constant advance of
Chinese archaeology may yet clarify many more aspects of the appearance of the cheng-huang
cultsperhaps, indeed, it has already done so. Any reconsideration of the problem will, however,
surely have to take into account the materials presented here. They, for their part, have largely been
located by means of the existing indexes and concordances to texts of the period covered, rather than a
comprehensive search of all conceivable sources: doubtless a number of passages of potential value
have been missed.
25

The Five Points and the Origins of the Buddhist Schools

L.S. Cousins

1. The historical background

The history of Buddhism in India between the death of the founder and the beginning of the
Suga period is remarkably little known. Apart from the account of the first two communal
recitations (sagti) or Councils and a certain amount of information relating to the reign of the
Emperor Aoka, we are largely dependent upon traditional Buddhist accounts of the origin of the
eighteen schools. As Frauwallner has commented: These accounts are late, uncertain and
contradictory, and cannot be relied upon blindly.1 The number eighteen is probably symbolic in
nature and should perhaps not be taken too seriously. Nevertheless it is clear that there is a
generally accepted tradition that in the course of the second and third centuries after the Buddhas
mahparinibbna the sagha divided into a number of teachers lineages (cariyakula)2 or
doctrines (vda;3 cariyavda4) or fraternities (nikya).5 At a later date these terms became in
effect synonymous, but this may well not have been the case earlier.
In the early centuries AD the Sinhalese commentators and chroniclers assembled the data
available to them and constructed a consistent chronology of the early history of Buddhism and of
the kings of Magadha. The absolute chronology which they created has not proven acceptable as
it places the reigns of the Mauryan Emperors Candragupta and Aoka more than sixty years too
early. However, the general account they provide has been reconciled with other data, mainly
from the Puras, to create a widely accepted chronological framework for the history of India
during this period. For our purposes, the essential points of this account are that the accession of
Aoka occurs in 218 BE and all eighteen schools were already

E. Frauwallner, The Earliest Vinaya and the Beginnings of Buddhist Literature, Rome, 1956, 5.
e.g. Kv-a 23.
3
e.g. Dp V, 51.
4
e.g. Kv-a 3.
5
ibid.
2

27

in existence by 200 BE.6 This we will call the long chronology, to use a convenient term of
Lamottes.7
A number of works of Sarvstivdin origin (and later works influenced by them) date the
accession of Aoka to 100 BE. In fact it seems clear that during the early centuries AD the
Vaibhika commentators attempted to create a chronological framework for the early history,
probably using a version of the Aoka legend as their starting point. Of course, many of the
Sanskrit texts simply give isolated statements, which could not be called a chronology. However,
we do possess a work on the doctrines of the eighteen schools which does go some way towards
achieving a unified framework. This is a treatise attributed to Vasumitra, extant in three Chinese
and one Tibetan translation. In fact the verses naming the author as the bodhisattva Vasumitra
are absent from the earliest Chinese translation (beginning of the fifth century AD) and were
clearly added in India at a later date. The first translation would hardly have failed to mention his
name, if its attribution to one of the famous figures of Sarvstivdin history bearing the name of
Vasumitra had been known at the time. Probably it is a work of the third or fourth centuries AD.
For our purposes the essential points to note are that for Pseudo-Vasumitra divisions begin during
the reign of Aoka in the second century BE.8 By the end of the second century the
Mahsghikas had eight new branches but the Sthaviras were still undivided. During the course
of the third century BE nine new branches of the Sthaviras emerge and the Sautrntikas arrive in
the fourth century BE. This we will call the short chronology.
The difference between the two chronologies is rather considerable. According to the
long chronology all eighteen schools existed eighteen years before the accession of Aoka.
According to the short chronology divisions among the Sthaviras do not begin until 100 years
after the accession of Aoka. We do not know whether other major schools than the Theravdins
and the Sarvstivdins had created their own chronologies. The riputraparipcch, a
Mahsghika work translated into Chinese between AD 317 and 420, follows more or less the
same chronology as Pseudo-Vasumitra.9 Bhavya preserves various traditions which may be old,
but it seems dangerous to rely on material only collected as late as the sixth century AD.

Dp V, 53; Kv-a 3.
E. Lamotte, Histoire du Bouddhisme indien, Louvain, 1958, 1415.
8
Later translations mention 116 BE, but it seems clear that originally the work, like the Dpavasa,
specified only the century. See A. Bareau, Trois traits sur les sectes bouddhiques attribus
Vasumitra, JA, 1954, 236ff.
9
E. Lamotte, Histoire du Bouddhisme indien, 310; 5878; A. Bareau, Les sectes bouddhiques du Petit
Vhicule, Saigon, 1955, 17.
7

28

A number of scholars have expressed doubts as to whether we can still accept a version
of the long chronology as authoritative.10 At present it does not seem possible to decide the
question. Here only a few of the relevant issues can be addressed, since our concern is to examine
the nature of the earliest divisions in the Buddhist community and of the earliest schools of
thought. However, some points cannot be avoided entirely. One of our earliest sources relates the
first schism of all to the second communal recitationusually known as the Council of Vaili.

2. The Council of Vaili

An account of the first two communal recitations is contained in all surviving recensions of the
Vinayapiaka. We possess one version in Pli, parts of two in Sanskrit, one in Tibetan and five in
Chinese. There is also a summary of the Vinaya of the Haimavata school in Chinese.11 This
material has been conveniently collected in French by Hofinger.12
The date of the events described is given as 100 BE in the Pli Vinaya and in the Vinayas
of the Mahsakas, Dharmaguptakas and Haimavatas. These schools are closely related as
regards their Vinayas.13 The Vinayas of the Sarvstivdins and Mlasarvstivdins give the date
as 110 BE. No doubt this divergence is due to a wish to reconcile the account of the second
communal recitation with the tradition found in Sarvstivdin works that the accession of Aoka
took place in 100 BE.14 The rather brief account in the Mahsghika Vinaya gives no date at all.
In any case it seems likely that the figure of 100 years was known in the last centuries

10

E.J. Thomas, Theravdin and Sarvstivdin Dates of the Nirva, B.C. Law Volume, Part II,
Poona, 1946, replied to by J. Filliozat, Les deux Aoka et les conciles bouddhiques, JA, 1948, 189
95; E. Lamotte, Histoire du Bouddhisme indien, 1315, however, adopts the long chronology as a
working hypothesis; H. Becherts several recent articles: The Date of the Buddha Reconsidered, IT,
1982, 2936, A Remark on the Problem of the Date of Mahvra, IT, 1983, 18790, Die Lebenszeit
des Buddhadas lteste fest stehende Datum der indischen Geschichte?, Gttingen, 1985, Remarks
on the Date of the Historical Buddha, Buddhist Studies, 1988, 97117.
11
According to E. Mayeda, Japanese Studies on the Schools of the Chinese gamas, in H. Bechert,
Zur Schulzugehrigkeit von Werken der Hnayna-Literatur, Gttingen, 1985, 101, most Japanese
scholars take this to be a Dharmaguptaka work. A. Bareau, Les sectes bouddhiques du Petit Vhicule,
201ff. suggests Kyapya which seems plausible.
12
M. Hofinger, tude sur le concile de Vaili, Louvain, 1946, usefully criticized by P. Demiville,
propos du concile de Vaili, TP, 1951, 23996.
13
M. Hofinger, op. cit., 167; E. Frauwallner, The Earliest Vinaya and the Beginnings of Buddhist
Literature, 55.
14
The Vinaya of the Mlasarvstivdins is well known to have been revised at a late date, while the
portion of the Sarvstivdin Vinaya which contains the account of the councils is an addition
translated at a later timeP. Demiville, propos du Concile de Vaili, 242ff. See also P.H.L.
Eggermont, New Notes on Aoka and his Successors, II, 88, and H. Bechert, Die Lebenszeit des
Buddha, 160.
29

BC, whether or not it is original. This would still be far earlier than most of our historical
information for the early period.
One hundred years is a round figure, and was almost certainly not intended as an exact
count of years. It is more interesting to examine the accounts of the event to see what they tell us
as to its likely dating. What is immediately striking is the paucity of claims to direct connection
with Buddha.15 Yet even as late as 60 BE there would have been monks in their eighties who
would have received upasampad in the lifetime of the Buddha (even perhaps some in their
seventies who were novices at a young age). Given the emphasis upon seniority in the sagha,
such monks would have played a major role (ceremonially even if not in fact) and their
connection with the Buddha would have been mentioned in all extant accounts. They are not
mentioned. We can assume therefore that the second communal recitation did not take place
much before about 70 BE.
On the other hand every account we have emphasizes the connection with nanda
(except the Mahsghika).16 The very brief Mahsghika account is however one of the few to
claim a direct relationship with the Buddha. At the very least it seems likely that in the original
version the presiding monk (very probably the oldest living monk)17 was specifically claimed to
have been a pupil of nanda. No early tradition survives as to the date of the death of nanda,
but it seems reasonable to suppose that he might have lived until around 20 BE.18 In this kind of
context being a pupil of nanda does not necessarily involve a long period of contact. In his old
age nanda would no doubt have been the head of a large group of monks and even the pupils of
his pupils would have had nanda as their nominal teacher so long as nanda was still alive.
At the traditional date (taken literally) of 100 BE it would just about be possible for the
most senior monk alive to be reckoned a pupil of nandahe would have to be an active
centenarian. A date ten or so years earlier would be more likely. In the form in which we have the
tradition, however, it is quite impossiblea whole group of active centenarians is not believable!
A group of active octogenarians is certainly possiblewe are after all dealing with a group of
elders selected precisely because of their age.19

15

M. Hofinger, op. cit., 26, 146, 147 and also the list of years of upasampad on page 124. Only the
Mahsaka and Mahsghika accounts in fact make such a claim.
16
M. Hofinger, op. cit., 27, 48, 50, 51, 57, 80, 92, 93, 99, 101, 133, 139, 140, 143.
17
Pahaviy saghatherosee M. Hofinger, op. cit., 9093.
18
According to Th 103943 nanda attended the Buddha for 25 years. He could not, therefore, have
been less than 45 years old at the time of the parnibbna.
19
It might be argued that life expectancy would have been lower at the time. However, we are dealing
with a group of individuals who are teetotal, non-smoking and celibate. They would have had plenty
of exercise and would usually be regarded as noncombatants in situations of conflict. Data on life
expectancy from Egypt in the early centuries AD suggest a 50% mortality rate for each decade of life
after adolescence, but this would be for the general population. See N. Lewis, Life in Egypt under
Roman Rule, 54. Even the later Dp IV 50, 52; V 23 claim that they had all seen the Tathgata is not
entirely ruled out. A small child could well have been taken to see the Tathgata at a very young age
and told about the event when it was older. As late as 80 BE the oldest monk alive would very likely
have some such memory.
30

What emerges from this is that a date of around 7080 BE is implied by the accounts as
we have them.20 Two further points should be noted. Firstly, the early traditions do not mention
the name of the king, presumably because it was of no interest and because he played no special
role in these events. Secondly, all the early accounts (including that of the Mahsghikas) leave
us to understand that the decisions taken were accepted by all parties.

3. The First Schism

The earliest accounts we have of the first schism in the Buddhist order are quite late. Even by the
short chronology we are speaking of sources between four and six centuries subsequent to the
event. By the long chronology we could be dealing with sources no earlier than eight centuries
after. The earliest source is possibly the Mahvibh, which is posterior to Kanika in date.21
However, the relevant passage is absent from the earliest translation into Chinese of this work.22
It could therefore be a later addition made in India. This account claims that the first schism was
the result of doctrinal controversies over the Five Points advanced by a monk named
Mahdeva.23 Let us note that Mahdeva is not named in this context in any other early source and
is therefore not certainly named before the fifth century ADnearly a thousand years later (by
the long chronology)!
Pseudo-Vasumitra, also a Sarvstivdin source, likewise attributes the schism to doctrinal
disputes over Five Points. The earliest Chinese translation refers to three monks named Nga,
Pratyaya (?) and Bahuruta. The Tibetan translation is similar. The two later Chinese translations
refer to four groups of monks.24 This is clearly related to a later passage from a work attributed to
Bhavya (sixth or seventh century) which attributes the schism to a worthy monk (unnamed or
named Bhadraka), subsequently supported by two learned (bahuruta) Elders named

20

This line of thought was first suggested to me by Richard Gombrich, but my conclusions differ
slightly from his. See R. Gombrich, The History of Early Buddhism: Major Advances since 1950,
Indological Studies and South Asian Bibliographya Conference, Calcutta, 1986, 17.
21
E. Lamotte, Le Trait de la Grande Vertu de Sagesse, Louvain, 1944, 111n.
22
P. Demiville, propos du concile de Vaili, 263n.
23
E. Lamotte, Histoire du Bouddhisme indien, 303ff.
24
A. Bareau, Trois traits sur les sectes bouddhiques attribus Vasumitra, Bhavya et Vintadeva,
1954, 236; Les premiers conciles bouddhiques, 98ff. See also E. Lamottes Histoire du Bouddhisme
indien, 302, and E. Frauwallner, Die buddhistischen Konzile, 243ff
31

Nga(sena) and Sthiramati (according to Bu-ston Valguka).25 Trantha infers from the contents
of the subsequent list of the propositions attributed to the different schools that this is a tradition
of the Sammitya school.
The same source (quoted by Bhavya) dates these events to 137 BE under the kings Nanda
and Mahpadma and mentions that the work of the Elder Vtsputra took place in 200 BE. This
date for the origins of the Pudgalavda is too late in terms of the long chronology, but reasonably
compatible with the short chronology which dates the beginning of divisions among the Sthaviras
to 200 BE. The first date is more in line with the long chronology. Probably the Sammityas had
their own chronology.
By contrast the Sinhalese tradition knows nothing of a doctrinal cause for the first
schism. The oldest source is the Dpavasa which probably dates from immediately after the
reign of Mahsena when its account ends. This would be early fourth century AD.26 It traces the
origin of the schism to the defeated party at the second communal recitation and is followed in
this by later Sinhalese chronicles.27 Noticeably, however, Buddhaghosa does not give an account
of the origin of the eighteen schools in the Samantapsdik. The commentary to the Kathvatthu
does.28 Its account is closely related to that in the Mahvasa, but also quotes the Dpavasa in
full. This strongly suggests that no account of the eighteen schools was preserved in the
commentarial tradition of the Mahvihra.
This can also be inferred from the Dpavasa. The first part of Chapter V is given a
separate title cariyavda. It contains the account of the schools preceded by an account of the
first two communal recitations or dhamma recensions (sagaha). Since Chapter IV had already
given an account of these it is obvious that the Dpavasa is drawing on a second older source,
presumably in Sinhala Prakrit. We can go further than this. That older source has clearly taken a
list of schools of northern origin and added to it an introduction giving an account of the two
communal recitations based on the Mahvihra commentarial tradition. That it is a list of northern
origin emerges clearly from its close relation to the lists given

25

A. Bareau, Trois traits sur les sectes bouddhiques attribus Vasumitra, Bhavya et Vintadeva,
1956, 172; C. Vogel, Bu-ston on the Schism of the Buddhist Church and on the Doctrinal Tendencies
of Buddhist Scriptures, in H. Bechert, Zur Schulzugehrigkeit von Werken der Hnayna-Literatur,
Calcutta, 1984, 104.
26
AD 274302 (G.C. Mendis, The Chronology of the Early Pli Chronicles of Ceylon, UCR, 1947,
54). Mendis, following Paranavitana, rejects the notion that an era based on 483 BC was known in
ancient Ceylon. H. Bechert, The Date of the Buddha Reconsidered, 32, agrees but R. Gombrich,
Theravda Buddhism, London, 1988, 141n., prefers to retain the traditional dating of Wickremasinghe
(followed by Geiger). For Mahsena this would give dates of 334361/2.
27
Dp V 30 ff; Mhv V 34; Mhbv 96.
28
Kv-a 25.
32

by Pseudo-Vasumitra and the riputraparipcch.29 In fact it is possible to infer that it derives

from a Sarvstivdin original, probably mediated by a Mahsaka source.
The reason this can be inferred is that the first schism in the Theravda is attributed to the
Mahsakas from whom the other divisions descend. This is the position where one would
expect the Sarvstivdins who are found conversely in the position where one would expect the
Mahsakas (i.e. in close connection with the Dharmaguptakas). The list gives details of minor
Sarvstivdin branches such as the Suttavdins and clearly lacked information on the later
Mahsghika schools of Amarvati and Ngrjunikoa. On the other hand the Sinhalese were
well aware of the Andhakas. Their views are often referred to in the commentary to the
Kathvatthu. There is inscriptional evidence of the presence of the Sinhalese school at
Ngrjunikoa in the third century AD.30 One of Buddhaghosas sources is an
Andhakahakath.31
In these circumstances it is easy to understand why the list of schools given in
Kathvatthu-ahakath does not relate very well to the attributions given in the body of that very
text. Not surprisingly, therefore, the Ceylon texts add a further list of six Indian schools.32 These
do relate to the Kathvatthu and are obviously based upon the Mahvihra commentary to that
work. We may note the mention of schools such as the Rjagirikas and the Siddhathikas, hardly
mentioned in Indian literature but known from inscriptions at Amarvati. Even more suggestive is
the presence of the otherwise unknown Vjiriyas.33 It is not then surprising that Kathvatthu
Commentary often feels the need to add the word etarahi nowadays when it attributes particular
views to particular schools.34
Like the Ceylon tradition, the eclectic Sriputraparipcch gives a list of the eighteen
schools of northern origin. It too knows nothing of a first schism due to discussion of doctrinal
points. Neither, however, does it describe the origin of the Mahsghikas as deriving from the
defeated party at the second communal recitation. Rather it sees the Mahsghikas as the
conservative party which has

29

A. Bareau, Les sectes bouddhiques du Petit Vhicule, 16ff.

EI, XX, 22.
31
E.W. Adikaram, Early History of Buddhism in Ceylon, Colombo, 1946, 12; K.R. Norman, Pli
Literature, Wiesbaden, 1983, 1212.
32
Dp V 54; Mhv V 1213; Kv-a 5; Mhbv 97; cf. Kv-a 52.
33
Probably the later term Vetullaka has been substituted for them in the extant version of Kv-a, just as
the term Vetulyavda (Mhv XXXVI, 41) replaces the earlier Vitaavda (Dp XXII 4344).
34
Etarahi occurs throughout vaggas 2 and 3, predominates in vaggas 1 and 4 and peters out in vagga
5. Apart from one occurrence in the eighth vagga it does not occur again except in vaggas 17 and 18
where it is always applied to the Vetullakas. This may be because one is intended to take it as read
after the first few vaggas. Alternatively, it is possible that the original information available for these
did not make sense and so the commentator has substituted a reference to the contemporary situation.
30

33

preserved the original Vinaya unchanged against reformist efforts to create a reorganized and
stricter version.35 Like the Dpavasa it sees the origin of the name partly in a council where the
Mahsghikas were in the majority and their opponents included many senior monks. This must
however be largely a myth based upon a folk etymology. Clearly the Mahsghikas are in fact a
school claiming to follow the Vinaya of the original, undivided sagha, i.e. the mahsagha.
Similarly the theravda is simply the traditional teaching, i.e. the original teaching before it came
to be divided into schools of thought.36 The Dpavasa makes this clear when it explicitly
identifies the term theravda with the term aggavda in the sense of primal teaching.37
We have then two accounts of the origins of the first schism. The first is of Sarvstivdin
origin. Known from two sources of around the third and fourth centuries AD and in many later
sources based on these, it attributes the origin to doctrinal disputes over the Five Points. The
second is of Theravdin and Mahsghika origin. Known from two sources of around the third
and fourth centuries AD, and in many later sources based on these, it attributes the origin to
Vinaya issues. It is obviously important to examine carefully the evidence for the content of the
doctrinal disputes. As we shall see, it is very much earlier in date than the evidence for the
eighteen schools.

4. The Five Points

The most detailed account we have of the Five Points is contained in a canonical Pli text, the
Kathvatthu. Traditionally this work is attributed to Moggaliputta Tissa in the reign of Aoka, i.e.
the latter part of the third century BC. Although some scholars have supported the traditional
view, it is in fact clear that it is not a unitary work in the form in which we have it.38
If the authenticity of the Ceylon tradition that the Canon was closed in the first century
BC is accepted, then even the latest portions would not be subsequent to the first century BC.
This cannot in any case be far wrong. The Kathvatthu on the one hand contains arguments
against some Mahynist or proto-Mahynist notions and on the other clearly does not know the
developed Mahyna. A good example would be the assertion in one of the final sections of the
Kathvatthu that Buddhas stand in all directions.39 The supporter of this view denies that they are
in any of the recognized heaven realms but is not able to name any such Buddhas

35

E. Lamotte, Histoire du Bouddhisme indien, 189.

So M i 1645.
37
Dp IV 13; V 14.
38
K.R. Norman, Pli Literature, 1035; E. Frauwallner, Abhidharma-Studien IV, WZKS, 1972,
124.
39
Kv 608609.
36

34

when challenged to do so. Such an argument would not have been possible once the developed
Mahyna literature was known.
We can in any case be certain of an early date for the oldest portions of the Kathvatthu.
The first vagga (known as the mahvagga) discusses mainly but not exclusively the views of the
person and of sabbam atthi all exists; it contains a number of anomalous linguistic forms.40
These are not quite absent in the remaining vaggas but relatively few. Norman has convincingly
established that these cannot be due to influence from Sinhala Prakrit but must be of North Indian
origin.41 He has also suggested that there was originally a dialect difference between the two
speakers in the framework of the puggalakath (the first portion of the first vagga).
This gains support from the fact that a canonical Sarvstivdin abhidharma work, the
Vijnakya, devotes its first two chapters to defending the doctrine of sarvam asti and criticizing
the notion of the pudgalathe same two topics that we find in the mahvagga but in reverse
order.42 In the first chapter the opponent of sarvam asti is named as Maudgalyyana (Mou-lien).
As was pointed out by La Valle Poussin, this must refer to Moggaliputta Tissa, the author of the
Kathvatthu.43 The earliest portion of the Kathvatthu is then likely to date from the third century
BC or very soon thereafter.
It is worth noting at this point that this suggests a three-way split. Party A would oppose
both the puggala and sabbam atthi. Led by Moggaliputta they would be Vibhajyavdins and
ancestors of the Ceylon tradition among others. Party B espouses sarvam asti and opposes the
doctrine of the person, preferring its own teaching referred to by the Vijnakya as
nyatvda. They would be the ancestors of the Sarvstivda. Party C would be the
Pudgalavdins who presumably rejected the doctrine of sarvam asti. This three-way split gains
some support from a Pli commentarial passage which treats puggalavda and suatavda as
extremes to be avoided.44 In any case it is not clear whether these were yet distinct fraternities
(nikya) or merely schools of opinion. Nor is it clear what the relationship of these three schools
would be to the Mahsghikas.

40

Mgadhisms outside the puggalakath are particularly prominent at Kv 119-120 and 159-162 i.e. in
discussions related to sabbam atthi.
41
K.R. Norman, Mgadhisms in the Kathvatthu; also K.R. Norman, Pli and the language of the
heretics; cf. H. Bechert, ber Singhalesisches im Plikanon, 7175.
42
Louis de La Valle Poussin, La controverse du temps et du pudgala dans le Vijnakya, EA,
1925; F. Watanabe, Philosophy and its Development in the Nikyas and Abhidhamma, Delhi, 1983,
174ff., and next.
43
Koa, (= LAbhidharmakoa de Vasubandhu, tr. L. de La Valle Poussin), I, xxxiv; so also E.
Frauwallner, Die buddhistischen Konzile, ZDMG, 1952, 258.
44
Mp ii, 30910.
35

The very next section of the Kathvatthu deals precisely with the Five Points.45 This
portion of the text must also be quite early. It seems to represent a genuine debate with a real
opponent. The tone of it is still very similar to the mahvagga. It is probably part of the original
core of the text. Even if not, it cannot plausibly be dated later than the second century BC.
It is a matter of some surprise that most scholars have in fact given more weight to much
later accounts than to the actual content of the Kathvatthu itself. Let us note that by the short
chronology the relevant portions of the text would be close in time to the original disputes. Even
by the long chronology they would only be a century or so later. This contrasts sharply with
sources belonging to the commentarial period some five centuries later. Moreover, such sources
mostly represent a genre of literature which handed down supposed views of different schools in
short statements. Out of context in this way they are subject to error and reinterpretation. In some
instances it is quite clear that this has been the case. Such works do not constitute a good source
for the understanding of controversial points. Wherever possible, these must be understood in
their original context, that is to say in the actual abhidhamma literature itself.
It is by no means clear that most of the views we are given as sectarian views were ever
the positions of clearly defined schools. Many of them are surely constructed dilemmas, intended
as debating points to sharpen understanding of the issues. They could never have been the cause
of serious sectarian division. It is much more probable that they, like much else in the canonical
abhidhamma, are simply the distant ancestors of the dialectic of the Mdhyamikas.

5. The Five Points in the Kathvatthu

The thing that stands out most clearly about the treatment of this subject in the Kathvatthu is that
it is closely related to the earlier discussion as to whether an arahat can fall away. The same
structure is applied to each of the first four points as is applied in the earlier discussion. The
parallel is so close that it is difficult to doubt that they are part of one and the same discourse.46
The view that an arahat

45

First identified by La Valle Poussin, The Five Points of Mahdeva and the Kathvatthu, JRAS,
1910, 41323. See also on the Five Points: P. Demiville, Les versions chinoises du
Milindapaha, BEFEO, 1924, 60-62; Lorigine des sectes bouddhiques daprs Paramrtha, MCB,
1932, 3040; propos du Concile de Vaili, TP, 1951, 262ff.; E. Lamotte, Buddhist Controversy
over the Five Propositions, IHQ, 1956, 14862; J. Nattier and C. Prebish, Mahsaghika Origins,
HR, 250257; A. Bareau, Les controverses relatives la nature de larhant, IIJ, 1957, 24150.
46
Compare even Kv 6970 with 195 for the fifth point but mainly Kv 79ff. with 168ff., 175, 182,
189ff. (ten kilesas); 8586 with 169, 1756, 182, 190 (bodhipakkhiya-dhammas, followed by a stock
phrase on the arahat); 86ff. with 1702, 1768, 1824, 1902 (samayavimutta/asamayavimutta
parallelled by sadhammakusala/paradhammakusala); the parallelism continues with citations from
suttanta. This structural similarity is badly obscured by the translation.

36

can fall away is standard in the Sarvstivda and the orthodox Vaibhika position on the subject
is recounted at length by Vasubandhu in the Abhidharmakoa.47 The Kathvatthu is clearly
criticizing a very similar position, i.e. one in which the arahat, never-returner and once-returner
can fall away, but the stream-enterer cannot. The opponent in the Kathvatthu and the Vaibhika
both support their case by reference to the obscure distinction between the samayavimutta and the
asamayavimutta.
The context in which we should see the Five Points is then that of the abhidhamma
debates which refine the interpretation of some of the more recondite points of suttanta teaching.
We shall see that such a context gives little support to notions which see the Five Points as
involving some kind of downgrading of the arahat as against a Buddha. This is not the issue. If
there is a downgrading, it is rather a devaluing of the arahat who has not developed the abhi.

6. The arahat has doubt

The simplest of the Five Points to understand is certainly the proposition that the arahat has
doubt. The first thing to notice is how remarkable this proposition is. It is a frequent declaration
of the suttanta literature that the stream-enterer has overcome doubt. So basic is this notion that
the statement that an arahat has doubt must be intended to startle. In fact when the argument is
examined in detail it is clear that it has been carefully constructed in order to generate a
challenging proposition.
In the first place, the word used for doubt is kakh. Now this is just slightly less
specialized in its usage in the earlier literature than the more technical vicikicch. It is
immediately agreed by both parties that the arahat does not have either vicikicch or kakh in
the technical sense of doubt as to Teacher, Dhamma, Sagha, etc. It is equally agreed by both
parties that an arahat may be in doubt as to name and family, as to right and wrong roads and as
to ownership of grass, wood and trees, but cannot be in doubt as to the four fruitions (phala). In
this restricted sense the proposition cannot really be disputed; so an initially counter-intuitive
thesis achieves the aim of both stimulating the hearer and sharpening the understanding. Clearly
all that is at issue is at most a terminological question, if

47

Koa, VI, 5665, etc.; cf. A. Bareau, Les controverses relatives la nature de larhant, 244; it
does not seem that this can be a Mahsghika view since they appear to have held the reverse thesis
that it is the stream-winner and not the arahat who can fall away. It is just possible that Pudgalavdins
could be meant here.
37

that. It is significant that the Satyasiddhistra gives a parallel account of the nature of doubt.48
Of special interest is the distinction made between an arahat skilled in his own dhamma
(sadhammakusala) and an arahat skilled in paradhamma. Only the latter is free from doubt in
both senses. The commentary is probably right to equate this to the distinction between
pavimutta and ubhatobhgavimutta. In this context that is equivalent to the distinction
between an arahat without higher knowledges (abhi) and one who has developed such
abilities. Interestingly this is not a standard term in the Pli abhidhamma and appears to be drawn
from the terminology of the opponent.

7. The arahat has ignorance

Hardly less surprising is the proposition that an arahat has ignorance. Here again a slightly less
specific termaa is used for ignorance rather than the more usual technical termavijj,
but the conclusions are practically identical. In fact the whole course of the discussion is on the
same lines as in the case of doubt.
A number of scholars have followed the Mahvibh in interpreting this as referring to
unafflicted (aklia) ignorance.49 It is perhaps worth noting that this, if correct, would situate the
discussion very much in the context of the Sarvstivdin tradition. Such a terminology is absent
from the Pli abhidhamma literature. Of course the substantial point is very similar. However, the
Jnaprasthna appears to have understood that an arahat could be ignorant as to his own
liberation.50

8. Paravitra

This is the fourth proposition in all the extant lists. Paravitra51 can mean:
A.
B.
C.
D.

induction of comprehension by others;

induction of investigation by others;
being made to overcome by others;
being made to complete by others.52

48

N.A. Sastri, Satyasiddhistra of Harivarman, Baroda, 1978, II, 288ff. Ki-tsang gives a similar
interpretation: P. 32, but the Jnaprasthna appears to apply it to doubt on the part of arahats as to
their own liberation (ibid.).
49
P. Demiville, op. cit., 32n.; E. Lamotte, Buddhist Controversy over the Five Propositions, 148;
Nattier and Prebish, op. cit., 253.
50
P. Demiville, op. cit., 35 n.
51
One MS has parivitra. Some such reading is probably the source of Bhavyas interpretation: la
connaissance parfaiteA. Bareau, Trois traits sur les sectes bouddhiques attribus Vasumitra,
Bhavya et Vintadeva, JA, 1956, 173.
52
See PTC s.v. traa and treti as well as Pli-English Dictionary, s.v. vitarati. Some of these senses
are more plausible in the abhidhamma context. Later interpreters have tended to take the primary
Sanskrit meaning of crossing over; cp. also ti kakh D ii 276; 279; 2813.
38

One suspects that a deliberate wordplay of the kind so frequent in the

Paisambhidmagga is intended.53 The Kathvatthu seems to take it in the first two senses. The
context54 suggests sense C which recalls the notion of kakhvitaraa overcoming doubt.55 It is
clear that the variations in the translations of Pseudo-Vasumitra, etc. are simply the different
options. The Jnaprasthna probably had the same term as the Kathvatthu.56
Again we have a superficially startling notion. The whole point of being an arahat is to
have an independent knowledge of truth such that no assistance would be required from others.
Note that this is the first point raised in the Kathvatthu and the opponent immediately concedes
that an arahat is not dependent on another and does not lack wisdom in the sense of knowledge of
the Buddhist path.
In fact each of the four senses given above requires abhidhamma analysis. Sense A is true
if what is meant is comprehension of mundane information. It is false if what is meant is the
liberating knowledge. Sense B is false if what is meant is the arousing of insight since the arahat
must have active wisdom at the time of realization. It would be possible, however, to argue that
someone might attain arahatship, but not label their experience: this is arahatship. If the
question were raised, they would be able to identify it.57 It is also possible to argue that not all
ariyas would have the relevant reviewing knowledge.58 Indeed this would be generally agreed for
stream-enterers (cf. the story of Mahnma); some would only be able to identify themselves as
stream-enterers after being told the relevant criteria and investigating to establish the absence of
doubt, etc.
Sense C, however, implies the existence of arahats who can only overcome defilements
after a stimulus from someone else and sense D implies arahats who can only complete the path,
etc. after such a stimulus. The need for such a stimulus (parato ghoso) is of course standard for
stream-enterers and reasonably widely exemplified for arahats.59 It would, however, be felt in the
Theravdin abhidhamma and other ekbhisamaya schools that the individual concerned was not

53

Pais is certainly another text of this formative period. See A.K. Warders introduction to Pais tr.
The second of the Five Points is precisely kakh in the Kathvatthu. Most other sources reverse
the order of the second and third points, which means that kakh immediately precedes paravitra.
This may be earlier, but one late source, Vintadeva follows the Pli orderA. Bareau, op. cit., 194. It
is also possible that the verse cited by Pseudo-Vasumitra, etc. has changed the order for metrical
reasons (see n. 71 below).
55
Cf. Pais ii, 63.
56
P. Demiville, op. cit., 32n.
57
ibid., 3940.
58
This would not be acceptable in later Theravda, since all arahats are held to have reviewing of
defilements abandoned (in contradistinction to sekhiyas who need not), cf. Vism 676ff.
59
P. Masefield, Divine Revelation in Pali Buddhism, Colombo, 1986, collects the data on this.
54

39

yet an arahathe would perhaps have experienced the ordinary (lokiya) path of arahatship but
not yet the transcendent (lokuttara) path. Such a view would be more appropriate to an early
version of the gradualism of the Sarvstivda.60
It is perhaps significant that the final point made in the Kathvatthu is an
acknowledgement that arahats are not made to comprehend the (fruit of) arahatship by others. La
Valle Poussin61 is misleading here. The opponent accepts this point. No-one is arguing that an
arahat can be mistaken as to his fruition. This possibly implies a school in which experience of
magga is not necessarily immediately followed by the phala. Again, I suggest, an early version of
anuprvbhisamaya.62
In the seminal article in which he identified the Five Points in the Kathvatthu, La
Valle Poussin offers three possible translations of Points 24. The third, which he considers, to
be probably (?) the original meaning of Mahdeva, is: being ignorant and subject to doubt, an
arhat ought to receive instruction. To my mind, this is unfortunate. La Valle Poussins article
has been extremely influential and widely followedin particular in his view that the general
import seems to be a strong depreciation of the arahats. In fact the other two translations which
he offers are more to the point. The first refers simply to acquiring mundane information while
the second is the case of an arahat unaware of his arahatship who gets certitude from the
asseveration of another.
What we have here is a constructed dilemma which clarifies the distinction between the
knowledge of dhamma which every arahat must have and the more mundane knowledges of name
and family, etc. which are only known to some arahats. There is no depreciation of arahats as
such, here. At most it is only arahats without higher attainments and higher knowledges who are
being (slightly) depreciated. Why then did La Valle Poussin think there was? Partly it must be
because of relying on the accounts associated with the name of Mahdevaaccounts which we
now know to be late and probably subsequent to the period of conflict between Mahyna and the
early schools which seems to have occurred around the third century AD.63 Even more important
was his interpretation of the first of the Five Points to which we must now turn.

9. Parpahra
Unusually there are two terms given for the first Point in the Kathvatthu. In the uddna we
find parpahra. This, in isolation rather cryptic expression, is found

60

Prior to the development of the theory of the nirvedhabhgiyas which contains elements of a
synthesis with ekbhisamaya views.
61
The Five Points of Mahdeva and the Kathvatthu, 420.
62
Item 9 in the same vagga of the Kathvatthu.
63
This is attested both for Ceylon (e.g. Mhv XXXVI 41; 1112) and for Central Asia. See Z.
Tsukamoto, A History of Chinese Buddhism, Tokyo, 1985, index s.v. Hnayna, Khotan, Kucha, etc.
40

also in Pseudo-Vasumitra. Demiville64 points out that the different Chinese translations must
derive from different interpretations of the term.65 The earliest translation and also the Tibetan
translation interpret it in the sense of providing. Bhavya clearly had the same word but the
Tibetan translators appear to have resolved the compound as providing for another instead of
being provision by another.66 Unfortunately both Lamotte and Bareau have chosen to follow
Hsan-tsang and translate this point as the arahat can be seduced by others.67
In the body of the text of the Kathvatthu the proposition is put at first as an arahat has
emission of impure seminal fluid. Demiville renders the Jnaprasthna version as: Il y a
chez lArhat, molest par le dieu Mra, mission dimpuret. Just as with the other Points the
proposition is very startling. The question of the emission of semen is extremely important in the
Vinaya literature and hence in the practical life of the bhikkhu. It is discussed there not
infrequently and the emphatic statement in the Mahvagga68 that it cannot occur that an arahats
semen would be released would have been well-known.
La Valle Poussin suggested that the notion here is that of a succubus. The Kathvatthu
refers to the opponents claim that divinities of the Mra class (Mrakyik devat) bring about
the arahats emission of seminal impurity. The Jnaprasthna also attributes this to the activity
of Mra. According to Paramrtha, Mahdeva claimed that all bodily outflows (tears, phlegm,
etc.) in an arahat are the work of Mra.69 The same source attributed to Mahdeva a stra in
which occurs the statement: Le roi Mra et ses femmes divines, afin de faire dchoir laaika,
souillent dimpuret son vtement. What is important to note is that no source claims that this
could occur as a result of a dream. Of course it is suggested that a dream occurred in the case of
Mahdeva, but this is precisely because he is, according to the Mahvibh, a false arahat.
Arahats do not dream.
The key to the interpretation of this passage lies in the presentation of the opponents
argument at the end.70 The Kathvatthu often allows the opponent to make a telling point near the
end of the discussion. Here the point made is that others may provide (upasahareyyu) the five
requisites; therefore there is parpahra for an arahat. This is textually slightly clumsy as it
stands. The reason

64

P. Demiville, Lorigine des sectes bouddhiques daprs Paramrtha, 31n.

See also A. Bareau, op. cit., 242 n.
66
A. Bareau, ibid., 172n. The second list in Bhavya (ibid. 174) and Vintadeva must be interpreting
upahra as providing teaching.
67
So E. Lamotte, Buddhist Controversy over the Five Propositions, 148; cf. A. Bareau (Les
controverses relatives la nature de larhant, 242) sduit par autrui; A.K. Warder, Indian
Buddhism, 216 an arahat may have erotic dreams due to visitations by goddesses.
68
Vin i 295.
69
P. Demiville, Lorigine des sectes bouddhiques daprs Paramrtha, 35.
70
Kv 172.
65

41

is clear. In the Five Points as they originally stood what was asserted was the proposition that an
arahat can be provided (with material things) by others. This is obviously closely analogous to the
provision of mundane information as envisaged in the following Points. As we have seen, it is
precisely this original proposition which is preserved by Pseudo-Vasumitra and Bhavya,
undoubtedly because it was enshrined in a verse.71
The Kathvatthu and Jnaprasthna have focussed on what they see as the weak point
in the opponents argument in a kind of reductio. One may guess that there really was a stra in
which Mra was depicted as doing some such thing. This would not be so far out of line with
some of the other things Mra is shown as doing in the Canon. The logic is after all clear: deities
can provide the requisites for monks.72 If supernatural beings can create food and robes for
arahats, then they can create other things. If so, Mras can create undesirable things. We have a
sutta to support this.
It is interesting to see how the Kathvatthu seeks to oppose the point. Initially it
establishes agreement that arahats do not have passionate attachment (rga) and implies that
seminal emission is appropriate only for those who do. Then it seeks to establish the origin of the
seminal fluid produced by deities of the Mra class. The opponent agrees that it is not from those
deities nor the arahats own nor from other people. Deities and arahats do not have seminal
emissions in the ordinary way. If from other people, how does it get into the body? The opponent
agrees that it is not provided through the pores of the body. This rules out either a source from
other human beings or a creation by those deities outside the body.
The question is then asked why these deities do this, and we learn that it is in order to
produce doubt (vimati). It is established that this is not doubt in the Teacher, etc. Presumably,
then, it is some kind of mundane doubt. This topic is then leftpresumably because it will be
taken up in discussion of the subsequent Points. Then we return to the question of the origin of
the seminal fluid. The point to note is that Mras are Paranimmitavasavattin deitiesthey have
power over the creations of others, they do not merely create. The opponent is clearly working on
the basis of traditional Indian medicine in which seminal fluid (sukka) is one of the seven
elements beginning with chyle (rasa) into which food is successively transformed. The objection
is raised that not all who eat have emissions of seminal fluid (e.g. boys, eunuchs and deities). It is
also objected that

71

The first line must be, in Middle Indian form, something comparable to:
parpahro aa / kakh paravitra //
Note that this requires the change in the order of the second and third points. Compare Sp iv, 874.
(See notes 76 and 84 below.)
72
M i 243: deities offer to introduce food by means of the pores of the body (dibba
ojaajjhohareyyu); cp. later the ehibhikkhpasampad.
42

the case of excretion is not analogous, since there is no reservoir (saya) for seminal fluid as there
is for digested food.73
All this seems a little out of harmony with the next section which is an abhidhamma style
circulating discourse. First it is established that an arahat has completely and utterly made an
end of passionate attachment. Then the same is established for each of the other nine kilesas. Next
it is established that the path has been brought into being in order to abandon passionate
attachment (rgappahnya maggo bhvito). The same is then established one by one for each of
the other six sets which make up the bodhipakkhiyadhammas. This whole process is then gone
through one by one for each of the other nine kilesas (which include both delusion and doubt). A
fine mnemonic chant! What is its purpose? The answer must be, to emphasize the thoroughness
with which an arahat has accomplished his task in order to counter the suggestion that an arahat
may fall away.
What follows is a quotation emphasizing the qualities of the arahat. Then comes the
distinction between the two kinds of arahat. Then the whole process involving the ten kilesas and
the seven sets applied to each of the ten is applied to the two arahats. In fact a catukoi is
employed in each case to point out the oddity of the proposition. It is this circulating discourse
which recurs for the next three Points and is found in very similar form in the earlier discussion
of an arahats falling away.

10. The original form of the Five Points

Analysis of the Kathvatthu gives, then, clear evidence of a historical development in the
materials from which it is composed. We can divide this into three phases.
Phase One is the development of a literature consisting of constructed dilemmas. Of
course, some of these were probably very old but a fashion, as it were, for them would be
associated with the rise of abhidhamma. They would not really be a radical departure of any kind,
just a stimulating formulation for purposes of clarification. It is material of this sort which has
been used as the basis for many of the kathvatthni points for discussion. No doubt, too, they
continued to be composed.
Phase Two would be slightly later than, but overlapping with, Phase One. This would be
the period of the three-way doctrinal discussions between Pudgalavdins, Sarvstivdins and
Vibhajyavdins. It is just these three schools for whom we have a coherent doctrinal structure
emerging from the early period and no others.74 In

73

See J. Jolly, Indian Medicine, Poona, 1951, 65 for the list of the seven reservoirs, which does not
include one for sukka.
74
For the Pudgalavda, see now P. Skilling, The Sasktsaskta-Viniaya of Daabalarmitra,
BSR, 1987, 323, and T.T. Chau, Le personnalisme du Bouddhisme ancien, ICO, 1973; The
Literature of the Pudgalavdins, JIABS, 1984, 716; Les rponses des Pudgalavdin, JIABS, 1987,
3353 as well as K. Venkataramanan, Smmityanikya stra, VBA, 1953, 153243, La Valle
Poussin, La controverse du temps et du pudgala dans le Vijnakya; LAbhidharmakoa de
Vasubandhu, ch. 9, S. Schayer, Kamalalas Kritik des Pudgalavda, RO, 1932, 6893 and my
forthcoming article Person and Self to appear in the proceedings of the Buddhism 2000 conference
(Bangkok, 1990). E. Frauwallner, Abhidharma-Studien III, WZKS, 1971, 69121 makes plain the
origin of the Sarvstivda as a coherent whole. For the Vibhajyavda it is manifest in the Pli
canonical abhidhamma.
43

this period, however, we must certainly think in terms of schools of thought rather than separate,
organized sects. In the Kathvatthu this would be evidenced by the Puggalakath and the sections
in the early chapters dealing with the Sarvstivda.75
The original version of the Five Points, if it was originally a set of five, would be:
a. provision by others (parpahra)
b. lack of knowledge (aa)
c. uncertainty (kakh)
d. induction of comprehension/investigation by others (paravitra)
e. the arahat falls away (parihyati arah).76
This would be a mnemonic for the following argument. There are certain individuals who
attain a temporary liberation. They require an external stimulus. How do we know that arahats of
any kind may require external aid? It is agreed that they can receive material aid from others.
Equally they can be in doubt as to the correct road to take on a journey and can lack knowledge of
mundane things. In such cases they require external information if they do not have psychic
powers. Similarly certain individuals can momentarily achieve arahatship but external
confirmation or an external stimulus to stabilize their achievement is required if they do not have
sufficient concentration.
Phase Three in the development of the Kathvatthu would represent a subsequent
reshaping in a changed historical situation. The northern Sarvstivdin tradition has receded from
awareness. Its centres in Kashmir, Gandhra and Mathur are far away. Contact now is with the
Mahsghika traditions further south. It is to this period that we should attribute the work of
Mahdeva. Pseudo-Vasumitra describes the origin of three schools as due to the work of
Mahdeva.77

75

Pudgalavdin are: 169; 93115; Sarvstivdin are: 6993; 1039; 11551; 21220; 2257; closely
related are: 15155; 15963. If the first four Points were originally Sarvstivdin (i.e. 16395), there
can have been very little in the first two vaggas concerned with schools other than these two.
76
If there was originally a mnemonic verse (note 71 above), then the pdas of the second line might
have been either:
arah parihyati / eta Buddhna
ssanam // or
arahatt
parihni
/
eta
Buddhnussanam //
77
E. Lamotte, Histoire du Bouddhisme indien, 30910.
44

The riputraparipcch refers to the school of Mahdeva in the same context.78 It is with the
same group of schools that the Kathvatthu Commentary associates the Five Points. These and
later schools are grouped by the Kathvatthu Commentary under the name of Andhaka and it is
precisely in inscriptions from Amarvati and Ngrjunikoa that we meet them.
According to the earliest translation of Pseudo-Vasumitra we should date this
Mahdevas work to the period before 200 BE. This is a short chronology work; so it must refer
to a date about one hundred years after the accession of Aoka, i.e. the early second century BC.
Since Pli sources also imply a date after the reign of Aoka, it is probably safe to date the
formation of these later Mahsghika schools to the second century BC.79 What I wish to argue
is that the Kathvatthu was expanded and reshaped precisely at this time in response to ideas
coming from these schools. In fact the commentary attributes the bulk of the views in the
Kathvatthu either to the schools it calls Andhakas or to the Uttarpathakas. This must be a
recollection of the situation at an earlier date. Probably many views originally of Sarvstivdin
origin have been transferred to the more familiar Andhakas.80 The term Andhaka itself is a
reflection of tavhana times.
Mahdeva would then have taken up the Five Points and reformulated them for his own
purposes. It is this reformulation which is evidenced in the Kathvatthu. Probably it is at this
stage that the first Point was transformed from a simple statement that arahats can receive
material aid from divinities to a claim that (some?) arahats are subject to physical interference by
divinities of the Mra class. Very possibly the subsequent points were also reinterpreted in a
stronger sense. What then of the fifth Point?

11. The fifth Point

The early Mahsghikas appear to have rejected the idea that an arahat could fall away.81 This
must be the reason why Mahdeva has changed the fifth Point. It might have seemed natural
simply to transfer it to the stream-enterer, but this has

78

ibid.
See Nattier and Prebish, op. cit., 25864 for the view that Mahdeva and the Five Points must be
associated with southern Mahsghika schools.
80
A good example of this is at Kv-a 60 where the distinction between appaisakh-nirodha and
paisakh-nirodha is attributed to the Mahissakas and the Andhakas. Yet it must surely be
Sarvstivdin.
81
See A. Bareau, op. cit., 244; Les Sectes bouddhiques du Petit Vhicule, 66. This seems to be what is
said in Pseudo-Vasumitra, although the earliest translation differs: A. Bareau, Trois traits sur les
sectes bouddhiques attribus Vasumitra, Bhavya et Vintadeva, 243 n. Bhavya is silent, but
Vintadeva (idem, page 194) attributes the view that there is no falling away from either arahatship or
stream-entry to the Lokottaravdins. Bareau cites the Vibh. Kv-a 37 attributes the view that arahats
can fall away to some Mahsghikas. Probably this idea was admitted later in the Andhaka schools.
79

45

not been done explicitly. Instead, other questions related to stream-entry have been taken up,
which could approach the same question more obliquely.
The second vagga of the Kathvatthu preserves two items. In fifth place we have the
proposition that there could be verbal utterance (vacbheda) on the part of someone in a
meditative attainment (sampannassa), while in sixth place we have another statement in cryptic
form: dukkhhro maggaga.82 Within the text of the first item we have the question: When
knowing suffering, does one utter the word suffering?, while in the text of the second is the
question: Do all those who utter the word suffering bring into being (bhventi) the path?
Bhavya has the same two items, but in reverse order, in his account of the views of the
Ekavyavahrikas.83 It is the second item which corresponds with the fifth Point of the Sanskrit
sources.84 It is suitably ambiguous. At first sight it could easily be taken to mean: the nutriment
of suffering is a factor of the patha rather unexpected notion.85 It could mean evoking
[knowledge of] suffering is the cause of the path, but this would not be at all controversial. What
must be in fact intended is: pronouncing [the word] suffering is the cause of the path or it
could be is a sign of the path.86
Paramrtha and his interpreters preserve two explanations.87 One is that repeating a verse
can provide the stimulus required to arouse the path of stream-entry as in the case of Sriputta.
The point here seems to be that attainment of stream-entry normally requires some form of
teaching from the Buddha or a

82

Maggapariypanna must be an intrusion into the text of Kathvatthu from the commentary.
A. Bareau (op. cit., 174)dukhahni has probably been translated in place of dukhhro,
presumably a manuscript error.
84
VintadevaA. Bareau, Trois traits sur les sectes bouddhiques attribus Vasumitra, Bhavya et
Vintadeva, 194 is very close. The other two occurrences in Bhavya must be related (idem, 173 and
188). No less than three alternative versions of it have been added in the later translations of PseudoVasumitra (idem, p. 243). In the version of the Mahvibh and in the actual list of the Five Points
given in Pseudo-Vasumitra a version is given in which an expression meaning verbal enunciation
seems to have replaced enunciating dukkha. Certainly if the pda could be replaced easily by one
meaning: Ce sont l tes paroles dmentes, as Ki-tsang tells us, then some word from the root vac
must have been introducedcf. P. Demiville, op. cit., 36). Nevertheless it seems fairly likely that the
verse attributed to Mahdeva would in Pli form be similar to:
parpahro aa /
kakh paravitra //
dukkhhro ca maggaga /
eta Buddhna (or nu) ssana //
85
Later interpreters have ingeniously understood that suffering is the food that keeps beings alive in
the nirayasJ. Masuda, Origin and Doctrines of the Early Buddhist Schools, AM, 1925, 25n.
86
CPD s.v. aga.
87
P. Demiville, Lorigine des sectes bouddhiques daprs Paramrtha, 323; 36; 40.
83

46

disciple.88 The second explanation, derived from the Mahvibh, is that the deliberate repetition
of the word suffering can act as the necessary impulse to arouse that path. By themselves such
explanations seem fairly straightforward. What does the Kathvatthu have to say?
The first thing to notice is that there are an additional three related topics. The question as
to whether one can hear sounds while in an attainment89 is closely related both conceptually and
in literary form to the question as to whether one can make utterances. Similarly the question90 as
to whether the knowledge this is suffering occurs for one uttering the words this is suffering
is clearly another formulation of the same issues. More interesting than either of these is a third
point, which emerges when the literary form of the discussion of dukkhhro maggaga is
examined.
The treatment of this topic is brief, but the identical form is repeated later in the second
vagga.91 The immediate question is: Do all those who hear the utterance (vohra) of Lord
Buddha bring into being the path? This is part of the larger question as to whether the utterance
of the Lord Buddha is transcendent (lokuttara). This is important and must be examined, but for
now it is sufficient to note that the issue in this topic is partly the question of momentariness. Can
different things go on at the same time or do they occur in a rapid, sequential process? That of
course is precisely the question of suttanta versus abhidhamma.
This is the hallmark of the Kathvatthus treatment of many of the views which later
tradition associates with the Mahsghikas. They are again and again criticized for overgeneralizing, for lack of precision or for excessive enthusiasm.92 Of course, the criticism is
usually in the form of asking questions rather than overt criticism but it is no less real for that.
This is what one would expect if the views current among them were suttanta formulations
lacking in abhidhamma exactitudea rather conservative doctrinal approach. In this context it is
interesting to notice that the Vinaya of the Mahsghikas seems to define abhidharma as the
ninefold strnta.93 This suggests that the early Mahsghikas (or some of them) may have
rejected the abhidharma developments.

12. Mahsghika origins

If the Five Points and Mahdeva were not involved in the First Schism, then we are left with
vinaya issues as the cause. It has been realized for some time that it is

88

P. Masefield, op. cit.

Kv 572573.
90
Kv 453455.
91
The bottom 15 lines of p. 223 correspond very closely to the top 14 lines on p. 204.
92
A good example of the last is the irony which greets the notion of the fragrance of the Buddhas
excrement (Kv 563)due to inappropriate affection for the Lord.
93
G. Roth, Bhiku-vinaya, Patna, 1970, 248n.
89

47

unlikely that the Mahsghikas are directly descended from the defeated party at the second
communal recitation.94 They would hardly give a favourable account of their own defeat! It is of
course quite possible that they, or some of them, originated in the same geographical area as the
Vajjiputtakas and were associated with them in the minds of their opponents.
Human nature being what it is, it is perfectly credible that the Mahsghikas believed
that they had preserved the original form of the Vinaya which had been altered by others. Their
opponents are unlikely to have agreed. They probably felt that things had become lax and it was
necessary to restore the pristine teaching. In such a dispute historians should not take sides.95 We
may be sure that each party was able to make a case for its position.
What is important is that the picture which now emerges96 is one in which the earliest
division of the sagha was primarily a matter of monastic discipline. The Mahsghikas were
essentially a conservative party resisting a reformist attempt to tighten discipline. The likelihood
is that they were initially the larger body, representing the mass of the community, the
mahsagha. Subsequently, doctrinal disputes arose among the reformists as they grew in
numbers and gathered support. Eventually these led to divisions on the basis of doctrine. For a
very long time, however, there must have been many fraternities (nikyas) based only on minor
vinaya differences. They would have been very much an internal affair of the sagha and the laity
would have been hardly aware of them. Geographical differences and personalities would have
been more important than doctrine.
What then of the early schools within the Mahsghikas? According to the Sammitya
tradition preserved by Bhavya the Mahsghikas divided into two schools, at a point subsequent
to the origination of the Pudgalavda.97 The Dpavasa and other Pli sources mention the same
two schools as the first division of the Mahsghikas. The two schools concerned are the
Kaukkuikas and the Ekavyavahrikas. A few sources connected with the North West mention a
third: the Lokottaravdins. This may be due to the later prominence of that school in the area of
modern Afghanistan. In fact, however, it seems likely that the Lokottaravdins and the
Ekavyavahrikas are two names for the same school.
The Pli form (Gokulika) and the various translations make it clear that three distinct
interpretations of the name of the Kaukkuikas were current. The first gives the Pli form, but is
almost certainly an error or popular etymology based on the

94

e.g. M. Hofinger, op. cit., 178-9; A. Bareau, Les premiers conciles bouddhiques, 86ff.; C.S. Prebish,
A Review of Scholarship on the Buddhist Councils, JAS, 1974, 251ff.; A.K. Warder, Indian
Buddhism, Delhi, 1970, 214; G. Roth, op. cit., x.
95
Nattier and Prebish, op. cit., 26570 accept the Mahsghika account too readily.
96
H. Bechert, Zur Schulgehrigkeit von Werken der Hnayna-Literatur, 2044.
97
A. Bareau, Trois traits sur les sectes bouddhiques attribus Vasumitra, Bhavya et Vintadeva,
1956, 173.

