VOLUME 20 (May 2021)

ISSN: 2047-1076

Journal of the
Oxford
Centre for
Buddhist
Studies

The Oxford Centre for

Buddhist Studies
http://www.ocbs.org
JOURNAL OF THE OXFORD CENTRE
FOR
BUDDHIST STUDIES

volume 20

May 2021
 Journal of the Oxford Centre for Buddhist Studies
Volume 20
May 2021
ISSN: 2047-1076
Published by the Oxford Centre for Buddhist Studies
www.ocbs.org
Wolfson College, Linton Road, Oxford, OX2 6UD, United Kingdom

Authors retain copyright of their articles.

Editorial board
Dr Alexander Wynne (Editor): [email protected]
Prof. Richard Gombrich (Editorial Consultant): [email protected]
Prof. John Holder: [email protected]
Dr Tse-fu Kuan: [email protected]
All submissions should be sent to: [email protected]

Production team
Operations and Development Manager: Steven Egan
Development Consultant: Dr Paola Tinti
Journal production and cover illustration by Ivan de Pablo Bosch (www.ivancious.com)

Annual subscription rates

Students: £20
Individuals: £30
Institutions: £45
Universities: £55
Countries from the following list receive 50% discount on all the above prices:
Bangladesh, Burma, Laos, India, Nepal, Sri Lanka, Thailand, Cambodia, Vietnam,
Indonesia, Pakistan, all African Countries
For more information on subscriptions, please go to www.ocbs.org
 Contents

Contents 4

List of Contributors 6

Editorial Alexander Wynne 8

Obituary: Peter Dennis Masefield Alexander Wynne 11

The Asokaparinibbānakathā
Peter Masefield and Jacqueline Filliozat 22

On the Buddha’s ‘Kammic Fluff’:

The Last Meal Revisited
Peter Masefield and Nicolas Revire 51

Did the Buddha address the monks in Māgadhī?

Ole Holten Pind 83

Toward a Computational Analysis of the Pali Canon

Dan Zigmond 107

Nissāraṇīya: A Codified Term Updating the Development

of the Pāli Vinaya, Part I
Juo-Hsüeh Shih 136
BOOK REVIEWS

Wisdom as a Way of Life. Theravāda Buddhism Reimagined

By Steven Collins
Reviewed by Alexander Wynne 163

Gender and the Path to Awakening:

Hidden Histories of Nuns in Modern Thai Buddhism
By Martin Seeger
Reviewed by Sarah Shaw 180

Reading the Buddha’s Discourses in Pāli

By Bhikkhu Bodhi
Reviewed by Niels Schoubben 187

Genèse d’un apocryphe bouddhique: le Sūtra de la pure délivrance

By Costantino Moretti
Reviewed by Frederick Shih-Chung Chen 193
 List of Contributors

Jacqueline Filliozat is a long-time research fellow of the École française

d’Extrême-Orient in Paris, specializing in cataloguing Pāli manuscripts of
various scripts and types from different provenances, in both European
and Asian collections. Since her retirement in 2007 she has worked on
critical editions of rare texts transmitted in so called “Indochinese Pāli”.
[email protected]

Peter Masefield (1943-2020) was a former lecturer at the University of

Sydney and Mahachulalongkornrajavidyalaya University in Bangkok, and
a translator of a number of Pali commentaries for the Pali Text Society. His
monograph Divine Revelation in Pali Buddhism (1986) was a pioneering book
within the study of early Buddhism.

Ole Holten Pind (1945-2018) was co-editor of A Critical Pāli Dictionary (Vol.
III.1-6), and an author of a number of its articles. He produced the critical edition
of the Kaccāyana and Kaccāyanavutti for the Pali Text Society (2013), and wrote
numerous papers on diverse aspects of Indian Buddhist thought and Pali grammar.
He was awarded the degree of Doctor of Philosophy at the University of Vienna
in 2015 for his work on the fifth chapter of Dignāga’s Pramāṇasamuccayavṛtti.

Nicolas Revire is Assistant Professor at Thammasat University, Faculty

of Liberal Arts, Bangkok, and holds a doctoral degree from the Université
Paris 3–Sorbonne nouvelle in France. He specializes in the Buddhist art and
archaeology of South and Southeast Asia, with a research focus on pre-modern
Thailand, Myanmar and Cambodia. He is general editor of a collective volume
titled Before Siam: Essays in Art and Archaeology and co-published by the Siam
Society (Bangkok, 2014). [email protected]

6
Niels Schoubben currently works at Leiden University as a PhD student in the
ERC project “The Tocharian Trek”, for which he investigates language contact
between Niya Prakrit, Tocharian and Middle Iranian languages, especially
Bactrian. This linguistic research is complemented by an interest in more
philological approaches towards Buddhist texts in, for instance, Gāndhārī and
Pāli. [email protected]

Sarah Shaw is the Khyentse Foundation Reader in Buddhist Studies at

the University of South Wales. She has been a member of the Faculty of
Oriental Studies, Oxford, since 2008, and is a fellow of the Oxford Centre
for Buddhist Studies. She works on Pāli Suttas, narrative, meditation and
chant. [email protected]

Juo-Hsüeh Shih is a freelance researcher living in Taiwan, and the

author of Controversies over Buddhist Nuns (PTS, 2000). She has a
D.Phil. from the University of Oxford. Her research interests are the
comparative study of the different Vinayas, and early Buddhist doctrine
and meditation. [email protected]

Frederick Shih-Chung Chen holds a doctoral degree from the University

of Oxford, and is a Visiting Research Fellow in Chinese Buddhism in the
Department of Philosophy at the National Chengchi University, Taipei. His
main academic interests are the history and archaeology of early and medieval
Chinese Buddhism. [email protected]

Alexander Wynne is the editor of this Journal, and Senior Academic

Consultant at the OCBS. His work focuses on the intellectual history of
Indian Buddhism, and the Pali manuscript tradition of Theravāda Buddhism.
[email protected]

Dan Zigmond is a Sōtō Zen Buddhist priest and Guiding Teacher at Jikoji Zen
Center in Los Gatos, California. He is also a writer and technologist working
in Silicon Valley. His most recent book is Buddha’s Office: The Ancient Art of
Waking Up While Working Well (Running Press, 2019). [email protected]   

7
 Editorial

Alexander Wynne

The world of Pali studies has recently lost a number of elder statesmen: L. S.
Cousins, Steven Collins, Ole Pind, K. R. Norman and Peter Masefield. The
current issue of JOCBS could be regarded as a memorial to this generation of
Buddhist scholars. Most contributions deal with the Pali tradition, and there
are posthumous articles by Ole Pind and Peter Masefield, as well as a review
of Steven Collins’ final book, Wisdom as a Way of Life. All of these pieces
exemplify K. R. Norman’s claim that ‘Everything that has not been done needs
to be done. Everything that has been done needs to be done again.’
The first of two articles by Peter Masefield is an edition and translation of
the Asokaparinibbānakathā, in collaboration with Jacqueline Filliozat. Because
the post-canonical Pali literature of mainland Southeast Asia has been so little
studied, this article serves as an example of an area which for the most part
has yet to be tackled. Many more of Peter Masefield’s works on this genre of
Pali literature will appear in future issues of JOCBS. Peter Masefield’s other
contribution in this issue, in collaboration with Nicolas Revire, revisits the
Buddha’s last meal. This article shows that what has already been done needs
doing again, and also demonstrates the utility of studying the Pali commentarial
literature in conjunction with previously unknown South East Asian sources,
textual and art-historical.
Dan Zigmond’s article on computational approaches to the language of the
Pali canon opens up a new avenue in the study of the Pali canon, one of the
many things yet to be done. So too does Juo-Hsüeh Shih’s study of the term
nissāraṇīya/nissaraṇīya in the Pali Bhikkhunī Vinaya. Although it is now almost
150 years since the founding of the Pali Text Society, it might appear surprising
that new discoveries about the Pali Vinaya can still be made. But the study of
the Pali canon is really still in its infancy; most studies need to be done again.
 Editorial