48

Middle Indian form. The second interpretation explains it as connected with the Pli kukkua (Skt.
kukla) a burning ember or a chaff fire. The only view that the commentary to the
Kathvatthu attributes to this school is that all constructions without exception are burning
embers (kukkua). The Kathvatthu criticizes this as an over-generalization.98 If this is a genuine
recollection of the teachings of this school, as its context in the second vagga might suggest, then
this school could have been promulgating some teachings related to insight meditation.99
However, this too may well be a popular etymology. Most probably the name Kaukkuika
originated from the name of the Kukkurma100 in Paliputraa monastery associated in some
sources with the Mahsghikas. This would be a school centred on that monastery. Possibly the
connection became unfamiliar when Paliputra ceased for a while to be the effective capital of
India or after some destruction in that city.101
We can, I think, say more about the Ekavyavahrikas. To do so, we must return to the
question as to whether the utterance (vohra) of the Lord Buddha is transcendent (lokuttara). As
we saw, this is closely related to Mahdevas new version of the fifth Point in the Kathvatthus
treatment. What is also interesting is that it in fact deals with two distinct views. With the first, all
utterance on the part of the Buddha is transcendent, just as Both a heap of corn and a heap of
gold can be pointed to with a golden rod.102 For the second view, the Buddhas utterance is
ordinary (lokiya) when he makes an utterance about ordinary things, but transcendent when he
makes an utterance about transcendent things. The commentary remarks at this point that this
is one view; it is the view nowadays of some Andhakas.
It can then be clearly understood that the Ekavyahrikas or One-utterancers are so
called because they held the belief that Buddhas have only one kind of utterance, i.e. a
transcendent utterance. Hence too their alternative name of Lokottaravdins those whose
doctrine is transcendent or those who affirm the transcendent speaking (of the Buddha). The
Kaukkuikas on the other hand must have espoused the alternative proposition that the Buddha
had two kinds of speech. This

98

Kv 208212.
cf. A iii 443444.
100
Possibly the inhabitants of that monastery interpreted its name as derived from the Mgadh
equivalent to Kukla. Bhavyas first list includes mention of a school called Kurukula supposed to be
another name for the Sammityas. This list does not include the Kaukkuikas; so Kurukula is probably
a rendering of their name. In BHSD we also have Kurkurma.
101
This could be due to invasion, but note that the Aokvadna and other sources attribute the
destruction of this monastery to PuyamitraE. Lamotte, Histoire du Bouddhisme indien, 42530.
102
Kv 224.
99

49

seems very appropriate if we examine the two schools into which the Kaukkuikas appear to have
divided at an early date.
The commentary does not identify any of the views found in the Kathvatthu as
belonging to these schools, but there is some information in later sources. Taking first the school
of the Bahurutyas, Pseudo-Vasumitra tells us that they distinguish between the transcendent and
the ordinary teaching of the Buddha. The former consists of five words which have the power to
lead out of sasra: impermanence, suffering, emptiness, no-self and the peace of nirva. All
other words uttered by the Buddha are his ordinary teaching. This is clearly a development of the
thesis of those opposed to the One-utterancers. It is not clear how their views differed from
those of the second school, the Prajaptivdins. Their name could refer to some kind of doctrine
concerning descriptions or concepts, but it is perhaps more likely in the context that it
concerned the Buddhas making known of some aspect of the teaching.103
The Dpavasa knows only one further school among the Mahsghikasthe Caitya
school. According to the Sammitya tradition given by Bhavya it is this school which was
founded by Mahdeva. It was probably the mother school, based at Amarvati, of the later
schools which the Sinhalese know as the Andhakas.104 The fuller form of their name means either
those with a doctrine about shrines, i.e. stpas or those who honour shrines.105 The latter is
supported by archaeologythe remains at Amarvati certainly testify to an interest in stpa
symbolism. Pseudo-Vasumitra tells us that this school held that honouring stpas does not bring
much merit, which would rather support the former interpretation. Perhaps it is also relevant that
there is some evidence of deprecation of the stpa cult in certain of the early Mahyna stras.
What then is the significance of Mahdevas, if Mahdeva it was, alteration of the fifth
Point? To understand this, we need to turn to another aspect.

13. The experiential dimension

As it is presented in the suttanta literature, the enlightenment experience is the result on the one
hand of meditational practice (including devotion and study) and on the other of immediate
triggering events. Traditionally, these immediate causes are expressed as the two conditions for
the arising of the ariya path: teaching of dhamma by someone who has already experienced it
(parato ghoso) and appropriate bringing to mind (manasikra) on the otherthe external and
internal conditions

103

Compare the series at Kv 3156 where we learn that disciples do not make known the aggregates
(khandhapaatti), bases, elements, truths, faculties and persons.
104
E. Lamotte, op. cit., 5823.
105
Dp V, 42; Mhv V 5; Kv-a 2; 4 indicate the Pli as Cetiyavd(). Inscriptions give both
Cetiyavadaka and Cetiavadaka (Lamotte, idem, p. 500).
50

which combine at an opportune moment. When such a moment arrives, the enlightenment
experience can occur quite suddenly.
An individual who has had such an experience and stabilized it is an ariya, a person who
is genuinely noble as opposed to merely noble by birth. His experience is referred to as
transcendent (lokuttara) and when, subsequently, he acts or speaks on the basis of that
experience, his speech or action are also referred to as transcendent.106 Presumably the notion is
that the experience he has had and continues to have somehow suffuses and transforms his
speech. This must obviously be even more true in the case of a Buddha or an arahat.
As a description of how it should appear in practice, this is not controversial for any
school of early Buddhism. The problem arises when the attempt is made to give a more exact
formulation. This attempt was made in the abhidhamma literature. Here the mind is defined as
momentary and intentional in nature; a given mental event involves the knowing by a single mind
of a single object. The enlightenment experience was defined as the moment in which a
transformed and hence transcendent mind, in association with the mental structuring of the
path,107 takes as its object the element (dhtu) which is unconstructed (asakhata), i.e. its basis is
an experience of an aspect of reality which is uncaused and which does not construct new mental
and physical events. Yet this aspect somehow acts as the support for the transformed and newly
harmonious balance of mental events.
Obviously the notion of an intentional consciousness experiencing an object which is
effectively without boundaries or limits raises some philosophical problems and there are
differences between the various abhidhamma systems precisely at this point. Fortunately these
issues can be disregarded for the present purpose. The important thing to note is that in general
the abhidhamma systems of the Vibhajyavda and the Sarvstivda do not allow the simultaneous
occurrence of different consciousnesses. In the present context this means that the experiences of
hearing or speaking or bodily action or experiencing the dhamma which does not construct must
all involve different objects. Speaking or hearing cannot therefore be transcendent in strict
abhidhamma terms.
We should not misunderstand this. Seeing and hearing do not occur simultaneously in
abhidhamma terms. Obviously, however, we seem to experience them as occurring together and
in ordinary language we can speak of them as occurring at the same time. In just the same way
the experience of the transcendent and sensory activity are not simultaneous. However, we could

106

M iii 74.
R. Gethin, The Path to Awakening. A Study of the Thirty Seven Bodhipakkhiya-dhamm in the
Nikyas and Abhidhamma, Ph D thesis, University of Manchester, 1987, gives a full account of the
development of the theory of the magga and associated ideas.

107

51

experience them in alternation as effectively occurring at the same moment. The suttanta way of
putting things is not wrong from the abhidhamma perspective. It is simply that there is a more
exact form of expression which is more appropriate for the development of insight.

14. The reformulation of Mahdeva

We can now return to Phase Three in the evolution of the Kathvatthus treatment of the Five
Points. There could be no objection to the proposition that repeating the word dukkha might
sometimes act as the necessary stimulus to enlightenment and it is not obvious why the notion
that its utterance might occur spontaneously at such a time would be unacceptable. Nor could the
claim that the Buddhas speech was transcendent be rejected as such, especially not if it was
limited to his speech on dhamma topics. These things can only be objected to from the
abhidhamma point of view.
Not surprisingly, then, the Kathvatthu rarely criticizes these points as such. Usually it
simply attacks them as generalizations. Not everyone who pronounces the word dukkha
immediately gains enlightenment regardless of their previous behaviour, nor even everyone who
has developed insight! Quite often the typical abhidhamma emphasis on the impossibility of two
simultaneous consciousnesses occurs.108 What is interesting, however, is the precise position
which is being commented on. The opponent is making a very specific claim. The spontaneous
utterance of the word suffering occurs only in one case. It does not occur in ordinary jhna,
whether of the form or formless realms. Neither does it occur in an ordinary path attainment
(strong insight of the later terminology). Nor does it occur if the path attainment, although
transcendent, is higher than the first jhna in level. The commentary even understands that it is
restricted to the path of stream-entry on the grounds of the denial that it occurs in all cases.
However, it would seem difficult to justify this position from the text.
This restriction to the first jhna is very suggestive. It immediately recalls the pure
insight worker who achieves the jhna level of concentration only at the moment of stream-entry
and perhaps the arahat who is pavimutta. This places the reformulation of the five Points
firmly in the context of the distinction between the arahat skilled in paradhamma and the one
skilled in his own dhamma. Probably then this too is part of Mahdevas reformulation. There are
a number of reasons why this should be so.
Firstly, it seems odd to have a difference between the case of the arahats falling away
and the other four cases. Secondly, it is easy to replace the references to paradhammakusala, etc.
with those to asamayavimutta, etc. but the converse is not possible. Only the question of
temporary versus non-temporary liberation is

108

See Kathvatthu Index s.v. samodhna.

52

appropriate to the issue as to whether an arahat falls away. This of course explains why the
substitution could not take place in that case. Thirdly, as suggested above, this is an unfamiliar
terminology. It must come from the opponent. Yet it is not, as far as I am aware, a Sarvstivdin
usage; it may very well, then, be Mahsghika. Fourthly, it suggests a later period when an
emphasis on concern for others as a higher spiritual motivation is beginning to be formulated
more specifically. Finally, it seems to be associated with an emphasis on the value of practising
the higher jhnas and the abhis. This is perhaps not especially characteristic of the
Sarvstivdins.
It is certainly characteristic of the Yogcrins and it may be suggested that this may be a
feature in which they were influenced by the Mahsghikas. There is some reason to believe that
practice of the jhnas is of great antiquity109 and the Mahsghikas, or this branch of them, may
well have been conservative in this respect as well as others. Frauwallner has suggested that the
Yogcrins must have taken over many of the non-Sarvstivdin aspects of the Mahynist
abhidharma system from an earlier system.110 It would not be very surprising if that source
proved to be the Mahsghikas of central India, an area that seems to have gone over to the
Mahyna en masse at a relatively early date.111
The two key features of Asagas abhidharma are the acceptance of the possibility of
more than one consciousness at a time and the introduction of the notion of the layavijna. The
former might very well have been part of Mahdevas formulation, to judge by the Kathvatthus
criticisms, while the latter was attributed by the Yogcrins precisely to earlier concepts of the
Sinhalese school and of the Mahsghikas.112 It would not be at all unexpected if the
Vibhajyavdin concept of the bhavaga consciousness, already current in the later canonical
Abhidhamma period, was taken over or shared in some form by their neighbours, the southern
Mahsghikas.
Can we then assess precisely how and why the Five Points were reformulated by
Mahdeva? I think the answer is yes. His argument must have run something like this. There are
two ways of practisinga selfish one in which you are concerned with getting your own
enlightenment as quickly as possible and a more altruistic approach with more concern for others.
In the latter case you must

109

J. Bronkhorst, The Two Traditions of Meditation in Ancient India, Stuttgart, 1986.

E. Frauwallner, Abhidharma-Studien III, 103.
111
A. Bareau, Les sectes bouddhiques du Petit Vhicule, 296305. However, I would not wish to rule
out an association of the Madhyamaka with the Sarvstvda. Both schools are largely insightorientated and Madhyamaka dialectic can be seen as emerging from abhidharma debate. If so, we
would expect Mahyna literature of a Madhyamaka orientation to be of northern origin and reach
China more rapidly. This does seem to be the case, but such a suggestion is speculative at present.
112
L.S. Cousins, The Pahna and the Development of the Theravdin abhidhamma, JPTS, 1981,
22; L. Schmithausen, layavijna, Tokyo, 1987, II, 255, n. 68.
110

53

develop the jhnas and the higher attainments. There are serious snags to the selfish approach.
You can be subjected to material assistance, even harassment by Mras. You can lack crucial
understanding and have doubt as to your own achievement. You may also lack the information
you need to help others. You could well require the aid of others in order to reach your goal or at
any rate to finalize it. Because your concentration development is limited, you may need to
verbalize your insight meditation in order to stimulate the necessary absorption or to compensate
for the absence of teaching by another person when it is required. None of this will be necessary
if you develop the jhnas in order to become an arahat skilled in paradhamma.
Clearly there must have been more to it than this. Obviously the fact that it was felt
necessary to reorganize the Kathvatthu treatment of the Five Points indicates at the least that
the old formulation had lost relevance, presumably because of the success of Mahdevas new
version. We may guess however that a more substantial development of some kind would be
required. Most probably a Mahsghika (or Andhaka) version of abhidhamma had been created
on the lines suggested above. Very probably many of its key features are recorded in the
Kathvatthu.
It may eventually be possible to reconstruct it but the task is formidable. The attributions
of the commentary cannot be trusted without confirmation. The later literature on the schools
reflects a later situation when the Mahsghikas had largely adopted the Mahyna.
Sarvstivdin writers may attribute Mahynist notions to the Mahsghikas in order to discredit
one or both. Mahynist writers of a later date (e.g. Paramrtha) associate the two in order to
show the antiquity of the Mahyna. Probably most later Mahsghikas believed that their
particular tradition had always been Mahynist. It is however clear that the Mahyna cannot be
this early.113 That is to say, Mahyna as a movement distinct from and opposed to the early
schools cannot be. Undoubtedly some of the tendencies which led to the Mahyna literature
were already extant. To reconstruct the ideas of the early Mahsghikas we will have to discount
this material and draw instead on the Kathvatthu and the early Sarvstivdin literature.

15. Chronological aspects

The three phases in the development of abhidhamma discussion which have been identified
(section 10 above) can be approximately located in time. The Sammatya tradition cited by
Bhavya would suggest that Phase One might correspond to the period of debates at and just
before the Mauryan period. Phase Two would be

113

See now G. Schopen, The Inscription on the Kun Image of Amitbha and the Character of the
Early Mahyna in India, JIABS, 1987, 99137 and P. Harrison, Who Gets to Ride in the Great
Vehicle? Self-image and Identity Among the Followers of the Early Mahyna, JIABS, 1987, 6789.
54

during the Mauryan period and Phase Three at the end of the Mauryan period. The Pli sources
would locate the second phase in the reign of Aoka. The third phase must then be later. The Pli
sources and the Sammatya tradition are in that case approximately in line. There is, however, no
way in which this can be reconciled with the Sarvstivdin sources according to which the
divisions among the Sthaviras do not begin until a hundred years after the accession of Aoka.
It does not seem possible in the present state of historical knowledge to reach a firm
decision either way. Perhaps, however, the balance of advantage still lies with the long
chronology. Certain things follow, it seems, from whichever choice is made. If the long
chronology is correct, then the Sarvstivdin traditions as to the date of the works contained in
their own Abhidharmapiaka may not be correct. We should probably date some of the later
works earlier than tradition claims. Their dates will have been brought down in time to fit a
shorter period than was actually the case.
Conversely, if the Sarvstivdin tradition is correct, then certain aspects of the Sinhalese
tradition cannot be accepted. In particular it will be difficult to accept the claim114 that the Pli
canonical texts were set in writing for the first time at the end of the reign of Vaagmai Abhaya
(8977 BC) after a Tamil invasion leading to a period of Tamil rule and soon after the separation
of the Abhayagirikas from the Mahvihra. As Bechert has commented,115 beginning with that
period [second century BC] the Ceylonese chronicles can be considered as highly reliable sources
of historical information. They are in fact often confirmed by archaeological evidence. Given
that this is the case, it is difficult to reject their testimony about events in Ceylon.
Bechert has recently revived the suggestion that there are indications of the presence of
the short chronology in Ceylon at an early date.116 This, I think, is mistaken, but there is evidence
of a slightly different version of the long chronology. Most Ceylonese sources date the accession
of Aoka to 218 BE and the third communal recitation to 236 BE (i.e. 218 + 18). The
commentary to the first book of the Abhidhammapiaka, the Ahaslin three times states that
Moggaliputta Tissa promulgated the Kathvatthu in 218 BE.117 This strongly suggests that there
may have been an earlier tradition which dated the third communal recitation to 218 BE. The
precise authorship of the Ahaslin is debated118 but it is clear that, whether it was an early work
of Buddhaghosa

114

e.g Dp XX, 2021.

H. Bechert, The Date of the Buddha Reconsidered, 3435.
116
H. Bechert, Die Lebenszeit des Buddha, 1469; Remarks on the Date of the Historical Buddha,
1012.
117
As 34; 6. The first occurrence is attributed to a Vitaavdin, the other two to a prophecy of the
Buddha.
118
P.V. Bapat & R.D. Vadekar, Ahaslin, xxxiii ff.; Norman, Pli Literature, 1225.
115

55

himself or the work of an associate, it is less carefully edited than most of the other commentaries
and sometimes preserves earlier traditions which have been normalized elsewhere.119
It we turn to the Samantapsdika, we find an account of the legend of Moggaliputta
Tissa. This begins with the Elders of the second communal recitation searching the future to
see if the ssana will have such a scandal again. They see that in the 118th year from now a
king named Dhammsoka will arise, will generously give support, and many non-Buddhist
mendicants (titthiya) will enter the ssana and cause such an affair. The Elders decide to visit the
future Moggaliputta who is at that time dwelling in the Brahm world. They inform him that there
would be a great scandal in the ssana in the 118th year from now. So we see that both the
accession of Aoka and the third communal recitation are attributed to 218 BE. Very probably
this is the tradition that the Sinhalese found in the old commentary to the Abhidhammapiaka
when they set out to determine the chronology of past events.
120

The Mahvasa gives an account of the life of Aoka first and so only refers back to the
elders beholding the future, but it then goes on to the story of their visit to the future
Moggaliputta and gives the same prediction of a time of trouble after 118 years.121 The
Dpavasa simply begins with the prophecy regarding Moggaliputta: That monk, an exemplary
samaa, will arise 118 years in the future.122 It is clear that the reason that no introductory
account is given is that the cariyavda has been inserted between the prophecy and the first
account of the second communal recitation. Nevertheless it clearly belongs in the context we find
in the Samantapsdik. It must belong in the same context here, since the parinibbna of the
Elders of the second communal recitation is immediately mentioned, which would be
unnecessary if the prophecy was by the Buddha. This cannot then be evidence of the presence of
the short chronology. It is simply that the earlier prediction of the time of trouble has become a
prediction of the arising of Moggaliputta.
The other passage in the Dpavasa which is cited as evidence for the short chronology
occurs in the first chapter. The first communal recitation is mentioned; then the next loka
declares: 118 years after that will be the third recension.123 As Oldenberg points out in his
edition, the simplest explanation for this is that a loka which mentioned the second communal
recitation has dropped out.124 This

119

L.S. Cousins, op. cit., 389.

Sp 35ff. The Chinese version is almost the same; see P.V. Bapat, Han-chien-pi-po-sha, 2021.
121
Mhv V, 100.
122
Dp V, 55.
123
Dp I, 25.
124
ibid., 15n.; cf. J. Filliozat, Les deux Aoka et les conciles bouddhiques, 190.
120

56

passage, then, like the Ahaslin passage mentioned above is evidence for the date of 218 BE for
the third communal recitation. The only other evidence known to me for the short chronology in
Ceylonese sources is a verse attributed to the Ancients (Por) in the late fourteenth century
Saddhamma-sagaha.125 However, this text refers to verses from the Clavasa as by the
Ancients; so it is not evidence for an early date. Moreover, it has not been critically edited and
the verse concerned is easily amended.126
There is, then, no reason to believe that the short chronology was known in ancient
Ceylon, but considerable support for the existence of a tradition that the third communal
recitation took place 118 years after the second. One might guess that originally the commentorial
tradition recorded the same figure for both the king and the recitation. Subsequently it was
realized that this was unlikely and the date of the recitation was moved a further eighteen years
on. It seems better to adopt the reverse procedure. This would suggest that the accession of Aoka
took place about a hundred years after the second communal recitation (assuming that the third
recitation took place about eighteen years later).127 However, it is more likely that the figure is
notional and slightly exaggerated as with the second communal recitation. In this case the
accession of Aoka should have taken place between about 140 and 160 BE (70/80 + 70/80).
This has the virtue of bringing the Sinhalese traditions into line with Bhavyas
Sammatya account. If we date Aokas accession at 52 years after the accession of Candragupta
in c. 313 BC,128 then the work of the founder of the Pudgalavdins will take place around 261 BC
with Moggaliputtas response and the third communal recitation, if there was one, at c. 243 BC.
The beginning of the controversies would be 63 years before Aoka, i.e. c. 324 BC under
Mahpadma Nanda. We know of course that a Nanda was ruling in Magadha at the time of
Alexanders invasion (327324 BC). This would imply a date for the beginning of

125

Saddhamma-s 47.
J. Filliozat, op. cit., 191. Two other verses attributed to the Pur are also relevant. Saddhamma-s
35 gives a date for the third communal recitation of 228 BE, while Saddhamma-s 44 gives the date of
238 BE for Mahindas Fourth Council.
127
K.R. Norman, Aokas Schism Edict, Buddhist Seminar, Kyoto, 1987, 1618, summarizes the
various Pli accounts relating to the third communal recitation. The figure eighteen is probably
notional for a number of years. See G. Obeyesekere, Myth, History and Numerology in the Buddhist
Chronicles, to appear in the volume mentioned below (note 129).
128
J. Filliozat, La date de lavnement de Candragupta roi du Magadha (313 avant J.-C.). Filliozats
arguments are not conclusive. However, since Candraguptas accession must be between Alexanders
departure from India in 325 BC and the return of Seleucus from India to the battle of Ipsus in 302 BC,
it represents a convenient median date. Magas of Cyrene probably died in c. 250 BC.see F.
Chamoux and further references in Peremans and Vant Dack, Prosopographica Ptolemaica VI; E.
Will, Histoire politique du monde hellnistique, Nancy, 1979, I, 66ff; 264ff.
126

57

the Buddhist era between 400 and 420 BC. Other evidence would also seem to support a date
close to the end of the fifth century BC.129
It may be suggested that one reason why the length of time has been increased is an
attempt to fit a king list with the Buddhist traditions. It seems most unlikely that the Buddhist
sagha would have preserved a list of the kings of Magadha together with their regnal years.
When the Sinhalese found a need for such a list, there is only one place they could have got it: the
brahmins. If there is a general similarity between the Sinhalese tradition and that in the Puras, it
is because the Sinhalese got it from the Puras or from where the Puras got it. In fact we
have no certainty of the existence of any other source from which they could have got it.
Summary
Sections 14 examine the historical problems and background, suggest a date of around 7080
BE for the Council of Vaili and discuss the available sources of information on the early
Buddhist schools. The significance for this of the Five Points is indicated. A discussion of the
date of the Kathvatthu indicates a stage in which there was a three-way controversy:
Sarvstivda, Pudgalavda and Vibhajyavda.
Sections 510 examine the first four Points in detail and seek to show that in their
original form the fifth Point was the question as to whether an arahat can fall away. The logical
structure of the original Five Points is indicated and it is suggested that in this form they were
probably Sarvstivdin. Three phases in the development of the Kathvatthu are proposed.
Sections 1114 examine the fifth Point and explore its connection with the Mahdeva
associated with the development of the later (southern) Mahsghika schools. Evidence from the
Kathvatthu is brought to bear on the nature of the earliest Mahsghika schools. The new
formulation of the Five Points is examined and suggestions are made as to the nature of the new
developments among the Mahsghikas. In particular, trends to emphasize the altruistic value of
developing the higher jhnas and a new formulation of a Mahsghika abhidharma seem likely.
Section 15 examines the chronological implications. Evidence in the Ceylon sources is
advanced to support the existence of an early tradition dating the Third Council to 218 BE. The
suggestion that there is evidence for a short chronology tradition in the Pli sources is refuted.

129

See K.R. Norman, Observations on the Dates of the Buddha and the Jina (to be published in a
volume on the date of the Buddha edited by H. Bechert).
58

Addendum
In late 1989 Professor Richard Gombrich circulated a paper on the date of the Buddha.130 He has
kindly given permission for it to be referred to here prior to publication. In this paper he has
offered an ingenious reinterpretation of the data given in the Dpavasa and has convincingly
shown that the information given there on the ages of the teachers in the vinaya lineage of
Mahinda (traditionally interpreted as their age since ordination) is better and more consistently
interpreted as their age since birth (or conception). This produces a date for the accession of
Aoka of c. 136 BE (with a margin of uncertainty due to the addition of a series of life-spans
given in figures rounded to whole years).
Gombrich takes the date of the accession of Aoka to be c. 268 BC and therefore suggests
that the Buddhas death took place within six or even five years of 404 BC. His argument can, I
believe, be taken one step further. Gombrich discards all data given in the Pli chronicles as to
regnal years. This seems in general appropriate. However, the information in chapter five of the
Dpavasa about the date of accession of Candragupta is likely to have been handed down as part
of the vinaya lineage.131 If so, Candragupta ascended the throne in c.100 BE.132
Taking the accession of Candragupta to occur in c. 313 BC, the following approximate
chronology arises:
BC
413
343/333
331
326/5
313
277
271
259
245

Mahparinibbna of the Buddha

Second communal recitation
Birth of Moggaliputta Tissa133
Alexander in India
Accession of Candragupta
Accession of Aoka134
Ordination of Mahinda
Third communal recitation
Death of Moggaliputta Tissa

130

BE
0
70/80
82
87/88
100
136
142
154
168

The title of the manuscript was Dating the Buddha: A Red Herring Revealed.
Dp V 69.
132
Could this be the source of some of the short chronology traditions? The later more familiar name
of Aoka could have been substituted for that of Candragupta.
133
The story in Sp of the Elders of the second communal recitation visiting Moggaliputta Tissa in the
Brahm world and requesting him to take birth now fits in very well (see pp. 4748 above).
134
The five Greek kings mentioned in the 13th Rock Edict would then be:
1. The Seleucid Antiochus I (281/280261 BC) or Antiochus II (261246 BC)
2. Ptolemy II of Egypt (285/283246 BC)
3. Antigonus II of Macedonia (276239 BC)
4. Alexander II of Epirus (from 272/271 BCdate of death not known)
5. Magas of Cyrenaica (c. 275c. 250 BC).
The Edict could not have been inscribed before the accession of Alexander of Epirus in 272/271 BC
nor much after 250 BC.
131

59

Although these dates are only approximate, they offer a real possibility of establishing a
definitive chronology, if new archaeological or other information should come to light.
The reason why the Ceylon chronicles went astray is now clear. They must have had
access to brahmanical traditions on the regnal years of the kings of Magadha (as well as to a
northern account of the development of the eighteen schools). They constructed (in the
Mahvasa or its sources) a new, more consistent chronology in an attempt to reconcile their
own traditions (which must have been based on the lineage of Mahinda) with the new data.
Ironically, it transpires that they would have been better advised to be less open to overseas
influences and keep their own tradition.

60

Some Formative Influences in Mahyna Buddhist Art

P.T. Denwood

The importance of art in Vajrayna Buddhism

Religions differ in the importance they place on representational art. Somesuch as
orthodox Islam, one phase of Zoroastrianism, and some forms of protestant
Christianityhave actively discouraged it. Otherscatholic and orthodox Christianity
and some forms of Hinduismencourage it but can live happily without it. Vajrayna
Buddhism appears to be a religion in which representational art is essential. The tantras
themselves, presented by their adherents as the fundamental documents of their system,
place great importance on iconographical descriptions of the deities and of their
maalas, which constitute a representational simulacrum of totality, and form the basis
of most of the practices of the religion. Although the philosophical argumentation taken
for granted by the tantras could be said to be equally or perhaps more fundamental, it is
hard to see how a Buddhism stripped of iconography could possibly function in the
distinctive way which the Vajrayna does. The investigation of artistic elements and
motifs, their meanings, associations and symbolism, origins and historical development is
therefore of peculiar importance in the study of Vajrayna Buddhism, because it touches
on the very core of the religious system. In this paper I shall indicate what I think may be
some formative influences on the development of the maala, which will necessitate
going back to a presumably pre-Vajrayna phase of Mahyna Buddhism, and further
afield into the history of India and of western Asia.
The first few centuries of the first millennium AD saw the rapid spread of the
universal religions of Christianity and Buddhism in western and central Asia, territories
ruled by the imperial powers of Rome, the Parthians, Kushans and Sasanians, in parts of
which Zoroastrianism also flourished. Imperial and universal ideologies contributed to a
burgeoning of new iconographic and architectural forms at a time of great cultural
interchange between different regions.

Byzantine architecture
The architectonic form of the classical, pre-Christian temple of Greece and Rome is wellknown with its axial plan, its inner cella, surrounding columns and gabled

61

pitched roof. When from the third century AD the Christians of the Roman Empire (then
centred on Constantinople) embarked on ambitious church-building programmes, they
perhaps deliberately avoided the form of the classical temple, and developed two types of
church design. The first, the basilica, is familiar in western Europe as the normal type of
axial layout with a nave flanked by aisles and leading up to the apse with its altar. The
type was borrowed from an existing type of building used for many secular purposes
meeting halls, lawcourts and markets.1
The second type, or rather range of types, now more familiar in the lands of
orthodox Christianity, has been called the centralised congregational church, usually
built around or constituting a memorial shrine for a saint or martyr called a martyrium.2
Although existing in many forms, a common factor of the plan is fourfold symmetry
around a central point. (see figs. 1, 2, 3). Thus the plan may be round, square, hexagonal,
octagonal or cruciform (or some combination of these). Sometimes there are entrances to
the four quarters. (An example of a cruciform ground plan with four such entrances is
shown in fig. 2.) The main body of the building is of one storey, though this might be
quite tall, while over the central part, supported by columns, walls or piers, is an upper
storey in the form of a low tower, spire or dome. This additional storey is not intended for
any practical use and is not usually separated from the lower storey by any ceiling of
the latter. Its purpose is purely symbolic or decorative. I shall leave further discussion of
this building type to a later section, mentioning only the fact that somewhat similar
architectonic forms are known from certain Roman palace building3 (and in one or two
cases, such as the audience hall of al-Mundhir at Resafa,4 shown in fig. 3,
archaeologists have found it hard to decide whether the remains represent a church or a
royal audience hall).

Iran
Moving further east to the Iranian sphere, we can also find many buildings displaying
fourfold symmetry in a square or cruciform plan, during the Parthian and Sasanian
periods (2nd century BC to 7th century AD). Again both religious buildings and royal
palaces are in question. The so-called square house excavated by the Russians at Nysa
seems to be part of the palace complex built by Mithridates I in the second century BC. It
consists of a large hall some 20 metres square with a main entrance and other entrances
on two of the other sides. Four massive columns supported the roof, though whether they
also held up any superstructure is not known. In the upper register of the walls were
niches holding

C. Mango, Byzantine Architecture, New York, Abrams, 1975, 81.

C. Mango, op. cit., 87.
3
C. Mango, op. cit., 85ff.
4
C. Mango, op. cit., 89.
2

62

huge clay statues, today very fragmentary but probably representing deified ancestors.5
The Parthian palace at Assur/Labbana on the Tigris (1st century AD) incorporates a
roughly square space flanked on all four sides by iwans or open-fronted rooms facing into
the centre. It is thought that the central space was not roofed over, but there was also a
four-columned hall nearby similar to the one at Nysa, and roofed with brick barrel vaults6
(see fig. 4).
Sasanian palaces also incorporated rectangular or quasi-cruciform audience halls
with entrances in the centre of all four sides: for example the palace of Ardashir I (3rd
century AD) at Firuzabad (fig. 5); the palace of Shapur I at Bishapur (also 3rd century);
the palace at Sarvistan (?4th6th century) (fig. 6), and the palace built by Khusrau II at
Qasr-i-Shirin (6th7th century). Where the roofing system of these Sasanian audience
halls is known, it takes the form of a central dome, set back slightly from the edge of the
building to leave a narrow surrounding flat terrace7 (fig.7).
Some of the characteristics of these Parthian and Sasanian audience halls are
shared with temple buildings of the same periods. Temples possibly as early as
Achaemenid date (pre-4th century BC) at Persepolis, Susa and elsewhere; of Parthian
date at Hatra; Graeco-Bactrian date at Ai Khanum and Sasanian date at Bishapur and
Takht-i Sulaiman and elsewhere show the square plan, sometimes with four internal
columns and four entrances, and with another feature shared with several of the palace
buildings mentioned above: corridors surrounding the main chamber which might form a
circumambulatory passageway.
Much of this Iranian material has been discussed in a useful article by D.
Stronach8 on the evolution of the early Iranian fire temple. (Many of the buildings have
been seen as fire temples by different scholars, though in many cases this identification is
open to serious doubt. In some cases the temples are more likely to have housed iconic
images.) In his conclusion Stronach points to a transference, from about 200 BC
onwards, of certain architectural concepts from what might be termed the east Iranian
world to the west Iranian world, associated with the rise of the Parthian dynasty.
Many of the same buildings are discussed by A.U. Pope in his Survey of Persian
Art.9 Here again there is sometimes cause for dispute as to whether certain buildings were
royal audience halls or fire temples (e.g. the Chahar Qapu at Qasr-i-Shirin10). Pope also
touches on the question of the relationship between

G. Herrmann, The Iranian Revival, Oxford, Elsevier, 1977, 34.

G. Herrmann, op. cit., 57.
7
G. Herrmann, ibid.
8
D. Stronach, On the Evolution of the Early Iranian Fire Temple, in Papers in Honour of Professor
Mary Boyce, Leiden, E.J. Brill, 1985.
9
A.U. Pope, ed., A Survey of Persian Art, Oxford, 19389, vol. 2, 533559.
10
ibid., 552 ff.
6

63

Zoroastrian fire temples and early Christian buildings, mentioning the building referred to
above at Resafa (here regarded as a mortuary church; the Praetorium of Musmiya, and
the Armenian type of domed church.

The two phases of Buddhist art and architecture

It is useful now to turn to the works of the Austrian scholar G. Franz, who has
distinguished two main phases in the art and architecture of Buddhism. The first phase,
lasting down to the 1st century AD, was characteristed by aniconic art forms as cult
objects, and by what he calls anarchitectonic forms as cult monuments: stpas, treeshrines, and pillars. Associated building forms were normally circular or partly circular
(apsidal) in plan at ground level.11
The second phase began in northern India and the lands to the northwest in the
late 1st century AD, and many of its developments came to be shared by Hinduism. It
was marked in architecture by the appearance of the rectangular, particularly square (and
also, I would add, cruciform) groundplans of buildings designed to house iconic cult
images.The rectangular plan was also applied to the base of the stupa at this time. Further
architectural features associated with this phase were a circumambulatory passage within
the square temple-building, and the tendency to vertical development of both stpas and
temples (also, the placing of the building on a substantial rectangular podium or terrace).
Franz is in no doubt about the origins of the new style, which he traces to the 1st
century AD in present-day Afghanistan and Pakistan, where there arose a combination
of Buddhist art with the forms of the Hellenistic-Parthian-Iranian repertoire. He provides
abundant evidence to support this claim, describing many of the buildings I have
indicated above, among others.

The maala-palace
There are various ways of envisaging or representing the maala in two dimensions, but
perhaps the commonest, given much prominence in the tantras themselves, is with a
buildingcalled a palaceshown somewhat diagrammatically in the centre, and housing
the main divinity with his/her retinue.
In an article written some years ago and shortly to be published,12 I argued that
the Tibetan and Sanskrit descriptions of the maala-palace to be found in the tantras
and the related two-dimensional drawings and paintings could profitably be studied
alongside the descriptions of Brahmanical temples in the Indian ilpa-stras, to their
mutual benefit. Much of the technical terminology of the two genres seems related, and at
bottom a similar or even near-identical concept of an ideal building is shared by both.
Further study, helped by personal communications

11
12

G. Franz, Von Gandhara bis Pagan, Akademische Druck, Graz, 1979, 7ff.
By the Percival David Foundation.
64

from other scholars,13 reveals that the Tibetan commentarial, oral and practical ritual
traditions can be of great help in delineating the features of the buildings concerned.
Furthermore, the palace-building which is envisaged in the Tibetan maala texts as a
cubical or quasi-cruciform building with a smaller upper storey (fig. 8).
In another article to be published shortly14 I have compared the literary and ritual
material with actual full-size buildings to be found in Tibet, India and elsewhere, and
have argued that the texts and buildings can in some cases be shown to correspond
(though of course it is not easy to say which came first).
As a result of these studies I have postulated that in the late Kushan and Gupta
eras in many parts of north and northwest India there had been developed a distinctive
concept of a religious building (no doubt only one among others). Erected on a square or
cruciform ground plan, it essentially comprised a chamber of roughly cubical
proportions, housing any images or cult objects. Over it as an upper storey was a second
chamber of reduced width and height, and either square or circular in plan. This upper
chamber might be shut off from the lower one by its floor, or it might form an upward
continuation of the central part of the lower storey. In either case it was not intended to be
regularly entered or to have any other function than to mark the building as something
special. Practical necessity would commonly require the walls of the upper chamber, set
inwards from the edge of the building, to be supported by either columns or walls below,
thus dividing the lower chamber to create either side aisles or even a closed
circumambulatory passage at ground level.
Once created, the basic schema could be modified in many ways to accommodate
different religious, economic and practical considerations. The original idea involved a
flat terrace around the edge of the upper storey. This might be practicable in arid regions
of low rainfall, or where stone was the sole constructional material to be used. In rainier
districts, particularly if brick or wood were to be used, some sort of sloping or curving
roof was necessary at this point, as also at the top of the building. The shape and
materials of these roofs would depend on local techniques, materials and resources,
giving rise to quite a number of variations on the basic theme. Once this principle has
been grasped, one can readily see how the schema was elaborated in different parts of
Greater India, from the stone-built Dravia-style temples of central and south India to
the tiered-roof temples of Nepal, Kulu, Kerala and related types in Bengal, Tibet and
Southeast Asia. The main later development in all these areas was often the
multiplication of further stories beyond the original two, but with no radical change of
principle in that the upper tiers nearly always continue to diminish in size and do not
fulfil any practical function. (see figs. 9, 1 0, 11).

13
14

Particularly M. Murray and M. Boord.

In South Asian Studies, no.6, 1990, 95104.
65

I would contend, then, that there arose, perhaps in the eastern parts of the Iranian
cultural world during the 1st millennium BC, a type of building exhibiting in plan
fourfold symmetryprobably originally squareand often with internal columns,
domed or other superstructure over the central part, entrances to the four quarters and
circumambulatory passages. This was developed within Iran and exported perhaps to the
eastern Roman Empire and certainly to India, in neither of which regions anything like it
had previously been known.

Royal cults
My next task is to explore some aspects of the religio-cultural context in which these
building types spread. Presumably foreign building styles are not adopted without some
rationale over and above mere fashion or desire for change. Since the western end of the
ancient world has been on the whole more thoroughly studied than the eastern, I shall
start with the Byzantine Empire and refer first of all to Cyril Mango,15 who in his book on
Byzantine architecture puts the question, where did the inspiration (for the centralised
church) come from? His answer is worth quoting at some length:
The most likely answer is that it came from the audience halls of palaces, where
the earthly monarchs were surrounded by a liturgy in many ways comparable to
that of the heavenly King. Indeed, what could have been more natural? Imperial
art was older than Christian art, and it is generally accepted that in the realm of
triumphal iconography it exerted a decisive influence on the latter. The Byzantine
mind imagined Gods habitat as a vastly expanded and more splendid version of
the emperors Sacred Palace, so that Gods house could logically be cast in the
same mould.
There is considerable evidence to suggest this line of derivation. The centralized
triclinium or reception hall was a regular feature of Roman palace architecture
as far back as Neros Domus Aurea (mid 1st century AD).
The context here adverted to is that of the ritual surrounding the king as he
receives his subjects in audience. One feature of oriental kingship in many periods from
Alexander on was the divinisation of kings who were worshipped as gods. This practice,
with its associations of universal sovereignty, was widespread in Hellenistic times in
kingdoms with rulers of Greek or Roman origin, and could become conflated with an
ancestor cult in which the ruling house and its divine origins were the object of
veneration. Perhaps the most remarkable monumental expression of such a cult is the
Hierothesion built by Antiochus of Commagene at Nimrud Dagh in eastern Anatolia
(1st century BC). Here colossal stone images

15

C. Mango, Byzantine Architecture, 87ff.

66

of the king, his family, ancestors and divine beings are arranged in the open air around
three sides of a square, and show a unique blend of Iranian and Greek religious ideas. A
recently excavated palace of Claudius now submerged in the Bay of Naples (1st century
AD) is a Roman equivalent of this ancestral portrait gallery with strong religious
overtones.
As far as I know the Parthians and Sasanians did not adopt the cult of divinised
kings, though the latter certainly had an elaborate court ritual enacted in specially
designed buildings. The Parthians at least seem to have had galleries of ancestral statues
at Nysa and also Toprak-kala in Soviet Central Asia, and at Shami in Iran. The Kushans,
with whom the development of the new styles of Buddhist art and architecture in
northwest India is closely associated, seem to have had both ancestor cults and divine
kingship. Their royal cult under the new title of devaputra and their remarkable ancestral
portrait galleries at Surkh Kotal in Afghanistan and Mat at Mathura have been ably
discussed by J. Rosenfield.16
It seems possible that this type of royal gallery of statues (which, at least in
Parthian Nysa was combined with the centralised type of royal audience hall) could
have contributed to the idea of the array of carefully positioned divinites around the
central figure in the Buddhist maala. Royal associations are certainly not absent from
Mahyna and early tantric Buddhism, as D.L. Snellgrove clearly showed in an article on
the notion of divine kingship in tantric Buddhism.17 Thus:
There need be no doubt that a form of royal consecration had been adopted in
certain Buddhist circles already by the 7th century AD.
[Translated from the Majurmlakalpa]
They [those to be consecrated in the maala] should be kshatriyas, those who
have been consecrated great kings, or their sons or daughters, those who are
unacquainted with the ways of common folk(Ratnaketu) the great cakravartin
chief is to be placed in the centre [of the maala]. He has the colour of saffron
and is like the rising sun he is like a great king with his palace and his
decorations
Snellgrove also drew attention in that article to a possible association between the
maala and the Buddhist temples of Nepal, without I think being aware of the
antecedents discussed in this paper. U. Wiesner has more recently emphasised the
intimate connection between king, temple and national divinity in later Hindu

16

J.M. Rosenfield, The Dynastic Arts of the Kushans, Berkeley & Los Angeles, University of
California Press, 1967.
17
D.L. Snellgrove, The Notion of Divine Kingship in Tantric Buddhism, in International Congress
for the History of Religions, The Sacral Kingship, Leiden, E.J. Brill, 1959, 204218.
67

Nepal: thus For the power of the king to be legitimated, it seems to have been essential
for the supreme deity of the country to reside near the seat of the ruler.18

Conclusion
The equating of universal Buddhahood with universal sovereignty and its imagery in
Mahyna Buddhism is well enough known by now. What I have tried to show in this
paper is that the notion of the monarch as potentially universal sovereign (and often of
divine ancestry) was associated with a particular type of building. This was the
centralised audience hall whose features might include fourfold symmetry in plan,
entrances to the four quarters, an upper tier of reduced proportions, and
circumambulatory passages. Over the whole area from India to Europe the type came to
be shared by certain religious buildings, at a time when emperor-cults are known to have
formed models for the cults of developing universal religions, with their rapidly growing
iconography. Another phenomenon, the royal gallery of ancestral and divine statues,
might exist separately or in combination with this type of building. In the case of
Buddhist India this complex of ideas and images is clearly in part an import from the
Iranian lands to the northwest.
That the complex was readily adopted in India is no doubt partly due to the preexistence there of related, if differently developed ideas. The cakravartin or world-ruler
was an ancient concept in India. Already in Vedic times the maala, probably in simple
two-dimensional form, existed; thus in the Ghyastras:19
For studying the ranyaka teacher and student go to a circle (maala) in the
north-eastern direction a king who is to be consecrated sits down in a specially
prepared circle (maalam)
As a matter of fact, as noted by Snellgrove, the conscious royal imagery evident
in the Majurmlakalpa had already waned in importance even by the time of the
Sarvatathgatatattvasagraha:
Already in the Sarvatathgatatattvasagraha the followers of the Buddha were
far beyond the stage of conscious adaptation of the symbols of sovereignty, for
they had now become an accepted part of the tradition. From now on the central
divinity resides in his own right in the

18

See Wiesner, Nepalese Temple Architecture, Leiden, E.J. Brill, 1978, 3ff. This association is a very
ancient one. Compare the following: The cosmological aspect of the city (in the ancient near east)
has its roots in the idea of the city as the abode of the god, the ruler of cosmos and nation, because the
temple, as the visible expression of his domain, was, at the same time, the kings property, the capital
was the ruling center of both the god and his vice regent the king. Therefore, temple and palace
should be seen as two aspects of the same phenomenon; together they constituted the essence of the
state. (G.W. Ahlstrom, Royal Administration and National Religion in Ancient Palestine, Leiden,
1982, 3ff.)
19
J. Gonda, Vedic Ritual: the Non-solemn Rites, Leiden, E.J. Brill, 1980, 233, 244.
68

temple-palace in the centre of the maala and he appears spontaneously wearing

full royal regalia.
Similarly, I suggest, the royal associations of the building type of the maalapalace and the idea of a formal group of royal statues fade into the background of tantric
Buddhism in the course of time, though the architectural and statuary forms live on, to be
given other layers of symbolism. So it was thousands of miles away in the Christian
world, where the royal associations of the centralised church were largely forgotten
though the building form lives on to this day. At both extremes, the decay of imperial
ideologies, due probably in large part to the very success of the new universal religions
and to other historical factors, led to the loss of the architectural and sculptural forms
which were the original model.

69

The rNying-ma Interpretation of Commitment and Vow

Gyurme Dorje

1. Preamble
The persecution of Buddhism in 838 marked a great watershed in the history of Tibet,
one which may well turn out to be even more divisive than the events of the last forty
years. It heralded a dark age which was to endure for over a century until the Buddhist
restoration of 953, and which effectively ended the powerful Yarlung dynasty in 869. In
retrospect, historians have considered this period to demarcate decisively the earlier
phase (snga-dar) and the later phase (phyi-dar) of Buddhist propagation in Tibet.1 Those
who adhered to the transmissions and teachings introduced during the earlier phase
became known as the rNying-ma-pa, in contrast to the proponents of the various new
translation schools, who were known as gSar-ma-pa. However, the hardening of attitudes
during the tenth and eleventh centuries, which is clearly evident in the rNying-ma
writings of Rong-zom Chos-kyi bZang-po, as well as in the ordinances of Lha bLa-ma
Ye-shes od, and in the polemical broadsides of Gos-khug-pa Lhas-btsas and Bri-gung
dPal-dzin, more accurately reflects certain differences between the Indian Vajrayna
Buddhism of the seventh/eighth centuries and that of the tenth/eleventh centuries.2 The
texts transmitted during the earlier period, notably the tantras of Mahyoga, Anuyoga and
Atiyoga, are referred to as the ancient translations (snga-gyur) when contrasted with
the later translations (phyi-gyur) of Anuttara Yogatantra. As distinct genres of
Buddhist literature, these texts undoubtedly exhibit differences in terminology, style,
versification, and maala formation, with the former in many cases appearing to
conform iconographically to even earlier developments within the stra and tantra
traditions of Indian Mahyna Buddhism and the latter appearing to incorporates more
non-Buddhist imagery.3

e.g. Gos Lo-ts-ba, tr. by G.N. Roerich, The Blue Annals, Delhi, 1976, 60 ff.
G. Dorje, The Guhyagarbha-tantra and its Commentary, PhD thesis, SOAS, University of London,
1987, 1417, 6172.
3
D.L. Snellgrove, Indo-Tibetan Buddhism, London, 1987, 147160.
2

71

Historians have sought to explain this development within Indian Buddhism as a

reaction either to the resurgence of Hinduism in India, or to the Islamic incursions and
persecutions, and it is indeed interesting to note that the Indian Buddhist tradition itself
describes Anuttara Yogatantra as the preferred mode of practice during periods of social
degeneration.4 Nonetheless, the internal evidence of the rNying-ma philosophical systems
and literature reveals that such differences as there are arose, not in an abrupt divisive
manner, but rather by way of a gradual transition.
From the Tibetan perspective also, it is clear that the events of 838 act as a
watershed not because the earlier lineages of Padmasambhava, Vimalamitra and
Buddhaguhya disappeared but because the eleventh-century Tibetans responsible for the
restoration of Buddhism were exposed to a new generation of Indian mastersNaropa,
Gayadhara, Virpa and Atia, with their own distinct lineages, in addition to the earlier
traditions. Tibetan polemists of the rNying-ma school have sometimes sought to
distinguish the ancient and new translations on the basis of their translation style rather
than their textual content, attributing the earlier to be meaningful and the later as lexical
interpretations of the Sanskrit originals.5
While such differences in translation style are undeniably present, we know from
the sGra-byor-bam-gnyis that, with the notable exception of the inner tantras, the
Tibetan Buddhist terminology had been subjected to continuous revision since the time of
the religious kings.6 Again the process seems to be one of gradual transition. The
eleventh-century scholar Rong-zom-pa appears to have been well-versed in both ancient
and new translation styles.7 Conversely, certain proponents of the new translations
schools have attempted to disprove the authenticity of a number of rNying-ma tantras
which were not included in the bKa-gyur by Bu-ston because he had no access to their
Sanskrit originals. More often than not, these controversial arguments have been eclipsed
by the subsequent discovery of the relevant Sanskrit manuscripts.8
It is not the purpose of this paper to elaborate further on the Indian background or
the Tibetan polemics which gave rise to the rNying-ma/gSar-ma division. This would
merit more attention than could possibly be given in the space available. Rather, it will
focus on the concept of commitment or sacrament (dam-tshig, Skt. samaya) as found in
the tantras according to the rNying-ma tradition, and its relationship with the vows
maintained by adherents of the stras.

D.L. Snellgrove, op. cit., 149-50; bDud-joms Rin-po-che, The Nyingma School of Buddhism,
London & Boston, forthcoming, vol. 1, Book I, 26969.
5
G. Dorje, The Guhyagarbha-tantra and its Commentary, 17, notes 1213.
6
ibid., 28, note 55.
7
bDud-joms Rin-po-che, op. cit., vol. 1, Book II, 7039.
8
e.g. G. Dorje, op. cit., 6970.
72

2. Definition of commitment and the distinctions between commitment and vows

In his Indo-Tibetan Buddhism, Professor Snellgrove indicates the subtle range of
meanings which this term conveys.9 The Tibetan term dam-tshig, which translates the
Sanskrit samaya has three traditional definitions:10
Firstly, the commitments are laid down because benefits are obtained when they
are guarded. As it is said in the sGyu-phrul-rdo-rje:11
If one who has obtained empowerment and consecration
In the intention of the genuine (or committed) ones
Acts precisely according to the genuine commitments (dam).
Discrepancies will be burnt away (tshig).
Secondly, commitments are so-called because retribution is exacted when they
degenerate. The same text indeed says:
One who swears to maintain the injunctions
And blessings of the genuine (dam) ones
Naturally becomes accomplished.
One who transgresses these commitments (dam)
Is burnt away (tshig), even when belonging to
(The vehicle of) indestructible reality.
Thirdly, commitments are so-called because a conventional oath of allegiance is
taken. It says in the Jnottaraparipcch-stra:
In order that the most secret amazing
Doctrine of the conquerors
Might be firmly held without degeneration,
Pledges of allegiance are steadfastly upheld
In accordance with the teaching given by the guru.
This is explained to be commitment.
Giving this definition, kLong-chen Rab-byams-pa equates the keeping of
commitments with the virtuous nature of one who perseveres with body, speech and
mind not to break the pledges one has undertaken concerning the renunciation of what
should be renounced, and the attainment of what should be attained.12 As for the
distinctions between commitments (dam-tshig, Skt. samaya) and vows (sdom-pa, Skt.
savara): Vows depend on individual mental control, whereas commitments are held by
maintaining the Buddha-body, speech and mind without degeneration. In terms of the
observance of vows, there are three kinds, namely, the vows of prtimoka, Bodhisattva,
and awareness-holder (vidydhara). The eight classes of prtimoka vows should be
guarded by one who desires peace and

D.L. Snellgrove, op. cit., 16566.

G. Dorje, op. cit., ch. 19, 119091.
11
rNying-ma-rgyud-bum, vol. 15.
12
ibid., ch. 19, 1191.
10

73

happiness for oneself alone, for the duration of ones life. The Bodhisattva vows bind the
mind with moral discipline which has a dual purposethey cause one to attain
realization and extraordinary enlightened attributes through the gathering of the virtuous
doctrine, and they benefit others by actions on behalf of sentient beings. The vows of the
awareness-holders bring a great wave of benefit for others and transform conflicting
emotions (klea) into pristine cognition (jna). Commitments, on the other hand, are
said to guard the indestructible nature of the Buddha-body, speech and mind without
degeneration, so that one is consequently and exclusively devoted to activity for the sake
of others.

3. Commitments and vows and the views of the Nine Vehicles

The holding of vows and commitments has always had great importance for the social
cohesion of Buddhism and indeed for those intent on spiritual progression within it, to the
mystification of many a student confronted by the apparent trivialities which gave rise to
early schisms within the sagha.13 A study of the nine vehicles, into which the rNyingma-pa have traditionally classified the diverse levels of Buddhist experience, reveals that
at each sequence of the path, the philosophical view maintained by its adherents requires
the concomitant observance of specific vows or commitments.
The distinctions between the nine vehicles, or sequences of the vehicle, are
discussed in the many philosophical treatises of the rNying-ma school which focus on
philosophical systems.14 The present account is derived from Lo-chen Dharmar, gSangbdag-zhal-lung and bDud-joms Rin-po-che, bsTan-pai-rnam-gzhag who, following the
Anuyoga text sPyi-mdo-dgongs-pa-dus-pa, make a fundamental distinction between the
first three stra-based vehicles, i.e. rvakayna, Pratyekabuddhayna and
Bodhisattvayna, which control the cause of suffering (kun-byung dran-pai theg-pa),
adopting a causal approach to enlightenment or Buddhahood, and the last six or tantrabased vehicles which maintain the resultant view that Buddhahood is atemporally or
primordially attained, and realized as such by the removal of the obscurations covering
enlightened mind. As for the remaining six tantra-vehicles, the outer Kriytantra,
Ubhayatantra and Yogatantra are known as vehicles of the outer tantras of austere
awareness (phyi dka-thub rig-pai rgyud-kyi theg-pa), while the inner tantras of
Mahyoga, Anuyoga and Atiyoga are known as vehicles of overpowering means
(dbang-bsgyur thabs-kyi theg-pa). Lo-chen Dharmar emphasizes that this enumeration
of nine is itself provisional because the structure

13

e.g. The second council, on which see M. Hofinger, tude sur le Concile de Vail, Louvain, 1946,
and J. Nattier and C. Prebish, Mahsghika Origins: The Beginnings of Buddhist Sectarianism,
HR, 1977, 23772.
14
S.G. Karmay, Origin and Early Development of the Tibetan Religious Traditions of the Great
Perfection, PhD thesis, SOAS, University of London, 1986, 254314.
74

may be simplified, e.g. into the twofold classification of Hnayna and Mahyna, or
extended, e.g. by adding the mundane Manuyayna or Devayna as preliminaries.15
Despite the rather stereotyped presentation, we should bear in mind that the nine
sequences of the vehicle are viewed as a single dynamic momentum, because a rvaka
or Solitary Buddha in one situation could well be a Bodhisattva or yogin in another.