The article by Ole Pind, on so-called ‘Māgadhisms’ in the Pali canon, is a

striking example of this. This paper continues a recent debate in JOCBS on
the language of early Buddhism, following the articles of Karpik (JOCBS 16),
Levman (JOCBS 17) and the editorial of JOCBS 17. In agreement with Karpik,
Pind argues that so-called ‘Māgadhisms’ in the Pali canon, such as bhikkhave,
are in fact a regular feature of Pali, and need not be regarded as a remnant of an
earlier linguistic stratum. As Pind puts it, ‘it is necessary to study the language
of the Tipiṭaka as a language sui generis and not as a random patchwork of
borrowings from other linguistic environments’. While not necessarily agreeing
with Karpik (and Gombrich)1 that Pali was the language of the Buddha, Pind’s
argument offers support for that view. In this case, redoing what has already
been done means rethinking all our previous assumptions about Pali.
Not all scholars agree with Norman’s point that everything which has been
done ought to be done again. Some, such as Steven Collins, seem to have
believed that certain areas of study should be shut down. In Wisdom as a Way
of Life, Justin McDaniel comments (p.lii) that ‘While Gombrich, like many
other scholars of Steve’s generation, believed that scholars could and should use
limited textual evidence to help speculate on and reconstruct the ideas, practices,
and even daily lives of early Buddhists (loosely 500 BCE to 1200 CE), Steve
found this project increasingly useless and even intellectually dangerous.’
This comment is most peculiar and quite misleading. If the early Buddhist
period includes evidence as late as 1200 CE, it is hard to see how scholars
of early Buddhism could speculate on the ‘daily lives’ of early Buddhists,
starting with the Buddha and his disciples in the 5th century BC. The Pali
canonical sources, of course, are not nearly as recent as the 12th century AD;
but although they are considerably older than this, they are not precise enough
to reveal details about ‘daily lives’. Nevertheless, our ancient sources are a
vital window into the past and as the articles in this issue of JOCBS show, they
can be used to expand our knowledge of early Buddhism. Norman’s adage
that ‘everything that has been done needs to be done again’ keeps the door to
new discoveries open; the view that the study of early Buddhism is ‘useless’
and ‘dangerous’ slams it firmly shut. There is no question which is the better
option for Buddhist studies.

1
Richard Gombrich, Buddhism and Pali (Oxford: Mud Pie, 2018), p.72ff.

9
 Peter Dennis Masefield
1943-2020

Alexander Wynne

Dr Peter Dennis Masefield, who passed away in Bangkok on September 7th 2020
at the age of 77, was one of the world’s leading scholars of Pali and Theravāda
Buddhism and a translator for the Pali Text Society for a number of decades.
Born in 1943 in Birmingham, England, Peter’s path to Buddhist Studies began
in his mid-twenties when he left England to travel the East. Travelling the
Indian subcontinent by rail, and hitchhiking when necessary, Peter visited the
 Obituary

caves of Bamiyan, saw the Buddhist monuments of Gandhara and experienced

Himalayan Buddhism in Kathmandu and Sikkim. Spending periods also in Sri
Lanka and Burma, Peter returned to England in 1969 and was admitted into
the fledgling Religious Studies programme at Lancaster University. Despite not
having the correct academic qualifications, Peter could read French and German
well, and impressed Ninian Smart with a short piece on India.
Peter graduated with honours from Lancaster with a BA in Philosophy &
Religious Studies in 1972, and spent the period 1973-75 conducting fieldwork
in India and Sri Lanka, supported by a Spalding Trust Fellowship. Despite a
bout of elephantiasis, contracted, he claimed, during a long trek up Adam’s Peak
to see the Buddha’s footprint, during this period Peter taught himself Pali. On
his return to England Peter took up a temporary appointment, the first of many,
at the University of Manchester, where he lectured on Buddhism and mysticism
in the Department of Comparative Religion. After a spell as Temporary Lecturer
in Durham University in 1976, teaching courses on Indian Philosophy and
Religion, Peter returned to Lancaster as a PhD student. Initially wishing to work
on karma and rebirth, he gained his PhD in 1980 for a thesis entitled ‘Thus
They Once Heard – Oral Initiation in the Pali Nikayas.’ Peter said that Dr Karel
Werner (the external examiner) spent much of the oral examination arguing with
Prof. Ninian Smart (the internal examiner and Peter’s PhD supervisor) about
Indian philosophy.
During the course of his PhD Peter was again on the move, teaching at
the University of Edinburgh (1977-78) and then taking up a fellowship at the
University of Otago, Dunedin (1978-79). Since the latter was a post-doctoral
fellowship, upon arriving in New Zealand Peter spent three weeks non-stop
at a typewriter, at the end of which he had produced the manuscript of his
dissertation. After his PhD Peter settled in Sydney, where he remained until
2005, in the meantime becoming an Australian citizen. At the beginning and
end of this period Peter was employed by the University of Sydney: as a visiting
fellow and lecturer (1980-85), and as a visiting scholar and casual lecturer
(1998-2005). In between, and besides various teaching stints in Sydney, Peter
was a guest lecturer at the University of Hawai’i (1985), a research assistant at
Monash University, Melbourne (1993-94), a Numata visiting professor at the
University of Toronto (1995), a guest lecturer at the University of California, LA
(1999), a guest lecturer at Assumption University and at Mahachulalongkorn
University in Bangkok (2000), and a visiting lecturer at Mahachulalongkorn
University, Bangkok (2002-05).

11
 Obituary

When Peter finally obtained a permanent lectureship it came about

unexpectedly. While he was moving to Thailand in 2006 to work as
a translator for the World Buddhist University (in Benchasiri Park,
Bangkok), the position mysteriously disappeared, leaving Peter high and
dry. Fortunately, Peter became a permanent lecturer at Mahachulalongkorn
University, first at its Tha Phra Chan campus in Bangkok, and then at its
Wang Noi campus in Ayutthaya.
Peter was a popular teacher who made many friends among his students,
but after 2011 he worked mostly as editor of the Mahachulalongkorn Journal
of Buddhist Studies. When the University withheld his salary in 2015 for no
apparent reason – he was eventually paid in full, but only after petitioning the
Ministry of Education – he left permanent employment in September 2016
to live the life of a retired scholar in Bangkok. Increasingly poor in health,
with damaged nerves in his back causing considerable pain and problems of
mobility, Peter withdrew into his apartment in a quiet residential area near
Thong Lo, venturing out only occasionally.
In his last years Peter had been working on a translation of Dhammapāla’s
commentary on the Cariyāpiṭaka for the Pali Text Society. This project was
the culmination of decades of work on Dhammapāla, which had begun while
Peter was a research fellow at Dunedin in 1978-79. After the PTS published
his translation of the Petavatthu and its commentary in 1980 (a reworking of
a translation by U Ba Gyaw), translations of the Vimānavatthu, Udāna and
Itivuttaka and their commentaries appeared over the years.
Peter’s translation style could be said to be overly formal and slightly
idiosyncratic. But he argued, quite reasonably and perhaps correctly, that at
this pioneering stage of research, style is not all that important: it is better
to be consistent and literal. The great merit of Peter’s translations is that
the reader does not get lost in interpretation, and for this reason they are an
excellent place to begin studying this difficult genre of Pali literature. Another
outcome of Peter’s research on the Pali commentaries was his argument (2002)
that Dhammapāla is to be dated much later than Buddhaghosa, since his
commentaries often include material identical or parallel to that found in the
Ṭīkās on Buddhaghosa’s commentaries.
Peter’s second significant contribution to Buddhist Studies was his
monograph Divine Revelation in Pali Buddhism (1986, Second Edition
2008, Routledge), based on his PhD thesis of 1980. Divine Revelation is
the work of an autodidact, an individualist willing to look afresh at the

12
 Obituary

world of early Buddhism, and unafraid to form and express his ideas about
it. At the time the book was something of a bolt out of the blue, a direct
challenge to the rationalist presentation of the Pali tradition which had been
popularised by such books as Walpola Rahula’s What the Buddha taught
(1959). As Paul Harrison’s favourable review put it, ‘Divine Revelation in
Buddhism? – the very title of the book comes as a shock.’1 In fact the title
was invented by the publisher, and caused Peter a certain amount of trouble
later on. He was once accosted at a conference by a Sri Lankan angry at the
‘Christian’ title, but on further discussion it turned out that this person had
not read the book at all.
Divine Revelation is notable for the impressive number of primary texts
consulted, including the Nikāyas, the Pali commentaries and exegetical texts
such as the Milindapañha. In the days before computers were widely available,
and well before the advent of electronic resources, such a wide range of reading
was uncommon in Pali Studies. Few books on early Buddhism had hitherto
tackled the Pali canon in such detail, and even fewer with Peter’s originality and
insight. Some of his arguments have since become standard thinking on early
Buddhism, for example that numerous teachings were formed as a response
to Brahminism, and that the Buddha’s ‘skill in means’ is a standard feature
of the Nikāyas, one most clearly exemplified by the ‘gradual discourse’. It is
unfortunate that Divine Revelation has not been read widely enough for Peter to
be credited as a major source of these ideas.
The central theme of Divine Revelation is the importance of hearing the
Dhamma and being transformed by it. It shows that the Nikāyas consistently
present direct contact with the Buddha as an exceptional experience which
engenders ‘right view’, and so converts a person from being a puthujjana
into an ariya-sāvaka. Peter was right to note that the main distinction in early
Buddhism was not between monk and layman, but between the ariya-sāvaka
and the puthujjana, with the laity and monastics being found in both groupings.
The book’s attention to detail remains unusual in Buddhist Studies: given the
highly repetitive nature of canonical Buddhist texts, it is easy to gloss over terms
such as sutavant, sāvaka etc. without thinking about their meaning. With his
sharp critical eye, Peter was able to see that such terms indicate an elevated
religious status through hearing.

1
Paul Harrison, ‘Buddhism: A Religion of Revelation after All? À propos Peter Masefield’s
Divine Revelation in Pali Buddhism,’ in Numen, Vol. 34, Fasc. 2 (Dec., 1987), pp. 256-264.

13
 Obituary

The interpretation of the term sotāpanna as ‘one who has come into contact
with (or undergone) the hearing’ remains controversial. But Divine Revelation
points out that in the Nikāyas, stream imagery – like flood and ocean imagery –
is usually a negative metaphor for all that is wrong with the world. The Buddhist
path is that which sets a person ‘against or across the stream’ (paṭisota), and
texts such as MN 34 claim that all sāvakas have crossed it. Moreover, the Pali
commentaries do not interpret sota in the sense of ‘river’ (nadī), the similar
term dhamamsota refers to the Dhamma-ear (by which a sāvaka hears the
sound of the deathless), and the equivalent term in Buddhist Hybrid Sanskrit is
śrotāpanna (Skt. śrotas, ‘ear’).
Although Divine Revelation backs its arguments up with detailed textual
scholarship, sometimes it strays into slightly more esoteric territory, for
example the claim that the sound of the Buddha’s teaching is a transcendental
manifestation of the Dhamma. And yet this point, despite its apparent
peculiarity, draws attention to numerous canonical statements on the
importance of sound: the Buddha roaring a lion’s roar, or beating the drum
of the immortal and so on. It is easy to pass over such imagery without a
thought, but in doing so important features of the early Buddhist world-view
are missed.
Divine Revelation brings into clear focus a number of crucial features
of early Buddhism which are easily overlooked: the importance of oral
communication and spiritual encounters, the vision of a cosmos full of
‘hearers’ extending up to the divine realms (and including even tree-
spirits), and especially the exceptional role occupied by the Buddha in
this religious landscape. All of these points and many more were a major
concern of early Buddhism. Peter’s stimulating study makes it easier to
understand what actually happened: conversions, missions, the emergence
of Mahāyāna etc.
Peter’s third and final major contribution to the study of Pali was his
research into the indigenous Pali tradition(s) of mainland South East Asia.
This work is barely known, since Peter published very little of it during his
lifetime, although his article ‘Indo-Chinese Pali’ (2008) indicates the depth of
his reading and knowledge. After settling in Thailand, Peter began studying
the Khom manuscript tradition of Thailand and Cambodia, in collaboration
with Mrs Jacqueline Filliozat, then of the École française d’Extrême-Orient
in Bangkok. Peter and Mrs Filliozat produced a number of editions of Indo-
Chinese Pali texts, now on record in the internal database at the EFEO in

14
 Obituary

Paris. Translations of some of these will appear in forthcoming issues of the

Journal of the Oxford Centre for Buddhist Studies, beginning in JOCBS 20
with the Asokaparinibbānakathā.

*****

I first met Peter in 2008, not long after moving to Thailand to work at Mahidol
University. At first Peter was just somebody with whom I could talk about Pali,
but since we invariably met in the evening he quickly became a good friend.
Already in his 60s, Peter would often reflect on the past, sharing his memories
of the likes of Ninian Smart, Edward Conze, and I. B. Horner, whose use of a
long-stemmed cigarette holder amused him greatly and inspired a number of
limericks (‘Little Miss Horner, sat in the corner…’). Ninian Smart was a great
early supporter of Peter, but still refused to believe Peter’s claim that used tea
cups are habitually thrown out of moving trains in India. Peter was also tricked
into buying Ninian Smart’s old Morris Minor car for £100: as soon as he bought
it, Ninian asked to drive him down to Manchester on the very next day so that
he could watch the Test Match, and drink as much bitter as he liked.
In the 1980s, Peter somehow got involved with the Moonies. This
resulted in a number of all-expenses trips to Moonie conferences where
Peter witnessed mass marriage ceremonies. Thanks to his tongue-in-cheek
paper ‘The Muni and the Moonies’ (1985) Peter was an honoured guest,
although it wasn’t clear if Peter was poking fun at the Moonies or Buddhism.
A humorous episode from later on concerned a trip to England with his then
Indonesian wife, in around 2000. When visiting Lance Cousins in Oxford,
Peter let slip how his wife was famed as a spirit medium in her home town.
On hearing this, Lance had Peter’s wife remain in the car outside, lest her
psychic powers clash with his.
Peter was not a conventional person. Largely nocturnal, he had no time for
the comforts of bourgeois existence; my impression was that for the most part
he had passed through life as a happy wanderer. When his travels brought him
to Thailand, a favourite meeting place was the Queen Victoria in Bangkok. Peter
had an exemplary taste in pubs, and the Queen Vic offered a cool, wooden escape
from the heat of the city, where Peter could often be found with local friends,
such as Nicolas Revire, Stephen Evans, Mark Hoolahan, Larry the American,
occasionally Stewart McFarlane, Volkmar Enßlin, Arthur from Yorkshire, Giri
the Indian etc.