4. Commitments and vows in the Causal Vehicles

rvakayna: This vehicle is entered when one disillusioned with suffering comprehends
the four truths and aspires towards peace and happiness. In view, selflessness of the
individual (pudgalanairtmya) is realized, but selflessness of phenomena
(dharmanairtmya) remains unrealized because the substrata of objective atomic
particles and subjective time moments of consciousness are held to be ultimately real and
discernible, either in association with each other, as is postulated by the Vaibhika, or
forming a single sensum, as held by the Sautrntika. While in meditation the rvaka
practises tranquillity (amatha) and higher insight (vipayan) in order to experience
sixteen aspects of truth in sixteen successive moments, the spiritual path in question is
characterized above all by renunciation (spang-ba). In order to become a stream-winner,
or indeed an arhat, the rvaka is required to observe a number of vows, corresponding to
the eight classes of prtimoka. In the case of more advanced renunciates, the twelve
ascetic virtues (dvadaadhtagua), which entail further restrictions on diet, attire and
residence, will also be observed.16
Pratyekabuddhayna: This vehicle is entered when one of highest, mediocre or inferior
provisions is born without a master and alone comes to meditate on and understand
dependent origination (prattyasamutpda) and its reversal in a single sitting. In view, the
selflessness of the individual and the selflessness of objective atomic particles are
realized, but the ultimate reality of subjective time moments is maintained. For this
reason, they are said to be realized in one and a half parts of selflessness. In order to
become a full-fledged arhat with supreme bliss of purpose, the Solitary Buddha will
naturally maintain the same monastic vows which the rvaka strenuously observes since
his meditative abilities are enhanced by the fusion of tranquility, higher insight and
dependent origination. However, he is distinguished from the samyaksambuddha by an
inability to verbalize his realization for the sake of others, and instead must rely on
communication through symbolic gestures.17
Bodhisattvayna: This vehicle is entered by one who seeks enlightenment (bodhi)
through great compassion (mahkaru) in order to remove the sufferings of all

15

Lo-chen Dharmar, gSang-bdag zhal-lung, 1719.

bDud-joms Rin-po-che, op. cit., vol. 1, Book I, 22327.
17
ibid., vol. 1, Book I, 22731.
16

75

sentient beings, and who realizes through discriminative awareness (praj) the two
aspects of selflessness and the quiescence of all conceptual elaboration (prapaca). The
Bodhisattva comprehends the two truths (dvisatya) within the context of the Mahyna
philosophical schools. From the standpoint of Vijnavda abhidharma he equates the
relative truth (samvtisatya) with the subject-object dichotomy, and ultimate truth
(paramrthasatya) with the consciousness of the ground-of-all (layavijna), although
the objects of consciousness may or may not be independent existents. Through
Svtantrika reasoning, he analyses the relative to be either correct or false in terms of its
causal effectiveness, and the ultimate to be with or without synonyms. Through
Prsagika dialectic, he realizes the relative to comprise both worldly and non-worldly
appearances, and the ultimate to be without synonyms. Then, through the experiences of
Great Madhyamaka, he comprehends the relative as the disharmony of the abiding and
apparitional natures, i.e. of emptiness and the appearance of Buddha-attributes, and the
ultimate as the harmony of these abiding and apparitional natures.18 To sustain this
compassionate and discriminating vision, and to make possible the attainment of
samyaksambuddhatva, the Bodhisattva meditates on the so-called thirty-seven aspects of
enlightenment (bodhipakadharma). As a discipline, he is required to gather virtue (i.e.
the six pramit), to act on behalf of sentient beings, and to control malpractices by
avoiding the root downfalls and observing the Bodhisattva vows which lead to the
cultivation of enlightened mind (bodhicittotpda).
Second turning stras and commentaries enumerate nineteen or twenty root
downfalls (rtsa-ltung bcu-dgu/nyi-shu), as follows:19 There are five root downfalls certain
for kings, namely, to steal the wealth of the Three Precious Jewels, to punish disciplined
monks, to direct a renunciate away from his or her training, to commit the five inexpiable
sins, and to hold wrong views; five for ministers, namely, to subjugate towns,
countryside, citadels, cities, and provinces; eight for ordinary persons, namely, to teach
emptiness to those of unrefined intelligence, to oppose those who enter into the greater
vehicle, to join the greater vehicle having rejected the prtimoka vows, to uphold or
cause one to uphold the vehicles of rvakas and Solitary Buddhas, to praise oneself and
depreciate others, to speak of ones own patience as profound, to give and receive the
wealth of the precious jewels, and to give the riches of tranquillity to those who are loudmouthed; and one which is common to all, namely, to abandon the enlightened attitude of
aspiration (smon-pai sems-bskyed). The enumeration of twenty root downfalls

18

gnas-snang mthun mi-mthun. On the philosophical schools, from the rNying-ma standpoint, see
bDud-joms Rin-po-che, op. cit., vol. 1, Book 1, 15686.
19
kagarbha-stra, as summarised in ntidevas iksamuccaya, ch. 4; and kLong-chen-pa,
Grub-mtha mdzod, 200.
76

includes that which occurs when the enlightened mind of undertaking or entrance (jugpai sems-bskyed) is also abandoned.
The cultivation of enlightened mind requires observance of the six or ten
transcendental perfections (pramit), but monastic vows, as observed by rvakas and
Solitary Buddhas, are optional.20 As for the third turning commentarial tradition, the
Bodhisattva-savaravimaka mentions forty-six transgressions (nges-byas zhe-drug),21
which include thirty-four contradicting the gathering of virtuous doctrines and twelve
contradicting activity on behalf of others. The former comprise seven contradicting
liberality, nine contradicting moral discipline, four contradicting patience, three
contradicting perseverence, three contradicting concentration, and eight contradicting
discriminative awareness. The latter comprise those transgressions which separate living
beings from both general and particular acts of benefit.

5.Commitment in the Outer Tantras

In Buddhism, the term tantra is well known to refer not only to the various cycles of texts
which bear the name tantra, but to the three continua of ground, path and result which
underlie the experience of these same texts, indicating the abiding nature of reality (gnaslugs) which is to be realized, the means of realizing it (thabs), and the culminating
Buddha-body (sku) and pristine cognition (ye-shes) resulting from that realization. It is
this structure of ground, path and result around which the tantra-texts, both rNying-ma
and gSar-ma are developed.22
Tantra-texts are said to surpass the stras for a number of reasons, notably
because they are able to make the actual result of Buddhahood into the path and can bring
relative appearances into the path without renouncing them. They are therefore
distinguished by their limitless approaches, by their diverse skilful means, their swiftness,
bliss and so forth. The Rali-Cakrasavara-tantra enumerates fifteen such ways in which
the tantras are superior.23
A number of commentaries on the tantra-texts have identified ten categories or
topics, which form the actual subject matter of the tantras (rgyud-don-gyi dngos-po byebrag-tu phye-nas bcu).24 These are a view of the real (de-kho-na-nyid lta-ba), determinate
conduct (la-dor-ba spyod-pa), maala array (bkod-pa dkyil-khor), successive gradation
of empowerment (rim-par bgrod-pa dbang), commitment

20

e.g. kyars response to Nyang-rals sons in bDud-joms Rin-po-che, op. cit., vol. 1, Book II,
758.
21
See M. Tatz, trans., Candragomins Twenty Verses on the Bodhisattva Vow and Its Commentary,
and Jam-mgon Kong-sprul, Shes-bya kun-khyab mdzod, vol. 2, 114117.
22
bDud-joms Rin-po-che, op. cit., vol. 1, Book I, 26367, and for a specific example, The
Guhyagarbha-tantra and its Commentary, 5961.
23
Enumerated and explained in bDud-joms Rin-po-che, op. cit., vol. 1, Book I, 24356.
24
G. Dorje, op. cit., 114-23, which is based on Mi-pham rGya-mtsho, gSang-snying spyi-don od-gsal
snying-po.
77

which is not to be transgressed (mi-da-ba dam-tshig), enlightened activity which is

displayed (rol-pa phrin-las), fulfilment of aspiration (don-du gnyer-ba sgrub-pa),
offerings which brings the goal to fruition (gnas-su stobs-pa mchod-pa), unwavering
contemplation (mi-g.yo-ba ting-nge-dzin) and mantra recitation (zlos-pa sngags),
accompanied by the seal which binds the practitioner to realization (ching-ba phyagrgya). Within this structure, commitment is regarded as an important linch-pin, enabling
one who has received the appropriate empowerments (abhieka) to focus more clearly
upon the goal.
Kriytantra: This vehicle is entered by a yogin who receives the water and crown
empowerments, and who then performs ablutions and regards the meditational deity (yidam), sacraments of offering (dam-rdzas) and mantras as objects of purity. In view, he
understands ultimate truth to be mind-as-such (sems-nyid) or pristine cognition (jna)
beyond conceptual elaboration (prapaca) and extremes, and the relative truth to
comprise both the correct relative appearances of divine maalas and the incorrect
relative appearances of mundane perception. In order to attain the status of a rigs-gsum
rdo-rje dzin-pa within sixteen life-times, the yogin meditates symbolically on himself as
the visualized samayasattva, into which the jnasattva will be invited to confer
accomplishments, and non-symbolically on the alternation of the two truths.25
This practice is sustained by the basic commitments (rtsa-bai dam-tshig) not to
abandon the Three Precious Jewels, the bodhicitta, the mantras and seals, the vajra and
bell, or the deity and the guru; and the ancillary commitments (yan-lag-gi dam-tshig) not
to eat meat/garlic/radishes, drink ale, or sleep on a high bed. In other words, by relating to
the meditational deity as a servant to a master, the Kriytantra emphasizes external
observances of diet, cleanliness, attire, astrology and so forth.26
Ubhayatantra: This vehicle is entered by a yogin who receives the vajra, bell, and name
empowerments. Its discipline and conduct conform to Kriytantra, and its meditation to
Yogatantra, although there is more emphasis on the recitation of mantra. In result, the
yogin is said to become a rigs-bzhi rdo-rje dzin-pa.27
Yogatantra: This vehicle is entered by the vajrcrya empowerment. In view, the yogin
understands ultimate truth to be inner radiance (prabhsvara), emptiness (nyat), and
absence of conceptual elaboration (niprapaca) with reference to all phenomena, and
relative truth to comprise both the correct appearances of the vajradhtumaala and the
incorrect mundane appearances. In order to become a rigs-lnga rdo-rje dzin-pa within
three lifetimes, he meditates symbolically on the generation of the samayasattva by
means of five awakenings (pacbhisambodhi)

25

bDud-joms Rin-po-che, op. cit., vol. 1, Book I, 34852.

ibid., vol. 1, Book I, 35051.
27
ibid., vol. 1, Book I, 35253.
26

78

and on the invitation of the jnasattva to confer accomplishment by means of the four
miracles (cho-phrul bzhi).28 The union of samayasattva and jnasattva is then secured
by means of the four seals (phyag-rgya bzhi)the great seal (mahmudr) which secures
the body as Buddha-body and the mirror-like pristine cognition (darajna), the
doctrinal seal (dharmamudr) which secures speech as Buddha-speech and the pristine
cognition of discernment (pratyavekanajna), the commitment seal (samayamudr)
which secures klea consciousness (kliamanovijna) as Buddha-mind and the pristine
cognition of sameness (samatjna), and the action seal (karmamudr) which secures
the five senses as Buddha-activity and the pristine cognition of accomplishment
(ktyupasthnajna). In non-symbolic meditation, there is no dichotomy between nonsymbolic ultimate reality and its divine apparition or blessing.
In order to sustain this meditation, the yogin cultivates bodhicitta as in the
Bodhisattvayna, and observes the fourteen commitments associated with the pacajina,
along with the commitments to avoid contact with those who have violated their own
commitments. The fourteen commitments are to guard against: disparaging the teacher;
transgressing the three levels of vows; hostility to vajra brothers and sisters; rejection of
loving kindness for sentient beings; abandoning the enlightened mind; disparaging ones
own doctrine or that of others; divulging secrets to the immature; abusing the five
components (pacaskandha) which are primordially pure; narrow views concerning the
intrinsic purity of phenomena; lack of compassion for evil beings who harm the doctrine;
application of conceptual thought to wordless natures; belittling those who have faith;
violating the commitments that have been undertaken; disparaging women, the source of
discriminative awareness.29
We can see from this synopsis of the six lower vehicles that the observance of
vows and commitments is intimately connected with the view and meditation upheld at
each stage of development. The rvakayna which emphasizes the selflessness of the
individual requires vows of arduous renunciation. The Pratyekabuddhayna with its
subtler appreciation of dependent origination can sustain the same vows with less effort.
The Bodhisattvayna with its base in compassion and discriminative awareness requires
commitments of total sacrifice for the sake of others. The Kriytantra requires the yogin
to adopt an attitude of purity vis--vis the meditational deity, and Yogatantra requires the
yogin to reject those views which distance him from union with the meditational deity,
and at the same time maintain purity by avoiding contact with violators of the
commitments.

28

On the pacbhisambodhi and cho-phrul bzhi, see bDud-joms Rin-po-che, op. cit., vol. 1, Book I,
35357.
29
Avaghoas Mlpattisagraha, as quoted by F. Lessing and A. Wayman, Mkhas Grub Rjes
Fundamentals of the Buddhist Tantras, The Hague & Paris, 1968, 328.
79

6. General commitments of the Inner Tantras

The three classes of inner tantras are distinguished from the outer tantras in a number of
ways. sGro-phug-pa identifies five such distinctions, namely: the inner tantras identify
mind-as-such with the Great Identity (bdag-nyid chen-po, i.e. the central Heruka of the
maala); they consider attainment to be intrinsically present, they are entered through
the three higher empowerments; they retain elements of sasra through skilful means;
they may lead the yogin to the desired result in one lifetime.30 In particular, Mahyoga
emphasizes the ground, utpattikrama and ritual activities. Anuyoga emphasizes the path,
sampannakrama and contemplation (samdhi), while Atiyoga emphasizes the result, great
perfection (rdzogs-pa chen-po, Skt. mahsandhi), and view.31
The general commitments upheld by yogins of the inner tantras comprise five
basic observances which elaborate on the basic commitments followed by adherents of
the Kriytantra, with the addition of ten ancillary commitments.32
Basic commitments are said to be those which, in the manner of the roots of a
tree, are the source of attributes when they are guarded and cultivated but not
when they are unguarded. Ancillary commitments are said to be the skilful means
and aids through which those (basic commitments) are guarded.
The five basic commitments are: not to abandon the unsurpassed vehicle; to
venerate the guru; not to interrupt the mantras and seals; to have loving kindness for
those who enter the Mahyna; not to divulge secret truths to others.33
The first basic commitment, not to abandon the unsurpassed vehicle, means that,
according to the ground, the yogin should not abandon all sentient beings because they
abide primordially as the Three Precious Jewels. According to the path, he should not
abandon the two aspects of enlightened mind (bodhicitta), ultimate and relative, which
unify the creation stage (utpattikrama) and the perfection stage (sampannakrama); and
according to the result, he should not abandon the three Buddha-bodies (trikya), their
Buddha-speech which comprises the stra and tantra-texts along with their respective
realizations, and the Buddhist community.
The second basic commitment, veneration of the teacher, concerns the various
categories of teacher and the means of veneration. The former are five in number,
namely, those teachers who guide or introduce one of the doctrine or the renunciate vows
(pravrajy); who liberate ones mundane consciousness by

30

On these distinctions, see bDud-joms Rin-po-che, op. cit., vol. 1, Book I, 346-48.
ibid., 358-59.
32
G. Dorje, op. cit., ch. 19, 1191.
33
The source for this section on the five basic commitments is The Guhyagarbha-tantra and its
Commentary, ch. 19, 11921204.
31

80

expounding the doctrine; who teach the esoteric instructions (upadea) and, as spiritual
benefactors (kalynamitra), grant transmissions (gama) concerning the cultivation of
enlightened mind; who repair violations of the commitments; who grant the commitments
and empowerments. In addition to these types there is also explained a sixth, namely, the
teacher whom all venerate as a holy being, and from whom all obtain a little kindness.34
As for the means of veneration, all teachers in general are to be regarded as a
genuine object of offering. Consequently:35
The self is considered as a disease, the teaching as a medication, its experiential
cultivation to be efficacious in the mind, and the spiritual benefactor to resemble a
learned physician.
The master of indestructible reality (vajrcrya), however is to be venerated in
three special ways: as a fourth precious jewel, as the Buddhas equal or superior.
The third basic commitment, that the mantras and seals should not be interrupted,
has two aspects. The mantras are a skilful means for attaining the accomplishments of the
different deities, in which recitation is performed in five sequential steps.36 This
attainment is then actually secured by the four seals, namely, the doctrinal seal of
thought, the commitment seal of speech, the action seal of transformation, and the great
seal of the deitys body:37
The best way not to interrupt these (mantras and seals) is when their continuity
resembles the flow of a river, the mediocre way is when one abides in sessions (of
meditation) by day and night, and the worst is when they are not interrupted for a
set period of months or years (in retreat), beginning on the fifteenth or eighth day
of the month.
The fourth basic commitment is loving kindness for those siblings who enter the
Mahyna. In general, six types of sibling are recognized: all sentient beings who are said
primordially to be ones own universal siblings; all adherents of Buddhism who are
described as siblings in the unique teaching; all harmonious siblings who hold
identical views and conduct; dear siblings who have the same teacher; close siblings
who listen to the doctrine together; and intimate siblings or vajra-brothers who receive
empowerment together.38
The fifth basic commitment, not to divulge secret truths, indicates that the
profound view, meditation, conduct and result of the secret mantras are most
34

Vilsavajra/Llvajra, Clarification of Commitments (Samayavivyakti), TTP, vol. 83, 147.2.4

149.4.8.
35
G. Dorje, op. cit., ch. 19, 1196.
36
These are enumerated in G. Dorje, op. cit., ch. 9, 850. They concern the visualization of oneself, the
deity, the mantra, its recitation and the emanation and absorption of light.
37
G. Dorje, op. cit., ch. 19, 1197.
38
ibid., ch. 19, 1198.
81

secret to unworthy recipients. As stated in the Propensity for the Commitments (TTP.
4745), profound view, profundity of conduct, retention of the deitys name, and signs of
accomplishment are described as the four general secrets. The places, times, assistants,
and sacraments for attainment (sgrub-rdzas) are the four interim secrets. The
sacraments of commitment (dam-rdzas) including the first fruits of offering and tgtor-ma
are the worthy secrets which it is improper to display.39 The action of maintaining
secrecy is also known as entrustment (gtad-pa).
Concerning those from whom secrecy should be maintained, the same text says
they should be kept secret:40
From all those whose commitments have been violated,
From those who have erred in their commitments,
And those without commitments
Who have not seen the maala,
Whether they are intimate or not.
If such secrecy is kept, then, as is said in the Jncaryadyuticakra.41
The mind should not think to teach,
The body performs all its activities covertly,
And speech should not be expressed,
Even though one has the tongue of indestructible reality.
Among these basic commitments, the first three are commitments to be attained
because they enable the yogin to attain extraordinary enlightened attributes, while the
latter two are commitments to be guarded because they guard against contradictions
respectively of the teachers mind, and of the secret mantra.
Now these five basic commitments are said to be inherent in the three
fundamental Mahyoga commitments of Buddha-body, speech and mind.42 For example,
in the commitments not to abandon the unsurpassed, to venerate the teacher and to have
loving kindness for siblings, the yogin should practise veneration through body, praise
through speech and respect through mind. Then, the commitment not to interrupt the
mantras and seals also utilizes the body, speech, and mind in their entirety; while the
commitment to secrecy is itself maintained by activities of body, speech, and mind.
The ten ancillary commitments comprise five not to be abandoned (spang-bar mibya-ba lnga) and five to be acquired (blang-bar bya-ba lnga).

39

The Tibetan terms for these ten categories are given in The Guhyagarbha-tantra and its
Commentary, ch. 19, 1201, note 6; see also Lo-chen Dharmar, gSang-bdag dgongs-rgyan, 41920.
40
Vilsavajra/Llvajra, Samaynuayanirdea, TTP, vol. 83, 149.4.8150.3.8.
41
Quoted in The Guhyagarbha-tantra and its Commentary, ch. 19, 1202.
42
On these commitments of buddha-body, speech and mind, see below, 83.
82

The former are the five conflicting emotions (pacaklea), which are not to be
abandoned for three reasons, corresponding to the ground, path and result.43
Firstly, they are known to be the seeds of Buddha-nature and to be naturally pure,
without having to be accepted or rejected. Ultimately they are without inherent existence,
and even on a relative level they need not be abandoned since they resemble a mirage.
Secondly, when retained by skilful means, conflicting emotions assist the path. This is
because the yogin who knows them to be the nature of the five pristine cognitions
(pacajna), resorts to a short path by experientially cultivating them according to the
esoteric instructions. When known correctly, they do produce the enlightened attributes
of Buddhahood because they are experienced as pristine cognition. Nor do they bind
beings within negative existences, because the yogin is utterly unblemished. The same
object endowed with conflicting emotion, which is seen as an object of renunciation by
rvakas and Solitary Buddhas, is seen as an object of conduct by Bodhisattvas, as an
object of purity by ordinary adherents of the secret mantras, and it is seen as an object to
be acquired by the uncommon yogins:44
For example, just as things fashioned from gold have a single nature but are dirty
when made into a wash-basin, clean when made into a trough, ornamental when
made into bracelets, and become receptacles of offering when made into the
representative images of deities, although (conflicting emotions) are conditionally
perceived in that way by the different vehicles, their essence is primordially pure
pristine cognition. This is truly why the nature of conflicting emotions which are
transformed into the realities of ground, path and result are not to be renounced.
Thirdly, conflicting emotions are not to be abandoned because they themselves
are described as the resultant Buddha-bodies and pristine cognitions:45
In this way, delusion is the commitment of Vairocana, who has the pristine
cognition of realitys expanse and belongs to the enlightened family of the
Tathgata. Hatred is the commitment of Akobhya who has the mirror-like
pristine cognition and belongs to the enlightened family of the vajra. Pride is the
commitment of Ratnasambhava who has the pristine cognition of sameness and
belongs to the enlightened family of gemstones. Desire is the commitment of
Amitbha who has the

43

The source for this section on the ancillary commitments is The Guhyagarbha-tantra and its
Commentary, ch. 19, 120411.
44
ibid., ch. 19, 12089.
45
ibid., 120910, based on Vilsavajra, Samayavivyakti.
83

pristine cognition of discernment and belongs to the enlightened family of the

lotus. And envy is the commitment of Amoghasiddhi who has the pristine
cognition of accomplishment and belongs to the enlightened family of activity.
Therefore, it is taught that from the moment of the appearance of the five
conflicting emotions which arise in all the minds of living beings they are not to
be abandoned because they are present as the Buddha-bodies and pristine
cognitions.
As for the five commitments to be acquired: these are the five nectars
(pacmta). They are to be acquired for four reasons: Firstly, they are actually regarded
as a display of primordial reality where there is neither acceptance nor rejection.
Secondly, they are the essential nature of the Buddhas of the five enlightened families
(pacakula), as it is said in the sGyu-phrul thal-ba:
The five nectars are the bodies (flesh),
Excrement, urine, and seminal fluids
Of the five enlightened families.
Thirdly, they are sacraments for attaining accomplishment (sgrub-rdzas),46
fourthly, dependent on the five nectars and the five meats, the kins are gathered and
accomplishments are approached.47 The nectars are therefore to be acquired because they
assist the yogin in the conduct of ascetic discipline.
The five basic and ten ancillary commitments, which have been described are said
to have three hundred and sixty branches, of which one hundred and sixty are derived
from the five basic commitments, and two hundred are derived from the ten ancillary
commitments.48 Furthermore, the branches of these commitments are also regarded as
inconceivable since they may equal the number of ideas accumulated by sentient
beings.49

7. Commitment in Mahyoga
This vehicle is entered by the yogin who receives the eighteen aspects of empowerment,
known as beneficence, ability and profundity (phan-nus-zab-dbang). These comprise
ten empowerments of beneficence which employ the sacramental objects of crownornament, diadem, rosary, armour, victory-banner, seals, parasol, vase, food and drink,
and the five nectars (pacmta) as well as five inner empowerments of ability (nangnus-pai dbang-lnga) and three secret empowerments of profundity (gsang-ba zab-moi
dbang-gsum).50 In view, the
46

i.e. through the sacrament of the pacmta, accomplishments (siddhi) are received. See ibid., 1210.
The five meats (pacamsa) are those of human flesh, elephant, horse, dog, and cow (or in other
traditionslion and peacock). See ibid., 1211.
48
See ibid., 121113, where each basic commitment is subdivided into thirty-two aspects according
with prajopya and buddha-body, speech and mind, attributes, activities, etc., and where each
ancillary commitment has twenty aspects, comprising prajopya, the pacajna and the
pacadhtu.
49
See ibid., 1213, which cites verses from Vilsavajra, Samayayiyyakti.
50
The Guhyagarbha-tantra and its Commentary, ch. 10, 86581.
47

84

yogin realizes ultimate truth to be awareness (rig-pa) that is spontaneously present

without conceptual elaboration, and the relative truth to be the mental energy of this
awareness (rig-pai rtsal) manifesting as a maala of Buddha-body and pristine
cognition. Then the non-dual truth (advayasatya) is realized to be the unity of
emptiness/awareness and pure appearance.51
In order to realize the pacakya, either in one lifetime or in the bar-do (Skt.
antarbhava), the yogin applies the non-symbolic contemplation of ultimate reality and
symbolic meditation of the two stages. The latter comprises the creation stage
(utpattikrama) and the perfection stage (sampannakrama). In the creation stage, the
indivisibility of deity and thought is gradually visualized through three contemplations
(ting-dzin gsum), namely great emptiness (stong-pa chen-po) which purifies death, great
compassion (snying-rje chen-po) which purifies the intermediate states after death (bardo) and the seals and attainment of the maala-clusters (phyag-rgya dang tshom-bu
tshogs-sgrub) which purify life itself, from conception to old age. In the perfection stage
visualization focuses on the energy channels, currents and seminal points (rtsa-rlung
thig-le) in the bodyeither in the upper door of ones own body (rang-lus steng-sgo) or
the lower door (sexual centre) of ones partners body (gzhan-lus og-sgo).52
Twenty-eight commitments (rnal-byor chen-poi dam-tshig nyi-shu-rtsa-brgyad)
are upheld in relation to this meditative practice, renunciation and attainment.53 These
are, namely, the three basic commitments of Buddha-body, speech and mind, and twenty
five ancillary commitments, namely, five aspects of sbyor-sgrol, which are to be
practised (spyad-par bya-ba), five conflicting emotions of desire, hatred, delusion, pride,
and envy which are not to be renounced (spang-par mi-bya-ba), five nectars of semen,
blood, urine, excrement, and flesh to be adopted (blang-bar bya-ba), five aspects to be
known (shes-par bya-ba), namely, the components, elements, sense-objects, sacraments
of the five meats which are considered taboo by mundane beings, and propensities in
their pure nature, and five aspects to be attained (bsgrub-par bya-ba), namely, Buddhabody, speech, mind, attributes and activities.

8. Commitment in Anuyoga
This vehicle is entered by one who receives the thirty six basic and eight hundred and
thirty one ancillary empowerments which are derived from all nine sequences of the
vehicle, including the stras.

51

bDud-joms Rin-po-che, op. cit., vol. 1, book I, 35963.

See the section on the paths of Mahyoga in bDud-joms Rin-po-che, op. cit., vol. 1, Book I, 278
81.
53
These twenty-eight commitments are enumerated in Vilsavajra, Samayavivyakti, 14748, and
Jam-mgon Kong-sprul, Shes-bya kun-khyab-mdzod, vol. 2, 182185.
52

85

The view of Anuyoga is that all phenomena are the primordial maala of
Samantabhadr (ye ji-bzhin-pai dkyil-khor), the uncreated awareness if the
spontaneously present maala of Samantabhadra (rang-bzhin lhun-grub-kyi dkyilkhor), and the supreme bliss of their offspring is the fundamental maala of
enlightened mind (byang-chub sems-kyi dkyil-khor), without duality of expanse and
pristine cognition. In order to realize the body of supreme bliss (mahsukhakya) with its
twenty-five Buddha-aspects in one lifetime, the yogin practices meditation according to
the path of skilful means (thabs-lam) or the path of liberation (grol-lam). The former
utilizes the energy channels, currents and seminal points either with reference to ones
own body or in union with a partner, and the latter comprises the non-conceptual
contemplation of reality (dharmat) and symbolic contemplation of the deities, who are
said to appear instantly in the manner of a fish leaping from the water.54
To sustain this practice, Anuyoga requires the nine enumerations of commitments
which are described in the sixty-sixth chapter of the mDo dgongs-pa dus-pa. These are
derived from all the nine vehicles in common, and when abridged they comprise
commitments with and without limits to be guarded:55
(a) four commitments definitive to the important Anuyoga stras (gal-mdo nges-pai
dam-tshig bzhi), namely, purity of body, speech, mind and the entire perceptual
range.
(b) twenty-eight common commitments (thun-mongs-gi nyi-shu rtsa-brgyad) which
are identical to the commitments of Mahyoga.
(c) four superior commitments (lhag-pai bzhi) which derive from Atiyoga, namely,
there are no limits to guard because the essence of commitment is free from
transgression and violation; there is an attitude of apathy and evenness because
the forms of the subject-object dichotomy have been transcended; all diverse
commitments are gathered in the single expanse of mind-as-such; there is
commitment to reality (dharmat) itself.
(d) twenty-three commitments relating to ascetic discipline (brtul-zhugs-kyi nyergsum), which sustain the paths of the inner tantras in general. Since these are
described in fine imagery, I cite them here in full:
(1) In the manner of a fox (va) who has been trapped, and turns away without
regard for life itself, having had a limb torn off, the yogin guards the
commitments even at the cost of life itself. This is the skilful means which
destroys disharmonious aspects and enters into the power of the commitments.

54

On Anuyoga, see bDud-joms Rin-po-che, op. cit., vol. 1, Book I, 28489, 36369.
These nine enumerations of commitments are all given in Jam-mgon Kong-sprul, Shes-bya kunkhyab mdzod, vol. 2, 188192.

55

86

(2) In the manner of the all-knowing horse (cang-shes) who knows everything
and swiftly encircles everything in a moment, discriminative awareness is
unimpeded discipline with respect to all things with individual and general
characteristics that can be known.
(3) In the manner of a Gyiling steed (gyi-ling) which roams anywhere with
great energy, therespectful body disciplines itself with perseverence and
without idleness in the dance, mudrs and exercises.
(4) In the manner of a rutting elephant (glang-chen spyod), who, incensed,
destroys whatever enemies appear without investigating them, one who knows
sasra and nirva to be indivisible performs conduct which destroys the
four enemies of view and conduct.56
(5) In the manner of a tiger (stag) whose aggressive spirit is fierce,
overbearing and hostile, the powerful discipline of heroic contemplation
which realizes the abiding nature performs rites of liberation (sgrol-ba) and
transference of consciousness (sakrnti) for those students who are
aggressive.
(6) In the manner of a great garua (khung-chen) who glides effortlessly
through the sky and discerns all without special regard, the view is one of
effortless conduct, realized in the indivisibility of the expanse and pristine
cognition.
(7) In the manner of a bear (dom) who terrifies and crushes whatever it
focuses upon without hesitation, one who has plumbed the depths of the view
and conduct of yoga is disciplined in the rites of abhicra and sexual union
without hesitation.
(8) In the manner of an ocean (rgya-mtsho) whose golden depths are
unmoved, is the discipline of firm unchanging mind which is able (to
understand) the profound secret meaning and experiential cultivation.
(9) In the manner of a dumb mute (lkug-pa gti-mug-can) who neither accepts
nor rejects, is the discipline which reaches the limit of discriminative
awareness, realizing selflessness by impartial meditative absorption.
(10) In the manner of unmoving Mount Sumeru (ri-rab mi-g.yo-ba) is the
discipline of skilful means which depends on the unwavering antidote of
unchanging loyalty to teacher and friends, and on contemplative absorption.
(11) In the manner of the vast and extensive sky (nam-mkha) which
accommodates everything without acceptance and rejection is the

56

lta-spyod-kyi dgra bzhi. Perhaps these are to be identified with the caturmra, namely, the four
demons of devaputra/kmadeya, klea, mtyupati, and skandha.
87

discipline which is warm, and hospitable to fraternal yogins and the conduct
which, without shunning the vehicles of sasra and nirva, remains within
this view and conduct of the Great Identity.
(12) In the manner of a thunderbolt (thog-geng) which falls and destroys is the
discipline which unimpededly destroys all enemies and obstacles by forceful
contemplation.
(13) In the manner of Vajrapi (lag-na rdo-rje) who destroys all who hold
erroneous views, the yogin performs the discipline through which, having
meditated on the wrathful deity, one cuts through and destroys these views
without hesitation.
(14) In the manner of a crow (bya-rog) who looks out for both enemies and
plunder at the same time, is the discipline of skilful means which perseveres
simultaneously in constant renunciation and acceptance.
(15) In the manner of an elephant (glang-chen) who plunges into water
without regard for wetness and dryness, one who has plumbed the depths of
the view and conduct of Great Identity practises without the duality of
renunciation and acceptance; and practises the four rites of enlightened
activity without discriminating among those who require training.
(16) In the manner of a friendless lion (seng-ge) who sits alone, is the
discipline which sustains the view and meditation by abiding in solitude after
renouncing those disharmonious associations in view and conduct.
(17) In the manner of a swan (ngang-pa) who easily associates without
marriage, so is the discipline which associates without ties and the skilful
means causing sentient beings to reach the happiness of liberation through
compassion and loving kindness.
(18) In a manner of a magician (sgyu-ma-mkhan) who constructs illusions,
one who meditates and teaches, having understood the components and
activity fields to be the apparitional maala of the conquerors, enacts
discipline through skilful means.
(19) In the manner of a pig (phag), who eats everything without discerning
purity and impurity, are the discipline and conduct of sameness, without
accepting and rejecting the five sacramental substances.
(20) In the manner of a jackal (ce-spyang), who likes to kill without
impediment, is the discipline of skilful means which liberates heretical
thoughts through compassion experienced in view and conduct, and then
perfects the provisions by arraying such (deluded) consciousness in an
uncorrupted (realm).

88

(e)
(f)

(g)

(h)

(i)

(21) In the manner of lightning (glog), which illuminates everything swiftly

and simultaneously, is the discipline which perseveres so that ones own and
others benefit be swiftly attained through experiential cultivation of the path.
(22) In the manner of a vulture (bya-rgod), who avoids the taking of life as a
moral discipline, is the discipline which delights in and sustains commitments
associated with the Great Identity (i.e. Heruka) but appears not to indulge in
other vehicles connected with austere observances.
(23) In the manner of a modest king (rgyal-po bag-ldan), who rules the
kingdom and dearly protects his subjects rather than himself, the yogin
performs acts of pure delightful discipline, protects living beings by realising
all things on behalf not of himself but of others and overpowers the kingdom
by the discipline which strives through skilful means to experience and realize
the indivisibility of the expanse and pristine cognition as supreme bliss.
twenty commitments concerning the attainments (sgrub-pai nyi-shu), which
internalize the fourteen commitments of Yogatantra in combination with the five
basic commitments of the inner tantras.
four of daily conduct (spyod-lam rgyun-gyi bzhi) which are associated with
Ubhayatantra and Kriytantra, namely, to abandon sleep which interrupts
concentration; to abandon alcohol; to propound the symbolic language of the
secret mantras; to destroy idleness.
four enemies to be destroyed (dgra-bzhi gzhom-pa) which are associated with the
Bodhisattvayna, namely, artificiality of view, meditation and conduct; prattle
about lower views and trainings; violations of basic and ancillary commitments;
deprivation of the goal through speculation and idleness.
five demons to be renounced (bdud-lnga spang-ba), which are associated with the
rvakayna and Pratyekabuddhayna, namely, insecurity caused by divisive
thoughts; laziness with respect to equanimity; caprice with regard to pleasure and
social diversions; the sharp sword of harsh speech; fierce wrathful disturbances.
Lastly there is the commitment of the view (lta-bai dam-tshig), which unifies all
the previous commitments.

9. Commitment in Atiyoga
This vehicle is entered by one who receives the empowerment of the expressive power of
awareness (rig-pai rtsal-dbang) in its elaborate and unelaborate aspects. In view, things
of sasra and nirva are primordial Buddhahood in the unique seminal point (thig-le
nyag-gcig) or dharmakya. Liberation occurs in this primordial Buddhahood without
hope or doubt. In order to attain the goal at the

89

present moment on this level of spontaneously perfect Samantabhadra, there are three
classes of meditationmental, spatial and esoteric instructional (sems-klong man-ngag gi
sde-gsum)the last of which includes the esoteric instructions of Cutting Through
Resistance (khregs-chod) and All-Surpassing Realization (thod-rgal), which actualize the
dharmakya respectively as the youthful vase body (gzhon-nu bum-pai sku) and the
body of great transformation / rainbow-light body (ja-lus pho-ba chen-poi sku).57
In conformity with this goal, conduct is without discipline, including the
commitments of nothingness, evenness, uniqueness and spontaneity.58

10. Integration of the Three Vows

Aware of the dynamic momentum formed by these diverse vehicles and the particular
ways in which their philosophical views are associated with the holding of commitments
or vows, rNying-ma authors have sought to integrate and present them as a unified
blueprint for the practitioner. Perhaps the best known work of this type is mNga-ris Pachens Ascertainment of the Three Vows (sDom-gsum rnam-nges). The present account is
derived from kLong-chen-pa, Phyogs-bcu mun-sel, Chapter 19, which concerns the
gathering of the three trainings of the prtimoka, Bodhisattva and mantra vows without
contradiction.
Despite the apparent contradictions between the commitments of guhyamantra
and the prtimoka vows of the rvakas who regard the four inimical defeats
(catuparjikadharma) as the basic transgressions to be avoided, and maintain ancillary
prohibitions on the drinking of ale and so forth, there is in fact no such contradiction:59
When the defeat of murder occurs, four elements are required to complete it
the ground or the human being, the thought of hatred which freely incites murder
when it is not bewildered in other respects, the main part which is the impeding of
the life-breath with weapons and so forth, and the aftermath or satisfaction. In this
context (of the secret mantras) however, when the rites of liberation are
performed, there is no defect at all because the aspect of the ground (of murder) is
incompletethe form (of the victim) is visualized as the syllable h, and
because it is essentially uncreated, there is no human being, non-human animal or
other creature. The aspect of the thought (of murder) is incomplete because one
has compassion which desires to separate (the

57

On Atiyoga see bDud-joms Rin-po-che, op. cit., vol. 1, Book I, 294345, 36972.
On the textual references for these four commitments of Atiyoga, see H.V. Guenther, Matrix of
Mystery, Boulder & London, 1984, 238. The terms nothingness and apathy are reminiscent of, but
quite distinct in meaning from their usage in laukikadharma. See bDud-joms Rin-po-che, op. cit., vol.
1, Book I, 6367.
59
G. Dorje, The Guhyagarbha-tantra and its Commentary, ch. 19, 121820.
58

90

being) from suffering and the attitude of a wilful murderer is not present; the
aspect of the main part (of the murder) is also incomplete because there is neither
an object to be killed, nor a subject which kills, so that there is neither life nor the
taking of life; and the aspect of the conclusion (of the murder) is incomplete
because afterwards there is no satisfaction motivated by hatred.
Similarly when stealing occurs, the foundation is the need to acquire property as
ones own possession. Here however (in the secret mantras), because oneself and
others are realized to be without duality, one takes property which is selfmanifesting in the manner of a dream, and rather than the need for the thought of
theft, in this context there is no thought of theft because (the property) is simply
the appearance of ones own mind, and is a self-manifesting display. As for the
main part (of the action), quite apart from taking possession of (the property) by
conceiving elaborately of its full value, that too is incomplete. There is no
conceptual elaboration because in the manner of an apparition or a dream there is
no duality. As for the aftermath (of the theft), rather than having the thought that
one has acquired something, here there is no apprehension of an obtained object
or an obtaining subject.
Again, when the defeat of sexual misconduct occurs, the foundation is the need
for another mans woman and so forth. However in this context (of the rites of
sexual union) the individual (yogins and yogins) are creatively visualized as
deities. Rather than the need for one who has inserted the penis into the vagina to
experience a climax (bags-rim das-pa), (here) these two organs are visualized as
a vajra and a lotus, and they are not therefore complete with independently
existing characteristics. Even the aspect of thought (associated with sexual
misconduct), namely, the desire to experience bliss with an unbewildered
attention is incomplete (in this context) because here (bliss) is known within the
three maalas.60 Therefore there is no defect.
Then, when lies are told, one has to convey a distorted perception to other beings,
as when saying that one sees a deity though one actually does not. In this context,
however, one knows that all phenomena are considered to be lies, and may be
beneficially expressed for the sake of others. Therefore, there is no defect.
Similarly when ale is drunk, it is transformed (through the secret mantras) into
nectar. The statement that it is improper to taste ale is not contradicted.

60

Tib. dkyil-khor gsum, i.e. those of the male consort, the female consort, and their coalescence.
91

In order to avoid these four inimical defeats within the rvaka tradition, it is
necessary to be a monk or nun, unbewildered by objects and undistorted in perception.
However, according to the guhyamantra, the prtimoka vows are not contradicted
because all phenomena in this instance are transformed into the maala of deities.
Nor are the Bodhisattva vows contradicted by guhyamantra although they oppose
the performance of acts harmful to sentient beings. The great compassion and skilful
means which desire to benefit others are retained by both the great spiritual warriors
(mahsattva) of the stras and the yogins who follow the tantras.
It is mind-control and purity of vision with respect to the conflicting emotions
which enable the yogin to gather and integrate all these levels of training. The rvaka
vows are gathered to bind the mind in order that the yogin might be released from his
own suffering:61
At the time when life is taken (by the rites of liberation), there are no nonvirtuous thoughts of hatred and so forth. Indeed, the vow which renounces killing
is actually subsumed in that rite in order to control ones own corrupt components
with their individual characteristics. The remaining (prtimoka vows) are
similarly incorporated. However coarse the conduct of skilful means may appear,
all vows of the rvakas are essentially and absolutely gathered because the
corrupt components with their individual characteristics are under control. For
example, when certain doctors prepare a cooling medication for the sickness of
fever, they are opposed by some but benefit comes swiftly through the cure.
Although these two seem contradictory there is in fact no contradiction. Likewise,
although the skilful means of the path appear contradictory, they are actually in
complete harmony with the renunciation of conflicting emotions and the
attainment of virtue in ones own mind.
Similarly, the Bodhisattva vows are gathered in the following way:62
At the time when one engages in any conduct, the moral discipline of control is
present because the wilful indulgence of ordinary conflicting emotions is
controlled. Dependent on this, the moral discipline which gathers virtuous
doctrines is present because the enlightened attributes of the path are further
increased. And, through that skilful means the moral discipline of action on behalf
of sentient beings is present because others are benefited and taken into ones
following.

61
62

G. Dorje, op. cit., ch. 19, 122122.

ibid., 1222.
92

Benefits and retributions

Those who have successfully guarded their commitments are said to accomplish all their
purposes, provisional and conclusive. As is said in the Array of the Three Commitments
(Dam-tshig gsum-bkod):63
One who keeps the commitments of secret mantras,
(The vehicle of) indestructible reality,
Will fulfil all aspirations,
And will always be guarded by the deities.
The conquerors, supreme conquerors and their sons too
Assuredly think of that one as their son.
There are inestimable attributes of contemplation,
And one achieves the genuine awareness of Samantabhadra.
Conversely, the same text describes the retribution exacted when commitments
have been violated or divulged to others with a defective attitude:
When the commitments have been transgressed,
Ones happiness consequently declines.
One experiences diverse unpleasant things
And abides for ten million countless aeons
In the suffering caused by terrestrial fires.
Restoration of commitments
On the question of the repairing or restoration of broken commitments, kLong-chen Rabbyams-pa says:64
According to the rvakas, it appears that degenerations can be restored seven
times if not kept secret, but if kept secret they cannot be restored even once, just
like a broken clay vase. To Bodhisattvas, (this reparation) resembles a broken
vase made of gemstones, because degenerations can be restored by relying on a
spiritual benefactor, just as a craftsman can (use the gemstones) to make a
container superior to the previous one. According to the secret mantras,
(degeneration) resembles a container of gemstones which has been broken, but
can be restored by its own power. It is as if that which has been destroyed is
rectified by its own creative energy without the need for a craftsman (mgar-ba
mkhas-pa).
Therefore, while rvakas claim there are no means of restoring broken vows, the
secret mantras apparently do possess the means capable of repairing violations of the
commitments. Purification is said to occur when one has practised fulfilment and
confession, attempting to apply ones own body, speech and mind to the

63
64

ibid., 121314.
ibid., 121516.
93

foremost commitments of Buddha-body, speech and mind, which have been violated and
broken. It says in the Jncaryadyuticakra:
If, when degeneration of commitments has occurred,
One understands the importance of pleasing ones venerable lord,
This (degeneration) becomes a source of fulfilment.
If one has abused ones guru and siblings,
And if, pleasing them during this very lifetime,
One is remorseful, with fervent confession,
That (abuse) becomes a supreme mode of fulfilment.
If one passes away from this life
Without having fulfilled (violations of commitments)
Relating to oneself or to ones guru and siblings,
One will transgress the bounds of fulfilment.
If ones pledge of Buddha-speech has been torn to shreds,
Meditate on oneself as Vajradharma,
And recite many hundreds of thousands of times
The pledge which one has lost.
If ones pledge of Buddha-mind has been torn to shreds,
Meditate on oneself as Vajrasattva,
And without speaking for three years,
Be well united in meditative concentration.
If ancillary pledges have been torn to shreds,
They are fulfilled by means of the different
Respective enlightened families (of deities).
Conclusion
Future detailed studies of the integration of commitments and vows will derive much
information from mNga-ris Pa-chens Ascertainment of the Three Vows (sDom-gsum
rnam-nges), the most important practical manual for the living monastic tradition of the
rNying-ma school, which maintains the Lower Tibetan Vinaya Lineage in O-rgyan
sMin-grol-gling and its branch monasteries. The present discussion provisionally
indicates some aspects of the relationship between commitment and view, and between
commitment and vow, within the context of the nine vehicles in general. In particular, it
clarifies that the sbyor-sgrol practices of Mahyoga are fundamentally in harmony with
the prtimoka and Bodhisattva vows. In view of the anxiety which has occasionally been
expressed regarding the public reaction to the treatment of the rites of liberation (sgrolba) in Part Five of bDud-joms Rin-po-ches rNying-mai chos-byung, which will shortly
appear in its English version, I would argue that the clarity of kLong-chen-pas
interpretation should remove such doubts and anxieties. While a very strong case can
undoubtedly be made for keeping secret from unworthy recipients the actual

94

means for attainments (sdhana) and esoteric instructions which enable the yogin to
practice such rites, the publication of general works which clarify these practices from the
theoretical standpoint of the view can help remove many misconceptions. Indeed, as the
biographies of the great lineage-holders of the bka-ma tradition outlined in Part Five of
this text themselves reveal, such attainments and abilities are the products of limitless
years or lifetimes of intensive application, rooted continuously in compassion, and rarely
seen in this world. They are the preserve of a minute number of yogins, whose
attainments of the nine vehicles are beyond the perceptual range of mundane beings such
as the beggarly author of this paper, who on account of his clouded thoughts has not the
slightest chance of reaching even the first Bodhisattva level.

95

Savti, Vyavahra and Paramrtha in the Akayamatinirdea

and its Commentary by Vasubandhu
Ch. E. Freeman

1. In defence of dichotomies
Dichotomies are rather out of fashion at the moment. The traditional divisions into sacred
and profane, religious and secular, mystical and scientific and so on are being dismantled:
carving up the world in this way is no longer thought to be very helpful methodologically
(or ontologically for that matter) and approaches are being adopted which stress either the
non-difference of things previously presented as different, or the multiplication of
categories to provide a much more complex scheme than that afforded by the old binary
oppositions. This process undoubtedly has some value in Religious studies especially
where the divisions of the investigator (the etic side) have imposed a structure on the
investigated (the emic side) which, in failing to take into account the divisions
consciously made within the investigated itself, is unfaithful to it. If this process is taken
too far, however, the baby is in danger of being thrown out with the bathwater. This is
certainly the case with the Buddhism of the great Mahyna stras and their
commentaries and in the following investigations the dichotomy will once more take
centre stage.
The reason for this is simply that such dual divisions are to be found as elements
of the inner presentations, or self-descriptions, both of the canonical literature of the
Mahyna and of its hermeneutical tradition. They are not only to be found but are of
central importance as heuristic or soteriological devices. Indeed, it may not be too much
of an exaggeration to claim that the entire system of philosphical speculation combined
with spiritual practice (not intended as a dichotomy) of the Mahyna rests, in the final
analysis, upon one crucial fundamental opposition: the distinction between savti, the
conventional, and paramrtha, the ultimate. Or, more specifically, between savtisatya,
conventional truth; and paramrthasatya, ultimate truth.
This distinction is already made in the Hnayna schools. The Vaibhikas, for
example, distinguish between conventional truths as those entities, such as pots etc.,
which are capable of being broken up, either physically or by a mental

97

operation, and ultimate truths as those which cannot be so treated, such as space, form
and so on.1 This rather narrow, scholastic usage, and the failure of the rvakas to realise
the profound implications of these ideas mark them out, according to Vasubandhu
wearing his Mahyna hat, as inferior in understanding. For the Mahynanists, who
invest the notions with much deeper soteriological meaning, savtisatya, very roughly
speaking, refers to the relative world of appearancesthings not seen directly as they are
but mediated through language and obscured by dualistic and realistic thinking;2 and
paramrthasatya is the emptiness of all dharmas of inherent, or independent, existence.
While there is reasonable agreement among the various philosophical schools
Yogcra, Mdhyamika, Svtantrika etc.on the subject of paramrthasatya discord
arises over the exact sense of savtisatya, in what way the concept may properly be
applied, and its relationship to paramrtha. Vasubandhu makes an interesting
contribution to this debate in the Akayamatinirdea-k, the gist of which will be given
in Part 2.
Generally, where the words savti and paramrtha are used (appropriately
inflected) without the accompanying satya they may be rendered in English adverbially
as conventionally and ultimately, or as nouns, the conventional, the ultimate. No
great significance need be attached to the separation, it is merely a kind of shorthand.
Given, however, that two semantically distinct concepts are in fact present in the
compound savtisatya it is possible for an interpreter, in certain contexts, to stress the
difference between mere savti and savtisatya. Candrakrti does this in his
Madhyamakvatra-bhy3 in order to make what is essentially a doctrinal point. When
Vasubandhu comments on the constituent elements of a similar compound,
vyavahrasatya, his primary concern is

Louis de La Vallee Poussin, LAbhidharmakoa de Vasubandhu, Paris, 192331, ch. 4, 139 ff.
Three definitions of savti are given in Candrakrtis Prasannapad Madhyamakavtti, 492, 10.
According to the third by savti is meant a symbolic system (saketa): the ordinary means of
communication of the world involving verbal expression and referent; cognition and cognised etc. In
the commentary to Akayamatinirdea-stra Vasubandhu operates mainly with this more linguistic
aspect of savti rather than the epistemological or ontological ones under which it refers to the way
things appear so as to conceal the ultimate truth.
3
Madhyamakvatra-bhya, TTP, vol. 98, 124-1-8. This brief, but obscure, passage bristles with
interpretational difficulties. What it appears to say is that whatever is taken to be ultimate by ordinary
people (inherent existence) is mere savti for the ryas, for it is the essential emptiness of things
which is ultimate. As for savtisatya, it is because it deceives that it is not paramrthasatya. The
point is probably that while ordinary people accept unquestioningly the apparent inherent existence of
things as their ultimate nature, i.e. accept as true that which is false, and so do not really possess the
truth, only the ryas, or Bodhisattvas, perceive the distinction between paramrthasatya and
savtisatya and can know their relationship. See Michael Broido: Padma dKar-po on the Two
satyas, JIABS, 8.2, 1985, 15.
2

98

to establish a formal principle of interpretation. In other words Candrakrti answers the

question why did the Buddha teach so and so? While Vasubandhu answers the question
what do the words vyavahra and satya mean when considered separately? The
principle can be applied to other Sanskrit compounds of an analogous structure, e.g.
satyavacana and will be shown in some detail later on.
It is not easy to find a single English word capable of translating adequately the
Sanskrit vyavahra whose meaning includes both the broad notion of established social
conventions and practices, in particular, in the context under consideration, conventions
associated with communication in the natural languages, whose operations are decided by
common consent; and convention in the more technical sense of the ability of a word to
convey a sense of the function of its referent.4 Broadly, then, it denotes conventions
used in communicative acts based on what is heard, seen, judged and cognised,5 i.e.
based on everyday experience involving sense perception and conceptual thought. It is
the means of articulation, which Vasubandhu firmly dubs the work of Mra.6 The
correlation between the terms savti and vyavahra in the k is sufficiently close
savti is usually explained in terms of vyavahrato warrant the assumption that the
treatment given to vyavahrasatya, which is outlined in the extracts provided in Part 2,
may be applied to savtisatya too.
The division of truth into savti and paramrtha carries with it another
opposition which, while present in the stras couched in the poetic and allusive style that
is typical of them, finds a more systematic working out in the Akayamatinirdea-k.
This is the opposition between what is expressible in language together with language
itself as the instrument of expression, and what is wholly beyond words: the effable
versus the ineffable.
This distinction came under attack in a previous paper presented at the Buddhist
Forum where the following argument (if I have understood it rightly) was advanced:
William James use of the term ineffable to describe one of the qualities peculiar to
mystical experience is unsatisfactory because many other types of experience can
legitimately claim to possess this qualitythe drug-induced, the aesthetic, the nuclear
physical and so on. So either the term mystical must be extended from its narrow
Jamesian field of reference, which limited it to the purely
7

Anne Klein: Knowledge and Liberation, New York, 1986, 184185.

Darutamatavijtavyavahra. Sandhinirmocana-stra, TTP, vol. 29, 4-5-6. I have not use the
English word convention to translate vyavahra in this paper partly to avoid confusionI have used
it for savtiand partly because I wanted something more general, hence communication which is
not really very satisfactory either. However, on the conventionalist view taken by Vasubandhu
communication is simply a matter of various, chiefly linguistic, conventions.
6
Akayamatinirdea-k, TTP, vol. 104, 190-5-8.
7
B. Sikls, Topics in Mysticism: Words, Science and Methods, unfortunately not included in this
volume.
5

99

religious, to include these too, or it must be admitted that ineffability is not peculiar to
mystical experience. In any case a formal typological division of experience into
ineffable and effable is inaccurate since it fails to take into account the pragmatics of the
situation. It is these which dictate whether a particular experience is expressible in words
or not. Ineffability claims, because they rest on a confusion, are not especially helpful in
trying to understand what the great religions have to say.
Does this negative evaluation hold when we come to examine what is meant by
savti and paramrtha in the Akayamatinirdea-stra, one of the most widely quoted
and influential of the Mahyna stras, and in its commentary? It is perhaps worth a
reminder here that Buddhism provides room for numerous presentations of the Doctrine
and even more numerous interpretations of it: Vasubandhus view is only one among
many on this extremely complex subject.
He defines the two terms very succinctly thus: savti is whatever is expressible
and thinkable; paramrtha is the opposite of this for it can be reached neither by
expressions nor by thought.8 As such it is often described using the Sanskrit word
anabhilpya (Tib. brjod du med pa) which does mean simply inexpressible,
ineffable.9 Samvti, on the other hand, is almost always connected by Vasubandhu with
linguistic activitythe communication of something via the vehicle of language which is
bound up with the series of mental operations we call thought. This preserves the
traditional association made in Indian philosophising between language and thought
given, of course, that their systems allow for a different type of experience from this
thought, the yogic or meditational, where the aim is to quieten all mental activity and
attain a state of calm concentration. Buddhism has always valued the meditative states
above mere ratiocination dependent on the use of language,10 these alternative states of
mind are rarely taken into consideration in any Western discussions of the
thought/language nexus. So the fact that once more we are faced with a dichotomy within
the object of investigationthis time between the effable and the ineffable, correlated
with the original distinction between savti and paramrthaobliges us to take it
seriously.
The word anabhilpya, however, is used with a narrow reference that the English
word ineffable lacks. The latter is used on a number of occasions when

Akayamatinirdea-k, TTP, vol. 104, 193-1-8.