15
 Obituary

Peter was social and genial, with sparkling eyes and an endearing laugh; he
was kind and tipped the staff well, becoming affectionately known as ‘Achan
Chang’ after his Thai beverage of choice. Although he enjoyed his beer, he
drank very slowly; I cannot remember ever seeing him the worse for wear.
Appearances can be misleading. While not in any sense a Buddhist, in Sydney
Peter served a period as President of the Bulkwang Meditation Institute,
and was a co-founder of the Buddhist Council of New South Wales. He also
contributed to the ‘Dharma the Cat’ comic series, about a pious novice monk
whose wise cat take things at a more leisurely pace. Peter took his scholarly
work and his role as a teacher seriously; he took pride in his classes, for
which he prepared diligently. Peter was also surprisingly sensitive and fairly
conservative in his tastes. He liked to recite the poems of Betjeman which he
had learnt as a schoolboy, and enjoyed watching old clips of Monty Python,
the four Yorkshiremen being a particular favourite; its eccentric joke about
an English family living in ‘an ’ole in the ground’ would always have him
chuckling with laughter.
Peter was a good listener, unenamoured by the sound of his own voice, and
willing to change his mind (sometimes). He could also be very stubborn, and
was unwilling to play politics. Perhaps he was too honest for his own good,
and simply too rebellious to put up with the daily grind of university life. It
is not surprising that his only permanent lectureship was for a few years at
Mahachulalongkorn. Bangkok is the last place to speak one’s mind without
fear of causing offence, but in a way it was perfect for Peter, a welcoming and
friendly place for those who have drifted far from home, and know they cannot
go back (and do not wish to anyway).
Peter remained quietly cantankerous and witty until the end. In his last years,
when he was suffering from his bad back and finding it increasingly difficult to
look after himself, friends would bring him beer and ice, and quite often some
food too. When I returned from England, I would usually bring Peter some new
slippers and a few English delicacies, such as sausage rolls, pork pies and HP
fruity sauce; we both enjoyed the subversive irony of smuggling pork products
into Thailand. When the end finally came, Peter slipped away more quickly than
anyone expected, but then again, he always tended to confound expectations.
For his many friends, Bangkok will not be the same place without our Pali
scholar in residence.

16
 Obituary

Peter was fond of mentioning a dedication in a Pali manuscript he had

once read, in which the scribe said he hoped to have made enough mistakes to
avoid millions of years stuck doing nothing in the Brahmaloka. Although he
sympathised with the scribe’s plight, we hope that the same is not true of Peter,
whose mass of good merit has surely transported him to a comfortable ‘’ole in
the ground’ in Brahma’s heaven, and with a steady supply of cold beer. Cheers,
Peter, and thanks for all the treasured memories.

Publications

Translations
2008-09. The Itivuttaka Commentary, being a translation from the Pali
of Dhammapāla’s commentary on the Itivuttaka (Paramattha-dīpanī-
Itivuttakaṭṭhakathā). Two volumes. Oxford: Pali Text Society.
2000. The Itivuttaka, being a translation of the canonical text of the same name.
Oxford: The Pali Text Society.
1994-95. The Udāna Commentary, being a translation from the Pali of Dhammapāla’s
commentary on the Udāna (Paramattha-dīpanī-Udānaṭṭhakathā). Two volumes.
Oxford: Pali Text Society.
1994. The Udāna, being a translation of the canonical text of the same name. Oxford:
Pali Text Society.
1989. Vimāna Stories, being a translation from the Pali of Dhammapāla’s commentary
on the Vimānavatthu (Paramattha-dīpanī-Vimānavatthuaṭṭhakathā). London:
Pali Text Society.
1980. Peta Stories, being a translation from the Pali of Dhammapāla’s commentary
on the Petavatthu (Paramattha-dīpanī-Petavatthuaṭṭhakathā). London: Pali
Text Society.

Monographs and Festchrifts

1994. Aspects of Religion: Essays in Honour of Ninian Smart (co-edited with Donald
Wiebe). New York: Lang.
1986. Divine Revelation in Pali Buddhism. London: George Allen & Unwin,
hardback; The Sri Lanka Institute of Traditional Studies, 1987, paperback.
Second Edition: Routledge, 2008.

17
 Obituary

Contributions to Anthologies
2002. “The Composition of the Itivuttakaṭṭhakathā”, in Buddhist and Indian Studies
in Honour of Professor Sodo Mori. Kokusai Bukkyoto Kyokai (International
Buddhist Association), Hamamatsu, Japan, pp 103-115.
1990. “The Pursuit of Merit—Sacrificial Devotion in the Pali Nikayas”, in Indian
Devotional Traditions, eds. G. Bailey and I. Kesarcodi-Watson, Sterling
Press, Delhi.
1988. “The Origin and Development of the Preta in Early Buddhism”, in Religions
and Comparative Thought—Essays in Honour of the Late Dr. Ian Kesarcodi-
Watson, ed. Purusottama Bilimoria and Peter Fenner, Sri Satguru Publications,
Delhi, pp 47-69.
1986. “How ‘Noble’ is the Ariyan Eightfold Path?”, in Perspectives on Indian
Religion—Papers in Honour of Karel Werner, ed. Peter Connolly, Sri Satguru
Publications, Delhi, pp 161-174.
1984. “The Progressive Talk—the Buddha’s Inducement of a Vision of Nibbana
through an Altered State of Consciousness”, in Under the Shade of a Coolibah
Tree—Australian Studies in Consciousness, eds. R. A. Hutch and Peter Fenner,
University Press of America, Lanham, pp 123-143.
1983. “Mind/Cosmos Maps in the Pali Nikayas”, in Buddhist and Western
Psychology, ed. Nathan Katz, Prajna Press, Boulder, pp 69-93.

Contributions to Encyclopedias
2003. Articles on “Ghosts” and the “Milindapañha” in Encyclopedia of Buddhism,
ed. Robert E. Buswell, MacMillan, New York.

Journal articles
2010. “A brief note on the Meaning of Moha”, in Mahachulalongkorn Journal of
Buddhist Studies, vol 3, pp 5-12.
2010. “Kammapilotika”, in Mahachulalongkorn Journal of Buddhist Studies, vol 3,
pp 13-24.
2009. “A brief note on methodology in the transliteration of Indo-Chinese and other
Pali manuscripts”, in Mahachulalongkorn Journal of Buddhist Studies, vol 2,
pp 11-14.
2008. “Indo-Chinese Pali”, in Mahachulalongkorn Journal of Buddhist Studies, vol
1, pp 1-9.

18
 Obituary

2008, with Jacqueline Filliozat. “Two Indo-Chinese Pali Versions of the Petavatthu”,
in Mahachulalongkorn Journal of Buddhist Studies, vol 1, pp 11-18.
2008, with Jacqueline Filliozat. “Manuscripts of the Pāli Jāt ṭīkā (Jātakaṭṭhakathā–
Līnatthapakāsinī), Jātatthakīnidāna & Sotatthakīmahānidāna”, in
Mahachulalongkorn Journal of Buddhist Studies, vol 1, pp 27-37.
1985. “The Muni and the Moonies”, in Religion, vol 15, pp 143-160.
1979. “The Nibbana-Parinibbana Controversy”, in Religion, vol 9, pp 215-230.
1972-76. “The Mahakammavibhangasutta—An Analysis”, in the Vidyodaya Journal
of Arts, Science and Letters, University of Sri Lanka, Vidyodaya Campus, vol
5, nos. 1-2, pp 75-83.