F. Egerton, Buddhist Hybrid Sanskrit Dictionary, anabhilpya: inexpressible; that cannot be put
into words.
10
The Buddha manifests the Buddha-actions, foremost of which is preaching, without ever stirring
from his samdhi, but though he may make use of prapacas (discursive proliferations) and vikalpas
(conceptualisations) while delivering a stra his language and thought are purified and he does not fall
into the trap of hypostatisation that these two set for the ignorant.
9

100

anabhilpya, as a qualification of paramrtha, or any of its synonyms such as nyat,

tathat or nirva, would not be used. Its use in the texts is extremely specific; no other
things are so described, or could be, for reasons which I hope will emerge during the
course of this discussion.
It is necessary, then, to look first at some of the things or states which might be
said to be ineffable when English is spoken. For example, the general claim could be
made that all the objects that populate the inner and outer worlds are ineffable in the
sense that it is inherent in language that it can only ever be about these objects. All signs
point to, refer to, something beyond themselves and it is the function of language to
indicate, represent, whatever it might be from simple physical entities to the most highly
personal and profound experiences. It cannot be them. At first sight, given the gulf
according to this view between transcendent things and the words that designate them,
such a state of affairs appears ripe for assimilation to the duality of paramrtha (beyond
language) and savti (the sphere of language). But this would be to misconstrue what
exactly it is about things that is said to be paramrtha. Ultimate truth refers to their
emptiness, not to their thinghoodbe it the tableness of table or the quintessence of some
more abstract objects of thought or experiential states. Thinghood is precisely what is
denied of things in post-Prajpramit Buddhism. Or, to use an idiom more natural to
Buddhism, their essential nature is that they are empty of essential nature; their own mark
is that they are devoid of own-mark and so forth. The position outlined here, the most
extreme, can be reduced to a truism: we may trivially say that all language, all words, are
merely about something elsein the world, in the mind, or just other words. What is
needed is a sense in which savti can only be about paramrtha in a way that it is not
about what falls within its own domain: a non-trivial sense of about. It may be possible
to distinguish such a sense by making use of a hypothesis enjoying some influence in
Western linguistics at the moment which stands at the opposite end of the spectrum from
this complete ineffability thesis. This hypothesis will be considered after some other
uses of ineffable have been disposed of.
While most people, the above reflections notwithstanding, will cheerfully admit
that all sorts and conditions of human experience are perfectly well captured and
communicated by language, certain exceptions may be acknowledged in the anomalous
situations cited earlier. The rather peculiar status of qualia, for example, is currently
excercising both philosophers and psychologists alike. As Jerry Fodor remarks: If
qualitative contents are in some sense ineffable, it would hardly seem reasonable to
require a philosophy of qualia to eff them.11 What is it, then, about the perception of red,
the L.S.D. experience, any intense aesthetic

11

Jerry A. Fodor, Representations: Philosophical Essays on the Foundations of Cognitive Science,

Brighton, 1981, 18.
101

sensation or the most arcane reaches of scientific discovery, for which ineffability might
be claimed, that distinguishes them from the Buddhist inexpressible ultimate?
One way of answering this would be to invoke yet another key dichotomy
maintained by Buddhists of all denominationsthat between the mundane (laukika) and
the supramundane (lokottara)for it is closely allied to the other two under
consideration. It should be emphasised, however, that mundane and supramundane do not
map neatly onto our categories of profane and sacred since the mundane, from the
Buddhist point of view, includes the non-Buddhist (but to our mind sacred) activities and
postulates, such as the worship of deities or the immortality of the soul, while only the
activities of the Buddhas and of the ryas and the dharmas of the Path are classed as
supramundane. Even this classification is not immutable as Vasubandhu demonstrates in
one of the extracts given in Part 2.
Following this division the cases in question would be dismissed as irretrievably
mundane, and the mundane is never described as anabhilpya. On the contrary it is not
disputed that the ordinary means of communication in the world (laukikavyavahra, Tib.
jig rten gyi tha sad) is quite capable of conveying many kinds of experience or
knowledge: the truths of art, philosophy and science which belong to the realm of the
conventional. If language appears inadequate in this sphere, which it often does, whether
to express a deeply felt emotion, describe a colour, or explain some recondite point of
scientific theory, this would not be because it fails to express the ultimate truth. Werner
Heisenberg may feel that everyday language is wholly unsuited to giving an account of
the sub-atomic world but this does not prevent him from giving a successful account of it
by switching from the signs and symbols of natural language to the signs and symbols of
mathematicsjust another form of vyavahra. And although it is the case that in many
specialist areas less out on a limb than this ordinary language is made to undergo
considerable modification, and technical vocabularies are constructed impenetrable to the
layman (though intelligible to initiates), superficially appealing comparisons with the
esoteric symbolic languages generated by the spiritual traditions are not, in the end, very
illuminating for the intentionality of the latter is wholly opposite. The aim and purpose of
some sacred language is to give rise to a very different order of knowledge from mere
information about the world. Ordinary language is inadequate, here, in the special sense
that, in order to gain such knowledge one must go beyond language in a way that to say
that one must go beyond language or mathematics to know, understand or be
communicated the meaning of the concept electron as a term of quantum theory would
be nonsense. Prolonged training, theoretical and practical, is undertaken in order to
understand the language of physics not to attain individual spiritual insight

102

into the ultimate truth of electrons which transcends all words and concepts. That insight
requires quite another programme of training.
In the case of mundane discourse, therefore, to understand the linguistic
meaning of a word is sufficient; to understand the linguistic meaning of certain words
of Buddhist discourse is not sufficient, though through their possessing such a meaning
we are able to use them at an intellectual or philosophical level, within the sphere of
mind, mentality or consciousness, without necessarily having penetrated their deeper,
translinguistic meaning. And to understand the non-literal meaning of certain words of
ordinary language that have undergone a semantic transformation within a given
traditionto know their esoteric meaningis not sufficient either for the use of such
language, together with its interpretation, forms part of a spiritual practice which is
designed not to explain obscure and difficult words in terms of yet more words, symbolic
or metaphorical, but to facilitate the individual insight that does not depend on any
language or any verbal teaching. The mundane technical vocabularies are created for the
purpose of getting better words; the vocabularies of the sacred traditions are created for
the purpose of getting at something which the words alone are incapable of expressing.
However, the locating of our anomalies in the mundane sphere which, de facto, is
not described as ineffable in the relevant literature does not really offer a reason why they
could not be so labelled. A more powerful argument lies in the fact that being mundane
they are savti and so belong by definition within the limits of language and thought,
and this is where a brief resume of the hypothesis referred to above might help. This
hypothesis, advanced by J.J. Katz,12 is built around a principle he has entitled the
principle of effability based on the observation by Frege that the capacities of natural
language are such that a thought grasped by a human being for the first time can be put
into words which will be understood by someone to whom the thought is entirely new.13
It also has affinities with John Searles expressibility principle: that whatever a speaker
might want to communicate can be said; and the principle of universality of Tarski: that
natural languages can express whatever may be meaningfully spoken about. According to
Katz, it follows from the effability thesis that each natural language is capable of
expressing the same body of thoughts and this impliesthe claim that all natural
languages are intertranslatable in the sense that, for any sentence in one natural language
there is at least one sentence in every other natural language that expresses the same
proposition.14

12

J.J. Katz, Semantic Theory, New York, 1982, 1920.

ibid., 20.
14
ibid., 20.
13

103

This is a very strong claim and others have sought to dilute it,15 but since
effability is a principlelanguages are in principle effablethe fact that speakers are
often at a loss for words, or feel that their thoughts cannot be expressed by language is a
contingent one, as Katz observes. They simply lack the ability to use language to its full
capacity. So it is the speaker who fails and not language. In support of this contention he
points to the distinction made by Chomsky between speakers competence (the linguistic
knowledge) and speakers performance.
It seems quite possible to link up the effability thesis to savti (satya), vastly
different though their contexts of use may be. I have a strong feeling that Vasubandhu
would not be in disagreement with the notion that language, which for him is the most
significant aspect of savti, is in principle wholly adequate to convey the body of
thoughts comprising the Dharma of the Buddha, let alone the thoughts and sensations that
are the product of mundane experience. The Buddha first and foremost, but also the great
Bodhisattvas, who have perfected their correct knowledge of languages
(niruktapratisavid), and their correct knowledge of eloquence (pratibhnapratisavid)
represent ideal speakers whose performance matches their competence (in a rough
analogy). The Buddha, after all, encouraged the transmission of his teachings in the
various native languages of his human audience (and in the modes of speech of all classes
of sentient beings) presupposing his acceptance both of the universality of the Dharma
and its intertranslatability.
So if savtisatya implies complete effability then, in principle, there is no gulf
between any mundane thought or mental state and its expression in language, though
such a gulf may exist in practice. To say that the sentence all conditioned things are
impermanent is about the proposition that all conditioned things are impermanent (or
about impermanence) is just to say that it expresses that proposition (or communicates
that concept). In understanding the sentence we understand the proposition, we grasp the
thought. But since the understanding of emptiness does not, in the end, involve the
understanding of any proposition or concept but is the ultimate, direct, meditative
experience, the savti sentence all dharmas are empty both expresses a proposition to
be understood as a preliminary step, and, together with its propositional content is about,
i.e. indicates, points towards but does not express, the paramrthasatya that is
inexpressible.
We may assume that there are valid experiential grounds for claiming the
ineffability of the ultimate truth which give rise to the logical: it is contradictory to
suppose that what is in essence non-dual, the experience of emptiness, can be

15

Edward L. Keenan, Some Logical Problems in Translation, in F. Guenthner & M. GuenthnerReutter, eds., Meaning and Translation: Philosophical and Linguistic Approaches, London, 1978.
104

articulated by language as a system of signs, for all signs have dualistic structure; or
represented in thought which also entails duality.16

2. Conventional truth and ultimate truth in the stra and its commentary.
In this section I have compiled some of the most revealing passages both from the stra
and the k in which the terms savti, vyavahra and paramrtha occur. References to
these three concepts are far more numerous in the latter than in the former, where they
appear only sporadically. The one place where they are given more than a passing
mention is translated below and even then the explanation provided is extremely brief.
This does not mean, however, that the dichotomy between savti and vyavahra (the
conventional) on the one hand, and paramrtha (the ultimate) on the other is not in
implicit operation throughout the text as the chief classificational division according to
which the understanding of its content should be structured. In the exegis the three terms
figure as key elements of Vasubandhus hermeneutical armoury. They are used to
interpret many statements in the stra where they are not explicitly mentioned. I have not
added much sub-commentary or speculation of my own (remarks in brackets are mine)
partly for reasons of space but mainly because I hope that the texts, although dense, will
speak for themselves with greater eloquence. Translation verbatim appears in quotation
marks: double for the commentary and single for the stra. Otherwise paraphrases of
Vasubandhus comments are given without quotes.
The fullest treatment of the subject in the stra occurs in a passage17 which
exposes the meaning of truth, considered here as one of the endowments of the
Bodhisattvas. It is presented in four ways: as fourfold, threefold, twofold and as unique.
The truth as fourfold is the four Noble Truths. The truth as threefold consists of:
(1) Conventional truth (savtisatya, Tib. kun rdzob kyi bden pa), (2) ultimate
truth (paramrthasatya, Tib. don dam pai bden pa), (3) the truth of signs
(lakaasatya, Tib. mtshan id kyi bden pa). What, then is conventional truth? It
consists of all the teachings given through the ordinary, worldly means of
communication (laukikavyavahra): through syllables and verbal expressions
together with a semantic component. Ultimate truth is that which is devoid of any
mental activity, let alone syllables.

16

In Buddhism thought, by definition, is intentional; it is thought of something: the intentional object.

It is therefore binary in structure as is its correlate, language. This is why the Yogcra term
cittamtra (mind only) is, amongst other things, an (intentional) contradiction. The theory and practice
of objectless thought is explored by Vasubandhu in his Vijaptimtratsiddhi
17
Akayamatinirdea-stra, TTP, vol. 34, 53-5-1.
105

In commenting on the word truth Vasubandhu says:18

It is conventional truth in that the conventional itself is true. This is because
(conventions) operate in accordance with things merely as they appear, in the
form of illusions or mirages.
The emphasis, here, is on the conventional aspect of truth; why what is merely
conventional may qualify as true. Samvti represents the world of appearancesthings
mediated through language and conceptual thought. Insofar as it is possible to make a
formal distinction between the objective pole (appearances) and the subjective pole
(perception of appearances and the subjective processes of thinking and expressing which
name them, endow them with meaning, and make them communicable, intersubjectively,
to others) the word savti covers both. For the Yogcras, in particular, there is no
essential difference between them, they are all merely transformations of consciousness.19
But the world appears the way it does (incorrectly) precisely because it is so
mediated. Language and thought carry with them a mistaken ontological commitment to
a dualitythe inherent existence of persons (subjects) and the inherent existence of
dharmas (objects). But although these are purely imaginary from the ultimate point of
view, it must be admitted that they exist as facts of sasra and to the extent that
language, as defined in the stra, adequately reflects the subjective sphere of thought or
thinker, and the objective sphere of the objects of thought, and facilitates
communication among persons of unimpaired faculties, it is true. So even the ultimately
bogus is true if judged on its own terms. Truth becomes a matter of concensus within
this world.
The commentary continues:
It is ultimate truth in that the ultimate itself is true. This is because it is nondeceptive in its essential nature and without error. With regard to (savtisatya)
the ordinary worldly means of communication is the chief concept and the rest
is the explanation.20 Syllables refers to A, Ka and so on which constitute the
transmission of the sacred texts; verbal expressions refers to sentences found in
the sacred texts such as conditioned things are impermanent etc; and a semantic
component

18

TTP, vol. 104, 199-1-5.

The destruction of the subject-object duality is the goal; until this is achieved both experience and
the language that articulates it are, of course, structured in this way, so we are obliged to speak as if
subjects (streams of thought) and objects (what appears to these streams) existed, together with the
inner (skandhas, dhtus etc.) and outer (sprouts etc.) prattyasamutpda, and other conventional
dualities.
20
The chief concept (Tib. bstan pa) and the explanation (Tib. bzad pa) are technical terms of the
commentarial vocabulary used by Vasubandhu which serve to distinguish, within the root-text,
between an item, or list of items, to be explained and the subsequent exposition thereof.
19

106

refers to what belongs to these sentences which makes them meaningful and
enables them to be understood21 In short (the stra), means to say that all
verbal expressions that are part of the ordinary means of communication which
takes into account the various different capacities of the sentient beings to be
educated, are conventional truth. Thus in the Ratnamegha-stra it is said:22 Son
of Good Family, if a Bodhisattva is skilled in these ten dharmas he is called
skilled in the conventional. Which ten? They are as follows: although he may
make the designation form, ultimately, since he does not apprehend23 any form
he remains without attachment to it. Likewise, although he may make the
designations sensations, perceptions, motivational impulses and
consciousness, ultimately, since he does not apprehend any consciousness etc.
he is without attachment to them. (The same argument is applied to all the
buddhadharmas right up to Awakening.) Therefore, son of good family, these
names and designations which belong to the ordinary means of communication
are called conventional truth. Yet even if these conventional dharmas are not
ultimate, without these conventional dharmas it would not be possible to reveal
the ultimate.
The ultimate is the opposite of this and so in defining it (the stra) says ultimate
truth is that which is devoid of any mental activity, let alone syllables. The
ultimate, because it transcends all discursive proliferations such as arising and
non-arising, conceivable and inconceivable, permanent and impermanent, effable
and ineffable (!), cessation and non-cessation, knowable and unknowable etc. is
devoid of any mental activity. It therefore transcends the sphere of mind and
mental events and, since the ultimate and the pristine knowledge that knows it are
non-dual, it cannot be an object of thought let alone syllables.
I have included the truth of signs because although it is given separate treatment
in the stra its relationship to the ultimate is analogous to that of the conventional, and
signs, of course, form one of the elements of conventional truth.

21

Semantic component (saketa, Tib. brda): in traditional Sanskrit linguistics a word possesses
denotational power (abhidakti) which conveys to the understanding the meaning which belongs to
the word by common consent or convention (samketa), V.S. Apte, The Students Sanskrit Dictionary.
The stra, here, is speaking of vyavahra in terms of language in its phonological, syntactic and
semantic aspects.
22
Akayamatinirdea-k, TTP, vol. 104, 199-3-3.
23
Does not apprehend, if unpacked, would yield does not apprehend some referent/object existing
independently of his act of designating it, for which attachment or aversion might arise.
107

The stra says: all signs are but one sign and that one sign is a non-sign: this is
the truth of signs. Vasubandhu then explains that all signs (a sign here being the
particular identifying mark of a thing) such as the wetness of water, the resistance of form
etc. are really only one sign and that sign is suchness. (All dharmas have different
conventional signs but only one ultimate sign.) This is why the stra states that that one
sign is a non-sign, for suchness transcends anything that can be counted as a sign. Since
it is not a sign such as arising and ceasing, existence and non-existence, permanence and
impermanence etc. it is a non-sign. As this truth of non-sign is ultimate truth itself and
not different from it, why are two (truths) presented, namely, the truth of signs (which is
really the truth of non-sign) and ultimate truth? Although (the former) being true, is not
different from (the latter) it is called the truth of signs because by means of it the ultimate
is both indicated and finally understood. Therefore the truth of signs is established as the
means of approaching the ultimate.
The stra, when it classifies the truth as twofold, divides it into the truth of the
ordinary means of communication and ultimate truth. The first is the timely speaking of
the truth, viz: the truth of suffering, the truth of its origin, the truth of its cessation and the
truth of the path; and whatever truths of worldly communication may be taught through
syllables and verbal expressions, together with a semantic component. The ultimate is
nirva, endowed with the quality of ineffability.
Vasubandhu comments:24
It is the truth of the ordinary means of communication because these ordinary
means are true in the sense that they are true merely in accordance with
appearances. Thus it is said in the stras:25 whatever is true for the worldly ones
is true. It is ultimate (truth) in that the ultimate is indeed true because it is not
other than what it is. (Unlike appearances, which are not how they seem.) And so
it is also said in the stras: conditioned things are false because they are
deceptive, but the unconditioned is true. Why, then, are two truths set forth? Two
truths are set forth in order to teach both the Noble Path and the result of that
Path. The truth of the ordinary means of communication is declared in order to
teach the Noble Path; ultimate truth is declared in order to teach the result of the
Path (nirva). Furthermore, the two truths are proclaimed with reference to the
scope (viaya, Tib.: yul)26 of mundane knowledge and the scope of supramundane
knowledge.

24

Akayamatinirdea-k, TTP, vol. 104, 200-3-3.

The Hnayna gamas and suttas.
26
Finding an English equivalent to the Sanskrit viaya is extremely tricky. In many instances object
will do but it cannot be used consistantly here. I have used scope instead for this section, and
elsewhere, as this retains more of the objective flavour of viaya than the more frequently employed
spheresee Websters Dictionary: scope: goal, targetrange, extent. See also Michael Broido:
Padma dKar-po on the Two satyas, 16.
25

108

It is at this point, having explained why it is necessary to introduce two truths,

that Vasubandhu, comments separately on vyavahra and vyavahrasatya. He maintains
that:
Vyavahra refers to the numerous kinds of verbal expressions while
vyavahrasatya refers to the truth which is to be understood by means of these
expressions.27
He goes on:
This truth of the ordinary means of communication is also twofold: (the truth
which is) the scope of non-mistaken mundane knowledge, and (that which is) the
scope of mistaken mundane knowledge.
The former corresponds to the first half of the sentence in the stra up to the
truth of the path, i.e. the four Noble Truths, which are objects of the knowledge that is
attained subsequent to meditation (prthalabdhajna, Tib. rjes las thob pai ye shes).28
The following illustration is given: when a person, having emerged from the nirvikalpa
samdhi (the concentration devoid of conceptual thought) reveals the meaning of what he
has experienced to those who are to be educated, using his prhalabdhajna to expound
the Noble Truths correctly, this is timely speaking of the truth. The latterthe scope of
mistaken mundane knowledgecorresponds to the second half of the sentence where
whatever (truths of) worldly communication may be taught through syllables etc. are
mentioned. Interestingly, the commentary twice omits the word truth here when quoting
the stra: it has laukikavyavahra but not laukikavyavahrasatya. If the stra version is
correct it would seem to be drawing a distinction between Buddhist truths and other
mundane truths. If the commentary is correct (no scribal error), and the gist of
Vasubandhus remarks appears to support the reading it gives, then the emphasis is on the
deceptive qualities of the conventional modes of speech per se, though the messages
which they communicate may be non-mistaken (Buddhist) or mistaken (non-Buddhist).
Deceptive, not in the sense that they fail to convey some intended meaning, since they are
within the domain of effability, but because of their intrinsic duality and inability, ex
hypothesi, to convey paramrthasatya.
Again the commentary asks:
When it is said in the stra that vyavahrasatya is the timely speaking of the
truth (kle satyavacana) what is truth? It is the truth of

27

Akayamatinirdea-k, TTP, vol. 104, 200-4-1.

The status of prhalabdhajna, and the relationship of the meditational and non-meditational
states to paramrtha and savti according to various different opinions, are discussed in La Valle
Poussin, LAbhidharmakoa de Vasubandhu, ch. 6, 141, 142 ff.

28

109

suffering etc. And what is speaking? It is the ordinary worldly means of

communication. So, in fact, vyavahrasatya is the teaching of the four truths
through the syllables and verbal expressions, together with a semantic component,
appropriate to the means of communication in this world.
The distinction being made here is one between language itself as a means of
expression and what is to be expressed through it. This does not correspond to one of the
basic divisions of Linguistics which distinguishes between the formal syntactical, and the
semantic, aspects of language since language as defined by Vasubandhu includes both a
certain formal structure and a semantic component already given. The split, rather, is
between language as the form of expression, including whatever everyday meanings may
be attached to the words and sentences employed, and a very particular contentthe
spiritual truths of the Buddhist Path. Sentences in worldly discourse express propositions,
but only those sentences which serve as vehicles for the teachings of the Buddha can be
said to express true propositions.
There are actually two levels of vyavahra (satya) implied here which may be
illustrated in the following way. The form of words Granny takes snuff is the
conventional vehicle of the very statement whose truth depends on whatever the
contextually relevant Granny is snuffnut.29 The statement made (or proposition
expressed) by this form of words is true if, and only if, the corresponding state of affairs
in the world is true. This, according to the stra, would count as a mundane truth but it is
not on the level of the Four Noble Truths. According to Vasubandhus version, although
the statement may appear true by mundane standards, since it is not about one of the
Buddhist truths which are true also in the sense of having soteriological value it is within
the scope of mistaken mundane knowledge; to this extent it is mere vyavahra and not
vyavahrasatya. There is an abiding tension between the inherent falsity of language as
such, and its ability nonetheless to be used as a weapon of truth which impels the division
into expressible and inexpressible truths.
Vasubandhus claim that the Four Noble Truths are the scope of non-mistaken
mundane knowledge is more radical than the analysis given in the Madhyamakvatara,30
where the third truth at least, cessation, is classed as supramundane, let alone the more
usual stipulation that all the dharmas of the Buddhist Path are supramundane. For him,
however, in this context only paramrthasatyaemptinessqualifies as the scope of
supramundane knowledge, and emptiness is not one of the four propositions expressed by
the Noble Truths.

29
30

Jerry Fodor, Psychosemantics, Cambridge Mass., 1987, 81.

Madhyamakvatara-bhya, TTP, vol. 98, 118-4-8 ff.
110

Thus the commentary goes on:

Ultimate truth, nirva, endowed with the quality of ineffability, is ultimate
because it is the result of the Path, and the scope of the ultimate supramundane
knowledge It is ultimate truth because in its own nature it neither arises nor
perishes, partaking from the very beginning of the essence of clear light. It is
suchness, endowed with the quality of non-deceptiveness.
Paramrthasatya, here, although formally analysed into two separate concepts is,
at bottom, largely tautologous unlike savtisatya, or vyavahrasatya. The ultimacy of
this truth lies in the fact that it is the final result of the Path. The final result is nirva;
Nirva is suchness; suchness is truth. (We have shifted, here, from the propositional
form of the third truththere is cessation of suffering; cessation is nirvato nirva
as a purely symbolic term: savti about paramrtha.) The dichotomy between means
and end, or form and content implied by savtisatya is not applicable in this case. It is
clear, then, that the two expressions savtisatya and paramrthasatya function
differently. The former operates entirely on its own terms: it is an instance of ordinary
language which may be used to express some propositional content. With the latter, the
notion of propositional content is marginal. Ultimately, the words stand for, symbolise,
something wholly other. Since the words paramrthasatya are savti (conventional)
this means that two kinds of savti (as linguistic instrument) must be postulated: the
first is used entirely within the domain of ordinary non-symbolic discourse where
effability operates; the second, which is specifically about the inexpressible
paramrthasatya consists of, or includes symbol terms that point beyond their content to
non-content; in other words the language used to describe paramrtha , while still
savti, has a second aspect which ordinary savti lacks. Sthiramati has something to
say on this subject which is reproduced at the end of this paper.
An identifiable theme woven around the basic oppositions of conventional
versus ultimate, and expressible versus inexpressible emerges from these
discussions, which also encompasses other related dualities such as activity and repose;
ethical practice and spiritual insight; multiplicity and sameness and so on. Variations on
this theme are presented in the stra and further explored in the commentary. For
example, in the section on the four correct analytical knowledges (samyakpratisavid)
the stra speaks of the first two of theseknowledge of dharmas and knowledge of
meaningin the following way:31 dharmas are the infinite variety of dharmas,
beneficial and unbeneficial, defiled and undefiled etc. right up to samsra and nirva,
and their meaning is that they are empty, signless, wishless, unarisen, unorigenated,
substanceless and so forth.

31

Akayamatinirdea-stra, TTP, vol. 34, 62-2-8.

111

Vasubandhu remarks:32
Here, although the dharmas themselves and their meaning are not different,
meaning should be considered as that which is to be taught and understood (the
end), and dharmas should be considered as that which effects the teaching and
brings about understanding (the means) Understanding the mark of all dharmas
(emptiness) is the correct analytical knowledge of meaning; knowing that such
and such a dharma has such and such a name is the correct analytical knowledge
of dharmas.
Again, a distinction is being drawn on the savti level between the diverse forms
of the sasra represented in language, and the unity of their meaning. Meaning, in this
case, is clearly not the normal linguistic meaning of a name but refers to the
translinguistic meaning of words mentioned in Part 1the ultimate understanding of
which is beyond words/names.
The same pattern is discernible in the stras treatment of the first of the four
reliances (pratisaraa): to rely on the meaning, not to rely on the word. The text says:33
word is the teaching of the mundane dharmas and activities; meaning is the
understanding of the supramundane dharmas.
Vasubandhu explains:34
Word refers to the teaching of what is designated in accordance with the
ordinary means of communication (vyavahraprajapti): topics such as the
mundane dharmas of action and result and so on. Meaning refers to the
understanding of what is meant by emptiness which, in transcending the dharmas
designated in accordance with the ordinary means of communication of this
world, such as action and result and so forth, is without a sign. Whatever is
connected with the exposition of conventional dharmas is called word; its
opposite, the understanding of the ultimate object itself, is called meaning.
He goes on to say that while word represents learningthe grasping and
memorisation of all kinds of dharmas heard from kalyamitrasmeaning represents
the knowledge of what is meant by that which is empty is ineffable, in the sense that, on
the ultimate level, all dharmas transcend both exposition and learning. In short, word is
all the teachings concerning the set of eighty-four thousand dharmas, and meaning is
that, in reality, their essential nature is ineffable.

32

Akayamatinirdea-k, TTP, vol. 104, 242-3-5.

TTP, vol. 34, 63-5-2.
34
TTP, vol. 104, 245-1-1.
33

112

In dealing with the third relianceto rely on the stras of definitive meaning
(ntrtha), not to rely on the stras in need of interpretation (neyrtha)the Akayamati
asserts that35 the stras which teach the establishment of the conventional are said to be
in need of interpretation; the stras which teach the establishment of the ultimate are said
to be of definitive meaning.
Vasubandhu confirms,36 unsurprisingly, that:
The stras preached in order to establish what exits merely on the level of
ordinary communication, through the numerous kinds of conventional worldly
means, are in need of interpretation; the stras preached in order to teach the
ultimate, which bears the mark of non-arising and non-perishing, are of definitive
meaning.
In distinguishing between the neyrtha and the ntrtha stras in this way
Vasubandhu appears to have had in mind the division of the stras into rvakayna and
Mahyna, for in commenting on the root-texts assertion that the ntrtha stras teach
the profound, hard to perceive, hard to understand, he claims that it is the rvakas,
their knowledge being only partial, who find the profoundthe selflessness of persons
and dharmas, i.e. emptinesshard to understand. He also speaks of the neyrtha stras
as being concerned with worldly practices in putting the emphasis on verbal teaching
that delights the world. The ntrtha stras, on the other hand, are said to rely less on
wordy scholasticism and more on instruction in the techniques of meditation. This is a
fitting interpretation for a celebrated ex-scholastic, who repudiated his own rvakayna
background to become one of the luminaries of the Yogcra system that valued the
practice of meditation above all else.
There would appear to be enough evidence in these passages to support the
contention that the dichotomy of savti and paramrtha entails a principled division
into what can be said: a certain body of propositions governing the actual practice of the
Buddhist Path, together with their linguistic means of expression; and what cannot be
said at all: the ultimate goal, emptiness. All the teachings involving the Path in its
theoretical and practical aspects are of course savti, but even the words used in talking
specifically about paramrtha are counted as savti, as Sthiramati makes quite clear in
his commentary on the Madhyntavibhga37 which makes numerous references to
Vasubandhus own bhya. In the former the author says that while paramrtha belongs
to the subtle, meditational sphere, savti represents the coarse non-concentrated sphere.
As such it is classified in three ways: 1) savti consisting of designations; 2) savti
consisting of (incorrect) knowledge; 3) savti consisting of symbolic terms. The

35

TTP, vol. 34, 64-3-7.

TTP, vol. 104, 247-2-8.
37
TTP., vol. 109, 164-3-6.
36

113

last obtains when the fully accomplished (parinipanna = paramrtha) is taught by

means of such synonyms as emptiness, suchness, the immaculate and so forth,
even though it entirely transcends both conceptual thought and verbal expression. This
establishes the relationship between the essentially ineffable highest, i.e. not reducible to
any proposition, meditational object/state and the discursive thinking and language that
characterize the everyday states of consciousness, which provide the introduction through
teaching to the realisation of the ultimate.
Finally, I include for good measure a marvellous piece from the Ratnameghastra, quoted by Vasubandhu,38 which luminously conveys the Buddhas teaching on the
inexpressible ultimate truth:
The Bodhisattva Sarvnvaraavikambhin inquired: Lord, what is reality?
The Lord replied: Son of Good family, reality is synonymous with truth. The
Bodhisattva inquired again: Lord, what is truth? The Lord replied: That which
is what it is, for that which is what it is cannot be mistaken; that which is what it
is is not other than what it is. The Bodhisattva inquired again: Lord, what is that
which is what it is? The Lord replied: Son of good Family, these are to be
intuitively realised by each one individually, they cannot be designated by means
of words. Why is this? These dharmas transcend entirely all words; they
transcend entirely all modes of speech; they transcend entirely all avenues of
language; they are free from all discursive proliferations; they are beyond all
analysis, all limitations and all logic for they do not fall within the scope of
investigation by any logical reasoning. Because they are devoid of individual
marks they are wholly beyond the reach of fools and outside the domain of Mra.
Since it transcends the scope of all impurities and the scope of consciousness, that
which is the scope of the pristine knowledge of the ryas, the stillness which is
neither the support of, nor supported by, anything, is a matter of their individual,
intuitive realisation.

38

TTP, vol. 104, 187-3-8.

114

Monk, Householder, and Priest; What the

Three Ynas Mean to Newar Buddhists
David N. Gellner*

1. Introduction: why study Newar Buddhism?

My subject is the Buddhism of the Newar people of the Kathmandu Valley, Nepal. In this
paper I want to try and explain how it works as a ritual and ideological system. My main
argument is that the conceptual hierarchy of Three Waysor Vehicles as they are
conventionally translatedis central to this. An understanding of this conceptual hierarchy is
essential if one is to understand Newar Buddhism. Those authors who have failed to
understand it in the past have usually failed to do so for this reason.
The importance of Newar Buddhism can be summarized under two heads, historical
and anthropological. Let us take the historical first. Newar Buddhism is a survival of north
Indian Mahyna and Vajrayna Buddhism. Newar Buddhists are the only Buddhists left
whose sacred and liturgical language is Sanskrit (apart from a handful of priests in Bali).
They are the only South Asians who follow Mahyna Buddhism. It was from Nepal that
Brian Houghton Hodgson sent forth the Sanskrit manuscripts which enabled the Indological
study of Mahyna Buddhism to begin. Sylvain Lvi went to Nepal in search of ancient
Indian Buddhism. In his three-volume history of Nepal published seven years later in 1905 he
declared that Nepal was India in the making and that in the Kathmandu Valley one could
see as in a laboratory the conditions of the late first millennium CE in India.1
Now anthropologists tend to be scepticaland I think rightly soof scholars who
study other cultures as representatives of something else, whatever that something else may
be. They distrust anyone whose main motive for the study of

Fieldwork was carried out for two years, 198284, in Lalitpur, thanks to a Study Abroad Studentship
from Leverhulme Trust. This paper was first given in a lecture in Harvard, April 1988. It summarizes
several themes dealt with in greater detail in my forthcoming book, Monk, Householder and Tantric
Priest: Newar Buddhism and its Hierarchy of Ritual.
1
S. Lvi, Le Npal, Etude Historique dun Royaume Hindou, 3 vols, Paris, 1905, repr. 1986, vol. I, 28
ff.
115

another society is the search for survivals, since this tends to undermine the attempt to
understand it as it is today.
This then leads me to the second, more anthropological, reason for studying Newar
Buddhism. As a complex ideological system it is worth trying to understand it in itself and on
its own terms. It is especially interestingand challengingto us because it seems to go
directly against every preconception we have about Buddhism. (By we I mean we
Westerners, and I include in that all those who have been educated by modern western
means, wherever they come from.)
Did you think that Buddhism was a philosophy or an atheistic religion, or at least that
it is indifferent to the existence of God or gods? Newar Buddhism requires its followers to
worship numerous divinities at every level and in every religious context. Did you think that
Buddhism was a system of meditation intended for monks (and possibly also nuns)? Ritual
and worship are Newar Buddhisms most important activities and it has no monks or nuns, or
so it seems at first sight. Did you think Buddhism was otherworldly, disdaining to involve
itself in life-cycle rituals or in magical rites to overcome disease or drought? Newar
Buddhism provides a full set of sacraments from the cradle to the graveincluding a
complex wedding ritualas well as numerous apotropaic and curing rites. Did you think
Buddhism was hostileor at least indifferentto caste and that it recruited its holy order
from all classes of society? The Vajrcrya priests of Newar Buddhism are a hereditary
group, part of a caste. Sylvain Lvi and Steven Greenwold have called them, with justice,
Buddhist Brahmans.2 In short, Newar Buddhism seems to go against all the received ideas we
in the West tend to have about Buddhism. Some of these received ideas are wrong about all
forms of Buddhism, or at least all pre-modern forms of Buddhism. Certain others of these
ideas are right only about Theravda Buddhism, but wrong for Mahyna Buddhism. One of
the tasks of this paper is to explain how Theravda Buddhism, on the one hand, and the
Mahyna and Vajrayna Buddhism of the Newars, on the other, can both be Buddhism.
Newar Buddhism has of course been studied in the past, mostly by those who have
been interested in it for the first, historical reason.3 Consequently there is a
2

S. Lvi, op. cit., vol. I, 226; S. Greenwold, Buddhist Brahmans, EJS, 15, 1974, 101123.
The standard work on Newar Buddhism is J. Locke, Karunamaya: The Cult of Avalokitesvara-Matsyendranath in the
Kathmandu Valley, Kathmandu, 1980, and his recent article The Unique Features of Newar Buddhism in T.
Skorupski, ed., The Buddhist Heritage, Buddhica Britannica Series Continua I, Tring, Institute of Buddhist Studies,
1989, 71116 is an excellent brief introduction. Other works which avoid the merely a survival view are M.R. Allen,
Buddhism without Monks: the Vajrayana Religion of the Newars of the Kathmandu Valley, South Asia 2, 1973, 1
14; The Cult of Kumari, Virgin Worship in Nepal, Kathmandu, 1975; Girls Pre-puberty Rites among the Newars of
the Kathmandu Valley, in M. Allen & S.N. Mukherjee, eds., Women in India and Nepal, Australian University
Monographs on South Asia 8, 1982; D.N. Gellner, The Newar Buddhist Monastery: an anthropological and historical
typology in N. Gutschow & A. Michaels, eds., The Heritage of the Kathmandu Valley, Sankt Augustin, 1987;
Monkhood and Priesthood in Newar Buddhism, Pururtha, 12, 1989, 16592, and other articles given in References
and below; S. Greenwold, Buddhist Brahmans, EJS, 15, 1974, 101123; Monkhood versus Priesthood in Newar
Buddhism, in C. von Frer-Haimendorf, ed., The Anthropology of Nepal, Warmister, 1974; T. Lewis, The Tuladhars
of Kathmandu: A Study of Buddhist Tradition in a Newar Merchant Community, PhD thesis, Columbia University,
1984. Good as an introduction to the history of the Kathmandu Valley, but weak on Buddhism, is M. Slusser, Nepal
Mandala, Princeton, Princeton University Press, 1982. An important work of narrow focus but superb scholarship is B.
Klver & H. kya, Documents from the Rudravara-Mahvihra, Pan. 1, Sales and Mortgages, Sankt Augustin,
1985. The standard ethnography of the Newars, but also thin on Buddhism, is G. Toffin, Socit et Religion chez les
Nwar du Npal, Paris, 1984. On the general ethnographic and political context of modern Nepal, see M. Gaborieau, Le
Nepal et ses Populations, Paris, 1978.

116

persistent trend within Western writings about Newar Buddhism to see it merely as a
survival. The fact that Newar Buddhism contrasts so strongly with Western preconceptions is
then explained by Newar Buddhisms supposed corruption or degeneracy. My contention is
that one must first understand Newar Buddhism on its own terms and only then use the
understanding so gained to answer questions about Buddhist history.
Westerners in Kathmandu often tell the story of a foreigner who asks a Newar Are
you a Buddhist or a Hindu? to which the Newar replies Yes. The foreigners natural
reaction is to assume that the Newar is confused and knows nothing about his own culture.
The Newar often confirms this by saying something like: Yes, I know little about it; there
was a pait who really understood it, but he died twenty years ago. The foreigner
concludes that the Newar carries on an ancient tradition which he does not understand. Over
the generations the tradition has become corrupt, degenerate, and inauthentic. The Newars
are seen as the inheritors of unique ancient urban civilization which they themselves are
spoiling through ignorance. Paradoxicallyor perhaps not so paradoxicallythis is precisely
the view of western-educated Newar cultural nationalists. In fact all modern-educated
Newars seem to share this view, a reflex of the Westerners image of the Newars as exotic
and authentically traditional, and yet at the same time ignorant and inauthentic.
What I have tried to do in my study of Newar Buddhism is to get beyond these
Eurocentric views, shared by Westerners and western-educated Newars alike, and understand
it on its own terms. This does not mean that I am a total relativist: far from it. I do not think
that once I have explained the traditional Newar Buddhist view, I can pack up and go home.
(If I were really to restrict myself to presenting the Newar Buddhist point of view the most
authentic way of doing it would perhaps be to clear the chairs, smear the room with cowdung,
and make you all sit

117

on the floor and do the guru maala rite, the most basic ritual of Newar Buddhism.) What I
do believe, however, is that until one has made the effort to understand Newar Buddhism as
traditional Newar Buddhists see it, one will not be able to contribute anything to the
analytical and historical questions implied in the perspectiveNewar Buddhism as a
survivalwith which I began.
2. Context of Newar Buddhism
The Kathmandu Valley is an outstandingly fertile bowl-shaped Valley about 4,000 feet high
in the Himalayan foothills. This has protected it from many of the vicissitudes which have
affected the Gangetic plain. The Muslims invaded once in the fourteenth century but did not
stay, and the British never conquered Nepal. Consequently the Kathmandu Valley has
preserved old patterns of Indian cultureincluding Buddhismin a way other more central
areas have not. At the same time the Valley is not so far on the periphery that it could receive
those cultural forms only in translation. Culturallythough not politicallythe Valley has
always been a full part of South Asia. This combination of cultural conservatismtypical of
the peripheryyet authenticitydue to its proximity to the Ganges plainaccounts, of
course, for the historical perspective on the Newars which I have referred to.
From the fifth century CE there is evidence of a flourishing Sanskritic culture in the
Valley, supported by the Valleys extremely fertile soil and by its strategic position astride
trade routes between India and Tibet. In what is known as the Malla period (c. 1200 to 1769)
the Valley was dominated by three citiesKathmandu itself, Lalitpur, and Bhaktapureach
being the centre of a separate kingdom from the late fifteenth century until 1769, when they
were all conquered by Prithvi Narayan Shah, ancestor of the present king.
Nowadays Newars constitute only about half of the population of the Valley. Many
of them have migrated out of the Valley. They are now the main trading and shopkeeping
caste throughout the hills of the modern state of Nepal. Inside the Valley they are still the
majority in the three big cities and in the nucleated Newar villages which surround them. But
in between there are now many hamlets of other castes and ethnic groupsin particular the
Brahman and Chetri castes of the dominant Parbatiys (hill people). And the conurbation of
Kathmandu and Lalitpur, with its suburbs spreading out around the ring road which encloses
both cities, is a modern mix of many different groups. The Newars have their own language,
Newari, their own cultural practices, and they believe the Valley to be their homeland and
that all other groups are immigrants there (simultaneously their

118

own caste myths of origin also posit outside origin for most of them, but at an earlier period).
These factors keep the Newars separate from the rest.4
My fieldwork was conducted in Lalitpur which is well known to be the most
Buddhist of the three cities. According to official figures given in the 1971 census, 32.5% of
Lalitpurs population of 60,000 was Buddhist, compared with 24.5% of Kathmandu and 7.5%
of Bhaktapur. Even using the official figures it is clear that Lalitpur is the most Buddhist and
Bhaktapur the most Hindu city of the three. Kathmandu is somewhere in between. However
the census operates on the same logic as the Westerner: are you a Buddhist, or a Hindu, or a
Muslim? Only one answer allowed. One cannot sayas many Newars, particularly certain
middle castes, do say when given the choiceI am both Buddhist and Hindu or I am
Buddhist, i.e. Hindu. In fact in most cases respondentsunless their name is Ahmad or
Muhammedare not even asked this question, but simply entered as Hindu.
Consequently it would be a mistake to assume on the basis of these figures that the
religious life of Lalitpur is 32.5% Buddhist and 67.1% Hindu. In fact in many ways
Buddhism is the dominant religion in Lalitpur. Thus the really big festivals of the city are
essentially Buddhist, though Hindus observe them too. Examples are the Karumaya
(Matsyendrantha) festival and the Matay procession to all the caityas of the city. Most of
the population of the city have Buddhist, i.e. Vajrcrya, family priests, not Hindu, i.e.
Brahman, family priests. (In parenthesis, it is worth pointing out that this is where exclusivist
logic seem to apply, though even here some families manage interesting combinations: for
the purposes of carrying out the main auspicious life-cycle rituals, one may have only one
hereditary family priest and he must be either Buddhist or Hindu; one cannot have both.)
Most families in Lalitpur, then, have Buddhist, not Hindu, family priests and the
dominant idiom of most important festivalsat least those particularly associated with the
city itselfis Buddhist. Nonetheless, the families with the highest social prestige (high-caste
rehas such as the Amtyas and Rjbhars) are strongly Hindu and the rituals associated
with the royal palace at the centre of the city are Hindu. Furthermore, in the country as a
whole, which is an officially Hindu kingdom, Hinduism is overwhelmingly dominant. Thus
even within the city the ideological dominance of Buddhism and Buddhists is by no means
easy or assured. A fortiori their political dominance of the local city councilachieved in
recent (1986) electionsis strongly disputed.

On Newar ethnic identity, see G. Toffin, Socit et Religion chez les Newar du Npal, 58593; D.N.
Gellner, Language, Caste, Religion and Territory: Newar Identity Ancient and Modern, EJS, 27,
1986, 10248; D. Quigley, Ethnicity without Nationalism: the Newars of Nepal, EJS, 28, 1987,
15270.
119

The old city of Lalitpur is laid out in a grid pattern with the palace at the centre.
High-caste Hindus claiming descent from the courtiers of the old Malla kings tend to live
near the palace. Untouchables live outside the old city walls and are excluded from the water
fountains used by clean castes. (They are now free to disregard the traditional taboo on their
having tiled roofs or three-storey houses, but they continue to live where they always have.)
The Butcher caste is permitted to live within the city walls but is relegated to its own
localities at the citys edge. Other castes tend to cluster in particular areas. The Buddhist
sacerdotal caste, the kyas and Vajrcryas, tend to live in courtyards adjacent to Buddhist
monasteries (in Newari referred to honorifically using the Sanskrit word vihra, or more
colloquially as bh or bah). There are about 18 large monasteries and over 150 smaller
monasteries in the city. They tend to be set back from the road in an enclosed courtyard,
whereas Hindu temples are sited at crossroads for maximum exposure.5
Father John Lockes survey6 revealed a total of almost 7,000 kya and Vajrcrya
men as initiated members of monasteries in Lalitpur, out of an overall total of just over
15,000 in the Valley as a whole. Including women and uninitiated boys there may be 17,000
kyas and Vajrcryas in Lalitpur, perhaps a fifth or more of the total population of the city.
Together the kyas and Vajrcryas comprise the holy order of Newar Buddhism. Only
Vajrcryas may be priests for others. There are about twice as many kyas as Vajrcryas.
Although all Vajrcryas may be priests there were not, even in the past, enough Buddhist
lay people or enough rituals to go round.
Nowadays, according to a survey I carried out of 144 Vajrcrya men over the age of
15, only 22% of Vajrcryas actually practise as priests, and only a tiny 4.4% live from the
priesthood alone. Other Vajrcryas are mostly artisans, like kyas: in particular they are
god-makers, silver- and goldsmiths, carpenters, and tailors. Some are successful traders or,
especially in Kathmandu, civil servants and teachers.
Vajrcryas are accepted as higher in status than kyas because of their priestly
prerogative, but this is not a question of caste: they intermarry and interdine. All kya and
Vajrcrya men must be members of a monastery. They become members by going through
the rite of Monastic Initiation in that monastery at the

On the Newar Buddhist monastery and its architecture, see J. Locke, The Buddhist Monasteries of
Nepal: A Survey of the Bhs and Bahs of the Kathmandu Valley, Kathmandu, 1985, and D.N.
Gellner, The Newar Buddhist Monastery: an anthropological and historical typology.
6
J. Locke, The Buddhist Monasteries of Nepal, 515.
120

age of 3, 5, 7, 9 or 11.7 Only sons of kyas and Vajrcryas by kya or Vajrcrya mothers
may be initiated. And one may only be initiated in the monastery of ones father.
Consequently the membership of a Newar Buddhist monastery is defined by patrilineal
descent and caste purity.
In ritual handbooks and in the liturgy, the ritual for joining a monastery is known as
pravrajyvrata, the Observance of Going Forth. Note that this uses the ancient term for the
first ordination of a Buddhist monk, pravrajy, Going Forth, but refers to the ritual as a
whole as a vrata, that is as a temporary ritualized Observance or fast. In ordinary speech
kyas and Vajrcryas tend to call it ckarma, tonsure. Indirectly the rite is also known
as bare chuyegu, which means something like making a Bare. This is the name under which
the ritual has become known in scholarly works. Bare is the colloquial term for kya,
also used by high-caste Hindus to refer slightingly to the combined kya and Vajrcrya
caste. The word Bare derives from the Sanskrit vandya, respectsworthy person, a term used
in ancient India for Buddhist monks. Newar Buddhists know that it means monk.
This ritual of Monastic Initiation lasts for four days. During it the kya or
Vajrcrya boy is a monk. On the fourth day, a concluding rite takes placesometimes in the
monastery, sometimes in the boys home, depending on local customand the boy
renounces his monastic status and becomes a householder. Later he goes on to marry and
earn his living as a householder. Yet he retains the status of monk. He receives alms as a
monk once a year during the Pacadn festival. He takes his turn for duties in the monastery.
If for any ritual reason he has to shave his head, the whole of his head is shaved, leaving no
top-knot, the symbol of householder status. He is addressed by other castes with the Newari
honorific auxiliary bijyye (from the Sanskrit vijaya). This honorific is used only to gods,
kings, Brahmans, ascetics, and monks.
Many observers of Newar Buddhism have seen it as ridiculous. In a recent study,
Nepal Mandala, which is in many other ways exemplary, Mary Slusser has charged Newar
Buddhists with making a mockery of their religion with this rite of Monastic Initiation.8
How can one be a married monk? How can monks be priests and Buddhist priests a
hereditary caste? In order to explain how this is possible and, on its own terms, legitimate, a
short historical digression is now necessary.
3. The Three Ways in Buddhist History
It is common knowledge that the Mahyna, Great Way or Great Vehicle Buddhism, used
the doctrine of yna, Ways, to explain its own relationship to the

On this ritual, see J. Locke, Newar Buddhist Initiation Rites, Contributions to Nepalese Studies, 2,
1975, 123, and D.N. Gellner, Monastic Initiation in Newar Buddhism, in R.F. Gombrich, ed.,
Indian Ritual and its Exegesis, Delhi, 1988.
8
M. Slusser, Nepal Mandala, 296b.

121

Buddhism which went before it. The Mahyna did not reject earlier Buddhism but included
it as a lower level within its own spiritual hierarchy.
All Buddhism had always recognized different types of enlightened beings: arhats
who achieve enlightenment by following the teaching of a Buddha; Solitary Buddhas
(pratyekabuddhas), who achieve enlightenment on their own and teach no one; and
completely enlightened Buddhas (samyak-sambuddhas), who both discover how to achieve
enlightenment for themselves and teach others. This hierarchy of arhat, pratyekabuddha, and
samyaksambuddha the newly arisen Mahyna conceived in terms of three, hierarchically
ordered Ways to enlightenment (cf. Tables I and III): the Disciples Way leading to arhatship, the Solitary Buddhas way leading to Solitary Buddha-hood, and the Great Way leading
to complete enlightenment as a Buddha. Now, the Buddha in his previous lives, when he was
working his way towards enlightenment, was known as a Bodhisattvaso this way was also
known as the Bodhisattvayna, the Way of the Bodhisattvas.
Table I
The Different Ways of Newar Buddhism9
Sahajayna
Vajrayna (Tantrayna, Mantrayna)
Mahyna
Mahyna
(Bodhisattvayna,
Pramityna)
Mahyna
Pratyekabuddhayna
Hnayna
rvakayna
In the pre-Mahyna schoolsincluding the Theravda, the only one to survive
todaythe Bodhisattva is a retrospective and descriptive concept: it refers to a Buddha
before his enlightenment. In the Mahyna it becomes forward-looking and prescriptive:
every good Buddhist must strive to become a Buddha, to save all beings through teaching the
Buddhist doctrine and by other merciful acts. In other words all Buddhists should strive to be
Bodhisattvas.
As Buddhists of the other schools strove only to be arhats, not fully enlightened
Buddhas, this was the origin of the charge by the Mahyna that they were selfish: they were
supposedly only interested in their own, and not in others,

Terms given in parenthesis are alternatives. Terms underlined are the ones generally used by Newar
Buddhists. The broken line indicates that the Sahajayna is sometimes considered to be a higher Way,
above even Vajrayna. The term Mahyna has both greater and lesser applications. In its broader
sense it includes what is otherwise opposed to it, the rvakayna. Hnayna is not much used but is
included here as it is well-known from the western literature on Buddhism.
122

enlightenment. This is of course a very partial and unfair charge from the Theravda point of
view.10
The idea of the Bodhisattva was a very fertile one because it united in one concept
two sociologically very distinct things: first, the moral and soteriological ideal of Buddhist
striving and, second, certain saints or divinities who could be petitioned by the Buddhist
laity. Pre-Mahyna Buddhism had a very limited pantheon. It consisted of the Buddha
himself, certain previous Buddhas, and a supporting cast of converted Hindu deities.
Mahyna Buddhism developed an extremely elaborate pantheon, including numerous
Bodhisattvas, the two most important being Avalokitevara (called Karumaya by Newars),
Bodhisattva of compassion, and Majur, Bodhisattva of wisdom. The laity was encouraged
to petition these Bodhisattvas even for worldly goods; the Bodhisattvas, for their part, were
said to be so meritorious and compassionate that they would grant the good in question.
However, on a higher level the idea of reciprocity between Bodhisattva and worshipper is
replaced by the idea that the worshipper earns the good by their own merit, since performing
the worship of such a Bodhisattva is inherently meritorious.
At some time after the rise of the Mahynaexactly when and how is a
controversial mattera new form of Buddhism arose, the Vajrayna, or Diamond Way
Buddhism. In one sense it aimed to replicate the logical movement by which the Mahyna
had trumped what went before: so, just as the Mahyna incorporated but was superior to
what it termed the Solitary Buddhas and Disciples Ways, so also the Vajrayna
incorporated and subordinated the Mahyna.
However, in another sense the relationship between Mahyna and Vajrayna is not
the same as that between pre-Mahyna and Mahyna Buddhism. This is why, on the
diagram in Table I, Mahyna is shown as including Vajrayna and not vice versa. Very
briefly, Mahyna and Vajrayna became aspects of each other: Mahyna was the exoteric,
public face of Mahyna Buddhism, Vajrayna its esoteric inner truth, andas a ritual
systemits essential motor, the legitimation of priestly power. Thus, the Vajrayna became a
secret, specialized, and privileged path within the Mahyna, and not, or not merely, a later
revelation working at the same level as the Mahyna.
It is possible to have Mahyna Buddhism without Vajrayna Buddhism. There are
examples, notably from Japan and Korea, and it was evidently practised thus in some times
and places in India. In Nepal however and Tibet the twoMahyna and Vajraynaseem
always to go together. The Vajrayna is the

10

R.F. Gombrich, Precept and Practice: Traditional Buddhism in the Rural Highlands of Ceylon,
Oxford, OUP, 1971, 31827.
123

secret path for the pious and the committed. The Mahyna preaches the doctrine of universal
salvation openly to the greatest number.
One consequence of this double relationship between the Mahyna and the
Vajrayna is that the words are polysemic. At first sight this can be confusing. I have tried to
sort out the meanings of Mahyna in Table I. Where the Mahyna is based on scriptures
known as stras, the Vajrayna is based on tantras, hence it is also known as Tantric
Buddhism. Devotional ritual had been present in Buddhism from the beginning. It wasand
in Theravda Buddhism remainsrelatively austere and simple. With Mahyna Buddhism
it received a great boost. But it required the Diamond Way for the development of an
elaborate ritual system with thousands of possible permutations.
The Vajrayna systematized the pantheon of the Mahyna into pentads. So there
were now Five Buddhas and five Buddha-families into which everything else was classified.
Much work on the history of Tantric Buddhism remains to be done.11 It is clear however that
new tantras were continuously composed which attempted to supersede those which had gone
before. Each tantra teaches the worship of an esoteric deity. The searcher for enlightenment
and/or supernormal powers is supposed to progress through all other forms of Buddhism,
starting with the Disciples Way and ending with the various philosophical teachings of the
Mahyna, before proceeding to the realization of the ultimate esoteric deity from whom the
tantra takes its name. This realization of the deity means meditation on the deitya kind of
ritualized trancethrough which one becomes the deity and achieves identity with the
ultimate reality or emptiness. Just as the hero of the Great Way is the Bodhisattva, so the hero
of the Diamond Way is the siddha, an antinomian figure who rejects monasticism along with
all other worldly conventions and achieves enlightenment here and now (cf. Table I).
Although there were many different stages in the development of Tantric Buddhism,
two are particularly important for the study of Nepalese Buddhism. The first is the stage
represented by the set of Five Buddhas and the Tantric version of that represented by the
Guhyasamja-tantra, what is called the Yogottara Tantra level in studies of Tibetan
Buddhism. The second is the stage represented paradigmatically by the deity Cakrasavara
and his consort Vajravrh. In Newar Buddhism the gods of this level are esoteric gods.
Other tantric divinitiesparticularly Vajrasattva and Caamahroaaact as a kind of
external face or exoteric version of the esoteric gods. It will be clear from this that
correspondences are made between the gods of the different levels. As one receives

11

The most recent and up-to-date attempt to write it, though still not making use of many unedited and
unpublished manuscripts, is D.L. Snellgrove, Indo-Tibetan Buddhism, London, 1987.
124

tantric initiations so one learns the secret identities of the Buddhas and Bodhisattvas whom
anyone may worship.
So far I have said nothing about that aspect of the tantras which most shocked the
scholars who first worked on them and which has made them notoriousnamely their
antinomianism. The Tantric scriptures, especially the later ones, prescribe the breaking of
taboos, particularly sexual and other purity taboos, as a means to achieve and demonstrate
advanced spiritual states. This factonce knownhas had a very deleterious effect on the
process of understanding of how Tantric Buddhism actually works. Few seem to realize that
in a traditional society, where literacy is a minority accomplishment and study of the
scriptures is carried on only by a minority of that minority, most of the priestslet alone the
laityare simply ignorant of the contents of these scriptures. Priests learn liturgies by heart,
they know how to perform rituals, and often have a deep intuitive understanding of their own
tradition. But only a tiny number are actually paits, that is to say, spend their time reading
the original scriptures. Moreover paits spend most of their time reading and transmitting
Mahynist textsmostly Rebirth storiesfor mass consumption, just like Theravda
monks. Consequently, although it is of course important for Western scholars to know and
study the original scriptures, it is equally important to know how those scriptures are used. In
Nepal they are used as the ultimate legitimators of ritual handbooks and a ritual system; but
they are rarely read, and those who read them make no attempt to reconstruct ritual practice
in the light of the scriptures.
The doctrine of the different ynas is, we have seen, a way for Buddhism to explain
its own history to itself. It allows for subsequent revelations which improve on, while not
displacing earlier ones. At one level the later and earlier ynas are opposed to each other. At
the higher level the later yna incorporates the earlier within itself and represents the whole.
This, I would argue, is very similar to Dumonts theory of hierarchy expounded in the
Postface of the revised edition of Homo Hierarchicus.12 Dumont explains that just as at one
level in traditional Christian thought man is opposed to woman but at a higher level man
incorporates woman and refers to the species as a whole, so also pure is opposed to impure.
This same hierarchical logic is at work in the relationship of different Ways to each other.
4. The Three Ways in Newar Buddhism
Now at last we are in a position to ask what these ideas mean to Newar Buddhists. A
summary is presented in Table II. Table III shows the way in which the different ideals of the
Three Ways are translated into practice. Of the different

12

L. Dumont, Homo Hierarchicus, The Caste System and its Implications (complete revised English
edition), Chicago & London, 1980.
125

Ways which occur in canonical lists, known by the Newars too, these three are the
operational ones. Obviously, since the Solitary Buddha by definition teaches no one, his
Way is of theoretical interest, but need not concern the anthropologist or sociologist.
Elderly Newar Buddhists refer to Theravda Buddhism as rvakayna. This is
entirely consistent with their Mahynist viewpoint. Theravda Buddhism was introduced
into Nepal in the 1930s.13 Those Newars who became monks under the Maha Bodhi Society
in India and then returned to Nepal were expelled by the Rana regime. But since 1951 they
have become well established in the Valley. Some Newar Buddhists ignore them, including
most practising Vajrcrya priests, but many carry on both traditional Newar Buddhism and
support the Theravda monks and nuns to some extent. The Theravda movement in Nepal is
the mainthough not the onlyconduit for Buddhist modernism in Nepal.
Table II
Summary of the different practices falling under
the head of the Three Ways (yna) in Newar Buddhism.
The
Way
of
the monachism, ascetic rule-observance,
Disciples: (rvakayna) worship of the Buddha.
The Great Way:
(Mahyna)

worship of all the gods, fulfilment of

hereditary householder duties,
including festivals and life-cycle rites;
acquiring merit through donations
(dna), cultivation of the moral
perfections in accord with the ideal of
the Bodhisattva.

The Diamond Way:

Tantric (Vajrayna)

worship of Tantric deities, taking Tantric

Initiation (dk), acquiring magical
powers and advanced spiritual states by
strict rule-bound devotion to powerful
deities or their exoteric manifestations.

The Three Ways are built into the structure of the Newar Buddhist monastery.
Newars sometimes say that the shrine of the main Buddha image belongs to (or is) the
Disciples Way, the long hall on the upper floor belongs to the Great Way, and the Tantric
shrine, located above the Buddha-image or over the main

13

On the Theravda movement in Nepal, see R. Kloppenberg, Theravada Buddhism in Nepal,

Kailash, 5, 1977, 30121; R.C. Tewari, Socio-cultural Aspects of Theravda Buddhism in Nepal,
JIABS, 6, 1, 1983, 6793; D.N. Gellner, Language, Caste, Religion and Territory: Newar Identity
Ancient and Modern; H. Bechert & J. Hartmann, Observations on the Reform of Buddhism in
Nepal, JNRC, 8, 1988, 130.
126

entrance to the monastic complex, belongs to the Diamond Way. The same structure is
reflected in the ritual and iconography appropriate to each part of the sacred area.
The main focus of public ritual in the monastery, as well as the main focus of
devotion from local people and pilgrims (often Tibetans), is unquestionably the Buddha
image sited opposite the main entrance. In most cases it has the earth-touching (bhmispara)
gesture. It is considered to represent kyamuni, and is not usually addressed as Akobhya.
In popular monasteries visited by many devotees numerous other subsidiary shrines may be
set up, both on the ground floor and the first floor. These will be to Mahynist figures such
as Tr, Amitbha and Avalokitevara. Of the 363 monasteries of different sorts in the Valley
only 31 do not have Buddha figures in their main shrine: they have Padmapi Lokevara
(20), Maitreya (7), aakar Lokevara (2), Majur (1), and Mahvairocana (1).14
Table III
Soteriological ideals and social roles of the Three Ways.
Religious Level
Disciples Way

Soteriological Ideal
arhat (noble one)

Great Way

Bodhisattva
(wisdom-being)

Diamond Way

siddha (realized one)

Social Role
monk (i.e. during
Monastic Initiation)
Practising Mahyna
Buddhists;
Vajrcryas in their
purely
exoteric
aspect.
Vajrcrya

The main Buddha shrine is worshipped in a manner reminiscent of royal Buddha

images in Theravda countries15 with flowers, water, light, incense, fruit, and the waving of a
yaktail. However, many of the devotional verses used in this context are Mahynist in
inspiration (e.g. the Dnabalena), and others are Vajraynist (e.g. the Nmasagti).
The long first-floor room of the monastery is found only in larger establishments. In
larger monasteries, such as Kw Bh and Uku Bh in Lalitpur, it is known as gumb
(from Tibetan dgon-pa), because it has been done up in Tibetan style this century by traders
returning from long periods in Tibet. The main divinity is Amoghapa Lokevara. The main
ritual performances held here are Observances (vrata), especially the aam vrata or
upoadha vrata, a

14

J. Locke, The Buddhist Monasteries of Nepal, 516.

H. Evers, Monks, Priests and Peasants: A Study of Buddhism and Social Structure in Central
Ceylon, Leiden, 1972, ch. 4.

15

127

ritualized version of the Eight Precepts.16 Unlike in the Tibetan version of the ritual, which is
also conducted in the same place by local Newars ordained in one of the Tibetan traditions,
the participants (mostly women) do not actually utter the Precepts themselves, but their
content is built into the structure of the rite.
The Tantric shrine usually consists of a small curtained-off room, which only a senior
functionary called the Cakrevara (the seniormost Vajrcrya in monasteries with both
Vajrcryas and kyas) may enter. In front of it is also a long hall where rituals involving a
number of people are held. There are regular monthly rituals involving the elders (sthavira,
ju) of the monastery, yearly rituals established by endowment, as well as occasional rituals,
such as those performed on the initiation of a new member of the monastery, a new
Vajrcrya, or a new elder. The rituals performed before the Tantric deity necessarily involve
the offering of meat and alcohol, offerings which are normally taboo before the Buddha
shrine and during the course of an Observance. Likewise, the use of dance (vajrantya) and
song (carygti, cc mye) are in order before a Tantric shrine, but not in the rvakaynist
or Mahynist context.
All clean castes (over 90% of the population) may approach the rvakayna and
Mahyna divinities, make offerings and participate in rituals directed to them. Access to the
Tantric divinity is much more restricted. Tantric Initiation (dk) is precondition of being an
elder, for this reason. The foyer before the Tantric deity may be entered by some of those
without Tantric Initiation, but only if they are of sufficiently high status or belong to the
institution in question. Mediating between the esoteric and exoteric spheres are such deities
as Vajrasattva and Caamahroaa who may appear outwardly, without consorts, and
receive worship from all comers. Some monasteries have shrines to them on the ground floor.
Outside the monastery one can also identify different spheres in which one or other
yna is predominant. The yearly festival of Pacadn is a clear assertion of monastic ideals,
with all kyas and Vajrcryas supposed by custom to beg alms in at least seven different
places on the day. The twelve- or five-yearly festivals of Samyak in Kathmandu and Lalitpur
are similar, although there the emphasis is on the perfection of giving, a Mahynist twist.
The various rituals of the householder are viewed by Newars as being part of the duty of a
Mahynist. Each high-caste lineage also possesses a Tantric shrine where regular rituals
have to be performed (though this is very much on the decline today).
The Disciples Way is not considered by Newars to be a system of ritual. Rather it
consists of observing the rule of chastity and other rules of restraint associated

16

J. Locke, The Upoadha Vrata of Amoghapa Lokevara in Nepal, LEthnographie, 83, 1987,
15989.
128

with the Ten Precepts. There are no permanent celibates in traditional Newar Buddhism. The
akyas and Vajrcryas, as already noted, are a caste of married part-time monks.
Consequently, the practice of the rvakayna is contextualized. It is put in a ritual
framework derived from the Mahyna and the Vajrayna. Thus priests explain that the
worship of the goddess Kumr at the end of an Observance (vrata) marks the transition from
the rvakayna to the Mahyna and the Vajrayna. In other words, the perennial Buddhist
practice of the Preceptssomething Newar Buddhism shares with all forms of Buddhismis
placed in a framework of exoteric and esoteric ritual. Furthermore, in the life-cycle rituals
which a Buddhist priest performs for his parishioners he has to make use of his Vajraynist
entitlement to summon and visualize divinities which will be worshipped by the householder
and thereby sacrilize his familys progress through life.
It should be clear that there are two hierarchies in Newar Buddhism. There is a
hierarchy of statuses from ordinary layman up to tantric priest. The tantric priests can only
become an initiated Tantric priest by first becoming a monk and then a householder.
Secondly there is a hierarchy of deities, with the higher deities being considered as the secret
inner identity of the lower and outer ones.
To conclude this section let me quote what four relatively articulate Newar Buddhists
said to me about the Three Ways. An old kya woman explained them as follows:
There are three Ways: the Disciples Way, the Great Way, and the Diamond Way.
Theres also the Way of the Solitary (Buddhas). Paits say that one shouldnt mock
others. All are the Buddhas teaching, whichever you practise. The Great Way is
higher, though. The Disciples Way is easiest because they only have to worship the
Buddha, whereas in the Great Way there are Tr and all the rest, and one has to read
so many books. The [Theravda] monks say that getting to the stage of being an arhat
(arhatpad) is enough, but in the Great Way one has to attain the stage of complete
enlightenment (samyaksambuddhapad). To reach that one has to be a Bodhisattva
and fulfil the Ten Perfections. The monks say that all one has to do is meditate (do
bhvan). Thats all very well, but one has to attain the Four Trances (caturdhyn)
and the Perfections too.
An old kya man said:
There are four Ways, rvakayna, Pratyekayna, Mahyna, and Vajrayna, but I
dont know Vajrayna. Vajrcryas have a special initiation for this [he meant, I
think, the Consecration of a Vajra-Master]. Vajrayna is the highest but what I
practise is Mahyna.

129

He said this even though he had taken Tantric Initiation and was quite knowledgeable
about Tantric Buddhism. There are two probable reasons why, nonetheless, he identified
himself as practising Mahyna and not Vajrayna. The first is that Vajrcryas are
particularly associated with the Vajrayna (it could not survive without them) and so, as a
kya, he thought of himself as practising the Mahyna. The second is that Vajrayna is
frequently associated with the practice of magic, healing, and supernormal powers. Ask about
Vajrayna and one often receives elaborate and hair-raising tales about nocturnal visits to
powerful, but unpredictable, Tantric practitioners.
A pait, Asha Kaji Vajracharya, explained the Three Ways thus:
rvakayna [the Disciples Way] refers to monastic practice (bhikucary), and this
was preached by Lord Buddha at Srnth.
Pratyekabuddhayna [the Solitary Buddhas Way] is that practised alone.
Mahyna [the Great Way] teaches the practice of the Bodhisattvas: Lord Buddha
preached this on Gdhraka hill.
And on Vajraka hill he brought to light Vajrayna [the Diamond Way]. There are
two types of Vajrayna. The first is onefold (ekkar): this is for one who does not
marry and lives alone. The second is twofold (dvrkr): this what is known as
Tantrayna [the Tantric Way] or Sahajayna [the Way of Innate (Bliss)], in which a
couple (strpuru) eat from the same plate. It is called Sahajayna because they are
united (sajog juye) and eat each others polluted food. The practice of Sahaja is
Tantric Initiation; that is why those who come to take it on their own are sent way [at
this crucial point].
Guna Ratna Shakya, a disciple of the famous pait of Bu Bh, Ratna Bahadur
Vajracharya, explained it in this way:
There are four ways of doing pj:
(i)
focused on gods (devattmak);
(ii)
focused on the five constituents (pacaskandhtmak);
(iii)
focused on emptiness (nytmak);
(iv)
focused on yoga (yogtmak).
The first is the approach of those who believe gods are responsible for everything,
bringing rain, causing illness, and (they also act from) fear of the gods punishment
(do). The second is the approach of the Theravda, who do not worship gods but
only the five constituents. The third is the approach which asserts that nothing exists,
there are no gods; it is completely nihilistic (nstik). The fourth is the approach which
combines rvakayna and Mahyna, which combines (i), (ii), and (iii). It puts gods
in front, analyzes all things as the five constituents, and realizes that all is empty
(nya). This, combining

130

emptiness and compassion (karu), is the yogic way of worship and is called
Vajrayna.
Initially Guna Ratna had given a more light-hearted answer to my question, what is
the difference between Mahyna and Vajrayna, which had puzzled both me and the others
present. In the light of the above it should be easier to understand:
The Disciples Way is like making noodles from flour. The Solitary Buddhas Way
is making chapatis. The Great Way is like the repeated kneadings necessary to make
lkhmari [the magnificent and gigantic Newar wedding pastries]. The Diamond
Way is like going into a sweetshop, mixing all the pastries they have, crushing them,
eating them all, using everything and leaving nothing.
5. Conclusions
I want very briefly to present three conclusions. Firstly, something about the supposed
antinomianism of the Tantric scriptures. The scriptures contain two idealsOne is the
antinomian saint who observes no rules and is liberated here and now, the divine madman.
The other is the learned, courteous teacher who realizes that all is non-dual and that rules are
merely conventional, yet continues to act as if the distinctions they are based on are real
because he feels such compassion for suffering creatures that he wants to lead them on the
right path. The existence of this latter ideal shows that Tantric Buddhism was routinized, to
use Max Webers term, at a relatively early period. That is to say, institutionalized Tantric
Buddhism like that of the Newars is very old. Newar Buddhism is a genuine inheritor, and
not a deformation, of a type of Buddhism which is at least 8 or 900 years old.
My second conclusion also relates to this history. The form of Buddhism practised by
the Newars drew its inspiration from great Buddhist centres in what are now Bihar, Bengal,
Bangladesh, and Orissa. With their demise in the thirteenth century, Newar Buddhism turned
in on itself. In its relation to them it was a little tradition compared to their learned monastic
Great Tradition. Nowand no doubt then toothe local kyas and Vajrcryas represent
the Great Tradition vis--vis the Newar Buddhist laity. It is the practices of the kyas and
Vajrcryas which define what is traditional Newar Buddhist orthodoxy and orthopraxy.
These two terms Great and Little Tradition are very unfortunate and have often been
criticized by anthropologists. In this context they are perhaps even more so, since they
conjure up the idea of separate and separately legitimated traditionswhereas here we are
dealing with three sets of practices which are ultimately legitimated by appeal to the same
criteria. Unsatisfactory though these terms are, I nonetheless still feel compelled to use them,
since I know of no better ones to label what I take to be an anthropological reality.

131

Finallythe conclusion of my conclusionone has to distinguish between a

scriptural doctrine and the way it is used. All Three Ways I have outlined in the practice of
Newar Buddhism are conceived to be for the individual seeking enlightenment. The fact that
all three, in spite of the different ideals envisaged in each, are focused on the individual was
no doubt a factor enabling Buddhism to absorb Tantrism so thoroughly. However, we have
seen that in practice Vajrayna Buddhism is used for the restrictive secret face of Newar
Buddhism, whereas Mahyna Buddhism is its public one. Although in theory the Vajrayna
is open to allor at least all may aspire to itin practice its stress on secrecy and initiation is
used to exclude the low born and outsiders, to define patrilineages, and to maintain a caste
monopoly of Buddhist monkhood and priesthood.

132

Councils as Ideas and Events in the Theravda

Charles Hallisey*

Before giving a short survey of the

traditions relative to the Buddhist
Councils, it seems advisable to state
what these councils were.
Louis de la Valle Poussin1
It is a standard scholarly practice to begin a presentation of research with a definition, in
the strict sense of the word, of the subject which has been investigated. We are
encouraged to do this early in our education, with reminders to define your terms, and
we generally admire the clarity that a good definition can bring to an argument. We value
definitions, even working definitions, in the presentation of research so routinely that
we rarely consider the implications of this practice for research itself. All of us know by
hard experience that the actual processes of research are far messier than is suggested by
the way we present our research. Even so, we assume certain parallels. Our research
begins with a choice of a subject that seems to function like the initial definition in a
research presentation. But while we may begin with an attempt to define a subject, as a
practical way of limiting and focusing our research, in the course of investigation we
often discover a state of affairs quite different from what we had anticipated. This
common turn of events can dismay or discourage, but it can also delight. A new
discoverymajor, of courseis the stuff scholarly dreams are made of.
It would be one thing if the first definition, taken as the starting point for research,
were wrong, out and out wrong, and thus could be replaced by the new understanding.
This is actually quite rare, however, in large part because we usually adapt these first
working definitions from other sources. As Bernard Cohn has said, each piece of
research doesnt start as if it were year one, nor does the

An earlier version of this paper was presented as part of a panel on Rethinking Theravda
Buddhism at the Association for Asian Studies Annual Meeting, Washington, D.C., March 1989. I
have benefited greatly from the papers and comments of my co-panelists at that meeting (George
Bond, John Ross Carter, Steven Collins, Charles Keyes, and Frank Reynolds).
1
Louis de la Valle Poussin, Councils and Synods: Buddhist, ERE, vol. 4, 179.
133

scholar begin as a tabula rasa to be instructed by the native or the document, nor is he or
she merely a pencil which records in some fashion what is read or seen.2 In fact, what
we choose to study is probably set more by the units of study or theoretical assumptions
in our particular field of training than by the subject matter itself. The seemingly better
understanding of a topic ends up being merely another interpretation yielded by a
different theoretical perspective. One understanding does not negate another, nor do we
see ways that they might be related in a common schema. Different understandings are
allowed simply to co-exist, in mutual isolation, while we argue back and forth.
For a community of scholars this is disastrous. Our research presentations end up
bearing witness to our lives in an academic Tower of Babel.
It is easy to see how this occurred. As students of Buddhism, we may welcome
new approaches to the rich resources of the Buddhist traditions.3 There is more than
enough work to do, and labourers are still few. Perhaps the chronological and spatial
extent of the Buddhist traditions made the introduction of some new approaches
relatively unproblematic at first. For example, anthropologists and sociologists, seeking
to understand the workings of culture and society, were naturally drawn to the study of
contemporary Buddhist communities, fields of research which textualists and historians
generally preferred to ignore.
This division of labour appears neater than it actually is. It looks as if it is a
division of subject matter, with historian and anthropologist each examining what he or
she is best prepared to study. But it sometimes masks a more profound difference in
theoretical perspective. This difference becomes an obstacle when both anthropologists
and historians have their own definite ideas about a common subject. Such is the case
with the councils (sagti or sagyan) in the Theravda.
The purpose of this paper is propaedeutic. That is, I wish to follow La Valle
Poussins advice, given in the quotation at the start of this paper, and sketch out what the
councils were in the Theravda; this sketch is a preliminary to the survey of the councils I
am working on. My purpose is to define the subject by showing how different scholarly
understandings of councils may be combined to interpret their place in the Theravda as a
historical tradition. Moreover, in sketching out what the councils were, I hope to indicate
how they might be fruitfully studied. These programmatic comments, I think, will have
applicability to other areas in the study of Buddhism.
Charles Prebish evocatively referred to the first Buddhist Councils as problems
which have haunted western Buddhological research through almost all of its last

Bernard Cohn, History and Anthropology: The State of Play, in An Anthropologist among the
Historians and Other Essays, Delhi, Oxford University Press, 1987, 47.
3
Indeed this is one of the purposes of the Buddhist Forum. See Tadeusz Skorupskis Introduction to
The Buddhist Forum, vol. I, London, SOAS, 1990, 1.
134

one hundred years.4 Even though these first Buddhist councils might seem to be a
natural starting point for any investigation of the Theravdin councils, to begin with them
is actually to become embroiled in a complicated and ongoing debate.5 Much of this
debate, especially on the historical value of different accounts about these councils,
seems one-sided when viewed from the vantage point of later Theravdin cultural history.
To begin from the later Theravdin councils is equally problematic though, since these
events conventionally draw their sanction from accounts of the earlier councils.
Our difficulty is an academic chicken and egg problem: in order to understand the
individual councils, the parts, we need to have some idea of the councils as a whole, but
we generally only know the whole through the individual parts. Some way out of this
hermeneutical circle, in this particular instance, can be found by looking at a similar case,
the role of the Pli canon in the Theravda.
In recent years, two different orientations to the Pli canon have emerged in the
scholarly literature concerned with contemporary Theravdin communities.6 Both are
reactions against the interpretive prominence the canon has had in Buddhist studies. One
orientation emphasizes the actual possession and use of texts. Charles Keyes, for
example, has argued that
the relevance of texts to religious dogma in the worldview of any people cannot
be assumed simply because some set of texts has been recognised as belonging to
a particular religious tradition. It is necessary, in every particular case, to identify
those texts that can be shown to be the sources of dogmatic formulations that are
being communicated to the people through some medium. There is no single
integrated textual tradition based on a canon to the exclusion of all other texts
The very size and complexity of a canon leads those who use it to give differential
emphasis to its component texts. Moreover, even those for whom a defined set of
scriptures exists will employ as sources of religious ideas many texts which do
not belong to a canon [Finally,] for any particular temple-monastery in
Thailand or Laos, the collection of texts available to the people in the associated
community are not exactly the same as those found in another temple-monastery.
In brief,

Charles S. Prebish, A Review of Scholarship on the Buddhist Councils, JAS, 33, 1974, 239.
Bibliographic information on this vigorous and inconclusive debate may be conveniently found in
J.W. de Jong, A Brief History of Buddhist Studies in Europe and America, Varanasi, Bharat-Bharati,
1976, 3031, 67, and in Prebishs article, cited in note 4.
6
My comments here depend greatly on the work of Steven Collins in his paper, On the Very Idea of
the Pli Canon, JPTS, vol. 14, forthcoming 1990.
5

135

the relevance of textual formulations to religious dogma in popular worldviews is

problematic in each specific case.7
The orientation that emerges from this discussion is primarily concerned with
issues of transmission and distribution: who had what texts when?
A second orientation emphasizes the idea of the canon. Franois Bizot, for
example, has pointed out with respect to modern Khmer Buddhism that the term tipiaka
refers less to a collection of texts than to an ideological concept.8 This orientation is
concerned more with the internal constitution of the tradition: what makes the Theravda
valid from the point of view of Theravdins?.
The two orientations highlight different facts. The first highlights the presence
within the Theravda of standpoints which are geographically and historically very
particular. This particularity, however, may be obscured for Buddhist individuals and
groups by the phenomenon highlighted by the second orientation, a perspective which is
considerably loftier and less determinately located.
These two orientations, taken together, can and should replace an assumption that
was once more widely held than it is today, although it still has a pernicious influence in
scholarship. That is, it was once widely assumed that the Pli canonor the early
Buddhism which was reconstructed from the canonconstituted the Theravda in all its
essentials. With this assumption, almost all interesting questions about the Theravda as a
historical tradition remained unasked. The two, more recent orientations are clearly an
improvement on that assumption, and although they were developed in connection with
the study of contemporary Buddhism, they are still very useful as tools for historical
investigations.
In shorthand, I will call the first orientations focus event and the focus of the
second orientation idea.9 By calling the first event, I wish to stress how a

Charles Keyes, Merit-Transference in the Kammic Theory of Popular Theravada Buddhism, in

Charles Keyes & E. Valentine Daniel, editors, Karma: An Anthropological Inquiry, Berkeley,
University of California Press, 1983, 272. Compare Richard Gombrichs comments: The contents of
sacred texts are not simply reproduced in the doctrines of the religions which venerate them; there
must be interpretation and selective emphasis. This is obviously true when the corpus of sacred
literature is large, as in Christianity and Buddhism. Historians of these religions may therefore ask
why certain doctrines and certain scriptures have been emphasized at the expense of others.
Buddhist Karma and Social Control, CSSH, 17, 1975, 212.
8
Franois Bizot, Le figuier cinq branches, Paris, EFEO, 1976, 21. This orientation has been
elaborated very convincingly in the paper by Steven Collins cited in note 6.
9
Bernard Cohn, op. cit., 45, speaks of the same distinction in the following way: We write of an
event as being unique, something that happens only once, yet every culture has a means to convert the
uniqueness into a general and transcendent meaningfulness through the language members of the
society speak [For example], the death of a ruler may be mourned by rituals which turn the
biographic fact of a death into a public statement relating not only to a particular ruler but to rulership
per se. In many societies ritual transforms uniqueness into structure.
136

particular set of circumstances are largely accidental and thus unique. A scholar
employing this orientation as an interpretive tool will discover how this set of
circumstances came about and the impact that it subsequently had. When these multiple
sets of circumstances collectively change to a substantial degree, then one may speak of a
transition or transformation in the tradition.10 By idea, I mean to emphasize persisting
patterns of meanings and norms which mark the Theravda; this notion could equally
well be called structure. These patterns can sanction or even shape the actions of
individuals and groups.
The notion of event is more conventionally historical in its emphases, while the
notion of idea is more typical of anthropology. Each can be used as a heuristic tool
independent of the other, according to the research purposes of the scholar, but the
phenomena they refer to are inevitably interrelated. The reason for the transmission of
manuscripts of the tipiaka cannot be separated from the idea of the canon. And if we are
aware that the texts of the canon are variously interpreted in different circumstances, as
Keyes argues, at the same time we need to remember that the idea of the canon provides a
framework which gives relative meaning and significance to the reading or hearing of
other texts, or the performance of rituals.
I am not advocating that we should have recourse to ever-ready ahistorical
frameworks of meaning here. We need to discover frameworks of significance, like the
idea of the canon, within particular historical contexts; we should probably expect to find
that persons may employ more than one framework within any given context. But when
we are able to identify such frameworks in situ, we will then be able to see the
constructedness of the Theravda tradition. To put it another way, when we discern
events being given meaning by ideas and ideas being shaped by events in
particular contexts, we will be able to see the Theravda as a tradition whose identity is
continually being constituted and reconstituted, with its history and account of continuity
in difference.
These general lessonsa distinction between event and idea, and the correlation
of these two notionscan be applied specifically to the Theravdin

10

See Charles Keyes, The Golden Peninsula, New York, Macmillan, 1977, 86: If the true Buddhist is
one who seeks to become an Arahat, the fully perfected monk who attains enlightenment, then quite
obviously Buddhism could never be a popular religion. It would be a religion of only a small number
of adepts. Ancient Buddhism may have been such a religion, but it underwent a transformation first in
the third century B.C., when it was brought under the patronage of King Asoka who set an example
for other ruling elites. Theravada Buddhism was further transformed in the fifth century A.D. through
the theological interpretations of Buddhaghosa and several of his contemporaries. Finally, it went
through yet another transformation in the twelfth to thirteenth centuries when it became a universal
religion, a religion for peasant farmer as well as for monk and king.
137

sagtis. While the first lesson is already a major, if implicit, part of the scholarly
literature on the Theravdin councils, the second still needs to be learned.11
Buddhological investigation of the first three councils at Rjagaha, Vesl, and
Paliputta has generally been part of a larger scholarly project to shed light on the
Buddhist past as it really was, to use von Rankes phrase, with the result that the
councils were defined, almost a priori, as events. While the goal of recovering the past
as it really was remains unrealizedif indeed such an aim is even possible
comparative research on the various accounts of the First Councils found in Buddhist
literature did succeed in making it impossible to attribute historical accuracy to any single
description preserved by a particular Buddhist tradition. In short, one of the
accomplishments of a century of research on the First Councils has been to drive a wedge
between our perception of the councils as historical events and Buddhist ideas about the
councils.
Scholarly reaction to this distinction between councils as ideas and councils as
events gradually evolved. An initial and understandable reaction was to see the
distinction as offering a clear and sharp choice: the accounts either contain real history or
they are fiction. Commenting on the account of the councils in the Vinaya, Oldenberg
wrote, what we have here before us is not history, but pure invention.12 A tendency to
see the accounts as essentially fictions was perhaps strengthened by the development of
questions about the motives which could have

11

For an example of a failure to correlate the two notions of event and idea, see the critique by
Michael Carrithers of Stanley Tambiah on the subject of Parkkamabhus council in twelfth-century
Sri Lanka. Carrithers writes: In World Conqueror Tambiah, pursuing the relationship between kings
and monks, dwells at length on the purification of the Buddhist order carried out by Parkkamabhu I
of Sri Lanka. He argues that this was patterned after a similar act of the Emperor Aoka, preceded
other similar royal acts, and was therefore part of a pervasive pattern in the relationship between
royalty and the Buddhist order throughout Buddhist history. On this account all purifications were
analogous, the working out of a particular timeless relation between kings and monks. But such an
account leaves out the single most important feature of Parkkamabhu Is reform, namely that it was
a radically new interpretation of the kings role, an interpretation which set a new pattern for
Theravda and Theravdin kings. (Michael Carrithers, Buddhists without history, CIS, N.S. 21,
1987, 167.)
Tambiahs rejoinder to Carrithers, in the same issue of CIS, does show a movement towards
the combination of heuristic concepts that I have in mind: Carrithers seem(s) to have the
simplistic notion that there are only two kinds of historical interpretation possiblethere is either a
stasis and repetition of the past or there is a radical change. (He does) not seem to appreciate both the
complexity and the pervasiveness of a historical condition in which certain kinds of persistence
coexist with certain kinds of change of state, and such amalgams and syntheses of varying kinds and
varying degrees of cohesion and tension characterise much of the so-called flow of history. (Stanley
Tambiah, At the confluence of anthropology, history, and indology, 194.)
12
Hermann Oldenberg, Introduction, in The Vinaya Piaka, ed. by H. Oldenberg, London,
Williams & Norgate, 1879, xxvii.
138

led to the composition of the council narratives. Przyluski, for example, argued that one
(could) explain the diversity of the accounts of the (first) council (by saying that) there
are so many different recitations [sagti] as there are sects having a distinct canon. Each
school tries to prove that its canon dates back to the origins of the Church and that it was
codified by the assembly of Rjagha.13
Przyluskis comment illustrates the possibility of discussing Buddhist ideas about
the councils independently of any judgement about the historical incidents themselves.
He displays a significance in the ideas that is worth pursuing on their own terms even if
the accounts are not reports of real occurrences.
In a similar way, scholars have formulated questions about the events which can
be pursued in isolation from Buddhist ideas about the councils. La Valle Poussin
intimated this possibility in his entry on Buddhist councils in the Encyclopedia of
Religion and Ethics: While it is impossible to accept the Buddhist opinion, which views
them as ecumenical assemblies after the Nicene type, it is at the same time necessary to
explain how Buddhist monastic life, without the help of such solemn assemblies,
nevertheless resulted in a sort of catholicism, and secured the redaction and the
compilation of Canons of scripture very like one another.14 La Valle Poussins position
was that while acknowledging the possibility (even the probability) of synods, we are at
no loss to point out more certain and farther reaching causes of the facts to be explained,
viz. the formation of the body of the Scriptures, the general (if not strict) consensus of
the sects of the Hnayna as concerns Buddhas teaching, and conversely, the splitting of
the Order into sects.15
More recently, in an earlier series of the Buddhist Forum, Richard Gombrich
illustrated another way that the First Councils might be discussed as events independent
of the Buddhist accounts, although he offers a more positive evaluation of those accounts
than La Valle Poussin allowed. The discussion quoted here takes up the same question
as La Valle Poussin: how did the teachings of the Buddha, given over a long period of
time in many places, come to be collected into what eventually became the Pli Canon?
The Sagti-suttanta begins by recounting that at the death of Nigaha
Ntaputta his followers disagreed about what he had said. The same passage
occurs at two other points in the Pli canon; but it makes good sense in this
context, for it is the occasion for rehearsing a long summary of the Buddhas
teaching in the form of mnemonic lists. The text says that the rehearsal was led by
Sriputta, in the Buddhas lifetime. Whether the text records a historical incident
we shall probably

13

Jean Przyluski, Le concile de Rjagha, quoted in Prebish, 243.

La Valle Poussin, ERE, vol. 4, 179.
15
La Valle Poussin, ibid., 179.
14

139

never know. But that is not my point. I would argue that unless we posit that such
episodes took place not merely after the Buddhas death but as soon as Sagha
had reached a size and geographic spread which precluded frequent meetings with
the Buddha, it is not possible to conceive how the teachings were preserved or
texts were composed. By similar reasoning, something like the first sagyan
(communal recitation) must have taken place, otherwise there would simply be no
corpus of scriptures. Details such as the precise time and place of the event are
irrelevant to this consideration.16
The historical reasoning in this discussion is noteworthy. I would especially like
to draw attention to Gombrichs use of the historians knowledge of the outcome of the
past to provide an alternative perspective with which to view and reconstruct the
processes of early Buddhist history. His reasoning restores some balance to the scholars
choice of seeing the first Buddhist councils as either events or as ideas, as fact or fiction.
It is understandable that since there is no archaeological or epigraphical evidence actually
from the First Council, its historicity could appear quite suspect in the light of the all-tooobvious, vested interests expressed in the various council narratives. Gombrichs
reasoning makes us seriously consider the historicity of an event like the First Council as
a necessity.17
Is what we learn from this argument, however, transferable to later events which
are also compared to or described as councils? This question would apply not only to the
Second and Third Councils, but also to the events sponsored by Theravdins in the
medieval and late periods. Gombrich seems to suggest such a possibility when he says
without qualification in another context the Councils (sagyan), better termed
Communal Recitations, served the function of systematizing knowledge and perhaps of
organizing its further preservation.18 Similarly, K.R. Norman seems to project a pattern
from the First Council onto the events of the medieval Theravda:

16

R.F. Gombrich, Recovering the Buddhas Message, The Buddhist Forum, vol. I, 6. It is interesting
to note that Minayeff took the minor details of the accounts as to some extent historical (cited in La
Valle Poussin, 182). Thus both the plot and the details of the first councils have been described as
both fact and fiction. Cf. Richard Gombrich, How the Mahyna Began, The Buddhist Forum, Vol.
I, 26.
17
It is also the case that the general scholarly tendency now current is to give the Buddhist accounts
the benefit of the doubt, in contrast to the inclination of scholars around the turn of this century.
K.R. Norman, for example, writes: Although we may have reservations about the texts which were
dealt with at the first council, there is no reason to doubt the general way in which it was held. (K.R.
Norman, Pli Literature, Wiesbaden, Harrassowitz, 1983.) At one time, this would have been a highly
provocative statement.
18
R.F. Gombrich, How the Mahyna Began, 2526.
140

It is not inappropriate to talk of a Burmese or Siamese or Sinhalese tradition for

the transmission of a particular text, and the differences which we find between
the readings of the manuscripts belonging to the various traditions must go back
to the councils which have been held from time to time in the different countries.
[T]he value of each tradition (stemming from different councils) will depend upon
the care with which evidence for variant readings was sifted, and the criteria
which were adopted as the basis of the decisions which were made.19
A projection of patterns reconstructed from events, however, is misleading. The
Mahvasa, the great chronicle of Sri Lanka, records at least twelve councils in
medieval Sri Lanka, and it is notable that a communal recitation or recension of the
tipiaka is not mentioned as being part of any.20 In fact, I am not aware of any definite
evidence dating from the medieval period itself which indicates that communal
recitations were held, although events which did occur still claimed the Third Council as
a precedent.21
Thus, as much as I admire Gombrichs historical reasoning in connection with the
events of the First Council, I also think we should keep in mind that it is applied to a
specific body of evidence, in connection with a particular problem in reconstructing the
Buddhist past. How much this reconstruction can serve as a guide to other events is a
more difficult issue. On the one hand, the historical problems which confront a student of
the Theravda, whatever the period, are not quite the same as those facing the students of
early Buddhism, even when both may be concerned with similar issues. This difference
is, in part, due to the increasing complexity of the tradition itself; for example, the student
of the Theravda, aware of the bhaka system and the use of writing, must acknowledge
that communal recitations were not strictly necessary for the

19

K.R. Norman, Pli Literature, 13.

See Mahvasa, 39:57; 41:2; 44:46; 44:76ff; 48:71; 51:64; 52:10; 52:44; 73:11ff; 78:2ff; 84:7;
91:10; 100:44.
21
An event in the medieval period which does approximate the conventional functions usually
attributed to a communal recitation, such as preserving knowledge and transmitting texts, is
Vijayabhu IIIs patronage of a rewriting of the canon (Mahvasa 81:40-45). Significantly, the
participants in this event were laymen, and it is not described as either a sagti or a sagyan.
20

The event which perhaps comes closest to an actual communal recitation is the scripture revision
and recitation sponsored by King Tilaka at Chiang Mai in 14757; this event is described in the
Jinaklaml (London: PTS, 1962), 115. Again, this event is not described in the Jinaklaml as a
sagyan, although later texts in the Thai tradition (e.g. Sagitiyavasa) do accept it as such. It is
also significant that this event was probably held after the writing of the Saddhammasagaha, which
radically recast the traditional idea of a sagyan.
141

preservation of the Pli canon.22 On the other hand, simply projecting a pattern
reconstructed from one event onto other events avoids asking how Theravda Buddhists
themselves transformed unique events into ideas of general meaningfulness.
The Theravdas transformation of councils from events into ideas has been
brilliantly investigated by Heinz Bechert in two articles which may be read together
profitably.23 Becherts main purpose in the first article is to add to our knowledge of the
Third Council as a historical event, but as part of a secondary argument, he traces how
the events of that council were subsequently transformed in the Pli commentaries and
chronicles. In a manner reminiscent of Pryzluskis explanation of the diversity in the
accounts of the First Council, Bechert argues that the events at Paliputra were actually a
synod of a monastic sub-group (nikya), which later and for obvious reasons were
portrayed as a unification and purification of the entire Sagha. In a second article on
ssana reform, Bechert discusses how these ideas about Asoka and the Third Council
were used in the medieval Theravda, arguing that the transformation of the historical
Asoka into a Theravdin sectarian in the chronicles and commentaries provided a
foundation for ideology of state-Sagha relations in Theravda countries.24
Keeping Becherts insights, I would turn his statement around and say that
Theravdins preferred to convert unique events into phenomena of general meaning and
import by historicist transformations.25 The presence of historical consciousness in the
Theravda tradition has frequently been noted, but its full significance in the
development of the tradition still remains obscure.26 Even so, there is ample evidence that
one of the uses of history in the Theravda tradition

22

On the bhaka system, see E.W. Adikaram, Early History of Buddhism in Ceylon, Migoda, D.S.
Puswella, 1946, 2432. Adikaram traces this system of reciters, but also suggests that Buddhaghosa
saw the bhaka system as making sagyan unnecessary as a means of preserving and transmitting
the canon.
23
Heinz Bechert, Aokas schismenedikt und der Begrift Saghabheda, WZKSO, 5, 1961, 1852,
and, by the same author, Theravda Buddhist Sagha: Some General Observations on Historical and
Political Factors in its Development, JAS, 29, 1970, 761778. See also Heinz Bechert, The
Importance of Aokas So-called Schism Edict, Indological and Buddhist Studies: Volume in Honour
of Professor J.W. de Jong, Canberra, Faculty of Asian Studies, 1982, 6168.
24
H. Bechert, Theravda Buddhist Sangha, 764.
25
On the historicist transformation of the Pali canon, see the paper by Steven Collins cited in note 7.
26
See, for example, Heinz Bechert, The Beginnings of Buddhist Historiography: Mahvasa and
Political Thinking, in Religion and Legitimation of Power in Ceylon, edited by Bardwell L. Smith,
Chambersburg, PA, Anima, 1978. In another vein, Stanley Tambiah, in World Conqueror and World
Renouncer, New York, Cambridge University Press, 1977, says that one of the most important
features of [the] Theravada Buddhist politics is their active consciousness of historical continuity
(page 518, emphasis in the original).
142

was to give individual events a general significance with ideas that have the appearance
of being reports about previous events.
On the functioning of a previous event as an idea which can sanction other events,
David Lowenthal has written that the past validates present attitudes and actions by
affirming their resemblance to former ones. Previous usage seals with approval what is
now done. At the same time, precedent legitimates action on the assumption, explicit or
implicit, that what has been should continue to be or be again.27
The use of the past to provide a general order of meaning is common in
Theravdin literature and inscriptions. We see this use of the past, for example, in
connection with Parkkamabhu Is reform of the Sagha in twelfth-century Sri Lanka,
as when the Mahvasa explicitly compares that king to Asoka in its detailed description
of that council.28 The historicist transformation of the event of the Third Council into an
idea is even more prominent in Parkkamabhus Galvihara inscription which explains
his motives for purifying the monastic order of his day:
Now, His Majesty reasoned thus: Seeing over and over again a blot such as this
on the immaculate Buddhist religion, if a mighty emperor like myself were to
remain indifferent, the Buddhist religion would perish, and many living beings
will be destined to the apya. Let me serve the Buddhist religion which should
last five thousand years.
[His Majesty pondered that in days gone by] the great king Dharma Aoka,
enlisting the services of Moggaliputta Tissa, the Great Elder of the Buddha Cycle
acknowledged by the Buddha himself, crushed out the sinful bhikkhus; suppressed
the heretics; purged the religion of its impurities and brought about the holding of
the Third Rehearsal of the Dhamma. In like manner, His Majesty
[Parkkamabhu] also enlisted the services of those (Udumbara-giri) bhikkhus
and, removing from the Masters religion many hundreds of sinful monks,
brought about a rapprochement of the three fraternities and a coalition of them
into one single fraternity (nikya)a reconciliation which former kings, despite
their great efforts, were not able to effect, even though there were at the time
eminently holy personages endowed with aggregates of diverse faculties such as
the six psychic powers, etc.29

27

David Lowenthal, The Past is a Foreign Country, New York, Cambridge University Press, 1985,
40. On the past as a sanction in the Theravda tradition, see S. Tambiah, World Conqueror and World
Renouncer, 528ff.
28
Mahvasa, 78, 27.
29
Epigraphia Zeylanica, London, Humphrey Milford, 1928, II, 274275. Concerning this council, see
the exchange between Michael Carrithers and Stanley Tambiah cited in note 11.
143

In this inscription, Asoka and the Third Council are claimed both as precedent and
model for Parkkamabhus own efforts to unify and reform the monastic order. The
precedent is not an exact blueprint for the proceedings, however, since there is no
mention of a recitation of the canon as part of Parkkamabhus reform, even though the
inscription does refer to the Third Dhammasagti. Perhaps the more literal meaning of
the term sagti (communal recitation) was no longer noticed, and the Third Council
was simply taken as providing a precedent for combating the forces of decay within the
Sagha.30 Obviously, the purification of the Sagha is linked to the preservation of the
ssana too.
We should also note that the holding of a council is linked to a persona of the
king. Parkkamabhus readiness to purify the monastic order is an expression of his
desire to do service to the Buddhas ssana and to humanity in general. The council is
also linked to the establishment of unified, that is, valid monastic order, an issue that
preoccupied Theravdin kings and monks throughout the medieval period.31
All of this is at a level of ideas of general significance and meaning, and not
events. The actual event itself is absent in this passage. Just as it is a mistake to take
events directly as ideas, by projecting a pattern from one event onto other events, so it is
a mistake to take ideas directly as events, by assuming that a particular council is a
simple realization of the general idea. Events and ideas need to be correlated with each
other, not rendered identical. The councils of the medieval and late Theravda were still
unique events, even as they were shaped and given significance by Theravdin ideas
about the councils. Any event is both more and less than these ideas may suggest.
Michael Aung Thwin has shown that ssana reforms in medieval Burma, modelled on
the Third Council, had very tangible political and economic benefits for the kings who
initiated the purifications.32 The ideas about the councils affirm that some aspects of the
events are not as crucial as others.
I have spoken of ideas about the councils shaping specific councils as events, but
this occurred in only the most general manner. Andr Bareau, in his classic study of the
first Buddhist councils, remarked that as astonishing as it may be, Buddhist literature
tells us almost nothing about the ritual and ceremony of a

30

H. Bechert, Theravda Buddhist Sangha, 763.

Councils were held, it seems, for the purpose of reordination of monks in a valid monastic lineage.
See Franois Bizot, Les traditions de la pabbajja en Asie du Sud-Est, Gttingen, Vandenhoeck &
Ruprecht, 1988.
32
Michael Aung Thwin, The Role of Sasana Reform in Burmese History: Economic Dimensions of
a Religious Purification, JAS, 38, 1979, 671688. See also Victor B. Lieberman, The Political
Significance of Religious Wealth in Burmese History: Some Further Thoughts, JAS, 39, 1980, 753
69 and Michael Aung Thwin, A Reply to Lieberman, JAS, 40, 1980, 8790.
31

144

council.33 There was actually very little guidance available to later Buddhists on the
procedures for holding a council. Instead, Buddhist literature provides only normative
motives for holding councils and information about the general benefits that would
follow.
This lack of specificity about the procedures of a council allowed for some
surprising interpretations of a sagyan. Late Theravdin inscriptions from Cambodia
indicate that the word sagyan was used to designate a form of merit-making, in which
a small number of monks recited or discussed a few texts over a number of days.34
Another elaborate association of sagti with specific acts of merit is found in the
Saddhammasagaha, a Pli text probably composed in Thailand in the fourteenth or
fifteenth century.35 This text is often utilized as a historical source,36 although it presents
itself as an nisasa work, and describes the advantages that accrue to those who
themselves write (the piakas), those who make others to write, and those who approve of
it.37 It also gives an account of the benefits that come to those who listen to the
Dhamma. These discussions of nisasa are preceded by a very innovative history of the
councils in the Theravda.
The Saddhammasagaha describes seven comparable events, the usual three
councils in India, plus four others in Sri Lanka. These other four are the recitation of the
Vinaya by the monk Ariha at the time of the introduction of Buddhism to Sri Lanka
(which the Saddhammasagaha numbers as the fourth council), the writing down of the
canon at Aluvihara, Buddhaghosas editing of the commentaries, and Parkkamabhus
council which is characterized as a revision of the canon and the source of the Pli
subcommentaries.
The description of an event like Buddhaghosas translation of the ahakaths as
similar to a council may seem surprising when viewed with a literal definition of a
sagti, but it makes good sense when seen as a historicist strategy for giving general
significance to numerous small events. In the Saddhammasagaha, paying for even a
portion of a Buddhist book to be copied is classed as belonging to the

33

Andr Bareau, Les premiers conciles bouddhiques, Paris, Presses Universitaires de France, 1955,
137.
34
See the following articles by Saveros Pou (also known as Lewitz), Inscriptions modernes dAngkor
34 et 38, BEFEO, 62, 1975, 286, 290; Inscriptions modernes dAngkor 35, 36, 37, et 39 BEFEO
61, 1974, 316; Les inscriptions modernes dAngkor vat, JA, 260, 1972, 123. Although the
inscriptions are called modern, they are all from the 1418th centuries.
35
Saddhamma-sagaha, ed. Ven. N. Saddhananda, JPTS, 4, 1890, 2190. See also the translation by
Bimala Churn Law, A Manual of Buddhist Historical Traditions, Calcutta, University of Calcutta,
1963.
36
K.R. Norman calls it a bibliographic text, and utilizes it as a source for information about the
writing of different Buddhist texts; see K.R. Norman, Pli literature, 179.
37
B.C. Law, A Manual of Buddhist Historical Traditions, 99.
145

same category of event as the first writing of the canon itself. It is possible that in the
Saddhammasagaha and the Cambodian inscriptions, we see an illustration of Hocarts
thesis of nationalization by which the kings state is reproduced in miniature by his
vassals.38
Pli chronicles composed in Thailand and Burma during the eighteenth and
nineteenth centuries share with the Saddhammasagaha an interest in numbering
councils, although each differs in its final count. The Burmese Ssanavasa counts four
councils before a fifth which was held in the nineteenth century, while the Thai
Sagtiyavasa counts eight before a ninth held in Bangkok at the end of the eighteenth
century.39 This suggests that one strategy for reshaping the normative idea about sagtis
was to number councils.
The device of numbering councils allowed for a few councils to be given greater
importance than others, and thus out of this constructed sequence of events, a new
normative idea could be fashioned. The variations in the constructed sequences, however,
are quite striking, especially when they are viewed comparatively:
Mahvasa

Saddhammasagaha

(Sri Lanka, ca. 56 AD)

(Thailand, ca. 1415 AD)

1. Rjagaha
2. Vesl
3. Paliputta

1. Rjagaha
2. Vesl
3. Paliputta
4. Arihas recitation at Anuradhapura
[Writing of canon;
included but not numbered]
[Buddhaghosas editing of
commentaries;
included
but
unnumbered]
[Parakkamabhus council;
included but not numbered]

Ssanavasa
(Thailand, 18 AD)
1. Rjagaha
2. Vesl
3. Paliputta
4. Arihas recitation
5. Writing of Canon

Sagtiyavasa
(Burma, 19 AD)
1. Rjagaha
2. Vesl
3. Paliputta
4. Writing of canon
[Mindons council at Mandalay;

38

Quoted in S. Tambiah, World Conqueror, 74.

Ssanavasa, edited by Mabel Bode, London, Pali Text Society, 1897; translated as The History of
the Buddhas Religion, by B.C. Law, London, PTS, 1952; Sagtiyavasa, Bangkok, n.p., 1977.

39

146

6. Buddhaghosa

included but not numbered]

7. Parkkamabhus
at Polonnaruva
8. Tilakas at Chiang Mai
9. Rama Is at Bangkok
How might we explain these variations? Without going into the specifics of each
case, a general explanation can be offered. As we have already seen with the
Saddhammasagaha, the ideas about the councils in the Theravda were not fixed, but
were subject to alteration by the events with which they were subsequently associated. It
is thus not inconsequential that Ssanavasa and the Sagtiyavasa were written in
connection with some events that were also described as councils by their participants.
Both councils were held in a context of concern about the immediate survival of the
ssana, with Mindons Fifth Council held just after the Second Anglo-Burmese War, and
Rama Is Ninth Council after the complete collapse of the Ayudhya kingdom.40 Both
councils were quite concerned with questions of monastic governance and factionalism
within the Sagha, but neither was able to accomplish a complete unification of the
Sagha, the conventional purpose of the great medieval councils. In both cases, the
chronicles connected with these councils reconstituted the idea of a council, and shifted
the normative means for preserving the tradition from the maintenance of a pure and
unified monastic order to the possession and purification of an authoritative scripture.
Unlike so many councils held in the medieval Theravda, these two councils took as their
central purpose the revision and writing of the tipiaka, which was portrayed as following
a precedent. Theravdins, like the English, [and] no less than Indian villagers or the
faculty at the University of Chicago, act as if what was recently created and denominated
a tradition is part of their ancient heritage.41
We are finally in a position to answer the question with which I began: what were
the councils in the Theravda? They were events, unique occasions of considerable
variety. Councils were held to recite texts, settle monastic disputes, bring about monastic
unity, preserve the ssana, display the power of a king, earn merit, and so on. Some
councils succeeded in their aims more than others. But the councils are also a varied set
of ideas, all of which were used to establish the continuing validity of the Theravda as a
tradition.
Seeing the councils as both events and ideas also suggests how they should be
studied. We need empirical research into the actual accomplishments of each

40

See, for brief accounts of these councils, E. Michael Mendelson, Sangha and State in Burma: A
Study of Monastic Sectarianism and Leadership, Ithaca, Cornell University Press, 1975, 11213, 276
78, 335 and Craig Reynolds, The Buddhist Monkhood in Nineteenth Century Thailand, unpublished
PhD dissertation, Cornell University, 1973, 5055.
41
Bernard Cohn, History and Anthropology: The State of Play, 45.
147

council, as far as our sources allow. Some of these accomplishments may appear
decidedly impious, such as the economic benefits that fell to a king from monastic
reform. We also need to keep in mind the historical importance of the idea of a council
for understanding the collective actions of those persons who convened, participated in,
and accepted the authority of a council. In turn, we need to be alert to ways that events
left an imprint on these normative ideas; we need a history of the reception of these ideas
in subsequent contexts.
Only when we begin to trace the history of phenomena with a dual character as
events and ideas will we begin to see the Theravda as it truly is: not as an unchanging
conceptual system, not as a static structure, but as a complex human movement in a
perpetual process of constitution and reconstitution. With such a history, we will see the
Theravda yath bhtamas it was, as it became, as it is.42

42

I owe this formulation to John Ross Carter.

148

The Practical Implications of the Doctrine of Buddha-nature

S. Hookham

Contrary to certain currents of widespread opinion both among Eastern and Western
scholars, there are two fundamentally different views of the nature of man, the mind and
the spiritual path within the Buddhist tradition, each of which has equal claim to
orthodoxy.
In this paper, which is exploratory in nature, I shall briefly outline these two
views and then ask the question of what the psychological or social effects of holding one
or other of these views might be. The views I have in mind are expressed in the Tibetan
Buddhist tradition as the view of self-emptiness and the view of other-emptiness (rangstong and gzhan-stong).

The Buddhist doctrine of emptiness1

Although it would be a hopeless task to try to explain self-emptiness and other-emptiness
in a few words, roughly speaking self-emptiness is the empty nature of illusory
phenomena that are not actually there and other-emptiness is the empty nature of reality
which actually is there. Although the term gzhan-stong has become strongly associated
with Tibetan Buddhist polemics, originally it was coined together with its complementary
term rang-stong as a means of distinguishing two different kinds of scriptural statement
as regards emptiness. This is how the great (yet much maligned) Jonangpa master
Dolpopa Sherab Gyaltsen used it in thirteenth-fourteenth century Tibet.

a. Self-emptiness
The term self-emptiness applies when it is said that the ordinary common sense world
around us is empty like a dream or an illusion. One immediately focuses on the idea that
it lacks reality. One thinks of the fleeting nature of life, how things are insubstantial and
likely to disappear at any moment. Self-emptiness means the emptiness of things like this
in themselves of themselves as, for example, a dream

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
1

The following description is based on Khenpo Tsultrim Gyamtsos Progressive Stages of

Meditation on Emptiness which is largely based on Jamgon Kongtrul Lodro Thayes Encyclopedia
of Knowledge (Shes-bya-mdzod).