Reviews
1995. Beyond Language and Reason: Mysticism in Indian Buddhism, by Ilkka
Pyysiäinen, in Indo-Iranian Journal, vol.38, no.3, pp 266-269.
1992. “Reply to F. J. Hoffman”, in Numen, vol.39, fasc. 2, pp 256-260.
1989. Rationality and Mind in Early Buddhism, by Frank J. Hoffman, in Numen, vol.
36, fasc. 1, pp 135-138.
1988. The Threefold Refuge in the Theravada Buddhist Tradition, by John Ross
Carter, George Doherty Bond, Edmund F. Perry and Shanta Ratnayaka, in
Journal of the Institute for the Study of Religion and Culture, vol. 11, no.7, pp
101-170.
1987. Catalogue of the Pali Printed Books in the India Office Library, ed. T. C. H.
Raper, revised by M. J. C. O’Keefe, in Asian Studies Association of Australia
Review, vol. 10, no. 3, pp 165-166.
1984. Comparative Ethics in Hindu and Buddhist Traditions, by R. Hindery, in
Journal of Religious History, vol.13, no.2, pp 226-227.

Indo-Chinese Pali
2007. Le Mahākassapanibbāna pāli, editio princeps par Peter Masefield d’après
la translittération des manuscrits conservés dans les collections françaises
déchiffrés par Jacqueline Filliozat. Bangkok: École française d’Extrême-Orient,
Princess Maha Chakri Sirindhorn Anthropology Center.

19
 Obituary

2007. Asokaparinibbānakathā, editio princeps par Peter Masefield d’après la

translittération du manuscrit pāli conservé à l’École française d’Extrême-Orient
(Paris) avec l’apparat critique du manuscrit conservé au Wat Phra Chetuphon
(Bangkok) déchiffrés par Jacqueline Filliozat. Bangkok: École française
d’Extrême-Orient, Princess Maha Chakri Sirindhorn Anthropology Center.
2007. Aṭṭhakesadhātu, A Critical Edition by Peter Masefield based on a transliteration
of a Pāli manuscript housed at the library of the École française d’Extrême-
Orient (Paris) with the variant readings from two other manuscripts in the French
collections and a further manuscript from Wat Phra Chetuphon (Bangkok) by
Jacqueline Filliozat. Bangkok: École française d’Extrême-Orient, Princess
Maha Chakri Sirindhorn Anthropology Center.
2007. Naḷāṭadhātuvaṃsa, A Critical Edition by Peter Masefield based on
transliterations of Pāli manuscripts from European and Asian Collections by
Jacqueline Filliozat. Bangkok: École française d’Extrême-Orient, Princess
Maha Chakri Sirindhorn Anthropology Center.
2007. Vijādharajātaka, A Critical Edition by Peter Masefield based on the
transliteration of four Pāli manuscripts housed in the French Collections by
Jacqueline Filliozat. Bangkok: École française d’Extrême-Orient, Princess
Maha Chakri Sirindhorn Anthropology Center.
2007. À propos de la litanie Iti pi so, aperçu des documents conservés en Europe,
par Jacqueline Filliozat, avec la révision de textes inédits de la Ratanamālā par
Peter Masefield. Bangkok: École française d’Extrême-Orient, Princess Maha
Chakri Sirindhorn Anthropology Center.
2007. Les manuscrits en pāli indochinois du Dibbamanta dans les collections
parisiennes, par Jacqueline Filliozat; révision des notices par Peter Masefield.
Bangkok: École française d’Extrême-Orient, Princess Maha Chakri Sirindhorn
Anthropology Center.

Published Interviews
2007. with Stephen A. Evans, “Found in Translation”, in Tricycle: The Buddhist
Review, https://tricycle.org/magazine/found-translation/
2000. “The Price of Liberation”, in What Is Enlightenment? Issue 18, Fall/Winter,
pp 98ff.
1998. “The Roar of the Timeless Beyond”, in What Is Enlightenment? Issue 14, Fall/
Winter, pp 116ff.

20
 Obituary

Radio Talks
1984. ABC Roundtable Discussion following the assassination of Mrs. Gandhi,
November 1984.
1982. Series Focus on Buddhism, 2-SER-FM The Theravada (2.4.82) Tibetan
Buddhism (16.4.82) The History of Buddhism in Australia (7.5.82) A Hundred
Years of Buddhism in Australia (19.11.82).

21
 The Asokaparinibbānakathā

Peter Masefield† and Jacqueline Filliozat 1

Abstract
‘An Account of Asoka’s Parinibbāna’ (Asokaparinibbānakathā) is a
little known Pali text from mainland South East Asia. The edition and
translation reproduced here are based on one Khom manuscript from
Wat Phra Chetuphon in Bangkok, and one Mūl manuscript, originally
from Cambodia, but now kept at the École française d’Extrême-Orient
in Paris.

Little is known about the provenance of the Asokaparinibbānakathā. The

catalogues2 list only three manuscripts: one of Cambodian origin kept in Paris,
one of Lao origin in Copenhagen, and one of Siamese origin in Bangkok. It
does not receive mention amongst the various recensions of the legend of Asoka
recorded by Strong,3 and as far as is known, no edited or printed edition has
been published to date. The beginning of the text roughly follows the Sinhalese

1
The present work is the final outcome of collaboration, over the years, between the late
Peter Masefield and myself on this text. I was reponsible for the original transliteration of both
manuscripts, prior to their translation by Prof. Masefield. During the course of this work, I was
more than ably assisted by Phra Maha Vanly Khemaraputto, one of Prof. Masefield's former
Cambodian graduate students at Mahachulalongkorn Buddhist University in Bangkok, who
swiftly brought a fresh pair of keen, native eyes to both manuscripts, thereby helping us to clarify
a number of previously uncertain readings. We are both greatly indebted to him for his assistance.
2
G. Cœdès (1966); Skilling and Pakdeekham (2002).
3
Strong (1989).

. 21(20): 22–50. ©21 Peter Masefield & Jacqueline Filliozat

 The Asokaparinibbānakathā

chronicle on Asoka as depicted in the Mahāvaṃsa, but continues and concludes

with quite different episodes involving proper nouns found only in the
Dasavatthuppakaraṇa, Sīhaḷavatthuppakaraṇa and Sahassavatthuppakaraṇa,
none of which are documented in the DPPN.
This edition is based on the transliteration of two of the above manuscripts,
namely, that in the Rāma III Collection, housed in the Santivan library of Wat
Phra Chetuphon (Wat Pho) in Bangkok,4 and that in the library of the École
française d’Extrême-Orient in Paris. We have, to date, had no access to the third
manuscript, that in the Lao language, housed in the Royal Library in Copenhagen,
concerning which we give only a short notice below as to its typology, together
with a few remarks drawn from the Cœdès catalogue (1966: 95).

Description of the manuscripts

1. That of Siamese origin: Wat Phra Chetuphon (WPC) 6/ta.3. 1

Asokaparinibbānakathā (incomplete). Top and last olas decorated with
devatā gilded on black lac. Cartouche in ink on first ola recto: braḥ
Asokaparinibbānakathā. Begins, ola da b;5 ends, ola dū b line 4 (missing
the last ola only). A single phūk, 11 olas, da-dū, 580 x 52 mm, gilded edges,
2 cord holes, 5 lines, 60 characters per line—Khom script—Numb. Khom
letters—covers decorated in Chinese style with inlaid mother-of-pearl floral
and foliage motifs. No date [middle of 19th century]. This manuscript belongs
to the collection known as Deb Jumnum, donated to Wat Phra Chetuphon by
king Rāma III, Phra Nangklao (1824-1851).