149!

fire is empty of fire, a dream tiger is empty of tiger. In other words although there was a
fire-like or tiger-like experience, there was no fire or tiger as such present. The question
remains, however, as to what that fire-like or tiger-like experience was. For those
Buddhists who accept only the view of self-emptiness, it was a stream of dependently
arising moments of consciousness and mental objects of consciousness empty of the fire
or tiger which we imagined was there.

b. Other-emptiness
Other-emptiness (or, more correctly, emptiness of other) means Emptiness as a
designation for ultimate reality which is explained in the Tathgatagarbha stras2 and
elsewhere as the vividness of non-dual awareness/experience complete with all the
Buddha qualities. The other it is empty of is both dependently arising and imaginary
phenomena. Dolpopa elaborates by explaining that dependently arising phenomena,
though appearing real to our deluded mind, in fact lack reality (i.e. they are self-empty).
Although the fire-like or tiger-like experience in the dream seems to arise from the
dependent arising of a moment of consciousness and its object, Madhyamaka reasoning
shows that nothing actually ever arises, so not only is the fire-like experience and the
tiger-like experience empty of a fire or a tiger (imaginary phenomena) but the experience
itself is empty of a consciousness and an object of consciousness (dependently arising
phenomena). So for those Buddhists who accept the other-emptiness view, the fire-like or
tiger-like experience is not really a stream of dependently arising events that can be
conceptually analysed. It is a direct expression of the Buddha Wisdom Mind (jna),
albeit distorted by ignorant dualistic tendencies into seeming to be a fire or tiger or into
seeming to be a dream of a fire or tiger experienced by a dreaming consciousness. This
does not mean that Buddhists who accept an other-emptiness view reject the selfemptiness of imaginary or dependently arising phenomena, nor does it mean that they do
not understand that even ultimate reality, insofar as it is a concept, is self-empty. Their
concern is, however, to stress that ultimate reality (Buddha Wisdom Mind), though empty
from the point of view of the mental processes that try to understand it conceptually, is
not empty in itself. In other words it is not illusory, false, impermanent, compounded,
dependently arising, conditioned and so on.

The two basic models of orthodox Buddhist thought

In this paper I am going to extend the use of the terms self-emptiness and other-emptiness
to distinguish two quite distinct models of the nature of mind, of man, of the spiritual
path and goal.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2

Examples of these stras are the Agulimlya-stra, Avatasaka (Buddhvatasaka), Gaganagajaparipcch, Gaavyha, Ghanavyha-stra, Tathgatagarbha-stra, Tathgatamahkarunirdea-stra, Mahbheriharakaparivarta-stra; rmladevsihanda-stra, Mahparinirvastra.

150!

Of the living traditions found in the West these days the Gelugpa (dGe-lugs-pa) and
Theravda schools typically adhere to what we will be calling the self-emptiness model,
while Nyingmapas (rNying-ma-pa; followers of the rDzogs-chen tradition) and Kagyupas
(bKa-brgyud-pa) typically adhere to what we will call the other-emptiness one. The
Hua-yen schools of Chinese Buddhism that follow the Avatasaka and some Chan
schools seem to follow an other-emptiness model.3
(Let me stress again that I am leaving aside the objections of those who are used to
hearing these terms used in Tibetan Buddhist polemics. I am not saying that Theravda
and Gelugpas etc. claim that they hold a self-emptiness view nor that all Kagyupas and
Nyingmapas claim that they hold an other-emptiness view. Rather I am saying that their
view conforms, roughly speaking, to the self-emptiness or other-emptiness model as
outlined in this paper.)
It is important that a broad distinction be made between these two apparently
equally ancient, widespread and orthodox models of Buddhist thought in order to clarify
the whole area of Buddhist thought in general. Only then can we begin to compare
different scriptural statements about Buddha-nature (tathgatagarbha), absolute/ultimate
reality (paramrthasatya), emptiness (nyat), nirva, dharmakya and so on. Without
knowing which basic model (self-emptiness or other-emptiness) the scripture or
statement in question has in mind, there is no hope of arriving at any precision in ones
analysis of its meaning.
This spills over into the area of translation. The translator of a Buddhist text, in
order to make a coherent translation, has to have some kind of overall model in mind.
One often detects that the translator only accepts one or other of these models as valid
and this prejudices his use of terms in translation. Hence, one comes across textual
statements originally applicable to an other-emptiness model translated in self-emptiness
terms or vice versa. More often than not, however, the translation is simply not clear at
all because the translator is unaware of the distinction and is experiencing a tension
between what he thinks the Buddhist model should be and what his integrity as a
translator tells him the text is actually saying.4
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
3

See G. Tuccis Minor Buddhist Texts, Part II, Rome, 1958, 6465 for how the Nyingma text bKatan-sde-lnga refer to Bodhidharma and other Chan masters and preserve some of the ideas of the
Chan schools.
Generally speaking Professor Tucci and Professor Ruegg follow the Gelugpa line in trying to
make out that the Dzogchen and the Jonangpa other-emptiness doctrines differ radically from those of
other Mahyana schools as if a gzhan-stong type of model did not exist in the Mahyna main line
tradition. In my unpublished doctoral thesis Tathgatagarbha Doctrine according to the Shentong
interpretation of the Ratnagotravibhaga I show that this is an unjustified and misleading assumption.
4
A good example of this is found in Mahmudr by T.T. Namgyal, trans. & an. by L.P. Lhalungpa,
Boulder, Shambhala, 1986. Lhalungpas introduction is based on the Gelugpa self-emptiness model
while the main text usually follows Dezhung Rimpoches explanations based on the other-emptiness
model.
In Kindness, Clarity and Insight, trans. & ed. by J. Hopkins, Ithaca, 1984, the Dalai Lama
(H.H. Tenzin Gyatso) defends Dzogchen from criticism by Gelugpa geshes by explaining it in
accordance with the self-emptiness model. This is not, however, because the Dalai Lama is ignorant of
the fact that Dzogchen actually conforms to the other-emptiness model.

151!

a. The Self-emptiness model5

The self-emptiness model has the following features:
1. The mind is a stream of moments of consciousness.
2. A moment of consciousness only exists in dependence on the momentary
existence of an object of consciousness.
3. Purification of consciousness means the gradual replacement of impure
moments by pure moments.
4. Nirva is the cessation of impure moments.
5. The Buddhas awareness is a stream of pure moments that have pure objects of
awareness.
6. A being is a stream of dependently arising events belonging to no lasting selfentity.
7. A being can become Buddha by seeing that everything is dependently arising and
cultivating dependently arising Buddha qualities to replace his present
dependently arising bad qualities.
This model is roughly speaking that of the various Abhidharmist schools,
Cittamtrins, Svtantrika Mdhyamikas and Gelugpa Prsagika-Mdhyamikas.

b. The other-emptiness model

The other-emptiness model has the following features:
1. Mind is essentially non-compounded.
2. Awareness is essentially non-dual. The belief in objects external to awareness is a
mistake.
3. Mind is essentially pure. Purification means the gradual emergence of this pure
mind from the mists of confusion.
4. Nirva is the emergence of this pure, non-compounded, non-dual
awareness/mind.
5. The Buddhas awareness is this pure, non-compounded, non-dual
awareness/mind.
6. A being experiences this pure, non-compounded, non-dual awareness/mind that is
his true nature, as impure and compounded. In other words his dualistic habits
distort his awareness into a dualistic consciousness of himself versus the world of
external objects.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
5

This analysis derives from material gathered for my doctoral thesis (see n. 3 above) which is due to
appear in a book published by SUNY.

152!

7. A being can become Buddha by abandoning his dualistic habits of mind and so
allowing the true nature of his being/experience to shine through complete with its
inseparable Buddha qualities such as love, compassion, wisdom, vision, power to
liberate others etc.
My own research over the last ten years allows me to associate confidently the latter
model with the Tathgatagarbha stras and, more tentatively, the early canonical
doctrines concerning the clear light nature of mind (prabhsvaracitta), nirva as
permanent, bliss, an entity, a place, Dharma as eternal and absolute reality, enlightenment
as the purifying of ordinary consciousness until the Dharmadhtu is reached and so on.

The self-emptiness and other-emptiness models in terms of Buddhist practice

a. Self-emptiness model
This is the exploratory part of this paper. I suggest that the self-emptiness model is
associated with the following features in terms of Buddhist practice:
1. Emphasis on pure versus impure conduct and the need to cultivate good qualities
and abandon the badhence the shunning of home life and a high regard for
monastic discipline (a nuns status being lower than that of a monk).
2. Emphasis on learning and minute analysis in the somewhat intellectual way
typical of Abhidharmists and Gelugpas.
3. Faith seen as a lesser faculty to reason.
4. A tendency to feel Buddhahood remote and no longer accessible in the current
Dark Age.

b. Other-emptiness model
I suggest that the other-emptiness model is found in association with the following
features:
1. Greater status is given to lay practitioners (men and women equally). Since mind
and man is essentially pure, the important thing is the abandoning of dualistic
concepts of pure and impure, good and bad etc. Although monastic discipline has
its advantages, for the true practitioner there is no essential reason to shun the
home life.
2. True knowledge arises naturally when the mind relaxes its dualistic tendencies.
Much learning is not essential to this process although it can be very helpful for
removing doubts and confusion.
3. Faith is a faculty of openness and clarity that allow the practitioner to trust the
arising of insight to be none other than the functioning of the Buddha Wisdom
Mind. This gives the practitioner the power and

153!

confidence to abandon intellectual doubt. On this level faith is superior to reason,

which functions only at the level of intellectual doubt.
4. Since the nature of all experience when free from the distortion of mental effort is
the Buddha Wisdom Mind, emphasis is put on feeling the presence of Buddha in
ones life, ones being and ones mind. In other words the Buddha becomes
mystically accessible through faith and devotion even in the current Dark Age.

Possible historical antecedent6

Even before the advent of the Mahyna one can see the differences between the
Hnayna schools of the Mahsghikas and the Sthaviras. When they split up in around
140 BC we find the following features of resemblance to the above model:

Sthaviras
i. Strong emphasis on the monastic life.
ii. Rigid adherence to the prtimoka (monastic rule).
iii. Accent on Arhat rather than Bodhisattva.
iv. Buddha as human figure, who disappeared at the attainment of nirva.
v. Faith played down.
vi. Analysis of experience into atoms called dharmas.
Mahsgikas
i. Closer relation with laymen.
ii. More flexible attitude to monastic rules.
iii. Accent on Bodhisattva rather than Arhat.
iv. Buddha as transcendental, but compassionate and always existing.
v. Faith given a bigger part to play.
vi. The nature of things is nyat.
Though sketchy, enough has been said here to suggest very obvious parallels with the
self-emptiness/other-emptiness analysis given above.

Attitudes to the spiritual path

This makes me want to explore further to what extent a persons initial conception of the
spiritual nature of man affects his whole outlook on life and whether this has any
practical implications in terms of his psychological attitude towards the spiritual life, his
lifestyle as a spiritual aspirant and the social organization of religious institutions.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
6

These points have been extracted from A. Bareaus Les Sectes Bouddhiques du Petit Vehicule,
Saigon, 1955.

154!

a. Self-emptiness model
The self-emptiness model, for example, takes the attitude that mans nature is flawed and
in order to attain the qualities of a Buddha, such as wisdom, compassion, the power to
liberate others and so on, he has to rid himself of bad and gradually acquire good
qualities through a long period of systematic training.
In terms of a persons psychological attitude towards the spiritual life, he will
naturally think of himself as starting off as an inferior being, who gradually over years of
patient training develops into something better. One would expect the lifestyle of the
serious practitioner to be visibly more disciplined than that of ordinary, untrained
individuals and for the ideal religious institution to be naturally hierarchical with the most
disciplined and most learned of the long serving practitioners to be at the top.

b. Other-emptiness model
The other-emptiness model takes the attitude that mans apparent flaws are incidental to
his true nature which is the Buddha Wisdom Mind, a transcendant reality continuous with
the nature of the Buddha.
The psychological attitude of a person with such a view would be to look for
signs of that reality within his own experience. Such signs would be wisdom and
compassion manifesting in his experience (albeit in a confused and limited way because
of his obscurations) and the flash of direct experience of the Buddha Wisdom Mind
transmitted through the teachers oral instruction. In a sense he will never become better
and efforts at self-improvement are irrelevant or even harmful. In fact, it is only through
abandoning of rigid concepts such as good and bad, right and wrong, better or worse and
so on that the mind can relax and finally allow the splendour of its true nature to shine
forth. Therefore, a serious practitioner might follow any number of lifestyles and may not
be visibly more disciplined than anyone else. There is no inherent reason why a beginner
should feel either inferior or superior to other practitioners and authority might as easily
be found outside as within the spiritual hierarchical structure of religious institutions.

Tathgatagarba doctrine
I mentioned above that in my view the other-emptiness model fits exactly that of the
Tathgatagarbha stras. The difficulty in saying so categorically is that some influential
Tibetan Buddhist traditions interpret it according to the self-emptiness model.
Buddha-nature is one of the synonyms given for tathgatagarbha which means
literally the womb or the embryo of the Tathgata (an epithet of the Buddha). It is a
Sanskrit compound that can and has been interpreted in a number of ways. It could mean
the embryo/womb that is the Tathgata or the Tathgatas

155!

embryo/womb. In other words it could refer to a nature within us that is continuous with
the nature of the Buddha in its totality, or it could refer to a certain aspect of our nature
that could develop into a Buddha given the right circumstances. There is an immediate
and significant difference here. In the first case we are talking of the Buddha-nature as
some kind of absolutethe nature of reality that is dynamic and alive in the world and of
which we partake by nature (an other-emptiness view). In the second case we are talking
of the Buddha-nature as the empty nature of the dependently arising mind-stream (an
other-emptiness view).
In the former case Buddha-nature is Buddha in the sense that it is the nondependently arising Buddha Wisdom Mind. Attaining Buddhahood is simply to realize
what has always been the case. In the latter case attaining Buddhahood is the arising of
something new which then continues to act in and influence the world. In the first case
we are already Buddha but do not realize it. Our faults and imperfections are merely a
mysterious cover that only appears as real. In the latter case our faults and imperfections
are a measure of the absence, as yet, of the Buddha qualities in us and the event of
enlightenment is the end of a long process of development.
Thus in the other-emptiness model the tathgatagarbha is literally Buddha
complete with all the Buddha qualities present in all beings, but this makes no sense in
terms of the self-emptiness model. This means that although the Tathgatagarbha stras
state quite clearly that the Buddha complete with all the Buddha qualities is present in all
beings, those commentators who do not accept other-emptiness as a valid Buddhist
doctrine have to give some explanation for the occurrence of such statements. One
solution is to heavily interpret the texts, giving each line that seems to imply an otheremptiness view a far from obvious meaning conforming to the self-emptiness model. The
other solution is to take the texts at face value and say that it is presenting an inferior
view for those who cannot understand the highest view, which, of course, conforms to the
self-emptiness model. Either way the Tathgatagarbha stras pose tremendous problems
of interpretation.
Professor Ruegg has often drawn our attention to the kinds of solution such
commentators adopt. Although they may have different theories, in general they all adopt
the line that certain statements are not to be taken literally. The Buddha must either have
given such untrue teachings in order to win over those who could not bear to hear the
truth or they are the truth albeit expressed indirectly in order to appeal to those who could
not bear to hear it put any other way. This latter approach becomes quite convoluted.

The RatnagotravibhgaA Commentary on the Tathgatagarbha stras

The Ratnagotravibhga-Mahyanottaratantra-stra and its Commentary were written
in India perhaps in about the 3rd or 4th century AD. Both comment on

156!

the tathgatagarbha doctrine of such stras as the rmladev-stra, and the

Tathgatagarbha-stra. According to Wayman in his introduction to the rmladevstra the Tathgatagarbha stras were written for a largely lay Buddhist community
with powerful women patrons. The queen expounds a doctrine on the nature of
tathgatagarbha, emptiness and nirva which sounds very like our other-emptiness
model above and she is praised by the Buddha who announces that the rvaka and
Solitary Buddha Arhats do not understand the Dharma as deeply as the queen does. Only
the high level Bodhisattvas share her understanding. So we see here the associated social
features of an other-emptiness doctrinea queen given spiritual status above those
following the strict renunciate discipline of the Arhat disciples. The psychological effect
that this would have on lay men and women practitioners is obvious.

Ratnagotravibhga on psychological effects of the tathgatagarbha doctrine

The Ratnagotravibhga (1.156166) is quite explicit about the psychological effect
tathgatagarbha doctrine is supposed to have. These verses explain why it is necessary to
teach this doctrine of tathgatagarbha when, as the Ratnagotravibhga Commentary
points out, it is quite clear that not even Bodhisattvas at the highest level realize it fully
let alone rvakas and Solitary Buddhas. One might wonder what hope there is for
ordinary beings to understand it at all.
loka 1.156 raises the question by asking why, after having been taught in various
places in the scriptures that all knowable things (shes bya) are empty like clouds, dreams
and magical illusions, does the Buddha teach here that all beings have the buddhagarbha.
Ratnagotravibhga 1.157 explains that it is taught because it serves to overcome five
faults which develop if beings do not realize they have the Buddha-nature. The five faults
are:
Faintheartedness: Some beings may feel so inadequate that they would not presume to
aspire to supreme Buddhahood. Thus, they would not take the Bodhisattva vow to bring
all beings to Enlightenment, which is essential for developing bodhicitta. Knowing that
they and all beings have Buddha-nature gives them confidence and encourages them to
try to attain Buddhahood.
Looking down on others: Some beings may suffer the opposite fault. They bravely and
confidently take the Bodhisattva vow and then feel that since they now belong to the
Buddha lineage, they are naturally superior to those who do not. Knowing that all beings
naturally belong to the Buddha family undermines this kind of arrogance.
Attachment to what is not real: Those who do not know that the true nature of beings is
tathgatagarbha take their faults and imperfections to be their true nature. In this way
they fall into the position of setting up something that is unreal as real.
Denial of what is real: By being attached to this unreality, they overlook completely the
reality that beings have the Buddha Knowledge (jna). They deny

157!

that beings have the perfect and unspoiled qualities of the Buddha and thus fall into the
position of denying reality to what is real.
Not loving others as oneself: If one realized that the nature of all beings was
tathgatagarbha, one would love all beings including oneself equally. Equalness is an
important feature of both the Buddhas wisdom and his compassion. If one does not
know that all beings have tathgatagarbha, one automatically sees self and others as
unequal and separate. When one understands the doctrine of tathgatagarbha, one knows
that self and others partake of the one undifferentiable reality, and automatically love and
compassion for others arise.
Thus, by being taught about tathgatagarbha, one will have five remedies,
instead of the five faults. The five remedies are confidence, respect for others, praj (in
the sense of not taking the unreal to be real) and jna (in the sense of knowing the real
to be real), and Great Compassion. With these five virtues one is able quickly to attain
Buddhahood. Clearly the Ratnagotravibhga expects the effect of its doctrine to be a
psychological boost for those of low spiritual status and a social levelling in terms of the
hierarchy of religious practitioners. It would be too complicated to do justice to a selfemptiness type of explanation of the above verses. Suffice it to say, it is possible to
explain them in a way which fits a self-emptiness model although to my mind such an
explanation is clearly forced.

The situation in Tibetan Buddhism

We have seen how the features of an other-emptiness model are found in the
Mahsghika lay communities of 140 BC and how the tathgatagarbha doctrine of the
third and fourth centuries AD again had the features of the other-emptiness model and is
associated with lay practitioners. Looking to the present day, the features associated with
both a self-emptiness and an other-emptiness model are seen in Tibetan society.
There are the monasteries and retreat centres where strict discipline is observed
often throughout a persons life. There are colleges with strict regimes of study and
debate. There are examination systems, granting of degrees and so on. It is all very
ordered and fits very well with the self-emptiness model of the nature of the spiritual
path.
Side by side with this one finds lay practitioners and even women taking thrones
in the assembly above those of senior monks. Great siddhas with little learning live
outside monastic institutions altogether. Furthermore, there are the enlightened mad
yogins who break all social conventions and yet still command respect. All this fits the
other-emptiness model.
However, generally speaking, traditional Tibetan society gives higher status to the
celibate monk than to the lay practitioner. It likes to see monastic discipline strictly
adhered to and holds learning in high esteem. Furthermore, although it is

158!

part of their tradition, most Tibetans seem quite suspicious of outrageous behaviour on
the part of even well-recognized siddhas.
One would expect, therefore the self-emptiness model to find more favour in
traditional Tibetan society. The self-emptiness model conforms to a common sense view
of the world and to the way ordinary people think. Anyone can see that we are not yet
fully Buddha with all the Buddha qualities. If we were there would be no meaning to the
spiritual path and goalno need for religion and religious institutions. Traditional
Tibetan society wants religion, religious institutions and a spiritual hierarchy of learned
and older monks to act as the authority in religious matters. Religious institutions of this
kind give the ordinary man a sense of confidence that his spiritual welfare is being safeguarded as he pursues his lay concerns. It seems obvious that most Tibetans are very
content with that structure and are quite happy to postpone their own Buddhahood to a
future life. For most, it is sufficient for the present to make offerings and to store up
spiritual merit. Whether one is at the top or the bottom of the hierarchy it is congenial and
easy to manage such a system in social terms. The religious take care of religious
concerns and do not compete in the worldly sphere. Being celibate they do not have
children to compete for the family inheritance.
The other-emptiness model challenges common sense. Obviously there must be
some sense in which we are not already Buddha since if we were there would, indeed, be
no use for a spiritual path and goal. Taking it as given therefore that there is some
obvious difference between ourselves and Buddha, the advocates of other-emptiness say
that the difference is incidental and merely apparent. There is, in other words, no real or
essential difference between beings and Buddhas (in terms of having the countless
Buddha qualities).
In social terms this has the disconcerting effect of putting everyone on an equal
footing with equal responsibility in terms of the religious life. The true practitioners are
hard to recognize and may play any role in society. Such a view is a threat to the
established hierarchy of religious institutions. It is a threat to the layman who wants an
easy way of shifting the weight of spiritual responsibility on to someone else. It is a threat
to the holders of wealth and power because they can no longer buy the goodwill of the
spiritual community by supporting easily-controlled monasteries. The bizarre and
unconventional behaviour of mad yogis is a threat to the social order.
Not surprisingly, therefore, do we find the ruling powers of Tibet from the earliest
times advocating the gradual, disciplined, monkish path associated with a self-emptiness
model rather than the anarchical, sudden path associated with the other-emptiness view.
Tibetan official history attacks the Chinese master Hwa-shang Mahyana who seems to
have advocated a sudden path to Enlightenment based on simply relaxing all conceptual
thought. From the accounts in early

159!

Nyingmapa texts it seems that this view was a version of the other-emptiness model
based on the Tathgatagarbha stras and akin to the Nyingmapa rDzogs-chen view.7 The
Nyingmapas themselves have never aspired to political prominence in Tibet and have
always favoured communities of lay practitioners over large monastic institutions. The
Sakyapas held political power for a long time and are famous for their constant attack on
other-emptiness type doctrines. For the last few centuries the Gelugpa school has been
the main wielder of political power and they too attack the other-emptiness doctrine
especially that of the Jonangpa school, which they claim to be a non-Buddhist view that
will lead one to hell. The Kagyupa school has from time to time achieved political
prominencesome of its main lineage holders have been great advocates of the otheremptiness view while others have attacked it. It would be interesting to see to what extent
religious view and political aspiration coincided in the history of the Kagyupas.
Generally speaking therefore it seems that the social and political forces within
Tibetan society work against an other-emptiness view. Tibetans love to read and hear
about the outrageous and unconventional behaviour of the yogis of the past, but seldom
appreciate them in the present. It was noticeable, for example, that although lip-service
was paid to Trungpa Rimpoches status and behaviour as a yogi by the Tibetan
community in India, few actually approved of him or continued to follow him once he
adopted the outrageous behaviour so admired in the great Tibetan folk heroes of the past
such as Drugpa Kunleg,8
One could argue in its favour that such an attitude protects Tibetan society from
charlatans posing as mad yogis who might otherwise undermine social values, while the
oral and literary tradition extolling the virtues of the outrageous yogi protects it from too
deeply entrenched prudery.
I shall conclude on a personal note. Having spent ten years of my life conforming
to traditional Tibetan values as a Kagyupa nun, I can vouch for the fact that it instils in
one a sense of personal purity and aloofness from the common herd. There is a strong
sense of being on the lower rungs of a great spiritual ladder and a tendency to depreciate
ones own experience in the wish to conform to the gradualist party line. Yet what I find
most curious is that the underlying view on which my practice was based was otheremptiness. Nothing was stressed more than faith and the key role of the guru, his oral
instruction and direct mind-to-mind transmission. There was a sense of irony and play
going on between the two models of the spiritual pathoutwardly one adhered to the
self-emptiness model and all its associated features and yet all the time there was a

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
7

See G. Tucci Minor Buddhist Texts, Part II, 103. According to Chinese sources published by P.
Demiville the winner of the debate between the Indian Kamalala and the Chinese Hwa-shang was
Hwa shang.
8
K. Dowman, tr., The Divine Madman, London, Rider, 1980.

160!

sense that this was not actually the point. It was almost as if there were a smoke screen of
respectability behind which the real work of the yogi continued more or less in secret.

161!

Observations on the Tibetan Phur-ba and the Indian Kla

R. Mayer*

The Tibetan ritual implement the phur-ba, called kla in Sanskrit, has attracted the
interest of Western scholars for many years. R.A. Stein, J. Huntington, T. Marcotty, G.
Meredith, and S. Hummell have all contributed studies ranging from entire books to
learned articles. A consensus seems to have been reached by most of these scholars, that
although the kla in some form or other was certainly known and used in India, the
characteristic form of it that is so widespread in Tibet is not of Indian, but of
autochtonous Tibetan provenanceor, in the opinion of Hummell, of Mesopotamian
origin. Thus Huntington writes no phur-ba like instrument has been demonstrated to be
of Indian origin up to the present day.1 He loosely suggests that the phur-ba was
incorporated from Bon into Buddhism.2 Similarly, R.A. Stein was for many years very
interested in the origins of the phur-ba, but he also writes that although the Tibetan
lamas religious and philosophical explanations of their phur-ba are in Indian Buddhist
idiom, the actual form and shape of the phur-ba itself, as known in Tibet, seems to be
purely Tibetan. Stein says that no literary or archeological evidence for it has ever been
found in India; and although he concludes at the end of his research that there is no doubt
that some kind of kla was known in India, he feels he can not establish that the Indians
ever knew it in the form used in Tibet.3
The implications are that it was borrowed from Bon and the allegedly translated
from Sanskrit rNying-ma-pa texts describing it are apocryphal.4 Keith Dowman follows
these two scholars and says the phur-ba is one of several very important ritual
implements which the rNying-ma-pa yogis borrowed from the Bon.5 While conceding
that the kla was known in India in some form, he says that there is no

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
*

An incomplete and primitive presentation of this material (including some errors) is due to appear in
the next issue of the Tibet Journal under the title Tibetan Phur-bas and Indian Klas. In the present
contribution the material has been reworked and expanded.
1
J. Huntington, The Phur-baTibetan Ritual Daggers, Ascona, 1975, vii.
2
ibid., vii.
3
R.A. Stein, Annuaire du Collge de France, 197172, 499.
4
R.A. Stein, Annuaire du Collge de France, 197778, 648, 654.
5
K. Dowman, Skydancer, London, 1984, 302.

!
!

163!

representation of the kla as the Tibetans know it in Indian bronze or stone art or in the
literature of the siddhas. Therefore Dowman stresses its Tibetan or Himalayan
shamanistic uses as being the more significant.
The purpose of this paper is to show that this consensus is quite possibly mistaken
and that the question remains open.6 Not only was the kla known and used in India in
some form or another, but that characteristic form that we now call the Tibetan-style
phur-ba might also be of surprisingly orthodox Indian provenance, and this in turn
implies that the allegedly apocryphal texts describing it might also in fact be genuine
translations from Sanskrit. This is not necessarily to deny that the prominence the phurba received in Tibet might have been partly Bon inspired, since it seems ever more likely
that the Bon themselves were at least in major part a religion of Indic origin, perhaps the
result of a very early stratum of Buddhist [or even aiva?] proselytizing in west Tibet.7
Thus they might well have had their own Indian-derived phur-ba tradition in Tibet before
the Buddhists. Nevertheless, Padmasambhavas well attested association with the kla
rites provides a more probable explanation for its popularity amongst his Tibetan
devotees.8
It appears that the Tibetan phur-ba, or kla in Sanskrit, is only the most recent
manifestation of a long and varied line of Indian ritual items that have been generated
over the last three millennia or so, based on a few particular Vedic or Epic concepts. Two
of these notions are especially important with regards to the phur-ba or kla, one a myth
and the other a ritual implement. The age of the myth is controversial; was it Vedic or
Epic? For the purposes of this paper, it makes no difference.9

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
6

The generosity of John Irwin greatly facilitated the production of this paper. Mr Irwin has, over the
years, accumulated a large number of references to the indrakla, which he unstintingly shared with
me, thus saving me many hours of laborious searching in libraries. Thanks are also due to Richard
Gombrich, who made me aware of Lily de Silvas article, and to Charles Malamoud, who gave me
additional useful references. Philip Denwood has been immensely helpful in referring me to the
various ilpastras, as have Kevin Latham and above all Luke Lau and Stephen Hodge in reading
Chinese texts for me. Michael Aris devoted some hours to correcting the worst enormities of my style
and presentation, for which I shall remain very grateful for years to come.
7
D.L. Snellgrove, Indo-Tibetan Buddhism, London, 1987, 390391.
8
See the various hagiographies and the Tun Huang text Pelliot 44; also R.A. Stein, A propos des
documents anciens relatifs au Phurbu, Csoma de Koros Symposium, Budapest, 1978 and Bischoff
and Hartmann, Padmasambhavas Invention of the Phur-ba, Etudes tibtaines ddi Marcelle
Lalou, Paris, 1971; also the sBa-bzhed, in S.G. Karmay, The Great Perfection, Leiden, 1988, 6.
9
The controversy hinges on whether the Vedic adjective vtrahan, = victorious, and cognate with
the Avestan verethragan, = victorious, was falsely etymologized by brahmans to render smiting
Vtra, Vtra thus coming to appear as a proper name and the serpent Vtra being invented to make
sense of the new etymology. Thus the whole myth of Indra slaying Vtra would have arisen out of a
false etymology. But in the much much later period of historical concern to this paper, the myth
certainly did already exist, so the controversy as to its precise origin is irrelevant. See M. Boyce, A
History of Zoroastrianism, Leiden, 1975, 64.

!
!

164!

The myth introduces the crucial notion of the kla as the cosmic mountain, Meru.
It is an account of creation in which Indra slays the serpent Vtra, thus allowing the world
to come into existence. F.B.J. Kuiper has (controversially) called this cosmogonic myth
the basic concept of Vedic religion.10 He feels it is basic because Indra is arguably the
most important of Vedic gods, and, as Kuiper writes:
Indras mythical role remains limited to this single exploit. Again and again the
poets say that he slew the dragon, extended the earth, and lifted up the sky, but
that is about all they can tell us about him.11
Kuiper adds:
This myth owed its fundamental importance to the fact that every decisive
moment in life was considered a repetition of the primeval process. Therefore the
myth was not merely a tale of things that had happened long ago, nor was it a
rational explanation of how this world had become what it is now. The origin
myth constituted the sacred prototype of how, in an endlessly repeated process,
life and this world renewed themselves again and again.
According to Kuiper, Vedic ritual made much of this myth.
The story itself varies slightly in different versions. The basic outline, in Kuipers
words, is as follows:
In the beginning there was only water, but these primeval waters bore in
themselves the germ of life. From the bottom a small clod of earth rose to the
surface where it floated about. The clod spread on the surface and became a
mountain, the beginning of the earth, but it continued to float (unanchored) on the
waters In this first stage, the world was still an undifferentiated unity, none
of the contrasts that constitute our phenomenal world yet existed. There was no
heaven or earth, no day or night, no light or darkness, no male or female
This state of undifferentiated unity came to a sudden end with the birth of Indra
from out of nowhere, who then performed his great demiurgic acts. Firstly, Indra found
that the primeval mountain was still floating about unsecured on the waters, but to allow
creation to flourish, Indra had to fix it firmly to the bottom so it could no longer move.
Secondly, it also had to be opened up or penetrated, to allow the locked-in creation to
pour forth from out of it. But there was a strong force of resistance to both of these acts
personified as the serpent Vtra, whose name literally means obstruction or resistance.
Indra slew Vtra after a struggle, and

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
10
11

!
!

F.B.J. Kuiper, JHR, 1975, 107120.

ibid., 110.2
165!

he managed to burst open the mountain, from which all creation then poured forth. Now
the mountain spread on all sides until it became the whole expanse of the world. Indra
then performed another very important cosmogonic act, by becoming for a while the
pillar that props up the sky, which until then had been lying flat on the face of the earth.
The relevance of this myth to the kla is revealed by Kuiper in a 1970 article.
Kuiper writes:
Indra made (the mountain) firmly rooted in the bottom of the waters. Since this
mountain was the cosmic centre, the central point of the earth, the whole earth
thereby became firm and steady. Thus the cosmic mountain not only was the
origin of the earth, but also came to function as the peg which secured the earth a
firm support. This idea still survives in the later literature, where Mt Mandara (=
the unmoving) as the cosmic pivot is called Indras Peg (Indrakla), and the
concept of a mountain functioning as a peg is expressed by the term kldri.12
This myth, or something very like it, seems to be closely connected to the advent
of the kla as a popular symbol or motif in Indian religious thinking, and its various
themes are repeated endlessly over the centuries in a wide variety of literary and ritual
references, especially, as Kuiper points out, in the later literature.
Charles Malamoud says that Indrakla features as the name of a mythical
mountain in both the epics and the Puras.13 Other authors, such as Srensen14 and Lily
de Silva,15 also mention several epic citations of Indrakla as a mythical or cosmic
mountain.16
The modern Tibetan phur-ba, of course, is also invariably and ubiquitously
associated with the Mt Meru or Mandara, both in liturgy and also sometimes in
iconography. For example, even in a modern liturgy by bDud-joms Rin-po-che,
composed up to two or three thousand years after the formulation of these ancient myths,
the deity Vajrakla still rolls the Mt Meru kla in his two central hands;17 just as several
modern Tibetan prints of this deity in my possession still depict

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
12

F.B.J. Kuiper, JHR, 1970, 110.

Personal communication, April 1989.
14
S. Srensen, An Index to Names in Mahbhrata, 1904, 341.
15
L. de Silva, The Symbolism of the Indrakla in the Parittamaapa, SSAC, 1978, 241.
16
Mahbhrata, 1.3.3; 1.18.112; 2.10.413; 3.37.1497; 3.38.30; 3.39.1562. See also iva Pura, vol.
3, p.1233. (atarudrasahit, ch. 37, v. 651.)
17
bDud-joms Rin-po-che, n.p., n.d., blockprint, 3 fols. dPal rDo-rje Phur-bui rgyun gyi rnal-byor
snying-por dril-ba, fol. 1. See also R.A. Stein, Annuaire du Collge de France, 197677, 610.
13

!
!

166!

the top of the kla as Mt Meru, instead of in its usual three-headed form indicating the
deity. The rolling of the kla is clearly identified in Tibetan literature with the churning
of the ocean with Mt Mandara, for Meru and Mandara are frequently or even usually
conflated in Buddhist texts.
In the myth, the original Indrakla mountain emerged from out of the waters,
where it acted as a peg. In the same way, the Tibetan klas three blades below are clearly
designed as a peg. Mt Meru, writes Lily de Silva, is said in Buddhist cosmology to be
submerged in the ocean to a depth of 84,000 yojanas and to rise above sea level to an
exactly equal height.18 Thus Tibetan klas are customarily made with their three-bladed
lower pegging part, the equivalent of the submarine part of Mt Meru, exactly half their
length. They then mark the sea level, at their mid-point, by the ornamentation of the top
of the blade with a makaras head, from the mouth of which ngas pour down the length
of the blade. These constitute standard and ubiquitous symbols of the ocean in Indian
art.19 Above the makara head, i.e. above sea level, rises the other half of the kla, the
eight-faceted shaft, which is equivalent to the different higher worlds and heavens of
Meru. It is only in this upper part, says the Phur-ba bcu-gnyis, a Vajrakla root tantra, that
the enlightened deities dwell; while the lower three-bladed part is only to subjugate.20
This is entirely appropriate, since the Indian gods live on the upper realms of Meru, not
under the waters, where the ngas dwell.
Indian temples, palaces, maalas and buildings in general were often
conceptualized as a universe in microcosm. Little wonder then that in their construction,
the first stage had to be the ritual stabilization of the building site with the astrological
locating of the serpent below, and the pegging down of its head to stop it moving, in an
apparently microcosmic reenactment of the macrocosmic mythology discussed above.
This rite of ngabandha, involving the use of a kla both as a gnomon and a peg, is
ubiquitous in the Indian ilpastras.21 R.A. Steins attribution of it to an indigenous and
traditional Tibetan nameless religion is quite improbable.22 In the case of buildings, the
ngabandha was frequently done with a kla made of khadira wood. But the sites for
whole cities also had to be stabilized in the same way, and this seems to have sometimes
been done with iron, as we hear in an old song about Delhi:
All above a polished shaft,
All a piercing spike below.
Where they marked the ngas head,

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
18

Lily de Silva, op. cit. 243.

J.C. Irwin, Aokan Pillars: A Reassessment of the Evidence, BM, 1976, 737 ff.
20
Phur-ba bcu-gnyis, ch. 10, p. 106. (This modern edition uses arabic numerals rather than the
traditional folios.)
21
See, for example, ilpapraka, 1, vv. 5560.
22
R.A. Stein, Tibetan Civilization, Stanford, 1972, 203.
19

!
!

167!

Deep the point was driven down

Soon a castle clothed with might
Round the iron pillar clomb,
Soon a city etc.
Unfortunately, when this particular great iron peg was moved by a malicious
enemy, the subterranean serpent stirred and shook the earth with a disastrous
earthquake.23 Similar catastrophes are predicted if and when the kla buried by
Padmasambhava under bSam-yas in Tibet is moved.
Iron and khadira24 are precisely the materials that are widely used by Tibetans in
phur-ba manufacture to this day, and are recommended in a large number of Buddhist
texts such as the Phur-ba bcu-gnyis25 and the Mahvairocana-stra commentaries.26 A
similar but slightly different and less violent probing of the lto-phye or underground
serpents armpit, rather than head, but still requiring astrologers etc.,27 and still based on
Indian tantric literature, is described by Tibetan authors, for example Klong-chen-pa in
his Phyogs-bcu-mun-sel28 and mKhas-grub-rje in his rGyud-sde-rnam-gzhag.29 Another
simpler Tibetan equivalent, but still with Indian, is the rite of sa bzung-ba, or holding
the earth. To take an example of this from modern rNying-ma-pa ritual, the sGrub-khog
of the bDud-joms bla-ma thugs-sgrub says that a kla must be inserted into the ground at
the centre of the site where a maala is to be created, thus

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
23

A.K. Coomaraswamy, Symbolism of the Dome, IHQ, 1938, 18ff.; M. Eliade, Cosmos and
History, the Myth of the Eternal Return, New York, 1959, 19.
24
This hard wood, often identified with the thorny acacia catechu, is constantly associated in
Atharvanic and thus also later literature as a material specially useful for all kinds of abhicra and
protection. These usages apparently derived from the nirukta or etymology, khad = devour. (See M. &
J. Stutley, Ancient Indian Magic and Folklore, London, 1980, 117.) Although some Japanese scholars
have been tempted to associate the thorny acacia as a prototype kla, the textual ubiquity of the
khadira makes such an identification unreliable.
25
Phur-ba bcu-gnyis, ch. 10, 106.
26
N. Iyanaga, Rcits de la soumission de Mahevara par Trailokyavijaya, in Tantric and Taoist
Studies in Honour of R.A. Stein, vol. 3, MCB, Brussels, 1985, 686.
27
Stephen Hodge has pointed out to us how, in some Chinese Buddhist texts, the procedures of
establishing the site with klas can become quite complex, including such refinements as all the
nakatras delineated by 28 klas around the central gnomonic kla. R.A. Stein frequently mentions the
Kriysagraha and the Vidyottamatantra as especially detailed in these architectural uses of the kla
and also its use in other kriys such as controlling the ngas to make rain; Annuaire du Collge de
France, 1977, 1978, passim. The rites of smabandha and ngabandha are complex enough to warrant
an entire independent study. As far as I know, no such study has yet been made.
28
Gyurme Dorje, The Guhyagarbha-tantra and its Commentary, unpublished PhD thesis, SOAS,
London, 1987, 1366 ff.
29
F. Lessing & A. Wayman, Introduction to the Buddhist Tantric Systems, Delhi, 1983, 280.

!
!

168!

penetrating the earth and bringing the whole phenomenal world under control, before
the construction of the sacred maala can begin.30
But if a building was to represent accurately a microcosm of the universe, it had
also to demonstrate a reflection of the vertical, Mt Meru-like axial structure of the
universe envisaged in traditional cosmology. According to Ananda Coomaraswamy, in
many Hindu buildings, this would be represented by the empty vertical space within the
central dome or chamber which had the khadira wood serpent-piercing kla below, and a
finial called a ypa or stpikla directly above, at the highest point of the building. These
two represented the top and bottom points of the central axis;31 and thus while the
insertion of the kla for the ngabandha rites ritually marked the very beginning of the
building process, the rites for the insertion of the stpikla ritually celebrated its final
completion, a kind of consecration.
The classic Indian architectural writings or ilpastras, such as Mnasra etc.,
give detailed instructions on the shape and ritual importance of the stpikla that went
above. As to the shape, Mnasra says:
the length (i.e. body) of the kla is stated to be triangular, the base square, the
middle part octagonal and the top circular.
The width of the kla at the top should be one agula, and it tapers gradually from
base to top.32
This shape is very reminiscent in conception to the modern Tibetan kla, which
also has a triangular length, i.e. blade, a square base, an octagonal central shaft, and a
round top, often formed as the threefold deities head that is usually its distinctively
Buddhist emblem. Mnasra adds that the stpikla should be made of khadira wood,
iron or copper, again the materials favoured in the modern Tibetan phur-ba.33
Mnasra devotes no less than seventy-eight verses to the ceremonial erection of
the stpikla. This is a very major ritual occasion, requiring the presence of Brahmans and
a highly complex ritual which, exactly as in many allegedly apocryphal Tibetan rNyingma-pa rites, includes the extensive worship of the kla itself as the supreme deity,
ornamenting it by attaching cloths and leaves, honouring it with many precious offerings
and circumambulations, the performance of a homa, and so on.34 The celebration of this
ritual is said to be

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
30

Page 6, line 6. See C. Cantwell, An Ethnographic Account of the Religious Practice in a Tibetan
Buddhist Refugee Monastery in Northern India, unpublished PhD thesis, University of Kent,
Canterbury, 1989, ch. 5, 1.2.
31
A.K. Coomaraswamy, op. cit., 1819.
32
Mnasra, viii, vv. 1479; see P.K. Acharya, Architecture of Mnasra, Oxford, 1933, 205 ff.
33
Phur-ba bcu-gnyis, ch. 10, 106.
34
Mnasra, viii, vv. 340418. See P.K. Acharya, op. cit., 217 ff.

!
!

169!

absolutely indispensable to the success of the building project and the wellbeing of all
those involved in it and of the whole community. The stpikla, after all, represents the
pinnacle of the cosmic axis or Mt Meru, the palace of Indra or the highest gods, rising up
above from the serpent- piercing kla down below. Extremely similar passages are found
in other ilpastras, for example in the ilpapraka, where, significantly, the word
ypa is substituted for kla; and in Mayamata where the finial is called a sthpikkla.35
In the Ajitgama and Rauravgama, the words stpikla or stpidaa are favoured.
Mnasra and other ilpastras mention that the middle part of the stpikla
should be octagonal. This feature is a universal requirement in all standard Tibetan phurbas,36 and is very widespread in the extensive Pli sources on the indrakla. However, it
does not seem to derive from the myth of Indra slaying Vtra and allowing the world to
come into being. Instead, it is apparently derived from the other Vedic notion that is so
central to the form of the modern Tibetan phur-ba, namely, the important ritual
implement called a ypa, which was a sacrificial post to which animals were tied in
Vedic sacrifices. As Lily de Silva has attempted to show, at the latest by the time of the
Pli texts, the ypa and the indrakla were apparently conflated.37 It is certainly not
entirely clear to me as yet exactly how and when this process began, whether it was
already so in Vedic times or not. It is not impossible that they were identified with one
another from the start. Jan Gonda feels that on the basis of his readings of the
Atharvaveda and its commentarial stra the Kauika, he can see a clear functional
identification of the ypa as the axis mundi, and that it was considered essentially
identical with the fulcrum or pillar of the universe. Hence its conceptualization would
have been very close to that of the indrakla in many ways.38 Likewise, he writes, the
ypa is considered a thunderbolt (vajra) standing erect as a weapon against the enemy.
The erection of the ypa destroys evil and the powers of darkness.39 This too is one of
the functions of Indras feats with the kla. In another context, Gonda writes the ypa
sustains the components of the universe; all existence has entered it; being the frame of
creation, it enters the thousandfold aspect and components of the universe.40 Noting that
the ypa was frequently associated with Viu in Vedic texts, he adds:
It seems to be in perfect harmony with the character of the god Viu that the
ypa should belong to him. Traversing the parts of the universe and linking these,
and especially the sun and the earth, forming the

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
35

See ilpapraka, 2, vv. 68692; and Mayamata, vol. 2, ch. 18.

Phur-ba bcu-gnyis, ch. 10, 106.
37
de Silva, op. cit., 244246.
38
J. Gonda, The Savayajas, Amsterdam, 1965, 230.
39
ibid., 147.
40
J.Gonda, Aspects of Early Viuism, Delhi, 1969, 812.
36

!
!

170!

mystic centre of the cosmos, being the path which leads to the upper regions, the
sacrificial stake and other objects in connection with it belong to the god who
pervades the universe, is concerned with the transmission of light to the earth,
etc.41
Margeret Stutley describes the ypa as an intermediary between the divine world
and earthly life and points out the consecrated ypa is included among the pr deities,
i.e., the deified objects used at the sacrifice. She adds:
sometimes the ypa consisted of three stakes, either bound together or tied at one
end to form a tripod the ypa is usually octagonal and likened to an eight-sided
vajra, and hence was believed to protect the sacrificer against his enemies from
all sides if the sacrificer desires to cause the death of an enemy he should set up
a ypa on a base shaped like a grave, one end sloping to the south the various
portions of the ypa are symbolically differentiated; the base belongs to the pits,
above that, as far as the girdle, to men; the girdle itself to plants; above it to the
Vivadevas; the top to Indra; the rest to the Sdhyas.42
Nevertheless, despite such obvious similarities between the kla and the ypa,
there is no explicit lexical identification of the two in Vedic literature as far as I know.
But their formal assimilation is eventually, according to de Silva, made clear in a number
of Sinhalese and Pli sources, where the merged ypa-kla is in some forms described as
a Brahmanical object and in others as a Buddhist object. Most other modern authors, for
example J. Irwin and J. Miller, tend to identify the ypa with Mt Mandara or the
Indrakla.43
Although an entire hymn is devoted to it in the gveda, where it is addressed as a
deity,44 it is from the Yajurveda and the Brhmaas that we learn most about the ypa. In
these texts, we read again and again that the ypa must always be eight-faceted in its
shaft, because the gyatr has eight syllables.45 We also learn that at navel height above
the ground, the ypa must always be girdled with a three-fold rope girdle wound from
kua grass, to which the sacrificial victim is to be tied.46 Both these items, the eightfaceted shaft and the knotted rope, are standard features of the Tibetan phur-ba to this
day, as they were standard features of the ypa or indrakla in ancient Sinhalese stpa
architecture, as we shall see. The knot and eight-faceted shaft have of course had to be
totally reinterpreted in Buddhism,

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
41

ibid. 83.
M. & J. Stutley, A Dictionary of Hinduism, 1977, 351.
43
Most of Irwins writings present this theme. See also J. Miller, The Myth of the Churning of the
Ocean of Milk, in P. Connolly, ed., Perspectives of Indian Religion, Delhi, 1986, 72.
44
gveda, 3.8.
45
See, for example, atapatha Brhmaa, 5.2.1.8, SBE vol. 41, 31; etc.
46
A.B. Keith, The Veda of the Black Yajus School, 1914, 516520.
42

!
!

171!

since neither the gyatr nor animal sacrifice have any place there. But the main mythical
and symbolic meanings of the ypa remain remarkably little changed in modern
Vajrayna ritual, since these were from the start more general and less specifically Vedic
in nature.
The basic ritual meaning of the ypa was twofold. On the one hand, it served as
the pathway to the gods, and as the conduit by which the essence of the sacrifice rose up
to the gods.47 Thus the Yajurveda says the ypa is connected with all the gods; verily in
setting up the post he delights all the gods.48
We read in the atapatha Brhmaa, in Eggelings translation, that at the climax
of an important sacrifice called the Vjapeya, a ladder is leant against the ypa and the
sacrificer and his wife mount the post. As they climb, the sacrificer says we have
become Prajpatis children. As he touches the top, he says we have gone to the light,
O ye gods! Then he finally stands on the top, and says we have become immortal,
whereby he wins the world of the gods. He mutters, ours be your power, ours your
manhood and intelligence, ours be your energies.49
But it was precisely this function of opening the pathway to the gods that gave to
the ypa its other basic ritual meaning; which, paradoxically, was the diametric opposite,
namely to block the pathway to the gods.
In his recent book, Cuire le Monde, Charles Malamoud analyses the mythology
that gives rise to the ypas paradoxical double function. The myth is that in earliest
times, the gods themselves were the first to discover that they could attain heaven by
means of the sacrifice. Wanting to remain exclusive possessors of this unique advantage,
they plotted to deprive mankind of any knowledge of their wonderful new discovery.
According to the Aitereya Brhmaa (following Levi and Malamouds translation):
By means of the sacrifice, the gods raised themselves straight up to the celestial
realms. They became worried, thinking When they catch sight of this heaven,
which belongs to us, the humans and the is will want to follow in our
footsteps. So they made an obstacle (ayopayan) with the sacrificial post; and it is
for this reason that the post is called ypa. They came down again, and fixed the
ypa upside down, with its tip pointing downward, and then they disappeared
straight up again into heaven.
Then the men and the is arrived at the site where the gods had made their
sacrifice. Let us look for anything that could be a clue to this sacrifice! they
cried. But they couldnt find anything except the upside down ypa post. They
perceived that by this post, the gods had blocked

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
47

See, for example, C. Malamoud, Cuire le monde, Paris, 1989, 248 ff.
Black Yajurveda 6.3.34; A.B. Keith, op. cit., 519 ff.
49
SBE, vol. 41, 31.
48

!
!

172!

the path of the sacrifice. So they dug it out, and re-erected it the right way up,
with its point towards the top. Thus they became able to see where the celestial
realms were. If in the sacrifice offered nowadays by men the post is stood with its
point upwards, this is in order to recognize the route of the sacrifice, to reveal in
which direction lies the celestial realms.50
Malamoud continues by explaining the puns on the word ypa found in this and
several similar passages of the atapatha Brhmaa and the Yajurveda. According to
traditional analysis, the word ypa can clearly be derived from the verbal root yup,
understood in the sense of to obstruct, of which the word ayopayan, they made an
obstacle, is the causative imperfect. But discernable in the background are two other
opposed meanings: yu to unite and yu to separate. These, writes Malamoud, are the
functions of the post: it serves as the passage from the earth to the celestial realms, when
it is standing point upwards; but it prevents this connection when it is placed point
down.51 Another understanding of the pun can render the causative of the root yup as
disperse or eliminate the traces of. Thus another variant of the myth of the ypa
occurs in the atapatha Brhmaa 3, 1.4.3. According to Malamouds French translation,
it goes as follows:
It is by means of the sacrifice that the gods conquered this conquest which is
their conquest. They said to themselves: How to make it impossible for the
humans to ascend to this wellbeing which is ours? They sucked up the sap of the
sacrifice as bees suck honey, and when they had thus drawn out all the milk of the
sacrifice, they took the sacrificial post (ypa) and used it to efface the traces
(ayopayan) of the sacrifice, and then they disappeared.
It is from these myths, ultimately, that the two major functions of the modern
Tibetan phur-ba derive. The phur-ba, like the ypa, serves both as a direct conduit to the
sacred expanse, especially of a magically slain victim, and also as a boundary marker
keeping the unsuitable out of the sacred arena.
The first of these occurs in the rite known as sgrol in Tibetan and moka in
Sanskrit, meaning liberation by killing. This important rite, especially ubiquitous
among the rNying-ma-pa, involves the liberation, i.e. the stabbing to death with a
phur-ba, of a liga or effigy of a Hindu deity dubbed Matragara Rudra.52

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
50

C. Malamoud, op. cit., 2489.

ibid.
52
This means Rudra the Mothereater and is a satirical Buddhist play on the Hindu myth of Rudras
parthenogenesis, i.e. his being born of the seed of all the gods but with no mother. On the contrary,
say the Buddhists, Rudra did have a mothera prostitute who had copulated with numerous
demonsbut he ate her up as soon as he was born, so giving the illusion of being motherless, and thus
should properly be regarded as a demonic mother-eating bastard, not a divine god as the aivas
believe. At least, this is my interpretation of the myth. R.A. Stein, however, sees matragara as a
corruption of matagi. See Annuaire du Collge de France, 1972; 1971, 2, 505.
51

!
!

173!

For unlike the more ecumenical Newar Buddhists, who see Rudra as a form of the
Bodhisattva Vajrapi, in Tibet this Hindu deity of impurity and destruction nearly
always only functioned as the Buddhist devil. The rite of sgrol is most frequently
performed at the tshogs, or gaacakra ritual. The implication of it is that the ignorant or
demonic consciousness identified with the effigy of Rudra, whether ones own or
someone elses, becomes liberated into the pure land or dharmadhtu of wisdom by this
forceful method. Stephan Beyer53 quotes a typical liturgical example from a rNying-mapa gaacakra pj:
the Vajra Kla!
He casts down like hail.
He is quick like lightning.
Planted in the hearts centre
Of those to be liberated
Out through the kla
Is drawn their awareness
And flung to the Heart of Padmasambhava,
The depths of all-beneficent intention. Ah!
The Black Yajurveda54 describes animal sacrifice in very similar terms. The
animal is said to give its consent to die, while tied to the ypa, because its consciousness
will rise from there straight up to the heavens. Likewise, in the gveda, the sacrificial
animal is told: Truly you do not die, you do not suffer harm. By paths easy to traverse,
you go to the gods.55 In this Buddhist version too, the practitioner gives his consent that
his spiritual negativities and egotism shall be sacrificed, represented by the form of a
Rudra-like dough effigy, precisely because, like the animal victim of the Vedic sacrifice,
his consciousness too will be drawn through the ypa-like phur-ba, straight to the Pure
Land or Heart of Padmasambhava; although of course the Buddhist rite is merely a
symbolic daily liturgy, not a genuine blood sacrifice like the Vedic rite. Beyer continues:
The Master plants his kla in the hearts of the demons, slaying their bodies and
liberating their awareness to the Pure Land of Padmasambhava. Then the
assembly offers up the corpses of the evil spirits as food.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
53

S. Beyer, The Cult of Tr, Berkeley, 1973, 316.

A.B. Keith, op. cit., 526, n. 6.
55
igveda, 1.162.21, quoted in M. Boyce, A History of Zoroastrianism, 149. This was why sacrificial
animals in both Brahmanic and Zoroastrian rituals had to be conscious at the time of death.
54

!
!

174!

This is then divided up by a complex process and eaten. Similarly, at a Vedic sacrifice,
while the spirit of the animal went up to the gods, its body remained behind to be divided
up as food, after a complex system of division.
In this way we can see that the sacrificial allusions implicit in the ypa-like form
of the Tibetan phur-ba are highly appropriate. Indeed, it is not at all impossible that they
were originally intended to be quite explicit. It seems indisputable that Buddhist kplika
ritual is calqued upon aiva prototypes, and animal sacrifice and even human sacrifice
were undoubtedly standard features in aiva ritual, where victims were offered daily, as a
part of the bali offerings, to the great patron deity of aiva tantrism, Rudra, or his
entourage of kplika gods.56 Such rites were seen as particularly offensive by Buddhists,
hence in parrallel Buddhist Mahyoga ritual, the same theme of blood sacrifice is
continued symbolically in the equivalent bali and gaacakra sections of the Buddhist
rite, but completely and totally inverted. Now Rudra is not only hoisted on his own
petard, he is sacrificed at his own ypa or kla and served up as food to the Buddhist
Herukas who consume Rudra, digest him, and thus render him and all his perverted
Hindu kplika ways beneficial. Thus the esoteric Vajrayna or Buddhist kplikaism is
born of a sacrificethe sacrifice of the evil Hindu deity Rudra at his own sacrificial stake
by the victorious Buddhist Herukasand this great event repeated and celebrated daily in
the Mahyoga gaacakra pjs or bali rituals. It is the macrocosmic counterpart to the
microcosmic transformation of the individual practitioners spiritual negativities, and
goes a long way to explain the centrality of the sacrificial implement, the kla, as a
symbol in Mahyoga ritual and texts. Indeed, this topsy-turvy theme pervades all
Mahyoga ritual, which consistently turns the world of extreme kplika aiva ritual
upside down, outwardly adopting its most transgressive elements while inverting all its
meanings to a morality acceptable to Mahyna Buddhism.57 The consistent precision
and accuracy with which rNying-ma-pa ritual colonizes, digests and inverts the extreme
kplika cults deriving ultimately from the Hindu deity Rudra suggests very strongly
indeed a Sanskritic origin for rNying-ma-pa

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
56

See S. Gupta et al., Hindu Tantrism, Delhi, 1979, 1523.

This interpretation of the use of the ypa-like phur-ba to sacrifice an effigy of Rudra in the daily
gaacakra practice is entirely my own. Nevertheless, I feel it is well supported by readings of all the
classic Mahyoga texts such as the Guhyagarbha-tantra (ch. 15 especially) and the Phur-ba bcu-gnyis
and is made well-nigh totally explicit in numerous ritual manuals. Cantwell (An Ethnographic
Account of the Religious Practice in a Tibetan Buddhist Refugee Monastery in Northern India, 180;
and glossary, 53) shows also how the red or symbolically bloody gtor-ma offerings of Mahyoga
Buddhism are always construed, ultimately, as being parts of Rudras sacrificed corpse, representing
the practitioners negativities. Contrast this with the exactly equivalent actually bloody bali offerings
of aiva ritual, where, far from being the victim, it is Rudra who demands and receives the blood
offerings. I am publishing a paper on this theme in the near future.

57

!
!

175!

scriptures. This applies not only to the phur-ba tradition, but also to all aspects of rNyingma-pa religion.58
It is therefore obvious that it is not only the Buddhist tantras that have such rites
of liberation by means of the kla. For example, the Vikha-tantra, the sole
remaining text of the once widespread Vmasrota of aiva tantrism,59 contains such a
rite. Here a kla made of human bone is planted in the genitals of a liga,60 and when it is
drawn out again, the being represented by the liga will become liberated (mokam),
i.e. die. Similar references occur in several other Hindu tantras.61 There is not always
indication here of the kla being triangular in shape, but according to Goudriaan, in some
texts this is made explicitfor example, in the Tantrasrasagraha of Nryaa, ritual
killing or eradication is effected with a triangular kla made of nimba wood. So also in an
identical stanza in the naivagurudevapaddhati, and in several other texts.62 However,
Goudriaan makes no mention of whether or not such Hindu tantric klas have an eightsided shaft. Unfortunately, human and animal sacrifice have almost become proscribed
subjects for modern Indologists, and the standard reference works on tantrism often give
few details. It is therefore hard to ascertain if the kla or ypa was the usual implement of
tantric blood sacrifice, although we know without doubt that it was the standard
implement of Vedic blood sacrifice.
The second mythical purpose of the ypa, its function as the marker of sacred
boundaries, is equally ubiquitous in Tibetan Vajraynas use of its ypa-like phur-ba. In
Tibetan tantrism, ten phur-bas are customarily placed, or visualized, around the periphery
of a maala to prevent unwanted forces from entering. This rite is found, for example,
in the lower tantras of Kriy and Cary; thus Tsong-kha-pa writes in his sNgags-rimchen-mo:
The Susiddhi-tantra explains that the fierce (deity) Kilikla is always associated
with the fence, latticework and klas surrounding the house in the sense of abiding
there as protection.63

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
58

Compare, for example, the one-eyed, one-haired, jackal-emanating rNying-ma-pa protectress

Ekaja with her almost identical twin of the same name in the Klikpura, 63, 62, and the
ubiquitous aiva tradition of regarding jackals (iv) as emanations of the akti. Or compare the
frequent occurrence of that Sanskritic literary clich, the cuckoo, kokila, in the name of rDzogs-chen
texts. Only by an unfortunate ignorance of Sanskrit literature have Tibetologists been able to persist
for decades in construing these as autochtonous Tibetan developments.
59
T. Goudriaan & S. Gupta, Hindu Tantric and akta Literature, Wiesbaden, 1981, 16.
60
H. Brunner-Lachaux, IIJ, 31, 3, 1988, 248.
61
T. Goudriaan, My, Divine and Human, Delhi, 1978, 263, 374 ff.
62
T. Goudriaan, op. cit., 374; personal communication, 23 Jan 1990.
63
J. Hopkins, The Yoga of Tibet, 1981, 100. See also R.A. Stein, Annuaire du Collge de France,
passim.

!
!

176!

Similarly, it is found in such anuttara traditions as Yamntaka, Cakraavara,

and Sampua,64 and also of course in the Mahyoga and Tibetan gter-ma traditions. For
example, Cathy Cantwell gives the following example from the same rNying-ma-pa gterma ritual we have mentioned above in connection with the earth-holding rite: here, a
protective circle of ten phur-bas are ritually hammered into receptacles set in the ground
at the ten directions all around the periphery of the maala, prior to setting up the main
maala of the rite. These phur-bas constitute the secret or most esoteric aspect of
marking the boundaries of the sacred maala, the abode of the Vajrayna deities.
Having performed this ritual, Cantwell writes, the boundaries were protected so that
people could not come and go.
Again, this commonplace Tibetan use of the Vajrayna phur-ba makes perfect
sense in terms of the original myth of the ypa; here, the deities inhabiting the sacred
Vajrayna maala want to keep out profane influences, and they do this by marking
their boundaries with ypa-like phur-bas, thus excluding the uninitiated and the
unsuitable, just as the Vedic gods of the myth used the ypa to try to keep the humans
and the is at bay.
Needless to say, Hindu texts also describe the use of klas as markers and
protectors of ritual or sacred boundaries of various kinds, although in modern ritual it is
often only simple sticks or bamboos that are used. Older textual examples are a little
more elaborate. One can quote the Tantrasamuccaya of Nryaa, where pegs, this time
called akus, are employed. Likewise, the Somaambhupaddhati describes the use of
four klakas to protect the funeral pyre, the Klottara prescribes twelve klas to protect
the maapa, and the Garua Pura has a ritual to protect a site or field from all kinds
of harm, especially from thunderbolts and explosions, by the enclosing of the site with
eight klas made from khadira wood. If Garuas mantras are recited over the klas, the
site is said to become inviolable.65 But S.C. Banerjee, a contemporary author, describes
the use of klakas merely as small sticks.66
Of course, both these aspects of the ypa myth, the aspect of opening up the path
to the gods and the aspect of blocking the path to them, are interrelated. A palace door,
after all, can both open and close. There are numerous references in

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
64

For Anuttaratantra examples, see Kazi Dawa Samdup, The rcakrasabhara Tantra, London &
Calcutta, 1919, 13 ff., as quoted in K. Dowman, Skydancer, 350, n. 29; also, Vajrval, 3337 and
Kriysamuccaya, 7779. Also D. Gellners unpublished DPhil thesis, 455. For gter-ma examples, see
C. Cantwells thesis, section 5.1.5.3. Bulcu Sikls has also shown me three instances of kla rites in
the tantras connected with Vajrabhairava, where killing is the most notable theme.
65
Garua Pura, ch. 20, vv. 810, p. 82; Tantrasamuccaya, Part 1, 79; H. Brunner-Lachaux,
Somaambhupaddhati, Part 3, Pondichry, 1977, 592.
66
S.C. Banerjee, op. cit., 560.

!
!

177!

the Pli sources of the conflated ypa-kla serving such palace gate like functions,
generally marking the boundary between a more exclusive and a less exclusive space in
ancient Indian town planning and architecture.
A.K.Warder translates the Pli word indakhlo as royal stake, marking the royal
threshold.67 He adds, Inda is the name of the king of the gods, hence a title for any
king. Lily de Silva elaborates:
Indakhla, although explained in the commentaries as ummra, threshold, is
never used for any and every threshold. It only means a threshold to a settlement
[a city or village] under regal authority or to a kings household. Pcittiya rule no.
83 clearly states that if a monk, not announced beforehand, crosses the threshold
(indakhla) of an annointed king, he is guilty of an offence of expiation. (Vinaya,
vol. 4, p. 160) Here, indakhla is annotated as the threshold of the sleeping
room. (ibid.) Sleeping room means: there wherever the kings bed is made
ready, even if it is only surrounded by a screen wall (snipkra). (ibid.)
Therefore the indakhla can be conclusively established as a royal stake marking
the royal threshold.68
In her scholarly article, Lily de Silva has reviewed much of the Pli literature on
the indrakla. We can do no better than to quote at length from her research:
In the canonical literature the indrakla is sometimes referred to as esik or
esiktthambha. According to the Mahsudassana-sutta the prosperous city of
Kusavati had as many as seven esik wrought of gold, silver, beryl, crystal, agate,
coral, and all kinds of gems, standing at each of the city gates. Its commentary
explains that these pillars, as tall as 15 or 20 cubits, were erected at each and
every doorpost of the city gates. The Dgha-Ahakath-ka identifies them as
indrakla. A passage in the Aguttaranikya speaks of an esik which stands in the
frontier city of the kingdom, and it is said to be deeply embedded, well dug in,
immovable and unshakeable. The commentary on this passage furnishes us with
the following excellent description of an esik:
The esiktthambha is made of bricks, stone, or some hard well seasoned
timber like khadira. When it is erected for protection, it is planted outside
the city, for ornamentation it is planted inside the city. When constructing
it with bricks, a large deep pit is dug and filled with bricks up to the
ground level, and above ground level it is made octagonal in shape and it
is painted white. It is polished and painted to such perfection that when
elephants rub their tusks

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
67
68

!
!

A K Warder, Introduction to Pli, London, 1963, 363.

de Silva, op. cit., 239.
178!

against it, the paint does not chip off. Stone pillars are also octagonal in
shape. If the pillars are 8 cubits high, 4 cubits are embedded underground
and 4 cubits remain visible above ground. It is the same with pillars 16 or
20 cubits high. In all cases half the length of the pillar remains buried
underground and half remains visible above ground. They stand in a zigzag pattern (gomuttavak). Therefore it is possible to utilize the space in
between them for some purpose by flanking them with timber
[?padaracaya katv] These pillars are decorated with beautiful drawings
and flags are hoisted on them.
It is interesting to note that these pillars constitute one of the security measures of
the city, for they form a sort of fortress for protecting the inmates and for
warding off danger from outside. In other words, with these defence measures
the king assures safety for his subjects living within the kingdom demarcated by
these limits, and outsiders dare not cross this border without provoking offence
Therefore it is possible to conclude that the indrakla is a royal symbol generally
set up at the entrance to places where the kings authority must be recognized.
From this survey of Pli literary sources we can finally define the indrakla as a
firm pillar which stands as a symbol of royal authority at the entrance to a city,
village or palace.69
De Silva continues with a survey of the Sinhalese literary references to the
indrakla. While noting that most of the Sinhalese sources merely confirm what was
already written in the Pli sources, she found that both Sinhalese and Indian Tamil
sources made one significant addition to the Pli sources. They furnish conclusive
evidence of the practice of worshipping the indrakla in the manner that an image of a
deity is worshipped.70 She quotes, for example, the Sinhalese Saddharmaratnvaliya71
which says that even though small children dirty the indrakla erected at the city gates
by putting rubbish on it, discreet ones offer incense and flowers to it. Similarly, the third
or fourth-century Tamil poem, the Garland of Madurai says, according to Basham, that
the poet enters the city by its great gates, the posts of which are carved with
images of the goddess Lakm, and which are grimy with ghee, poured in oblation
upon them to bring safety and prosperity to the city they guard.72
De Silva adds:

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
69

ibid., 240.
ibid., 242.
71
ibid., 242.
72
A.L. Basham, The Wonder that was India, London, 1954, 203.
70

!
!

179!

In Pli literature, the idea of deities residing in the dvrakohaka of cities, and in
palaces of kings and setthi is quite common. These deities ward off disaster as is
revealed by the Kliyakkinnvatthu of the Dhammapada Ahakath, and they
sometimes even try to advise the inmates of the household under their care. They
are occasionally propitiated even with human sacrifices. Now the dvrakohaka
and indrakla are component parts of city gates and are so accepted even in the
Abhidnappadpik. We can therefore surmise that the idea of deities residing in
dvrakohaka may have been the historical antecedent of the indrakla being
regarded as an object of worship.73
Like Mt Meru, de Silva writes,74 these indraklas of the Pli sources must be
embedded in the ground to a depth equal to the part appearing above ground. In this way,
the kla is identified with Mt Meru, on the top of which, of course, lies Indras paradise.
Like the ypa, it must be eight faceted. And the ypa too has Indras paradise situated at
its top, according to all the sources in the Yajurveda and the Brhmaas. Thus the
conflated ypa-kla, known usually as an indrakla, can fulfil its functions as boundary
marker. It symbolises the pathway between the human realms at its base, and Indras
realm at its top.
Connected with the indrakla as ritual boundary marker is its more prosaic
function as a gate bolt, i.e. as the firm stake set in the ground at the middle of the gateway
to which the two doors are then fastened to keep them closed. Kauilyas Arthastra
describes the indrakla thus in the context of a discussion of fortress construction.75
Similar references to the indrakla as gate bolts occur in Buddhist Sanskrit texts such as
the Divyvadna, the Mahvastu and the Gaavyha-stra.76 But we can see that the
indrakla is in these sources no ordinary doorbolt; for apparently when the Buddha treads
on it, on first entering a particular city, the whole earth shakes and roars. Here the
indrakla as doorbolt seems to be a paradigm of immovability and stability.
It is from ancient Buddhist Sanskrit and Pli architectural sources that we find
further evidence of the conflation of the ypa and the indrakla. The Divyvadna
describes how a pillar called a ypa should always be set up as the central axial column
within a stpas dome.77 This is reminiscent of the erection of ypas as a magical device
by Brahmans, attested in the Pli sources.78 Paranavitana,79 working on archaeological
sites in Ceylon, found in exact accordance with the

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
73

de Silva, op. cit., 242.

de Silva, op. cit., 243.
75
Kauilya Arthastra, 2.3.36.
76
Quoted in de Silva, 241, note 53.
77
Quoted in de Silva, op. cit., 247.
78
Jtaka, vol. 6, 211214. Quoted in de Silva, 245.
79
S. Paranavitana, The Stpa in Ceylon, Colombo, 1947, 3539.
74

!
!

180!

Divyvadna, that large stone pillars, eight faceted and with the ypas knotted rope
carved at the appropriate place, were indeed characteristic of ancient Sinhalese stpa
sites. John Irwin has assembled evidence to show that by now decayed wooden ypas
might well have existed as the axial columns of many other ancient stpas too. But
Paranavitana reported that the modern Sinhalese monks always identify these vast stone
ypas as indraklas. In their eyes, the two items had apparently become totally conflated.
As the Hindu equivalent to these earlier Buddhist structures, the ilpastras stpiklas
also had the 8-faceted characteristic of the ypa, but also bore the name kla in some texts
and the name ypa in others.
The most important contemporary use of the kla in the Theravda tradition is
surely its function as a central feature of the paritta ceremony. De Silva writes,
The indrakla is a special ceremonial post which enjoys pride of place in the
parittamaapa. It is traditionally erected for all 7-day paritta ceremonies, and
sometimes for over-night ceremonies as well.80
The very popular paritta ceremony is used, as its name implies, to give protection
from diseases and other such misfortunes, and although it largely involves the recitation
of a set of particular canonical Pli scriptures, its function is decidedly apotropeic. The
association of the kla as a central feature of this most important of Theravda Buddhist
protection rituals is entirely consonant with the Vajrayna Buddhist notion of the kla as
the most powerful and favoured of specifically protective ritual implements. Like his
Sinhalese monastic counterpart, the contemporary Tibetan phur-ba expert is frequently
called out to use his Vajrakla rituals to protect the laity from disease and misfortune; yet,
like the paritta ceremony, the Tibetan Vajrakla rituals are very explicitly Buddhist in
nature, not merely magical. Likewise, in the parittamaapa, the kla is associated with
the cosmic central axis as the bodhimaa at Bodhgaya, the place where the Buddha
became enlightened. In the same way, the Tibetan phur-ba is esoterically associated with
the cosmic axis as the esoteric bodhimaa, i.e. the central channel in yoga, the place
where the tantric yogin becomes enlightened.
In conclusion, we can see that the myths of Indra slaying Vtra and the important
Vedic ritual device, the ypa, gave rise to a wide range of ritual developments over
subsequent millennia. In this very preliminary survey, necessarily only a small fraction of
them have been mentioned. Nevertheless, as we can see, Lily de Silva concluded from
her detailed study of the Pli sources that the kla and the ypa became conflated at some
stage.
In support of her conclusions, we can cite not only the ilpastras, where ypa
and kla are interchangeable terms for the same item, but we can add that the Tibetan
phur-ba, of which she was apparently entirely ignorant, both in its

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
80

!
!

de Silva, op. cit., 234.

181!

iconographical forms as well as in its ritual meanings, appears also very accurately to
embody all the important aspects of both the indrakla proper and the ypa. And if the
Tibetan phur-ba too seems to embody these two uniquely Indian devices of kla and
ypa, this tends to suggest an Indian rather than Tibetan origin for it. Since its precise
form seems to have close counterparts in the stpikla described by Mnasra and other
texts, and in the triangular kla of some Hindu tantras, the probability is therefore strong
that the Tibetan phur-ba or kla as we know it is a product of Indian culture. Every detail
of its shape is apparently accountable for from Indian sources. Its eight-faceted shaft
ornamented with knotted ropes derives ultimately from the Vedic ypa, and precisely
parallels the later ypa or kla of South Indian and Sinhalese religious architecture, both
Hindu and Buddhist. Its tripartite base is found both in a variety of ilpastras, and in
Hindu tantric rites of magical eradication or killing. One might strongly surmise that
something very like it was used in aiva human and animal sacrifice at the great
Brahmanic temples (but information on this is hard to find because such sacrifices have
become virtually a taboo subject among contemporary Indologists). Its makara
ornaments half-way along its length, at the junction between the triangular blade and the
eight-faceted shaft, with the ngas extending down the blade, are consonant with the
conceptualization in Indian Buddhist cosmology of Meru as semi-submerged, and
perhaps also find a further Indic counterpart in the related and frequent practice of
conceiving ritual pillars of many descriptions as rising out of mythical oceans.
If no physical survivals of Indian Buddhist klas of the so-called Tibetan model
have come down to us, this can perhaps be attributed to the rusting of iron, or the rotting
of khadira wood; or, more likely, to an insufficient search. After all, no tantric Buddhist
klas of any type whatsoever have come down to us from India, as far as I and the learned
staff of our larger museums know. Yet not even the most conservative scholars have
doubted that such items must have existed in India, widely attested as they are in a wide
range of Buddhist tantric texts of all classes, including such Kriy texts as the Susiddhitantra etc., Cary texts such as the Vajrapyabhieka etc., Yoga texts such as the
Sarvatathgatatattvasagraha etc., Mahyoga texts such as the Guhyasamja etc., and
Anuttara texts such as the Cakraavara tantras. The use of the single-bladed kla
continues to this day among the followers of the undoubtedly canonical
Mahvairocana-stra in Japan. Yet if none even of these has turned up in our museums
from Indian sources, why should we expect the three-bladed ypa-like variety of the
Mahyoga texts to have fared any better?
Finally, and most importantly, the widespread opinion that the triangular-bladed
Tibetan phur-ba is unattested in Indian Vajrayna texts is quite possibly mistaken. Even
if we are, like some contemporary Western scholars, to unquestioningly and

!
!

182!

uncritically follow a few ultra-radical strands of a later Tibetan (or was it Manchu?)
minority tradition81 and accept the extreme and completely unanalysed and unproven
assumption that all the Bon and rNying-ma-pa tantras without exception are (or are
considered by the tradition to be) apocryphal forgeries, nevertheless even R.A. Stein
mentions three canonically accepted texts, the Vajraml-tantra and the Heruka-tantra
and a Mahkla-sdhana by Vararuci, transmitted by Siddha avarip,82 all of which, he
says, describe the kla in exactly the form that the Tibetans use it. The Vajraml-tantra
is an explanatory tantra of the Guhyasamja, and thus much favoured by the dGe-lugs-pa
school, which might explain why they too favour a three-bladed phur-ba. We can
probably expect more such references to be discovered in the future. However, so
entrenched has become the notion of the Tibetan phur-ba as a Tibetan invention, that
Stein has found it prudent to defer to established academic tradition and suggest, albeit
somewhat unconvincingly, that the canonical references to the so-called Tibetan phurba he had unexpectedly found were perhaps forged Tibetan interpolations, written into
the venerable Indian texts at a later date by cunning Tibetans determined to legitimate the
use of their beloved three-bladed phur-ba, by fair means or foul.83

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
81

Despite the Chien-lung emperors (17351796) cynically manipulative efforts to divide and rule
Tibet by such devices as a Golden Edict (gser-gyi bka-lung) in which the dGe-lugs-pa school were
declared by decree of the Manchu state to be a purer and better form of Buddhism than the three other
schools, (Thu-bkwan, Collected Works, 1969, 601 ff., thanks to Dan Martin for these references) we
must not forget that the great majority of Sa-skya and bKa-brgyud hierarchs from earliest times until
now have been enthusiastic expounders and practitioners of rNying-ma-pa non-canonical tantras, as
have many important dGe-lugs-pas, most famously perhaps the Great Fifth and sLe-lung bzhad-pai
rdo-rje and the Sixth Dalai Lama etc. Western scholars like Per Kvaerne, for example, must therefore
ask themselves what precisely the Christian term non-canonical means in such a context. I feel
Kvaernes perception of the rNying-ma-pas as a persecuted religious underground whose scriptures
were rejected en bloc by all the other schools (see P. Kvaerne, Tibet, the Rise and Fall of a
Monastic Tradition, in H. Bechert & R. Gombrich, The World of Buddhism, London, 1984, 262) is
deceptive. On the contrary, they were the most widely respected and expansive of all schools in
nineteenth-century East Tibet, for example, and clearly enjoyed widespread affection and appreciation
before that time too among most Tibetans. As a doctrinal common denominator uniting the other three
schools, however, they appear to have been attacked by foreign powers presumably wishing to control
Tibet by using the dGe-lugs-pa as puppets, e.g. by the Dzungars in 171720, and by the Yung-cheng
emperor in 1726. Much of Tibetan history still remains obscure, however; the reason for these foreign
attacks on the rNying-ma-pas is not yet certain.
82
R A Stein, Annuaire du Collge de France, 197778, 648, 6534; See also his Gueule de Makara,
in the same serial 1977, 55, also note 8. Vajraml-tantra, ch. 54; Heruka-tantra, ch. 12 (actually, as
Stein points out, this is a rNying-ma-pa text that gained admission to the bKa-gyur later on);
Vararucis sdhan is in the Peking edition of the bsTan-gyur, vol. 86, 65-6.
83
R.A. Stein, Annuaire du Collge de France, 197778, 648, 654.

!
!

183!

But we must not forget that all Tibetan traditions, both the Bon and the totality of
the Buddhist schools, accept the so-called Tibetan phur-ba as authentic, i.e. as nonTibetan. The Bon see it as originating in the (Sanskritized) west, and the Buddhists as
originating in the Indian south. The Sanskrit-based Newar tradition likewise accepts it as
orthodox.
Some Western scholars might argue that they do so on suspiciously slender
textual evidence. In reply, one must point out that entire and once huge sectors of aiva
tantrism textually disappeared from their native India over a few centuries. For example,
only one Vmasrota text survives; and nothing textually survives at all of the great Saura
tradition of tantrism that produced, inter alia, Konarak temple in Orissa, and which once
boasted at the very least a corpus of 85 texts, the titles of which alone survive, listed in
the rkanthyasahit.84 So, if only a handful of Sanskrit versions of the highly secret
Buddhist Mahyoga tantras had survived the depredations of time and Islam by the time
Bu-ston published his bKa-gyur; or, more to the point, if only a handful had been
surrendered up to him and thus to certain exposure in the public domain by their
zealously secretive rNying-ma-pa proprietors in Tibet, we need not be surprised. If entire,
vast and wealthy aiva tantric traditions could so rapidly disappear leaving no textual
trace from their native India, we should have no problem accepting that only a small
section of the triangular-kla quoting texts of the Mahyoga tradition survived in Sanskrit
in Tibet, or were available to Bu-ston in Sanskrit. After all, five centuries and several
Islamic invasions of India separated the times of Bu-ston and Padmasambhava.
A close reading of the main rNying-ma-pa Mahyoga Vajrakla root tantra, the
Phur-ba bcu-gnyis in 199 pages, shows it to be, judging by its internal evidence alone,
typically Indic in every detail, revealing no sign at all of Tibetan material. Its mantras,
mudrs, maalas and rites all correspond with the mainstream Buddhist tradition of
such texts as the Sarvatathgatatattva-sagraha and the Guhyasamja. For this very
reason its evidence in regard to a Sanskritic origin of the so-called Tibetan phur-ba is
also not absolutely irrelevant. Even if it is a forgery (an axiom apparently assumed a
priori and with no further analysis by so many Western scholars), it is at the very least a
forgery so well and so early made as a perfect replica of a genuine Indian tantra that we
can, as scholars, still use it as a valuable source for Indological material.
But we should not forget that it was the ultra-orthodox Sa-skya Paita himself
who claimed ownership of the original Sanskrit manuscript of the rNying-ma-pa
Vajraklayamla-tantrakhaa, and these very same ancient folios themselves, believed
by Sa-skya Paita to have belonged to Padmasambhava personally, apparently survived
intact at Sa-skya until the recent Maoist Cultural Revolution.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
84

!
!

Verbal communication, Alexis Sanderson, Oriental Institute, Oxford, 15th Oct. 1986.
184!

Although accepted into the bKa-gyur through the Sa-skya Panditas influence,85 this 5page fragment of text is to all appearances an excerpt of crucial passages from the much
longer and non-canonical rNying-ma-pa Phur-ba bcu-gnyis, and thus lies at the heart of
all traditional rNying-ma-pa Vajrakla sdhanas, both ancient and modern. In similar
vein to the Sa-skya Paita, Bu-ston noted that his main teacher Nyi-ma rgyal-mtshan
and others claimed to have seen parts of a Vajrakla-tantra in Sanskrit in Nepal. Not
having the Sanskrit versions himself, and being constrained by political pressures, Buston could not include them in his bKa-gyur.86 Yet despite such reliable testimony,
absolutely the sole positive evidence cited by Western scholars in asserting a Tibetan
origin for the triangular-bladed phur-ba is its association with rNying-ma-pa and Bon
texts which were rather wildly alleged by early Tibetologists, on the basis of no serious
examination, to be Tibetan-composed en bloc. There is no other independent evidence.
However, the discipline of Tibetology has now matured to the extent that the guesses of
the pioneers can be subjected to more intense scholarly scrutiny.
Since it now seems increasingly likely that the triangular kla is of Indian origin,
perhaps we ought to reconsider the matter. Perhaps the probable Indic origins of the
triangular kla can be added to the accountsof the Tun Huang texts P44, of Lalou 349;
of the sBa-bzhed; of the testimonies of the Sa-skya hierarch and Bu-stons teacheras
additional contributory evidence to a highly probable Sanskritic origin for the Vajrakla
tantras preserved in Tibet. After all, most Tibetan historians of all schools have been
saying this all along. So also, in recent years, has David Snellgrove, who very correctly
puts forward in his Indo-Tibetan Buddhism an account of the Indic or Sanskritic origin
for most of the Bon and rNying-ma-pa material. Snellgroves version of events has been
favorably received inter alia by de Jong, in his review article in the Indo-Iranian Journal.
As evidence to support Snellgroves analysis, we can point out that the existence
of an independent Buddhist kla deity, called variously Kilikla, Vajrakla, or later on,
Vajrakumra, has never been in doubt. As the Buddhist deification of the ypa-like kla
(and we must remember these were deified in the Hindu tradition from the start), Kilikla
usually functioned as the deity of establishing the sacred site and protecting its
boundaries, i.e. of ngabandha and smabandha etc., hence becoming connected to the
daakrodha protecting the ten directions. Stein, for example, has discovered him in such
a role in a wide range of Buddhist texts such as several Kriytantra texts translated into
Chinese in the eighth century. These include among many others the
Amoghapakalparja, the Susiddhikara-mahtantra, and

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
85

TTP, vol. 3, no. 78., TTD, 439. sDe-dge rgyud ca-pa, pp.s 8690. See the colophon by Sa-skya
Paita.
86
Quoted by G.N. Roerich in The Blue Annals, 102, n. 1.

!
!

185!

the Guhyastra etc.87 Stephen Hodge informs me that the deity exists in some other
Chinese texts not mentioned by Stein, such as the Ekkaramahoa; also in another
longer Ua text in 10 fascicules which describes itself as from Nland [T945]; also
in the Guhyasamanya-tantra; and even in the tantric Perfection of Wisdom in 500 lines.
By consulting the Taish Indices, Hodge also discovered further occurrences of the deity
Kilikla or Vajrakilikla in T895, T974 [a female form here!], T1120, T1124, T1125,
T1132, T1225, and T1227.88
R.A. Stein has, however, above all failed to point out that there are at least two
completely independent texts devoted specifically to Vajrakumra in the Chinese Canon,
and these are both undoubtedly authentic translations from the Sanskrit. According to the
Chinese tradition, they were translated several decades before the conversion of Tibet to
Buddhism, thus predating Padmasambhavas arrival in Tibet. Their existence in Chinese
is significant, for it falsifies the assertion of some commentators who have tried to deny
that any such tantric tradition with an independent literature ever existed in India. The
texts I have found so far are the Vajrakumra-tantra [T1222], a very long text that will
probably comprise between 30 to 50 pages in English translation; and the shorter
*Vajrakumrajapayoga-sdhana [T1223].89 There is a further text, T1224, which was
probably composed in China or Japan. From a preliminary reading, much of the material
in T1222 and T1223 corresponds functionally [if not lexically] with the Tibetan tradition,
despite their being explicitly derived from an earlier South Indian tradition as opposed to
the explicitly Northern origins of the later Tibetan tradition. In both traditions, the central
deity has the same name and the same wrathful functions and kplika appearence. The
Chinese text says that Vajrakumra stands on a jewelled mountain arising out of the
ocean, which is just perhaps suggestive of Meru etc. In another passage, T1222 says that
Vajrakumra wields a single-pointed vajra, often conflated with a kla in far-eastern
texts. Also in both traditions there is mention of the Asuras cave as an all-important
location, and above all they share an absolute centrality of the theme of the use of the
abhicra system to convert forces hostile to Buddhism, and both include a particular
emphasis on the rites of moka in which an effigy is stabbed with a kla. They also both
have rites to discover hidden treasures (Skt nidhi, Tib gter-ma) protected by guardian
deities, and stress the special appropriateness of the cycle for monarchs

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
87

R.A. Stein, Annuaire du Collge de France, 197678 passim.

Personal communications, Stephen Hodge, 20 December 1989, and 25 February 1990.
89
Thanks to Stephen Hodge for helping me locate the Taish number of the longer text, T1222, and in
the process discovering the other two. Heartfelt thanks also to Luke Lau, of St Antonys College,
Oxford, and to Stephen Hodge for reading most of T1222 with me. It is unequivocally not an excerpt
from ch. 6 of the Susiddhikara, but a substantial independent text with striking parallels in function
and purport to the Tibetan apocryphal versions, although lexically it is quite different.
88

!
!

186!

and generals to protect the state (although these latter two themes are perhaps too
commonplace to be counted as significant parrallels).
A detailed reading of these texts is not yet complete, but findings from them will
be published in a forthcoming article. It is not yet entirely clear if the deity in these texts
is yet identified with Kiliklabut the fact that a Buddhist kplika deity specializing in
abhicra and called Vajrakumra was worshipped in eighth-century India is itself very
significant. This much has not yet been established or accepted by previous researchers.
Nor has adequate account been taken of the prominent role Vajrakumra plays in
Japanese esoteric traditions, both as an individual and as a class of guardian deity, as
Michael Strickmann has recently pointed out to me.
As well as such Chinese texts, many canonical Kriyatantra texts in Tibetan cite
the Kilikla deitye.g. the Vidyottama-tantra; the Kriysagraha; and a wide range of
other texts, such as four Sittapatr texts and their commentary. These texts also cite the
daakrodha in the same form that many apocryphal Tibetan Vajrakla texts do.90 The
incidence of the kla deity seems to be extremely frequent in Kriytantra texts in general.
However, he is also widespread in Cary texts such as the Vajrapyabhiekamahtantra; and in Yoga texts such as a Vajraekhara-tantra; and in famous Mahyoga
texts such as the Guhyasamja and the Vajraml; etc. In these Mahyoga texts, the kla
deity begins to take a much more central role, and to resemble the rNying-ma-pa deity
very closely, for example his consort Dptacakra appears in the Guhyasamja-tantra and
its derivative literature.91
Perhaps a special note should be made of the occurrences of the kla material in
Nepal. There are problems in assessing this material, because of the undoubted influence
of Tibetan Buddhism in the Kathmandu Valley. Nevertheless, kla rites occur at all levels
of Newar religion, both Hindu and Buddhist. Not only is the kla the standard implement
of the Nepalese jhkr, but the vajrcrya too performs a large number of kla rites,
ranging from esoteric initiatory ritual to the protection of domestic houses or the making
of rain.92 As a result, a very large body of Sanskrit kla material can be found in Nepal,
but so far none of the material I have looked at seems to be specifically of the rNyingma-pa model.93 It tends to

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
90

F. Lessing & A. Wayman, op. cit., 117, n. 18.

See R.A. Stein, Annuaire du Collge de France 1978, 197778, 648. Stein shows her to be the
personification of the rakcakra, which of course includes the protectors of the ten directions, the
daakrodha.
92
Thanks to David Gellner for this information.
93
Stephen Hodge tells me that a particular Nepalese text preserved on microfiche by the IASWR
seems to be a much longer Buddhist Sanskrit kla text, just conceivably a Sanskrit version of rNyingma-pa material, but interpretation of this text is not yet complete. While in Kathmandu, I found many
uncatalogued and unedited kla texts, some of them lengthy, too many to research exhaustively at that
time. Brahmans and Vajrcryas alike told me of a complex kla tradition in the Kathmandu Valley.
91

!
!

187!

correspond to the rakcakra rites of the Anuttara tantras, or to the Guhyasamja

material. Noteworthy is how similar some of the Hindu and Buddhist material is.
Nevertheless, all Newar Buddhists certainly use the same shape and form of kla as their
Tibetan counterparts.
The kla deity has therefore had a very long and distinguished history in many
undoubtedly canonical texts in Sanskrit, Tibetan and Chinese. Even more numerous,
even ubiquitous, are the mantras, mudrs, maalas and rituals involving the kla as
object; such occurrences pervade the whole of tantric Buddhist writings from the earliest
Kriytantras up to the latest Klcakra literature. It therefore seems to me unrealistic and
romantic to take as an unquestionable axiom (as do so many Tibetologists) that all
Tibetan texts and rituals devoted specifically to the kla deity must be apocrypha copied
by deceitful rNying-ma-pa charlatans from autochtonous Tibetan shamanswhoever
they might have been.
The rNying-ma-pas have always claimed that the Vajrakla texts they preserve
were brought by Padmasambhava from Nland to Tibet, in order to perform the kla
rites needed to remove the obstacles to the completion of bSam-yas Temple, and then to
consecrate it and the whole of Tibet. Tun Huang text P44 gives exactly the same account.
This scenario is, for anyone at all Indologically informed, far more probable than the
bizarre scenario alluded to above. We know, from Sanskrit sources, that all new temples
required such kla rites for their foundational and consecrational rituals; and that the deity
Kilikla was always associated, from earliest Buddhist tantric texts onwards, with the
removal of preliminary obstacles to such undertakings and the protection of the site. We
also know from Chinese sources that independent Vajrakumra tantras already existed in
India by the eighth century; moreover, texts with similar material or at least identical
functions to the Tibetan rNying-ma-pa Vajrakumra tantras. We also know that a kla
deity with many of the precise characteristics of the rNying-ma-pa version had begun to
emerge in the undoubtedly canonical Guhyasamja literature. Why then must we take it
as an article of faith that all Tibetan Vajrakla or Vajrakumra tantras of the same period
must be apocryphal forgeries? It is far more likely that a substantial core at least of the
rNying-ma-pa Vajrakla texts were genuinely translated from Sanskrit, establishing a
model upon which later Tibetan composition was based.
There really can be no reasonable doubt that bSam-yas temple was consecrated
with kla rites, since all ilpastras require this. Nor can there be any reasonable doubt
that such rites were extremely long and complex, marking both the first and the last
moments in the construction process, since, again, the ilpastras make this clear. Nor
can there be any reasonable doubt that a highly developed abhicra tradition used to
destroy or convert forces hostile to the buddhadharma and based on the deity
Vajrakumra was translated into Chinese from Sanskrit in the first

!
!

188!

decades of the eighth century. I therefore believe (pace Stein, Dowman et. al.) that the
texts describing these lengthy rituals and abhicras are preserved in Tibetan translation in
the non-canonical rNying-ma-pa collections; for I have read texts there that would seem
to precisely fulfil these specifications.
It is true that, quite unlike their Chinese counterparts many centuries earlier, the
first compilers of the Tibetan Canon refused to allow any Vajrakla or Vajrakumra
texts at all into their collection, except, eventually, for one minuscule fragment that was
newly retranslated in their own time, 500 years after the original translation, by the Saskya Paita in the thirteenth century. Earlier translations even of that fragment, they
excluded from their canon, despite, as Bu-ston points out, their certain knowledge of the
existence of its Sanskrit original. But this is surely a reflection of the religious politics of
the times where, as Bu-ston broadly hints,94 the purveyors of the New Anuttara tantras
fought bitterly to upstage and displace the proprietors of the Old Mahayoga tantras such
as the Vajrakla.
The merely political and therefore religiously insignificant nature of the conflict is
revealed when one considers that despite all the rhetoric surrounding their exclusion
from the canon, the rNying-ma-pa tantras were in practice almost universally loved,
taught, and propagated by the hierarchs of the newer bKa-brgyud and Sa-skya schools,
and also by a great many dGe-lugs-pas after their arrival on the scene- until, it seems to
me, the divide and rule policy of the Manchu hegemony of the eighteenth century, as
formulated in the Golden Edict (gser-gyi bka-lung), mentioned above, made this
unpolitic.95 But with the decline of Manchu influence in Tibet, the 13th Dalai Lama
reintroduced the rNying-ma-pa Vajrakumra rites to his personal chapel, the rNam-rgyal
drva-tshang, where they continue to be practised to this day.
Thus it is that such a high proportion of the leading commentators and
practitioners of the apocryphal Vajrakla tantras in Tibet have, throughout the last 900
years, been gSar-ma-pas rather than rNying-ma-pas; for example, the Fifth Dalai Lama,
or the seventeenth-century Sa-skya-pa A-myes-zhabs. The still popular Western
academic myth that it is only after the nineteenth-century ris-med movement that rNyingma-pa tantras in general became widely accepted by the other schools is demonstrably
false; those who doubt this are invited to read, for example, the life histories and works of
all the Karma-pas, who effectively controlled Tibet for so many centuries and were
nearly all ardent devotees of rNying-ma-pa traditions.
Interestingly but predictably, it seems probable that it was not only the Buddhists
who worshipped a deity such as the one under discussion. A popular

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
94
95

!
!

G. N. Roerich, The Blue Annals, 102, note 1.

See note 81 above.
189!

Hindu god exists called variously Kumra, Skanda, or Krttikeya, and famously
worshipped in the south as Muruga or Subrahmaya. Some versions of this deity seem
quite probably to have been Hindu parallels to the Buddhist Vajrakumra. The parallels
between the two deities are sometimes striking, although, as always, great caution is
needed to avoid jumping to false conclusions with such materials. But both deities are
often worshipped in the month of Krtikka; with six-segmented rudrka bead mls.
Both hold two different kinds of vajras in their hands.96 Similar to his Buddhist
counterpart, the Hindu Kumra is regularly worshipped in what appears to be the form of
a ypa, and is identified with protective circles made up of cosmic mountains.97 Both
deities are considered by their followers to be the main demon-slayers or protectors of
their respective religions. Both are married to their work in the form of their consorts
the Hindu Devasen (army or weapon of the gods) and the Buddhist Dptacakra ( circle
of fire, i.e. protection). Kumra has an entourage of 9 male heroes (navavra)98 and
Vajrakumra has an entourage of 10 male Herukas (daakrodha). In their respective
Northern versions, Kumra has a vast entourage of female animal and bird-headed
mtks,99 while Vajrakumra has a vast entourage of female animal and bird-headed
pics. Kumra is closely associated with a goddess of child illnesses, Revat, and
Vajrakumra is closely associated with a disease-controlling goddess called Remat, who
sits on the flayed skin of her own child as the saddle of her mule. Just as Vajrakumra is
the supreme son (sras-mchog), so also Kumra is invariably celebrated as the supreme
divine son of Rudra.
Even in the very earliest strata of Tamil literature, the early Cakam literature,
what little we know of the Hindu god is quite strikingly close to the Buddhist tantric
tradition: the use of an effigy or puppet and a three-bladed, chisel-like stabbing
instrument in exorcisms; the worship of the deity as a small pillar (Tamil: kantu)
connected with blood sacrifice; the worship of the deity as a mountain.100 But our
knowledge of early Tamil material is so vague that it would be rash to jump to any
conclusions on the basis of such slender evidence. Nevertheless, tantalizingly, Filliozats
translation of an allegedly very early Tamil Kumra-tantra shows that the two deities
even have the distinctive kili element in their mantra in common, and LHernaults study
of the original form of the Hindu gods vel or lance shows that it may just conceivably
relate to the Buddhist kla; it

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
96

F. LHernault, LIconographie de Subrahmaa au Tamilnad, Pondichry, 1978, 19 passim.

F.W. Clothey, Rhythm and Intent, Madras, 1983, 122 ff. The ypa is of course used in other Hindu
rites as well, so one should not place too much emphasis on this point.
98
See F. LHernault, op. cit., 1756.
99
ibid., 30.
100
See F.W. Clothey, The Many Faces of Muruka, The Hague, 1978, 25 ff. But Fred Hardy reads the
excorcism text as a scrutiny, not stabbing. See his Viraha-Bhakti, London, 1983, 139. Worship of
the deity as a small pillar is however quite certain.
97

!
!

190!

was a three-faceted short stabbing instrument identified as a type or form of Indras

vajra.101
Perhaps significantly, T1222, the Chinese Vajrakumra-tantra, makes more
frequent and prominent mention of the Hindu deity Kumra than of any other Hindu
deity, and includes rites directed to him as well as to Vajrakumra. The Hindu Kumra
first appears in Buddhist texts in the Janavasabha-sutta of the Dghanikya. Here called
Sanam-Kumra, he plays the leading role in this Pli sutta as the principle preacher and
supporter of buddhadharma among all the gods, rather like a divine counterpart to the
human arhats nanda or riputra.102 As LHernault has shown,103 in the Hindu tradition
this youth-god has twin functions of scholar and warrior. By the time of Mahyna
Buddhism, he seems (again one suggests all this extremely tentatively) to be considered
to have become enlightened as a Bodhisattva and to manifest as two separate deities
first as Krttikeyamajur (later simply Majurkumra)104 the scholar and then later as
the tantric Vajrakumra the warrior. The precise details of how this Vajrakumra became
identified with the deity of establishing the site and the periphery Kilikla and the kla as
object are not yet quite clear at the time of writing, but it does not seem unreasonable to
suppose that something along these lines happened.
Properly speaking, the history of the development of the Buddhist Kla deity
through the ages, his texts and rites, and his Hindu parallels, are the subject matter for a
subsequent paper which will follow the present one.105 I mention it in passing here only
in order to illuminate the need to examine more than merely Tibetan sources before
jumping to wrong conclusions about the Tibetan origins of the rNying-ma-pa Vajrakla
texts. For until they have been critically examined and compared with the parallel
Chinese and Sanskrit material, it is simply unscientific to pass judgement on their origins.
So despite being very properly acclaimed as a seminal Tibetan scholar (and he has
undoubtedly been a major inspiration to us all) by his own admission,106 R.A. Stein is not
a Sanskritist, and his investigations into the kla therefore quite legitimately and very
professionally could use no Indic sources whatsoever other than those in Tibetan
translation. However, we must recognize the limitations of

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
101

See F. LHernault, op. cit., passim, but especially pp. 1456, 1589, 169171.
See The Sacred Books of the Buddhists series, vol. 3.
103
See F. LHernault, passim.
104
See A. Macdonald, Le Maala du Majurmlakalpa, Paris, 1962, 1223, n. 3.
105
Unknown to me at the time of setting out to write this paper, Martin Boord with the help of some
scholars at SOAS was simultaneously accumulating information the deity Kilikla and the kla as
object from Far Eastern, Sanskrit, and Tibetan sources. I hope to be permitted to draw on this
considerable resource when it is complete.
106
R.A.Stein, A propos des documents anciens relatifs au Phurbu, 434, n. 34.
102

!
!

191!

his approach.107 In the light of the Indological material that has been presented here,
perhaps we can now begin to give the Tibetan tradition the benefit of the doubt, and
consider the possibility that our Western academic tradition has perhaps been a little
mistaken on this particular issue. Perhaps we have been over-hasty in condemning the
various Tibetan traditions as mere medieval foolishness; for they all clearly assert the
non-Tibetan origin of the three-bladed kla and its literature, seeing its source either from
the (Sanskritized) west in the Bon tradition, or from the Sanskritized south in the
Buddhist conception. In my present opinion, both traditions are very possibly correct.
The present research only scratches the surface of the issue of the kla in India.
For lack of space, I have omitted much material, for example the closely related
indradhvaja or dhvajastambha cults and a mass of village magical material that involves
impalation. But clearly the magical concept of sticking pins in effigies of ones enemies
is so pervasive (in India as elsewhere) that not all such occurrences can be considered to
directly relate to the ancestry of the Tibetan phur-ba, although some of them might. I
have also been unable to deal with a large number of magical uses for klas which also
exist in important Buddhist texts; for example, they are used to control vetlas in the
Majurmlakalpa.108 So a great deal remains to be researched on the subject. I am sure
answers will not be hard to find, if only some one, preferably an Indologist fluent in
Sanskrit, finds the time and inclination to do the research.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
107

S. Karmay seems to take a quite different line from my own. He has attempted a history of the
Great Perfection with almost no reference whatsoever to Indian or Sanskrit sources. Although I
personally have found his work most illuminating and excellent, I can not help but note that it remains
unconvincing to those with a detailed knowledge of equivalent eighth-century esoteric Indian tantric
philosophies, whether aiva or Buddhist. A growing body of opinion seems to hold that the
requirements of scientific method neccessitate an enquiry into Indian sources as well as Tibetan,
before attempting to arrive at any conclusions about the Great Perfection doctrines country of origin.
Some scholars doubt if a non-Sanskritist or non-Indologist is qualified to make such an investigation.
108
Thanks to Martin Boord for this information.

!
!

192!

Theravda Buddhism and Brahmanical Hinduism:

Brahmanical Terms in a Buddhist Guise
K.R. Norman

It is obvious that a teacher must be able to communicate, if his teaching is to be

understand by his audience. This creates difficulties if he wishes to teach something new,
for which terms do not yet exist. He has the choice of coining new terms or of using old
terms in a new sense. Both categories must be clearly defined, or his listeners may not
understood the first and may understand the second in their old sense. An investigation
into the terminology used by the Buddha shows how he coped with this problem.
There have been those who thought that Buddhism was simply an offshoot of
Hinduism, while there are others who maintain that there is no trace of Hinduism in
Buddhism. The truth, as always, lies somewhere between these two extremes. What is
certainly true is that Buddhism owes much, especially in terminology, to Brahmanical
Hinduism and much of the Buddhas preaching would have been unintelligible to those
who had no knowledge of Brahmanical teaching. Although some of the technical terms of
Buddhism are exclusive to that religion, e.g. paisabhid, much Buddhist terminology
is, in form, identical with that of brahmanism. At the same time it must be recognized
that, although the Buddha took over some of the terminology of Brahmanical Hinduism,
he gave it a new Buddhist sense. The change of meaning is almost always a result of the
fact that the Brahmanical terms were used in a framework of ritualism, while the Buddha
invested them with a moral and ethical sense.
I should start by making several points clear. First, for convenience I speak in this
paper of the Buddhas usage, without, for the most part, distinguishing between his usage
and that of his followers. Second, I quote Buddhist terms in their Pli form. This implies
nothing whatsoever about the form in which these words were first used in Buddhism.
Third, I aim to do little more than list a few of these terms, some already well known, and
point out briefly how the Buddha adapted them for his own purposes. The full
consideration of some individual items would merit a whole paper to themselves, while
the subject as a whole would merit an entire book.
I propose to deal with the Buddhas use of Brahmanical terms in three categories:

193!

1. Terms and structures taken over by the Buddha:

devas: their existence was accepted by the Buddha but they were not allowed any
causal role in the universethey were merely super-human,1 and like all others in
sasra were subject to death and rebirth. The Buddha, in fact, increased their number,
since each of the world-systems of Buddhism had its complement of devas. He did,
however, allow for three categories of devassammuti-devas conventional devas, i.e.
kings, etc., since deva can mean both god and king, upapattidevas rebirth devas,
i.e. the gods of Hinduism, and visuddhi-devas purity devasthe last of which one
included Buddhas like himself.2
myths and fables: in the Brahmajlasutta the Buddha jokes about the way in
which Brahm thinks that he has created other beings, and he makes reference to the
creation myth in the Bhadrayaka Upaniad. In his comments in the Aggaasutta on
the way in which brahmans are born the Buddha satirizes the Puruaskta of the gveda.
I include these myths, etc., under my general heading of Brahmanical terms in a Buddhist
guise because the Buddha is using them in a different way from the brahmans. His aim is
not to present a cosmogony according to the brahmans, but to use the stories as a source
of mockery and a means of attack upon the brahmans, as Richard Gombrich has shown.3

2. Terms taken over by the Buddha but used with new senses:
aggi: as part of the ritual prescription of the Vedic tradition the brahman
(hitgni) has to keep three fires burning. The Buddha stated that there were three fires
which should not be served, but abandoned, viz. the fires of rga, dosa and moha.4
hra: in Brahmanical thought we find the idea that food is required to sustain the
existence of the inhabitants of other worlds or in the next life. The gods needed sacrifices
as their food, the pits needed offerings to continue their existence, and good deeds were
seen as a sort of nourishment for the next life.5 The Buddha, however, speaks of four
sorts of food,6 which are instruments of continuity, and imply the future process of
rebirth.
amata: in Brahmanical thought amta is the world of immortality, heaven,
eternity, or the nectar which confers immortality, produced at the churning of

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
1

K.R. Norman, The Buddhas View of Devas, Beitrge zur Indienforschung: Ernst Waldschmidt
zum 80, Geburtstag gewidmet, Berlin, 197, 329.
2
K.R. Norman, Devas and Adhidevas in Buddhism, JPTS, IX, 1981, 154.
3
R. Gombrich, Recovering the Buddhas Message, T. Skorupski, The Buddhist Forum, vol. 1, 1990,
13 ff., and 14 n. 25.
4
R. Gombrich, op. cit., 17
5
S. Collins, Selfless Persons, Cambridge, 1982, 20810.
6
cattro hr: kabalikro hro oriko v sukhumo v, phasso dutiyo, mano-sacetan tatiy,
via catuttha, D III 228, 35 (solid (physical) food, sense-impressions, mental volitions,
consciousness).

194!

the ocean. The Buddha, however, uses the word as an epithet of nibbna, which is
described as the amata pada.7 This is not, however, the immortal place, but the place
(or state) where there is no death. There is no death in that state, because there is no birth
there, and therefore no old age leading to death.
aa, ia: in the Pli commentaries we find the word anaa explained as without
defilement.8 I would suggest that aa and ia are to be derived < a debt, and that
this was a Brahmanical term taken over by the Buddha, and interpreted in a Buddhist way
when its Brahmanical meaning was forgotten. The brahmans three debts were the study
of the Vedas, the begetting of sons, and the offering of sacrifices. One who became an
ascetic when he had paid his debts would be anaa, while an ascetic who had not
fulfilled the proper conditions would be sa. The requirements would be meaningless to
a Buddhist, who would therefore interpret aa in a general sense as defilement.
brahman:9 there seems to be no occurrence in Pli of the uncompounded neuter
word brahma in the sense of the Upaniadic brahman, but the word brahma is used in
compounds apparently in the sense of excellent, perfect.
brahma-cariy: in its basic Brahmanical sense this means the practice of a
brhmaa, i.e. to live a celibate life, learning the Vedas. The Buddha used the phrase in
the more general sense of to live a holy, celibate (or in the case of married couples, a
chaste and moral) life.
brahma-patha: in the Upaniads this means the way to brahman or Brahm.
The Buddha used in it the sense of the way to the best, i.e. nibbna, and it is explained as
being the same as brahma-vihra.10
brahma-vihra: It is possible that this was in origin a Brahmanical term.11 It
would literally mean dwelling in or with brahman or Brahm, and it perhaps shows a
trace of its original meaning in the Tevijja-sutta12 where the Buddha speaks to young
brahmans who were disputing the correct way to obtain brahma-sahavyat. In the context
this would seem to mean union with brahman, but the

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
7

If amata pada is a split compound for amata-pada, then it might well be analysed in a
brahmanical sense as a tatpurua compound the place of the immortals, i.e. of the immortal gods.
8
nikkilesa-vasena, Th-a III 41, 17 (ad Th 789); sabba-kilesna khatt, Th-a III 62, 36 (ad Th
882); kilesa-ia pahya ana, Th-a 9, 2 (ad Th 2); ana, niddos apagata-kiles, Th-a 107,
3233 (ad Th 110); kma-cchanddi-ipagamena anao, Th-a 245, 2 (ad Th 364).
9
For the occurrences of brahman in the Pli canon, see K. Bhattacharya, Brahman in the Pali Canon
and in the Pali Commentaries, Amal Praj: Aspects of Buddhist Studies, Delhi, 1989, 91102.
10
iriyamna Brahma-pathe ti catubbidhe pi brahma-vihra-pathe, brahme va sehe phalasampatti-pathe sampajjana-vasena pavattamna. Th-a, III 9, 911 (ad Th 689).
11
E.J. Thomas, The Life of Buddha, (third ed.), London, 1949, p. 126.
12
D III 23553.

195!

Buddha, perhaps jokingly, interprets it as meaning a state of union with the god Brahm.
He explains that someone who practises the four types of concentration13 called brahmavihra is reborn as a Brahm in the Brahma-world.14 It is to be noted that this means only
being born in the same heaven as Mah Brahm, not union with the Upaniadic brahman.
It is noteworthy that what we might suppose to be the ways to gain brahma-vihra
dwelling in brahman are in fact given the name brahma-vihra by the Buddha, whereas
the four means are appropriately called brahma-patha.
brhmaa: in Brahmanical Hinduism a brahman (< bh- to be strong) was a
brahman by birth, and was a kinsman of Brahm. This idea was known to the Buddha,15
but by adopting a different etymology (< bh- to destroy), he was able to justify his
view that a brahman was one who had destroyed evil.16 The Buddha points out that a
brahman does not become a brahman by birth, but by his actions.17 He gave a revised
version of the theory that a brahman was only a true brahman if seven generations before
him were pure-born brahmans,18 if he knew the Vedas, if he was handsome and of
brahma-colour and brahma-splendour, if he was virtuous and if he was wise. He was able
to persuade the brahman Soadaa that only the last two of these five conditions really
matter, and it is virtue and wisdom which make a true brahman.19
deva-yna: in Brahmanical thought this is the way leading to the gods.20 The
Buddha uses yna in the sense of magga, the way followed by the Buddhas, etc., which
leads to nibbna.21
jhna: Sanskrit dhyna is religious thought, meditation. For the Buddha, jhna
applies to a very specific type of trance, and it is only rarely employed with a

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
13

mett, karu, mudit, and upekkh.