4
This library was established according to the wish of Somdet Phra Ariyavamsakatayarn
(Poon Poonasiri Mahathera Barien Dhamma 6), the 17th Patriarch and the 11th Lord Abbot of
Wat Phra Chetuphon. It contains, amongst many other manuscripts, the collection known as Deb
Jumnum, donated by king Rāma III (Phra Nangklao 1824-1851) to Wat Phra Chetuphon.
5
This shows that this phūk formely belonged to a larger series of manuscripts. At present,
the previous phūk, containing olas ka-thaḥ, is missing, our text having been placed amongst
another set of vaṃsa texts having different features and dimensions. Note also that this
Asokaparinibbānakathā is found along with five other texts in the same manuscript under this
shelfmark containing a total of 12 phūk. It comprises: 2. Jinadantadhātuvaṃsa [Dāṭhāvaṃsa];
3. Pāḷisāvakanibbāna; 4. Aṭṭhakesadhātuvaṃsa; 5. Nalāṭadhātuvaṃsa; 6. Dantadhātunidāna.

23
 The Asokaparinibbānakathā

2. That of Cambodian origin: École française d’Extrême-Orient, EFEO

PALI 29
Asokaparinibbānakathā. Ola 1a, middle: Asokaparinibbānakathā | Margin,
in ink in Roman characters: Asokaparinibbānakathā (pāli); olas 1b-4a, bl.
Begins, ola 4b (da); ends, ola 10b de line 3; olas 11-12, bl. a single phūk, 12
olas, da-de, 580 x 60 mm, 2 cord holes, 5 lines, 55 characters per line—Mūl
script—Numb. Mūl letters. No date (probably copied circa 1914-1915 for
EFEO in Cambodia)—Former shelfmark: CO 225.

3. That of Lao origin. Royal Library, Copenhagen LAOS 79 (V)

Cœdès (1966: 95) describes the Asokadhammarājanibbāna, giving the
beginning and end of the text in Pāli with a summary of the Lao nissaya, as
follows (translated freely from the French):
Asokadhammarājanibbāna. Manuscript on 18 palm leaves,
51.5 x 5 cm, 4 lines on 45 cm. Writing quite accurate, good
preservation. Formerly belonging to Văt Pā Ḍēṅ. Provenance:
Tuxen collection Laos 2. Story of the passing of king Asoka. The
text starts with a passage in incorrect Pāli: pañcālāvattasampattiti
kuṇḍalattābyaghāth[e]rānaṃ nidinakathā ime panāyasmante
pubbabuddhānaṃ santike umpacittakusalamulā devesu ca
mānussesu va saṃsaranto amhākaṃ bhagavato parinibbānato
aṭṭhārasāthikānāddhinaṃ vassasatānaṃ tamatthake jammbudīpe
ca imasmim laṅkādīpe ca ma[hā]dhammasoka devatānaṃpiyatissā
ti laddhanāmā senā adiṭṭhamahāyābhātvā buddhasāsane
mahantaṃ puñarassī sañcinantādhasisu.
(Translation of the summary of this nissaya): Kuṇḍalatissa and
Byāghatthera, having accomplished meritorious acts in the
presence of the Buddhas of the past, and after having transmigrated
in the world of devas and humans, were reborn in Jambudīpa
and Laṅkādīpa 218 years after the nibbāna of the Lord, where
they received, respectively, the names of Mahādhammāsoka and
Devanāmpiyatissa, accumulating good deeds in the Buddhist
religion. Colophon (chī r°4): asokadhammarājanibbān gå lēv kōr
lē (the Asoka Dhammarājanibbāna is complete).

24
 The Asokaparinibbānakathā

The present transliteration and translation are those of the Wat Phra
Chetuphon (WPC) manuscript, with variant readings contained in the École
française d’Extrême-Orient manuscript (EFEO) given in the footnotes. The
foliation of both manuscripts is virtually identical, showing that both derive
from a common exemplar; the foliation of WPC is indicated by pointed brackets
and normal font size (e.g. <da a> = folio da, recto); the foliation of EFEO is
shown by square brackets and a reduced font size (e.g. [da b] = folio da, verso).
An Appendix is also given, quoting a partial, parallel account of the episode by
Buddhaghosa in his Samantapāsādikā (CSCD edition).

25
 The Asokaparinibbānakathā

braḥ6 asoka parinibbāna kathā

|<da b>|[da b] tena kho pana samayena pātaliputtanagare

bindusāro7 nāma rājā rajjaṃ kāresi | tassa ekasataputtā ahesuṃ |
asoko attanā saddhiṃ ekamātikaṃ tissakumāraṃ thapetvā sabbe te
ghāṭetvā cattāri vassāni anabhisito rajjaṃ kāretvā sakalajambūdipe
ekarājābhisekaṃ pāpuṇi |
[dā a] abhisekānubhāvena8 cassa imā rājiddhi <dā a> yo āgatā |
pathaviyā heṭṭhā yojanappamāne aṇāpavattati tathā9 upari ākāse
| anotattadahato aṭṭhahi kājehi soḷasapāṇiyaghaṭe divase divase
devatā āharanti |
aṭṭhaghaṭe bhikkhusaṃghassa10 adāsi | dve ghaṭe saṭṭhimattānaṃ
pitakataya11 [dā b] bhikkhusahassānaṃ dve <dā b> ghaṭe
aggamahesiyyā12 asandhamittāya | dve ghaṭe soḷasannaṃ nāṭakitthī
sahassānaṃ dve ghaṭe attanā paribhuñji |

6
EFEO omits.
7
EFEO bindhasāro.
8
EFEO abhisekānubhavena.
9
EFEO āṇāpāpavatti tathā; possibly āṇāpāpavattito.
10
EFEO bhikkhasaṃghassa.
11
EFEO piṭakataya.
12
EFEO aggamahesiyā.

26
 The Asokaparinibbānakathā

An Account of Asoka’s Parinibbāna

Now, on that occasion, the king named Bindusāra was ruling in

the city of Pātaliputta. And one hundred and one sons were his.
Asoka after slaying all of these, save for Tissakumāra, who had the
same mother as he did, ruled unconsecrated for four years, and then
became consecrated as sole king over the whole of Jambudīpa.
And13 through the majesty of his consecration, the royal potencies14
befell him; below the earth, his authority extended a yojana,
likewise in the sky above;15 devatās would each day fetch sixteen
pitchers of drinking-water from Lake Anotatta on eight pingoes.
(Of these) he would give eight pitchers to the order of monks,
two pitchers to as many as sixty thousand monks knowing the
three Piṭakas, two pitchers to his chief-queen, Asandhamittā, two
pitchers to his sixteen thousand dancing-girls, whilst he himself
would consume two pitchers.

13
There are a few places in the text where certain readings are either difficult or impossible
to understand or reconstruct. The fact that these are more or less common to both mss suggests
an early scribal confusion that must have occurred prior to either of our mss. This difficulty is
compounded by (a) the close similarities between certain graphemes, such as t/g and p/m in
the Khom script (especially when inscribed on a palm-leaf); and (b) the frequent manner in
which scribes tend to spell a Pali term phonetically in accordance with the local pronunciation,
suggesting in turn that some ‘copying’ might actually have been performed by way of dictation.
For a discussion of these idiosyncrasies in Khom mss see, for instance, Masefield (2008). I have
done my best trying squeeze out some sort of sense the original passages may have had, often with
little or no success. Under the circumstances, I have simply recorded the troublesome passages in
a footnote without comment, other than ‘unclear.’
14
At M III 176, it is said that the cakkavatti: (i) is handsome, comely, and graceful, possessing
the supreme beauty of complexion, and surpasses other human beings in this respect; (ii) lives
long and endures long, and surpasses other human beings in ths respect; (iii) is free from illness
and affliction, possessing a good digestion that is neither too cool nor too warm, and surpasses
other human beings in this respect; (iv) is dear and agreeable to brahmins and householders.
15
So the punctuation of WPC; Sp (CSCD) punctuates somewhat differently.

27
 The Asokaparinibbānakathā

himavante kira nāgalatā nāma dantakaṭṭhaṃ athi saniddhaṃ16

sādhukaṃ ojavantaṃ divase divase devatā āharanti |
aggadāmalakaṃ aggadāharitakaṃ ambapakkaṃ ca suvaṇṇavaṇṇaṃ
gandharasasampannaṃ devatā āharanti |
chaddantadahato17 pañcavaṇṇanivāsanaṃ devatā āharanti18 | taṃ
nivāsanaṃ adhovitabbavatthaṃ nāma | yadā raño nivāsanakāle
taṃ vivaṇṇaṃ hoti | tadā aṅgārarāsimhi19 tāpitaṃ suparisuddhaṃ
cupasahadavaṇṇaṃ20 hoti |
chaddhantadahato21 sañjātasāliyo navavāhasahassāni sucagaṇā22
āharanti | divase divase mūsikā nitthusakaṇena23 karonti | eko pi
khaṇḍataṇḍulo nāma nāhosi | ayam eva taṇḍulo raño paribhogaṃ
gacchati |
haritacandanaṃ himavantato24 devatā āharanti |[di a] rājā dhammaso
<di a> ko evarūpo mahānubhāvo ahosi |
gate cavanakālantare āyumhi khayamāne dānaṃ dātukāmo
bhaṇḍagārikaṃ pakkosāpetvā āha gaccha tāta koṭṭhāgārato
suvaṇṇādiratanāni khomakoseyyakādīni25 idha rāsiṃ karohi dānaṃ
dassāmi26 |

16
EFEO aṭisiniddhaṃ.
17
EFEO chandhantadahato.
18
EFEO omits.
19
EFEO aṅgāravāsimhi.
20
EFEO dhumasamāvaṇṇaṃ.
21
EFEO chandhantadahato.
22
EFEO suvagaṇā.
23
EFEO nitthusakaraṇena.
24
EFEO hivantato.
25
EFEO khomakoseyyakādīni vatthāni.
26
EFEO dassāmī ti.

28
 The Asokaparinibbānakathā

It is said that there is, on the Himalaya, the tooth-stick named

Nāgalatā,27 which is smooth,28 agreeable29 and possesses nutritive
essence; this the devatās would bring each day. The devatās would
fetch an antidotal myrobalan,30 an antidotal gallnut31 and a ripe mango
that was golden-coloured and endowed with fragrance and flavour.
From Lake Chaddanta the devatās would fetch a five-coloured,
lower-garment; this lower-garment was not in need of being
washed. Whenever it became faded at such time as the king was
wearing it, it would, when heated over a heap of embers, become
utterly purified, of a colour equal to that of the paduma(-lotus).32
Each day, flocks of parrots would fetch nine thousand cartloads of
sāli rice that was in season33 from Lake Chaddanta; mice would
divest it each day of its husks and powder.34 Not even a single grain
became broken. This same grain had been for the king’s use.
The devatās would fetch yellow sandal from the Himalaya. Such
was the great majesty of king Asoka. And when he became desirous
of giving alms, at the time his lifespan was waning during the phase
in which he would die, he had the storekeeper summoned, and said:
‘You should go make a heap here, my dear, of gold and jewels and
so on, and linen and silk and so forth, from the warehouse—I will
give alms.’

27
nāgalatā; the ironwood tree.
28
saniddhaṃ.
29
sādhukaṃ; Sp (and elsewhere) mudukaṃ, pliant.
30
phyllanthus emblica.
31
aggadāharitakaṃ (where aggadā is to be read as agada, ‘medicine, antidote’). This is not
the yellow myrobalan (terminalia citrina or chebula), as stated by PED sv harītakam, but the gall
nut with a hard shell and about the size of a nutmeg (Sinh. araḷu); the myrobalan (āmalaka) is
smaller, green and smooth, about the size of a medium-sized grape (Sinh. nelli). Both have a hard
stone inside. I am grateful to N. A. Jayawickrama for this information.
32
cupasahadavaṇṇaṃ; EFEO dhumasamāvaṇṇaṃ. Given the similarity of the graphemes c,
d, dh, and ph in the Khom script, I conjecture, especially on the basis of EFEO, that the original
reading was, in all probablity, padumansamavaṇṇaṃ—cp. padumasamaṃ at Mil 354 and
padumasamānavaṇṇatāya at Vv-a 35. If so, the error must pre-date both of the mss presently at
our disposal. Sp reads differently at this point.
33
sañjāta; Sp uṭṭhita.
34
nitthusakaṇena; Sp nitthusakaṇe.

29
 The Asokaparinibbānakathā

taṃ sutvā bhaṇḍāgāriko assupuṇṇehi nettehi rājānaṃ sañāpento

imaṃ gātham āha |
koṭṭhāgārañ ca kosañ ca
yañ c’ añaṃ atthi35 te dhanaṃ
sabbaṃ sulañcitaṃ deva
evaṃ jānāhi khattiyā ti |
taṃ sutvā rājā bhaṇḍāgārikaṃ pucchanto āha |
koṭṭhāgārañ ca kosañ ca
yañ c’añaṃ atthi dhanaṃ me
kena sulañcitaṃ sabbaṃ
taṃ me akkhāhi pucchito ti |
bhaṇḍāgāriko kathento āha |
koṭṭhāgārañ ca kosañ ca
yañ c’ añaṃ atthi te dhanaṃ
devaṃ dubbalaṃ maggaṃ disvā
sabbaṃ amhehi lañcitan ti |
taṃ sutvā rājā36 amacce pakkosāpetvā37 pucchi |[di b] tumhehi <di b>
tipa ma dānaṃ dātukāmassa38 koṭṭhāgārāni lañcitānī ti | evaṃ devā ti |
kim atthāyā ti | pacchīmassa raño atthāya devā ti | idaṃ kassa rajjan ti |
tumhākaṃ devā ti | yadi me rajjaṃ kathaṃ dānaṃ39 dātuṃ na dethā ti |
amaccā tuṇhi ahesuṃ

35
EFEO atti.
36
EFEO taṃ svā sutvā rāja.
37
EFEO pakosāpetvā.
38
EFEO tumhehi kira me dānaṃ dātukāmassa.
39
EFEO dāṃnaṃ.

30
 The Asokaparinibbānakathā

Upon hearing this, the storekeeper uttered this verse, informing the
king with tear-filled eyes:
‘The warehouse and the storeroom, and any other treasures of
yours there be, are all well-sealed, your majesty; please be aware
that this is so, O khattiya.’
Upon hearing this, the king, questioning the storekeeper in turn,
said: ‘The warehouse and the storeroom, and any other treasures
of mine there be, by whom have they all been well-sealed ? Inform
me of this when questioned.’
The store-keeper, replying, said: ‘The warehouse and the storeroom,
and any other treasures of yours there be, all have been sealed by
us, your majesty, upon seeing your own weakness.’40
The king, upon hearing this, had his privy councillors summoned,
and then asked: ‘It is said that the warehouses of mine, who am
desirous of giving alms,41 has been sealed by you.’
‘That is so, your majesty.’
‘For what purpose?’
‘For the sake of the next king, your majesty.’
‘Whose kingdom is this?’
‘Yours, your majesty.’
‘If this is my kingdom, why do you not allow me to give alms?’
The privy councillors became silent.