See the Tevijja-sutta (D I 23553). Cf. so cattro brahma-vihre bhvetv kyassa bhed para
mara Brahmalokpago ahosi, D II 196, 78.
15
So it is said of Agaikabhradvja: ito pubbe jti-mnena brhmaa-bhvato brhmana
samaya brahma-bandhu nma si, Th-a II 85, 45 (ad Th 221).
16
bhita-ppatt pana idni kho arahattdhigamena paramatthato brhmao amhi, Th-a II 85, 56
(ad Th 221). See Dhp 383423 (Brhmaavagga).
17
Not jtiy but kamman. See Sn 142, etc.
18
yato kho bho ubhato sujto hoti mtito ca pitito ca sasuddhagahaiko yva sattam
pitmahayug akkhitto anupakku ho jtivdena, ettvat kho brhmao hoti, Sn 115, 1316.
19
See the Soadaa-sutta (D I 11126).
20
E.M. Hare, Woven Cadences of Early Buddhism, London, 1945, 22 f.n., refers to the way of the
gods of the Vednta.
21
mahantehi buddhdhi pa ipannatt mahpatha, brahmalokasakhta devaloka ypetu
samatthatt devalokaynasakhta a hasampattiyna abhiruyha, Pj II 184, 2326 (ad Sn 139).
Cf. magg-ahagika-yna-yyin ti, ahagika-magga-sakhtena ariya-yyena nibbna-pura
yyin upagat, Th-a 257, 68 (ad Th 389).
14

196!

wider application. Absorption, rather than meditation, has been suggested as a more
appropriate translation.22
kamma: the word karman is used in a Brahmanical context to refer specifically to
the ritual act enjoined by Brahmanical ideology. The Buddha stated that he would
interpret act to refer to intention,23 with the result that there is a shift from ritual to
ethics. Whereas the performance of the ritual action of sacrifice gave an automatic result,
this development in the interpretation of kamma meant that the quality of the next life is
determined by the quality of the actions.24
khetta-jina: this word occurs in Suttanipta 52324. The commentary on that text
seems uncertain about its meaning, which on the face of it seems to mean conqueror of
the fields(s), and gives a double explanation based upon the verbs ci- and ji-, which
doubtless goes back to a dialect where both -c- and -j- became -y-.25 I have suggested
elsewhere that the second element of the compound is from -ja, not -jina. I see here a
connection with the word ketra-ja which occurs at Manu XII 12, where it is given
various explanations by the commentators, including the individual soul (jva).26 We
find the negative akkhetta, with a different development of -ja, at Ja IV 371,14*,
where it is explained as not knowing the (right) field for alms-giving (dnassa).
nhtaka: in its Brahmanical sense sntaka is used of a brahman who has carried
out the ceremonial bathing at the end of the brahma-crin stage of his life. The Buddha
rejected the efficacy of ritual bathing, and uses the term metaphorically of washing away
evil by means of the eight-fold path.27 Carrying the theory of ritual washing to its logical
conclusion, it is said that if water washed away sins, then fish, crocodiles, etc., would be
the purest of creatures and would all go to heaven.28
pua: the earliest meanings of puya seem to be auspicious, happy, beautiful,
good, but as part of the replacement of ritual by ethics the Buddha gave a new
29

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
22

See L.S. Cousins, Buddhist jhna: its Nature and Attainment According to the Pli Sources,
Religion, III, 2, 1973, 116.
23
cetanha bhikkhave kamma vadmi, A III 415, 7. See R. Gombrich, Notes on the Brahmanical
Background to Buddhist Ethics, Buddhist Studies in Honour of Hammalava Saddhatissa, Nugegoda,
1984, 91.
24
S. Collins, Selfless Persons, 5556.
25
tni vijeyya jetv abhibhavitv viceyya v aniccdibhvena vicinitv upaparikkhitv (Pj II 428, 27
29). The double explanation is repeated: etesa khettna vijitatt vicitatt v khettajino (Pj II 429,
6). See K.R. Norman, Notes on the Sutta-nipta, 106 (ad Sn 52324), referring to Monier-Williams,
s.v. ketra, and comparing Manu, XII, 12 ff.
26
See G. Bhler, The Laws of Manu, SBE XXV, 485, n. 12.
27
ahagika-magga-jalena suvikkhlita-kilesa-malatya paramatthato nahtako, Th-a II 85, 1213
(ad Th 221).
28
See Th 24142.
29
See M. Mayrhofer, Kurzgefasstes etymologisches Wrterbuch des Altindischen, II, s.v.

197!

ethical content to the word,30 and it could then be used in the context of kamma, in the
sense that merit could be acquired which would bring a good rebirth.
sla: in Sanskrit la means custom, habit, conduct, good conduct, moral
conduct. The Buddha used the word to mean the rules of behaviour: the five great rules
applicable to all Buddhists and also the rules of discipline of the monastic code. It has
been described as the monks successful role performance, or something like character in
Western society, which is built up by moral habit.31
sottiya: the Brahmanical sense of rotriya is acquainted with the Vedas,
knowing ruti. For the Buddha the connection with the verb to hear is retained,32 but
the connection with Brahmanical ruti disappears. The fact that the word sometimes
appears as sotthiya suggests that the Buddhist tradition believed that there was some
connection with sotthi < svasti.33
suddhi: in brahmanism, uddhi refers primarily to a ritual condition. The Buddha
made purity a strictly moral concept. The aim was purity of thought.34
tevijja: although the phrase tia vedna pragu (cf. Skt. vedapraga) is used
of a brahman and correctly understood by the Buddha,35 tisso vijj, originally the
knowledge of the three Vedas as applicable to a brahman, is explained in Buddhist terms
by the commentaries.36 Just as in Sanskrit traividya means one possessing the trividy, so
one who possesses tisso vijj is called tevijja.
uposatha: in Brahmanical terminology, upavasatha was a fast day, the day of
preparation for the Soma sacrifice. In Buddhism the fast day itself became the day of
reciting the ptimokkha (for monks and nuns) and listening to recitations (for laymen),
i.e. it was no longer part of a ritual for purity, but became the occasion for a confession of
moral and ethical transgressions.
veda: veda is used in Buddhism in its general sense of knowledge37 rather than
as the title of Brahmanical texts. The word vedagu, which in its Brahmanical sense meant
one who had gained competence in the Vedas, was interpreted as one who had gained
knowledge of release from sasra.38
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
30

R. Gombrich, Notes on the Brahmanical Background to Buddhist Ethics, 101.

R. Gombrich, op. cit., 100.
32
sutv sabbadhamma abhiya loke sottiyo, Sn 534; sutavatt sottiyo ti hu, Pj II 432, 27.
33
suvimutta-bhav-assda-dhamma-jjhnena paramatthato sotthiyo, Th-a II 85 1415 (ad Th 221).
34
S. Collins, op. cit., 112.
35
e.g. brhmana vijjsu nipphatti gato tia vedna prag, Th-a III 169, 2324 (ad Th
1171).
36
pubbe-nivsa-a, dibba-cakkhu-a, sava-kkhaya-a, tisso vijj, Th-a I 85, 2021 (ad
Th 24).
37
e.g. veda-sampanno ti, a-sampanno, Tha III 169, 20 (ad Th 1170).
38
veda-sakhtena maggaena sasra-mahoghassa vedassa catu-saccassa ca pra gatatt
adhigatatt tatt paramatthato vedag, Th-a II 85, 1719 (ad Th 221).
31

198!

yogakkhema: In the gveda yogakema means the security or safe possession of

what has been acquired, the safe keeping of property, welfare, prosperity, substance,
livelihood. In the Aitareya Brhmaa it is a dvandva compound, as can be seen from the
fact that it appears in two forms (yogakema and kemayoga),39 rest and exertion. The
Buddha took it as a tatpurua compound, used first in an agricultural context, of the ox
moving on towards rest from work,40 where the idea of freedom, release from the yoke
(of the plough) was probably implied. The idea of welfare was then applied to nibbna,
of which the word is used as an epithet. This was then interpreted as freedom from
bondage,41 i.e. the things which tie creatures to sasra.

3. Terms referred to but rejected

There are ideas referred to by the Buddha but rejected, but in such a way that the grounds
for his objection could only be understood by those who knew the Brahmanical
terminology:
att: The Buddhas rejection of the existence of the att, i.e. his view that
everything was anatt, was based upon the Brahmanical belief that the tman was nitya
and sukha. Hence the Buddha could refute this by pointing out that the world was in fact
anicca and dukkha.42
Besides the convenience of taking over terms which were already known to his
audience, albeit in a different sense, the Buddha possibly had other reasons for acting in
this way. In part it may have been due to his desire to show that Brahmanical Hinduism
was wrong in its basis tenets: a Brahmanical brhmaa was not as good as a Buddhist
brhmaa, Brahmanical uddhi was inferior to Buddhist suddhi, etc. If a teacher takes
over his rivals terms and repeat them often enough in his own meaning, he gives the
impression that he is using them in the correct sense, and the original owners are wrong
in their usage.
It must be made clear that we cannot prove that the Buddha (or the Buddhists)
was the first to make use of these Brahmanical terms in a new sense, since there is a
possibility that such a use of some of these terms was also common to other
contemporary religions. Some of the terminology of Buddhism is held in common with
Jainism,43 e.g. buddha, pratyeka-buddha, jina, nirva, tathgata, bhvan, dhuta, yoga,
kevalin, srava, karman, gati, moka, ramaa, pravrajy, pravrajita, tapas, i, t(d)in,
phsu(ya), and also certain epithets of the Buddha and the

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
39

See Monier-Williams, Sanskrit-English Dictionary, s.vv.

Sn 79, quoted by Collins, op. cit., 221.
41
ettha yogehi khematt yogakkheman ti nibbna vuccati, Pj II 150, 23 (ad Sn 79).
42
See K.R. Norman, A Note on Att in the Alagaddpamasutta, in SIP, Ahmedabad, 1981, 22, and
R. Gombrich, Recovering the Buddhas Message, 14.
43
To overcome problems arising from dialect differences, I quote most of these in their Sanskrit form.
40

199!

Jina.44 It is possible therefore that the use of Brahmanical terms in a non-Brahmanical

sense was taken from the general fund of vocabulary of ramaical45 religions.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
44

e.g. vs-candana-kalpa, sama-lo a-kacana, vyvtacchadman.

If the compound brhmaa-ramaa covers the whole range of Indian religion, then it is
appropriate to use the word ramaa for all those members of religious sects who were not brahmans.

45

200!

REFERENCES

Acharya, P.K., Architecture of Mnasara, Oxford, l933.

Adilcaram, E.W., Early History of Buddhism in Ceylon, Migoda, D.S. Puswella, 1946.
Ahlstrom, G.W., Royal Administration and National Religion in Ancient Palestine, Leiden, 1982.
Akao Eikei, Hz ni micru smoku jbutsu ni tsuite, IBK, 32.2, 1984, 404-11.
Allchin, F.R., A Cruciform Reliquary from Shaikhan Dheri, Aspects of Indian Art, Leiden, E.J.
Brill, 1972, 1526.
Allen, M.R., Buddhism without Monks: the Vajrayana Religion of the Newars of the
Kathmandu Valley, South Asia, 2, 1973, 114.
Allen, M.R., Girls Pre-puberty Rites among the Newars of the Kathmandu Valley in M. Allen
& S.N. Mukherjee, eds., Women in India and Nepal, Australian National University
Monographs on South Asia, 8, Canberra, 1982.
Allen, M.R., The Cult of the Kumari, Virgin Worship in Nepal, Tribhuvan University,
Kathmandu, INAS, l975, reissued Kathmandu, Madhab Lal Maharjan, Himalayan
Booksellers, 1987.
Aung Thwin, M., The Role of Sasana Reform in Burmese History: Economic Dimensions of a
Religious Purification, JAS, 38, 1979, 67188.
Ayscough, F., The Chinese Cult of Cheng-huang lao-yeh, JNCBRAS, 55, 1924, 13l155.
Banerjee, S.C., A Brief History of Tantric Literature, Calcutta, Naya Prokash, 1988.
Bapat, P.V. & R.D. Vadekar, Buddhaghosa: Aaslin, Poona, 1942.
Bapat, P.V., Han-chien-pl-pcrsha, Bhandakar Oriental Series, 10, Poona, 1970.
Bareau, A., Trois trails sur les scctes bouddhiques attribues Vasumitra, Bhavya et
Vinitadeva, Ie Partie: JA, 242, 1954, 229266; IIe Partie: JA, 244, 1956, 167200.

201

Bareau, A., Les premiers conciles bouddhiqucs, Paris, Presses Universit.aires de France, l955.
Bareau, A., Les sectes bouddhiques du Petit Vhicule, Saigon, EFEO, l955.
Bareau, A., Les controverses relatives la nature de larhant dans le Bouddrusme ancien, IIJ, l,
1957, 24150.
Bareau, A., Recherches sur la biographie du Buddha dans les sutrapiaka et vinayapaka anciens,
vol. I, Paris, 1963.
Bareau, A, Etude du Buddhism I in LAnnuaire du College de France, 73e Anne, Paris, 1972
73.
Barrett, T.H., Towards a Date for the Chin-so liu-chu yin, BSOAS, 52.2, 1990, 29294.
Basham, A.L ., The Wonder that was India, London, 1954.
Bechert, H., ber Singhalesisches im Plikanon, WZKSO, I, 1957, 7175.
Bechert, H., Aokas schismenedikt und der Begrift Saghabheda, WZKSO, 5, 1961, 1852.
Bechert, H., Theravda Buddhist Sangha: Some General Observations on Historical and Political
Factors in its Development, JAS, 29, 1970, 76178.
Bechert, H., The Beginnings of Buddhist Historiography: Mahvasa and Political Thinking,
in Bardwell L. Smith, ed., Religion and Legitimation of Power in Ceylon, Chambersburg,
PA, Anima, l978, 112.
Bechert, H., The Date of the Buddha Reconsidered, IT, 10, 1982, 2936.
Bechert, H., The Importance of Aokas So-called Schism Edict, Indological and Buddhist
Studies: Volume in Honour of Professor J. W. de Jong, Canberra, Faculty of Asian
Studies, 1982, 6168.
Bechert, H., A Remark on the Problem of the Date of Mahvra, IT, 11, 1983, 18790.
Bechert, H., Zur Schulgehrigkeit von Werken der Hnayna-Literatur, Gttingen, 1984.
Bechert, H., Die Lebenszeit des Buddhadas lteste fest stehende Datum der indischen
Geschiechte?, Gttingen, 1985.
Bechert, H., Remarks on the Date of the Historical Buddha, Buddhist Studies, 17, 1988, 97
117.
Bechert, H., & J. Hartmann, Observations on the Reform of Buddhism in Nepal, JNRC, 8,
1988, 130.
Bell, C., Religion and Chinese Culture: Toward an Assessment of Popular Religion, HR, 29.1,
August, 1989, 3557.

202

Berger, P.L., A Rumor of Angels, Garden City, New York, 1970.

Beyer, S., The Cult of Tr, Berkeley, University of California Press, 1973.
Bhattacharya, K., Brahman in the Pali Canon and in the Pali Commentaries, Amal Praj:
Aspects of Buddhist Studies, Delhi, 1989, 1531.
Bischoff, F.A & C. Hartmann, Padmasambhavas Invention of the Phur-ba, in Etudes tibetaines
dedies a Marcelle Lalou, Paris, 1971.
Bizot, F., Le figuier cinq branches, Paris, EFEO, 1976.
Bizot, F., Les traditions de Ia pabbajja en Asie du Sud-Est, Gttingen, Vandenhoeck & Ruprecht,
1988.
Boltz, J., A Survey of Taoist Literature: Tenth to Seventeenth Centuries, Berkeley, 1987.
Boyce, Mary, A History of Zoroastrianism, Leiden, E.J. Brill, 1975.
Broido, M., Padma dKar-po on the Two Satyas, JIABS, 8.2, 1985.
Bronkhorst, J., The Two Traditions of Meditation in Ancient India, Stuttgart, Steiner-VerlagWiesbaden, 1986.
Brunner-Lachaux, H., Somaambhupaddhati, Part 3, Pondichry, 1977.
Cantwell, C., An Ethnographic Account of the Religious Practice in a Tibetan Buddhist Refugee
Monastery in Northern India, unpublished Ph D thesis, University of Kent, Canterbury,
1990.
Carrithers, M., Buddhists without History, CIS, N.S. 21, 1987, 1658.
Chang, Garma C.C., The Buddhist Teaching of Totality: The Philosophy of Hwa Yen Buddhism,
London, Allen and Unwin, 1972.
Chang, Garma C.C., gen. ed., A Treasury of Mahayana Sutra, Selections from the Maharatnakuta
Sutra, tr. from the Chinese by Buddhist Association of US, IASWR Series, 1983.
Chamoux, F., Le roi Magas, RH, 216, 1956.
Chao lun, The Treatises of Seng chao, tr. into English with bibliography, introductory notes and
appendices by Walter Liebenthal, Hong Kong, Hong Kong University Press, 1968.
Chau, T.T., Le personnalisme du Bouddhisme ancien, ICO, VII, 1973.
Chau, T.T., Les sectes personnalistes (Pudgalavdin) du Bouddhisme ancien, thesis for the
Doctorat dEtat s Lettres, Sorbonne, Paris, 1978.
Chau, T.T., The Literature of the Pudgalavdins, JIABS, 7.1, 1984, 716.
Chau, T.T., Les rponses des Pudgalavdin aux critiques des coles bouddhiques, JIABS, 10.1,
1987, 3353.
Chen, K., Buddhism in China, Princeton, Princeton University Press, 1964.
Chiang Hsing-y, Chung-kuo yin-shih y Chung-kuo wen-hua, Shanghai, 1947.

203

Chin En-fu, ed., Kuei-ku-tzu, Beijing, 1985.

Ching, Julia, To Acquire Wisdom: The Way of Wang Yang-ming, Ithaca, Columbia University
Press, 1976.
Cleary, T., Entry into the Realm of Reality: I Text. Boston & Shaftesbury, 1987.
Clothey, F. W., The Many Faces of Muruka, the History and Meaning of a South Indian God,
The Hague, Mouton, 1978.
Clothey, F.W., Rhythm and Intent, Madras, 1983.
Cohen, A.P., Coercing the Rain Deities in Ancient China, HR, 17.3, 4, February-May, 1978,
244265.
Cohn, B., History and Anthropology: The State of Play, in An Anthropologist among the
Historians and Other Essays, Oxford & Delhi, Oxford University Press, 1987, 1849.
Collins, S., Selfless Persons, Cambridge, 1982.
Collins, S., On the very Idea of the Pali Canon, forthcoming in JPTS.
Cook, Francis H., Hua Yen Buddhism: The Jewel Net of Indra, Pennsylvania State University
Press, 1977.
Coomaraswamy, A.K., Symbolism of the Dome, IHQ, 14, 1938.
Cousins, L.S., Buddhist jhna: its Nature and Attainment According to the Pli Sources,
Religion, III, 2, 1973, 11531.
Cousins, L.S., The Pahna and the Development of the Theravdin Abhidhamma, JPTS, 9,
1981, 2246.
Dagens, B., tr., Mayamata, trait sanskrit darchitecture, Pondichry, 1976.
Dalai Lama, H.H. Tenzin Gyatso, Kindness, Clarity and Insight, tr. & ed. by J. Hopkins, co-ed.
by E. Napper, Ithaca, NY, Snow Lion Publications, 1984
Demiville, P., Les versions chinoises du Milindapaha, BEFEO, 24, 1924, 1258.
Demiville, P., propos du concile de Vaili, TP, 40, 1951, 239296.
Demiville, P., Lorigine des sectes bouddhiques daprs Paramrtha, MCB, I, 1932, l5-64;
repr. in Choix dtudes bouddhiques, Leiden, 1973.
Eggermont, P.H.L., New Notes on Asoka and his Successors, II, Persica, 4,1969.
DeJong, J.W., A Brief History of Buddhist Studies in Europe and America, Varanasi, BharatBharati, 1976.
Dorje, G., The Guhyagarbha-tantra and its Commentary = The Guhyagarbhatantra and its
Fourteenth
Century
Tibetan
Commentary
Phyogs-

204

bcu mun-sel, unpublished Ph D thesis, London, SOAS, University of London, 1987.

Dowman, K., Skydancer, London, 1984.
bDud-joms Rin-po-che, The Nyingma School of Tibetan Buddhism: Its Fundamentals and
History, [including the bsTan-pai mam-gzhag and rNying-mai chos-byung], tr. & ed.
by G. Dorje with the collaboration of M. Kapstein, 2 vols., London & Boston, Wisdom
Publications, forthcoming.
Dumont, L., Homo Hierarchicus: The Caste System and its Implications (complete revised
English edition), Chicago & London, University of Chicago Press, 1980.
Eliade, M., Cosmos and History, the Myth of the Eternal Return, New York, 1959.
Evers, H., Monks, Priests and Peasants: A Study of Buddhism and Social Structure in Central
Ceylon, Leiden, E.J. Brill, 1972.
Filliozat, J., Le Kumratantra de Rvaa, Paris, 1937.
Filliozat, J., Les deux Aoka et les conciles bouddhiques, JA, 1948, 189195.
Filliozat. J., La date de lavnement de Candragupta roi du Magadha (313 avant J.-C.), JS,
1978, 175184.
Finazzo, G., The Notion of Tao in Lao Tzu and Chuang Tzu, Taipei, 1968.
Fodor, J.A., Representations: Philosophical Essays on the Foundations of Cognitive Science,
Brighton, Harvester, 1981.
Franz, G., Von Gandhara bis Pagan, Graz, Akademische Druck, 1979.
Frauwallner, E., Die buddhistischen Konzile, ZDMG, 102, 1952, 240261.
Frauwallner, E., The Earliest Vinaya and the Beginnings of Buddhist Literature, Serie Orientale
Roma, 8, Rome, 1956.
Frauwallner, E., Abhidharma-Studien IIIDer Abhisamayavda & Der Abhidharma der
anderen Schulen, WZKS, 15, 1971, 69121.
Frauwallner, E., Abhidharma-Studien IVDer Abhidharma der anderen Schulen
(Fortsetzung), WZKS, 16, 1972, 95152.
Fukunaga Mitsuji, Issai shuj to smoku doseki, Bukky shigaku, 23.2, March, 1981, 103118.
Fung Yu-lan, tr. D. Bodde, A History of Chinese Philosophy, 2 vols., Princeton, Princeton
University Press, 195253.
Gaborieau, M., Le Nepal et ses populations, Paris, Editions Complexes & Presses Universitaires
de France, 1978.
Garua Pra, ed. by J.L. Shastri, tr. by a Board of Scholars, Delhi, 1978.

205

Gellner, D.N. Language, Caste, Religion and Territory: Newar Identity Ancient and Modern,
EJS, 27, 1986, 10248.
Gellner, D.N., Monk, Householder and Priest Newar Buddhism and its Hierarchy of Ritual,
unpublished D.Phil. thesis, Oxford University, 1987.
Gellner, D.N., The Newar Buddhist Monastery: An Anthropological and Historical Typology,
in N. Gutschow & A. Michaels, eds., The Heritage of the Kathmandu Valley,
Proceedings of an International Conference held in Leubeck 1985, Nepalica 4, Sankt
Augustin, VGH Wissenschaftsverlag, 1987.
Gellner, D.N., Buddhism and Hinduism in the Nepal Valley, in S. Sutherland, L. Houlden, P.
Clarke & F. Hardy, eds., The Worlds Religions, London, Routledge, Boston, G.K. Hall,
1988.
Gellner, D.N., Monastic Initiation in Newar Buddhism, in R.F. Gombrich, ed., Indian Ritual
and its Exegesis, Delhi, OUP, Oxford University Papers on India, vol. 2, Part l, 1988.
Gellner, D.N., Hodgsons Blind Alley? On the so-called Schools of Nepalese Buddhism,
JIABS, 12.1, 1989.
Gellner, D.N., Monkhood and Priesthood in Newar Buddhism, in V. Bouillier & G. Toffin,
eds., Prtrise, Pouvoirs et Autorit en Himalaya (Pururtha 12, 16592), Paris, Editions
de lEHESS, 1989.
Gethin, R., The Path to Awakening, A Study of the Thirty Seven Bodhipakkhiyadhamm in the
Nikyas and Abhidhamma, Ph D thesis, University of Manchester, 1987.
Gimello, Robert M. & P.N. Gregory, Studies in Chan and Hua Yen, Kuroda Institute Studies in
East Asian Buddhism No. l, Honolulu, University of Hawaii Press, 1983.
Gombrich, R.F., Precept and Practice: Traditional Buddhism in the Rural Highlands of Ceylon,
Oxford, Clarendon Press, 1971.
Gombrich, R.F., Buddhist Karma and Social Control, CSSH, 17, 1975, 21220.
Gombrich, R.F., Indian Cosmology, in Carmen Blacker & Michael Loewe, eds., Ancient
Cosmologies, London, Allen & Anwin, 1975.
Gombrich, R.F., Notes on the Brahmanical Background to Buddhist Ethics, in Gatare
Dhammapala et al., eds, Buddhist Studies in honour of Hammalava Saddhtissa,
Nugegoda, University of Sri Jayewerdenepura, 1984, 91102.
Gombrich, R.F., The History of Early Buddhism: Major Advances since l950, Indological
Studies and South Asian Bibliographya Conference, Calcutta, 1986.
Gombrich, R.F., Theravda Buddhism, London, Routledge & Kegan paul, 1988.

206

Gombrich, R.F., How the Mahayana Began, in T. Skorupski, ed., The Buddhist Forum, vol. I,
London, SOAS, 1990, 2130.
Gombrich, R.F., Recovering the Buddhas Message, in T. Skorupski, ed., The Buddhist Forum,
vol. I, London, SOAS, 1990, 520.
Gonda, J., The Savayajas, Amsterdam, 1965.
Gonda, J., Aspects of Early Viuism, Delhi, 1969.
Gonda, J., Vedic Ritual: the Non-solemn Rites, Leiden, E.J. Brill, 1980.
Gos Lo-ts-ba, Deb-ther sngon-po, tr. by G.N. Roerich, The Blue Annals, 2nd. ed. Delhi, Motilal
Banarsidass, 1976; typeset edition, 2 vols., Chengdu, Sichuan Minorities Press, 1984.
Goudriaan, T., My, Divine and Human, Delhi, Motilal Banarsidass, 1978.
Goudriaan, T. & Gupta, S., Hindu Tantric and kta Literature, Wiesbaden, 1981.
Graham, A.C., Chuang-tzu: The Seven Inner Chapters and Other Writings from the Book of
Chuang-tzu, London, 1981.
Greenwald, S., Monkhood versus Priesthood in Newar Buddhism, in C. von Frer-Haimendorf,
ed., The Anthropology of Nepal, Warminster, Aris and Phillips, 1974.
Greenwald, S. Buddhist Brahmans, EJS, 15, 1974, 101123, repr. as The Role of the Priest in
Newar Society, J.F. Fisher, ed., Himalayan Anthropology: The Indo-Tibetan Interface,
The Hague & Paris, Mouton, 1978.
Gregory, P .N., ed., Sudden and Gradual: Approaches to Enlightenment in Chinese Thought,
Honolulu, University of Hawaii, 1987.
Guenther, H.V., The Royal Song of Saraha: A Study in the History of Buddhist Thought, Berkeley
& London, Shambhala, 1968.
Guenther, H.V., Jewel Ornament of Liberation, London, Rider, 1959, repr. 1970.
Guenther, H.V., Matrix of Mystery, Boulder & London, Shambhala, 1984.
Hare, E.M., Woven Cadences of Early Buddhism, London, 1945.
Hakeda, Yoshito S., tr., Awakening of Faith Attributed to Avaghosha, New York, Columbia
University Press, 1967.
Harrison, P., Who Gets to Ride in the Great Vehicle? Self-image and Identity Among the
Followers of the Early Mahyna, JIABS, 10, 1987, 6789.
LHemault, F., LIconographie de Subrahmaya au Tamilnad, Pondichry, 1978.
Herrmann, G., The Iranian Revival, Oxford, Elsevier, 1977.
Hibi Sensh, Tdai Tendaigaku josetsu, Tokyo, 1966.
Hofinger, M., tude sur le concile de Vail, Bibliotheque du Muson, vol. 20, Louvain, 1946.

207

Holmes, K. The Changeless Nature, Kagyu Samye Ling, 1979.

Hopkins, J., Meditation on Emptiness, Part 1, Ann Arbor, Michigan, University Microfilms Inc.,
1973; London, Wisdom Publications, 1983.
Hopkins, J., The Yoga of Tibet, London, George Allen & Unwin, 1981.
Hooykaas, C. Santi, a Ritualistic Object from Bali, AM, N.S. 11, 1964.
Hsiang-chih, Y-ti chi-sheng, Taipei, 1962 (rep. of 1860 ed.).
Hsiao Tung, comp., Wen-hsan 21, Hong Kong, 1960.
Hs Chia-jui, Ta-li ku-tai wen-hua-shih kao, Hong Kong, 1979.
Hummell, S., Der Lamaistische Ritualdolch Phurbu und die altorientalische Nagelmenschen,
Bern, 1952.
Huntington, J., The Phur-ba-Tibetan Ritual Daggers, Ascona, 1975.
Hurvitz, L., Chih-i, 538597: An Introduction to the Life and Ideas of a Chinese Buddhist Monk,
MCB, 12, 1962.
Hymes, R.P., Statesmen and Gentlemen, Cambridge, 1986.
Irwin, J.C., The Stpa and the Cosmic Axis in Dallapiccola et al., The Stpa, its Religious,
Historical and Archaeological Significance, Heidelberg, 1980.
Irwin, J.C., The Stpa and the Cosmic Axis-Ypa/Yai, in Acarya-Vandana, ed., D.R.
Bhandarkar Birth Centenary Volume, Calcutta, 1984 .
Irwin, J.C., Asokan Pillars: a Reassessment of the Evidence BM, vols. 115118, London, 1973
76.
Ishii Masako, Ofuchi Ninji, comp., Rikuch, T, S no kobunken shoin Dky tenseki mokuroku,
sakuin, Tokyo, 1988.
Iyanaga Nobumi, Rcits de la soumission de Mahevara par Trailokyavijaya, daprs les sources
chinoises et japonaises, in M. Strickmann, ed., Tantric and Taoist Studies in Honour of
R.A. Stein, vol. 3, (=MCB, 22), Brussels, 1985, 633745.
Johnson, D., The City-god Cults of Tang and Sung China, HJAS, 45.2, December, 1985, 363
457.
Johnston, E.H., ed., The Ratnagotravibhga Mahynottaratantra stra, with Introduction,
Patna, Bihar Research Society, 1950.
Johnston, W., The Still Point: Reflections on Zen and Christian Mysticism, New York, Evanston,
San Francisco, London, 1971.
Jolly, J., tr. by C.G. Kashikar, Indian Medicine, Poona, 1951.
Jam-mgon Kong-sprul, Shes-bya kun-khyab mdzod, atapiaka Series, vol. 80, New Delhi, 1970,
typeset edition in 3 vols., Beijing, Minorities Press,1982.

208

Kanai Noriyuki, Sdai no sonsha to Bukky, Bukky shigaku kenky, 18.2, 1976, 3156.
Kangle, R.P., The Kauilya Arthastra, Bombay, 1972.
Kamata Shigeo, Chgoku Bukky shisshi kenky, Tokyo, 1969.
Kamata Shigeo, Shmitsu kygaku no shisshiteki kenky, Tokyo, 1975.
Kamata Shigeo, Chgoku no Bukky girei, Tokyo, 1986.
Kamio Katsuharu, Kittan bunkashi k, Dairen, 1937.
Karmay S.G., Origin and Early Development of the Tibetan Religious Traditions of the Great
Perfection, Ph D thesis, SOAS, London University, 1986.
Kazi Dawa Samdup, ed. A. Avalon, The rcakrasabhara Tantra, London & Calcutta, 1919.
Keith, A.B., The Veda of the Black Yajus School, Harvard Oriental Series, vols. 1819,
Cambridge, Mass., 1914.
Keyes, C., The Golden Peninsula, New York, Macmillan, 1977.
Keyes, C., Merit-transference in the Kammic Theory of Popular Theravda Buddhism, in
Charles Keyes & E. Valentine Daniel, eds., Karma: An Anthropological Inquiry,
Berkeley, University of California Press, 1983, 26186.
Kloppenberg, R. Theravda Buddhism in Nepal, Kailash, 5.4, 1977, 30121.
Khn, L., Seven Steps to the Tao, MSMS, 20, Nettetal, 1987.
Klver B. & H. kya, Documents from the Rudravara-Mahvihra, Pan 1, Sales and
Mortgages, Nepalica 1, Sankt Augustin, VGH Wissenschaftsverlag, 1985.
Koseki, A.K., Prajpramit and the Buddhahood of the Non-sentient World: The San-lun
Assimilation of Buddha-nature and Middle Path Doctrine, JIABS, 3.1, 1980, 1633.
Kriysamuccaya, atapiaka Series, vol. 237, New Delhi, 1977.
Kubota Ryon, Shina Ju-D-Butsu kshshi, Tokyo, 1943.
Kuiper, F.BJ., Cosmogony and Conception: a query, HR, 10, 1970.
Kuiper, F.B.J., The Basic Concept of Vedic Religion, HR, 15, l975.
Kuiper, F.B.J., Ancient Indian Cosmogony, New Delhi, 1983.
Kvaerne, P., Tibet, the Rise and Fall of a Monastic Tradition, in H. Bechert & R. Gombrich,
eds., The World of Buddhism: Buddhist Monks and Nuns in Society and Culture, London,
Thames & Hudson, 1984.
LaFleur, W., Saigy and the Buddhist Value of Nature, HR, 13.2, 1973, 93128; & 13.3, 1974,
227248.

209

Lamotte, E., Le Trait de la Grande Vertu de Sagesse de Ngrjuna, Louvain, 1944.

Lamotte, E., Buddhist Controversy over the Five Propositions, IHQ, l956 148162.
Lamotte, E., Histoire du Bouddhisme indiendes origines lre aka, Louvain, 1958; tr. by
Sara Webb-Boin, History of Indian Buddhismfrom the Origins to the aka Era,
Louvain-La-Neuve, Institut Orientaliste, 1988
Lati Rimpochay, tr. E. Napper, Mind in Tibetan Buddhism, London, Rider, 1980.
La Valle Poussin, L. de, The Five Points of Mahdeva and the Kathvatthu, JRAS, 1910,
41323.
La Valle Poussin, L. de, tr., LAbhidharmakosa de Vasubandhu, 6 vols., Paris
l923-3l; repr. Bruxelles, l97l.
La Vallee Poussin, L. de, La controverse du temps et du pudgala dans le Vijnakya, EA
(publies loccasion du vingt-cinquime anniversaire de lcole Franaise dExtrmeOrient), Paris, 1925, vol. 1, 34375.
La Valle Poussin, L. de, Councils and Synods: Buddhist, ERE. vol. 4, 17985.
Law, B.C., ed., Buddhist Studies, Calcutta, 1931.
Law, B.C., A Manual of Buddhist Historical Traditions, Calcutta, University of Calcutta, 1963.
Lessing, F. & A. Wayman, A., Mkhas Grub Rjes Fundamentals of the Buddhist Tantras, The
Hague/Paris: Mouton, 1968; repr. as Introduction to the Buddhist Tantric Systems, Delhi,
Motilal Banarsidass, 1983.
Levi, J ., Les fonctionnaires divins, Paris, 1989.
Lvi, S., Le Npal, tude historique dun royaume hindou, 3 vols., Paris, Leroux, 19051908,
reissued Kathmandu and Paris, Raj de Condappa and Toit du Monde and Editions
Errance, 1986.
Lewis, N., Life in Egypt under Roman Rule, Oxford, 1983.
Lewis, T., The Tuladhars of Kathmandu: A Study of Buddhist Tradition in a Newar Merchant
Community, Ph D thesis, Columbia University, 1984.
Lieberman, V.B., The Political Significance of Religious Wealth in Burmese History: Some
Further Thoughts, JAS, 39, 1980, 75369; Michael Aung Thwin, A Reply to
Lieberman, JAS, 40, 1980, 8790.
Lienhard, S., Problmes du Syncrtisme Religieux au Npal, BEFEO, 65, 1978, 239270; also
published in German in H. Bechert, ed., Buddhism in Ceylon and Studies on Religious
Syncretism in Buddhist Countries, Gttingen, Vandenhoeck & Ruprecht, 1978.

210

Lienhard, S., Nepal: The Survival of Indian Buddhism in a Himalayan Kingdom in H. Bcchert
& R.F. Gombrich eds., The World of Buddhism: Buddhist Monks and Nuns in Society and
Culture, London, Thames & Hudson, 1984.
Lienhard, S., Buddhistisches Gemeindeleben in Nepal, in H. Bechert. ed., Zur
Schulzugehrigkeit von Werken der Hnyna-Literatur, Gttingen, Vandenhoek &
Ruprecht, 1985.
Lienhard, S., Dreimal Unreinheit: Riten und Gebruchen der Nevars bei Geburt, Menstruation
und Tod, in B. Klver & S. Leinhard, eds., 1986, Formen kulturellen Wandels und
andere Beitrge zur Erforschung des Himalaya, Nepalica 2, Sankt Augustin, VGH
Wissenschaftsverlag,1986.
Li Pai-yao, Pei-Chi shu 32, Beijing, 1972.
Lo-chen Dharmar, gSang-snying spyi-don gsang-bdag zhal-lung, rNying-ma bka-ma rgyas-pa,
vols. 3334, ed. by bDud-joms Rin-po-che, 45 vols., Kalimpong, 1982.
Locke, J., Rato Matsyendranath of Patan and Bungamati, INAS Historical Series 5, Kathmandu,
1973.
Locke, J ., Karunamaya: The Cult of Avalokitesvara-Matsyendranath in the Valley of Nepal,
Kathmandu, Sahayogi, 1980.
Locke, J ., The Buddhist Monasteries of Nepal: A Survey of the Bhs and Bahs of the
Kathmandu Valley, Kathmandu, Sahayogi, 1985.
Locke, J., The Upoadha Vrata of Amoghapa Lokevara in Nepal, LEthnographic, 83, 1987,
100/101, 15989.
Locke, J., The Unique Features of Newar Buddhism, in T. Skorupski, ed., The Buddhist
Heritage, Buddhica Britannica, Series Continua I, Tring, The Institute of Buddhist
Studies, 1989, 71116.
Loewe, M.A.N., Imperial China, London, 1966.
Lo Shih-lin, Chen Li, Liu Wen-chi & Liu Y-sung, comps., Tang-shu ching-chi i-wen ho-chih,
New Haven and London, 1973.
Lowenthal, D., The Past is a Foreign Country, New York, Cambridge University Press, 1985.
kLong-chcn Rab-byarns-pa, Grub-mtha mdzod, ed. by mDo-grub-chen Rin-po-che, Gangtok,
Sikkim, c. 1969.
Macdonald, A. Le Maala du Majurmlakalpa, Paris, 1962.
Malamoud C., Cuire Je Monde, Paris, 1989.
Malek, R., Das Chai-chieh-lu, Frankfurt am Main, 1985.
Mango, C., Byzantine Architecture, New York, 1975.

211

Marcotty, T. & A. Heller, Dagger Blissthe Tibetan Phurba Cult, Delhi, 1987.
Marcotty, T., Phur-paTibetan Ritual Daggers, Arts of Asia, 17.4, Hong Kong, 1987.
Masefield, P., Divine Revelation in Pali Buddhism, Colombo, 1986.
Masuda, J., Origin and Doctrines of the Early Buddhist Schools, AM, 2, 1925, 178.
Mayeda, E., Japanese Studies on the Schools of the Chinese Agamas, in H. Bechert, ed., Zur
Schulzugehrigkeit von Werken der Hnayna Literatur, Gttingen, Vandenhock &
Ruprecht, 1985.
Mendelson, E.M., Sangha and State in Burma: A Study of Monastic Sectarianism and
Leadership, Ithaca, Cornell University Press, 1975.
Mendis, G.C., The Chronology of the Early Pli Chronicles of Ceylon, UCR, 5, 1, 1947, 3954.
Meredith, Georgette, The Phurbu: The Use and Symbolism of the Tibetan Magic Dagger, HR,
6.3, 1967.
Miller, J., The Myth of the Churning of the Ocean of Milk, in P. Connolly, ed.,
Perspectives on Indian Religion, Delhi, 1986.
Mi-pham rGya-mtsho, gSang-snying spyi-don od-gsal snying-po, rNying-ma bkama rgyas-pa,
vol. 27, ed. bDud-joms Rin-po-che, Kalimpong, 1982.
Miyakawa Hisayuki, Local Cults around Mount Lu at the Time of Sun Ens Rebellion, in H.
Welch & A. Seidel, Facets of Taoism, New Haven & London, 1979, 83101.
Muranaka Yush, Tendai shoki gyh no shsei ni tsuite , IBK, 23.2, 1975, 56l6.
McMullen, D.L., State and Scholars in Tang China, Cambridge, 1988.
McRae, J., The Northern School and the Foundation of Early Chan Buddhism, Honolulu, 1986.
Nalanda Translation Committee, The Rain of Wisdom, Boulder, Colorado & London, Shambhala,
1980.
Namgyal, Takpo Tashi, tr. & an. by Lobsang P. Lhalungpa, Mahmudr: The Quintessence of
Mind and Meditation, Boulder Colorado, Shambala, 1986.
Nanamoli, Bhikkhu, tr., The Path of Purification (Visuddhimagga) by Bhadantcariya
Buddhaghosa, Colombo, R. Semage, 1956.
Nanananda, Concept and Reality in Early Buddhist Thought, Kandy, Ceylon, Buddhist
Publication Society, 1971.
Nattier, J. & C. Prebish, Mahsaghika Origins: The Beginnings of Buddhist Sectarianism, HR,
16.3, 1977, 23772.

212

Norman, K.R., Pli and the Language of the Heretics, AO, 7, 1976, 11726.
Norman, K.R., The Buddhas View of Devas, in Beitrge zur Indienforschung: Ernst
Waldschmidt zum 80, Geburtstag gewidmet. Berlin, 1977, 32936.
Norman, K.R., Mgadhisms in the Kathvatthu, in A.K. Narain, ed., Studies in Pali and
Buddhism in Memory of Bhikkhu Jagdish Kashyap, Delhi, 1979, 27987.
Norman, K.R., Devas and Adhidevas in Buddhism, JPTS, 9, 1981, 14555.
Norman, K.R., A Note on Att in the Alagaddpamasutta, in Studies in Indian Philosophy,
Ahmedabad, 1981, 1929.
Norman, K.R., Pli Literature, Wiesbaden, Otto Harrassowitz, 1983.
Norman, K.R., Aokas Schism Edict, Buddhist Seminar, 46, Kyoto, 1987.
Norman, K.R., Notes on the Sutta-nipta. SLJBS, vol. I, 1987, 10016.
Norman, K.R., Pli Language and Scriptures, in T. Skorupski, ed., The Buddhist Heritage,
Buddhica Britannica, Series Continua I, Tring, The Institute of Buddhist Studies, 1989,
2953.
Nyanatiloka, Mahathera, Guide through Abhidhamma-Piaka, Kandy, Ceylon, Buddhist
Publications Society, 1971.
Obermiller, E., The Sublime Science of the Great Vehicle to Salvation, Being a Manual of
Buddhlst Monism: The work of Arya Maitreya with Commentary by Arysanga, AO, 9,
Parts ii, iii and iv., 1971.
Ofuchi Ninji, The Formation of the Taoist Canon, in H. Welch & A. Seidel, eds., Facets of
Taoism, New Haven & London, 197, 253267.
Orbach, C.Z., Seeing Chen-yen Buddhism: Traditional Scholarship and the Vajrayana in China,
HR, 29.2, Novembcr, 1989, 87114.
Osabe Kazuo, T-S Mikkyshi Ronk, Kobe, 1982.
Owen, S., The Poetry of the Early Tang, New Haven, Yale University Press, 1977.
Paranavitana. S., The Stpa in Ceylon, Colombo, 1947.
Park, Sung Bae, Buddhist Faith and Sudden Enlightenment, Suny Series in Religious Studies,
Albany, NY, 1983.
Peremans, W. & E. Vant Dack, Prosopographia Ptolemaica, VI, Louvain, 1968.
Phur-ba bcu-gnyis kyi rgyud, in rNying-mai rgyud-bum, vol. l9, Thimphu, 197375.
Pope, A.U., ed, A Survey of Persian Art, Oxford, l93839.
Prebish, C.S., A Review of Scholarship on the Buddhist Councils, JAS, 33.2, 1974, 23954.

213

Quigley, D., Ethnicity without Nationalism: the Newars of Nepal, EJS, 28, 1987, 15270.
Rabten, Geshe, tr. by S. Batchelor, The Mind and its Functions: A Textbook of Buddhist
Epistemology and Psychology, Mt Pelerin, Switzerland, Tharpa Choeling, 1978.
Reynolds, C., The Buddhist Monkhood in Nineteenth Century Thailand, unpublished Ph D
thesis, Cornell University, 1973.
Riccardi, T., Buddhism m Ancient and Early Medieval Nepal, in A.K. Narain, ed., Studies in
the History of Buddhism, Delhi, B.R. Publishing House, 1980.
Rivire, C., En Chine avec Teilhard, 19381944, Paris, 1968.
Robinet, l., La rvlation du Shangqing dans lhistoire du Taoisme, Paris, 1984.
Roerich, G.N. tr., The Blue Annals, repr. Delhi, Motilal Banarsidass, 1976.
Ropp, P.S., ed., The Heritage of China: Contemporary Perspectives on Chinese Civilization,
Berkeley, 1990.
Rosenfield, J .M., The Dynastic Arts of the Kushans, Berkeley & Los Angeles, University of
California Press, 1967.
Roth, G., ed., Bhiku-vinaya, Patna, 1970.
Ruegg, David Seyfort, The Jo nang pas: A School of Buddhist Ontologists According to the
Grub mtha sel gyi me long, JAOS, 83, 1963, 7391.
Ruegg, David Seyfort, The dGe lugs pa Theory of the Tathgatagarbha, Pratidnam Studies
presented to F.BJ. Kuiper, Janua Linguarum, Series Maior, 34, The Hague, 1968, 500
09.
Ruegg, David Seyfort, La thorie du Tathgatagarbha et du Gotra, Paris, EFEO, 1969.
Ruegg, David Seyfort, Le trait du Tathgatagarbha de Bu sTon Rin Chen Gru, traduction du De
bzhun gshegs pai snying po gsal zhing mdzes par byed pai rgyan, Paris, EFEO, 1973.
Ruegg, David Seyfort, The Meanings of the Term Gotra and the Textual History of the
Ratnagotravibhga, BSOAS, 39, 1976, 34163.
Russell, Bertrand, A History of Western Philosophy, London, George Allen & Unwin, 1961.
Samaynuayanirdea, Dam-tshig phra-rgyas bstan-pa, TTP, No 4746, vol. 83, 149.4.8.l50.3.8.
gSang-snying grel-pa gsang-bdag dgongs-rgyan, rNying-ma bka-ma rgyas-pa; vol. 32, ed. by
bDud-joms Rin-po-che, 45 vols., Kalimpong, 1982.
ntideva, ikmuccaya, tr. by C. Bendall & W.H.D. Rouse, 1st ed. 1922; 2nd ed., Delhi,
Motilal Banarsidass, 1977; rep. 1981.

214

Sa-skya Drag-pa rGyal-mtshan, tr. M. Tatz, Candragomins Twenty Verses on the Bodhisattva
Vow and its Commentary, Dharamsala, Tibetan Library of Works and Archives, 1982.
Sastri, N.A .. Satyasiddhistra of Harivarman, Baroda, 1978.
atapatha Brhmaa, tr. J. Eggeling, SBE, vols. 12, 26, 41, 43, 44, Oxford, 1882.
Saveros Pou [Lewitz], Les inscriptions modernes dAngkor vat, JA, 260, 1972, 10729.
Saveros Pou [Lewitz], Inscriptions modernes dAngkor 35, 36, 37, et 39, BEFEO, 61, 1974,
30137.
Saveros Pou [Lewitz), Inscriptions modernes dAngkor 34 et 38, BEFEO, 62, 1975, 283353.
Schayer, S., Kamalalas Kritik des Pudgalavda, RO, 8, 1932, 6893.
Schmithausen, L., Philosophische Bemerkungen zum Ratnagotravibhaga, WZKSO, 15, 1971.
Schmithausen, L., Zu D. Seyfon Ruegg Buch Theorie du Tathgatagarbha et du Gotra,
WZKSO, 16l7, 197273, 123160.
Schmithausen, L., layavijna, 2 vols., Tokyo, The International Institute for Buddhist Studies,
1987.
Schopen, G., The Inscription on the Kun Image of Amitbha and the Character of the Early
Mahyna in India, JIABS, 10.2, 1987, 99137.
Seidel, A., La divinisation de Lao tseu dans le Taoisme des Han, Paris, 1969.
Sekiguchi Shindai, Tendai Shikan no kenky, Tokyo, 1969.
Shively, D.H., Buddhahood for the Nonsentient: A Theme in N Plays, HJAS, 20, 1957, 135
161.
de Silva, Lily, The Symbolism of the Indrakla in the Parittamaapa, SSAC, vol. 7, ed. J.E. van
Lohuizen de Leeuw, Leiden, 1978.
iva Pura. ed. by A. Kunst & NA. Sastri, tr. by a Board of Scholars, Delhi, 1978.
Skilling, P., The Sasktsaskta-Viniaya of Daabalarmitra, BSR, 4.1, 1987, 323.
Slusser, M.S., Nepal Mandala: A Cultural Study of the Kathmandu Valley, 2 vols., Princeton,
Princeton University Press, 1982.
Smith, BL., ed., Religion and Legitimation of Power in Ceylon, Chambersburg, PA, Anima, 1978.
Snellgrove, D., lndo-Tibetan Buddhism: Indian Buddhists and their Tibetan Successors, London,
Serindia, 1987.

215

Snellgrove, D.L., The notion of divine kingship in tantric Buddhism, in International Congress
for the History of Religions, The Sacral Kingship, Leiden, E.J. Brill, 1959, 204218.
Srensen, S., An Index to Names in Mahbhrata, London, 1904.
Stcherbatsky,Th., The Central Conception of Buddhism and the Meaning of the Word dharma,
Prize Publication Fund, vol. 8, London, 1923.
Stcherbatsky, Th., The Conception of Buddhist Nirva, Leningrad, 1927.
Stcherbatsky, Th., Buddhist Logic, 2 vols., New York, Dover Publications Inc., 1962.
Stein, R.A., Tibetan Civilization, Stanford, 1972.
Stein, R.A., La gueule du Makara, in A. Macdonald (Spanien) & Y. Imaeda, eds., Essais sur
lArt du Tibet, Paris, 1977, 5362.
Stein, R.A., A propos des documents anciens relatifs au Phurbu, Csoma de Koros Symposium,
Bibliotheca-Orientalia Hungarica, vol. 23, Budapest, 1978.
Stronach, D., On the Evolution of the Early Iranian Fire Temple, in Papers in Honour of
Professor Mary Boyce, Leiden, E.J. Brill, 1985.
Stutley, Margaret & James, A Dictionary of Hinduism, London, 1977.
Stutley, Margaret, Ancient Indian Magic and Folklore, London, 1980.
Takasak.i Jikido, A Study on the Ratnagotravibhga, Being a Treatise on Tathgatagarbha
Theory of Mahyna Buddhism, Serie Orientale Roma, 33, Rome, IsMEO, 1966.
Tambiah, S., The Confluence of Anthropology, History, and lndology, CIS, N.S. 21, 1987,
187216.
Tambiah, S., World Conqueror, World Renouncer, New York, Cambridge University Press,
1977.
Tanaka Rysh, Tonk Zensh bunken no kenky, Tokyo, 1983.
Tatz, M., tr., [Sa-skya Drag-pa rGyal-mtshan], Candragomins Twenty Verses on the Bodhisattva
Vow and its Commentary, Dharamsala, Tibetan Library of Works and Archives, 1982.
Teiser, S.F., The Medieval Chinese Ghost Festival, Princeton, 1988.
Teng, S.Y. & K. Biggerstaff, An Annotated Bibliography of Selected Chinese Reference Works,
3rd ed., Cambridge, Mass., 1971.
Tewari, R.C., Socio-cultural Aspects of Theravda Buddhism in Nepal, JIABS, 6, 1, 1983, 67
93.
Thomas, E.J., The Life of the Buddha, 3rd ed., London, 1949.
Thomas, EJ., Theravdin and Sarvstivdin Dates of the Nirva, B.C. Law Volume, Part II,
Poona, 1946, 1822.

216

Thrangu Rimpoche, tr. b) Salya Dorjc, ed. by Michael Lewis, The Open Door to Emptiness,
Lhundrub Teng, Kathmandu, 1978.
Thuu-bkwan, Blo-bzang chos-kyi nyi-ma. Collected Works, Delhi, 1969.
Toffin, G., Socit et religion chez les Nwar du Npal, Paris, CNRS, 1984.
Tsukamoto, Z., A History of Chinese Buddhism, 2 vols., Tokyo, New York & San Francisco,
1985.
Tsultrim, Khenpo tr. by Shenpen Hookham, Meditation Progression on Emptiness, Oxford,
Longchen Foundation, 1986.
Tucci. G., Minor Buddhist Texts, Part II, Serie Orientale Roma, Rome, IsMEO,1958.
Twitchctt, D.C. & M.A.N. Loewe, eds., The Cambridge History of China, vol. l, Cambridge,
1986.
Unno, T. The Buddhat Theory of Fa-tsang, in Transactions of the International Conference of
Orientalists in Japan, 8, 1963, 3441.
Vaidya, P .L., ed., Gaavyhastra, Darbhanga, 1960.
Vajrval, atapiaka Series, vol. 239, New Delhi, 1977.
Venkataramanan, K., Smmityanikya stra, VBA, 5, 1953, 153243.
Verellen, F., Du Guangting (850933): Taoste de cour la fin de la Chine mdivale, Paris,
1989.
Vilsavajra/Llvajra Samayavivyakti, Dam-tshig gsal-bkra, TTP, No. 4744, vol. 83.
Vogel, C., Bu-ston on the Schism of the Buddhist Church and on the Doctrinal Tendencies of
Buddhist Scriptures, in H. Bechen, ed., Zur Schulzugehrigkeit von Werken der
Hnyana-Literatur, Gttingen, 1984.
Volpert, P.A., Tschng huang, Anthropos, 5.5/6, 1910, 9911026.
Waley, A., The Life and Times of Po Ch-i, London, 1949.
Wang Gungwu, The Middle Yangtse in Tang Politics, in A.F. Wright & D.C. Twitchett, eds.,
Perspectives on the Tang, New Haven & London, 1973, 193235.
Wang Mo, ed., Han-Tang ti-li shu chao, Peking, 1961.
Wang Shao-ymg, ed., Sou-shen hou-chi 5, Beijing, 1981.
Warder, A.K., Indian Buddhism, Delhi, Motilal Banarsidass, 1970; 2nd rev. ed., 1980.
Warder, A.K., Introduction to Pli, London, 1963.
Watanabe, F., Philosophy and its Development in the Nikyas and Abhidhamma, Delhi, Motilal
Banarsidass, 1983.
Weller, R., Unities and Diversities in Chinese Religion, Seattle, 1987.

217

Welch, H. & A. Seidel, eds., Facets of Taoism, New Haven & London, 1979.
Whalen Lai, Tao-shengs Theory of Sudden Enlightenment Re-examined, in Peter N. Gregory,
ed., Sudden and Gradual: Approaches Enlightenment in Chinese Thought, Honolulu,
1987, 169200.
Will, E., Histoire politique du monde hellnistique, 2nd ed., Nancy, 1979.
Williams, P., A Note on Some Aspects of Mi bskyod rDo rjes Critique of dGe lugs pa
Madhyamaka, JIP, 1983, 125145.
Wilson, H.H., ig Veda Sanhita, London, 1857.
Wright, A.F. & D.C. Twitchett, eds., Perspectives on the Tang, New Haven London, 1973.
Xinazzo, G., The Notion of Tao in Lao Tzu and Chuang Tzu, Taipei, 1968.
Yanagida Seizan, Shoki Zensh shisho no kenky, Tokyo, 1967.
Yen Keng-wang, Tang-jen tu-shu shan-lin ssu-yan chih feng-shang, Chung-yang yen-chiuyan Li-shih y-yen yen-chiu-so chi-kan 30.2, 1959, 689728.
Yoritomi Motohiro, Chgoku Mikky no kenky, Tokyo, 1979.
Zangmo, K. Yeshe (Bikkhuni Sik), An Index to the sDe dgc bKa gyur with Book Titles and
Volume Contents, Cambridge, 1981.

218