40
Reading deva for devaṃ, and dubbalam attaṃ for dubbalaṃ maggaṃ; as noted above
(n. 13), the Khom characters ta and ga are easily confused.
41
Reading EFEO kira me for WPC tipa ma: ki → gi (aural confusion) followed by gi → ti
(orthographic confusion); and ra me → pa ma (orthographic confusion).

31
 The Asokaparinibbānakathā

tasmiṃ kāle raño jālahatthe osatthatthāya aḍhḍhāmalakaṃ atthi |

rājā oloketvā gātham āha |
paccaṇavatasampatti42
vijjumālisamupamā43
mahaddhanaṃ pi sampatto
dalidhaṃ44 punamāgato
yadā me dātukāmo pi
dātabbaṃ natthi me dhanaṃ
khīṇapuñaṃ45 ahaṃ patto
sampatto maccuno mukhan ti |
tadā raño samīpe tikicchānatthāya eko vejo thito ahosi | so rājānaṃ46
byādhī pi dukkhataraṃ sampatti bhassamānaṃ disvā gātham āha |
dhīratthu47 rajjam aniccarūpaṃ
bhuñjantā jivhāliṅgatāmini
[dī a] ciratthutaṃ48 jivataṃ49 pata <dī a> m āhu
ussāvabinda va tiṇatamhi
mahādhano narindo pi
asoko50 rājakuñjaro
jambudīpissaro51 hutvā
aḍhḍhāmalakissaro ti |
taṃ sutvā bālhagilāno hutvā kodhavasena khaggaṃ gahetvā kosato
nikaḍhḍhitvā ukkhipitvā taṃ paharitukāmo52 |

42
EFEO paccalāvatasampatti. PED: paccana: boiling; āvāṭa: pit.
43
EFEO vijjumālīsamupamā.
44
EFEO daliddhaṃ.
45
EFEO khiṇapuñaṃ.
46
EFEO rājā.
47
EFEO dhiratthu.
48
EFEO dhiratthutaṃ.
49
EFEO jivitaṃ.
50
EFEO asokako.
51
EFEO jambudipissaro.
52
EFEO haritukāmo.

32
 The Asokaparinibbānakathā

At that moment, there was half a myrobalan for laxative purposes

on the palm of the king’s hand. Surveying this, the king uttered
this verse:
‘Reaching the burning pit, which is likened to a garland of lightning,
even though endowed with great treasure, I am come once more to
poverty.
‘Even though I am desirous of giving, there is no treasure of mine
that can be given; having reached my merit’s destruction, I have
arrived at Death’s jaws.’
At that time, there was a physician stationed in the king’s vicinity for
the purpose of curing him. Upon seeing the king drooping, having
met with a sickness of even greater dukkha, he uttered these verses:
‘Woe be those enjoying this reign, of an impermanent nature,
and now exhibiting no sign of life;53 this life,54 worshipped as
long-lasting, has (now) fallen like a dew-drop on the tip of a
blade of grass.55
‘Asoka, the royal elephant, one of great wealth, as well as lord
of men, having been ruler of Jambudīpa, is now ruler of half a
myrobalan.’
When he heard this, the king became severly sick, angrily seized
his scabbard, hauled it out of its sheath, and then raised it aloft,
desiring to strike him.

53
jivhāliṅgatāmini (unclear). Possibly jīva-aliṅga-gatāvin?
54
Reading EFEO jivitaṃ.
55
Reading tiṇag(g)amhi for tiṇatamhi; on the change k → g → t, see n.13 and n.40.

33
 The Asokaparinibbānakathā

tasmiṃ khaṇe nirassāso asoko ahosi | kodhavasena upapātiko

ajjaggaro hutvā ekadonikanāvappamāṇo56 hutvā balavasena gantvā
bhogehi parikkhipitvā satte vātetvā57 māreti |
yadā satte58 na labhi aciravatiyā attānaṃ olambi phaṇena59 udakaṃ
paharitvā macchakacchape khādi |
tena kho pana samayena āyasmā mahāmahindatthero buddhasāsane
sakapatiladdhajjhānasukhaṃ patiladdhajjhānasukhaṃ bhuñjanto
tamhā vuṭṭhāya kena dinnaṃ jjhānanasukhan60 ti cintetvā
asokamahājapatinā dinnan61 ti disvā kuhiṃ me <dī b> thanato62
etarahi [dī b] supadhārento63 kodhavasena64 kālaṃ katvā
ajagarayoniyaṃ65 nibbattibhāvaṃ disvā
aho dujātiko yatra hi nāma evarūpaṃ caturāsīti vihārasahassaṃ
caturāsīti cetiyassa patimaṇḍitaṃ kāretvā mahantaṃ dānaṃ
pariccāgaṃ katvā appamattakena maraṇāsannakodhena maraṇaṃ
patvā tiracchānayoniyaṃ66 nibbato | ki67 me kattabbaṃ| saccena
upekkhayyaṃ nassati | dāni yannunāhaṃ68 mahārājassa patisaraṇaṃ
karissāmī ti |
turitagamano yena rājā tenupasaṅkami | theraṃ upasaṅkamitvā
attano samipe thitaṃ disvā rājā susū ti kurumāno69 yena thero tena
dhāvati |

56
EFEO ekadoṇikanāvappamāṇo.
57
EFEO sagge pātetvā.
58
EFEO sagge.
59
EFEO phaṇe.
60
EFEO jjhānasukhan.
61
EFEO asokamahārājapatinādinnan.
62
EFEO janato.
63
EFEO etarahī ti upadhārento.
64
EFEO etarahī ti upadārento kodhavasena.
65
EFEO jagarayoniyaṃ.
66
EFEO tiricchānayoniyaṃ.
67
EFEO kiṃ.
68
EFEO yantūnāhaṃ.
69
EFEO karumāno.

34
 The Asokaparinibbānakathā

But at that instant Asoka became breathless and, owing to his

anger, arose spontaneously as a python70 the size of a boat of
one-doṇika.71 He went about forcefully, causing creatures to
fall72 by encircling them with his coils before killing them.
When he could not get any creatures, he hung his body down into
the Aciravati, struck the water with his hood73 and then devoured
fish and turtles.
Now, on that occasion, the venerable elder Mahāmahinda, whilst
enjoying the bliss of jhāna he had attained—the bliss of jhāna he
had himself attained during this Sāsana of the Buddha—emerged
therefrom, wondering who had granted him that bliss of jhāna.
Upon seeing that it had been given by his lord, the great king
Asoka, and reflecting as to where his father74 might be now, he saw
that, owing to his anger, he had finished his time and had come into
being in the python-womb, whereupon he thought:
‘Oh dear, he is of such bad birth; for although he performed such a
great gift consisting of generosity, in that he had built eighty-four
thousand vihāras adorned with eighty-four (thousand) cetiyas, he has
now, owing to such trifling anger when close to dying, come into being
in the animal-womb—what can I do? His peace of mind has truly been
destroyed. What if I were to act as a haven for the great king?’
The elder hastily approached the king. Once he had approached,
the king, seeing him standing in his vicinity, slid towards the elder,

70
ajjaggaro; cp ajagarayoniyaṃ below. According to PED, ajagara (literally ‘goat-swallower’)
denotes a boa constrictor, whereas CPD claims it denotes a python. But female pythons are
oviparous (lay eggs), which sets them apart from the boa family, most of which are ovoviviparous
(bear live young). In neither case do they have hoods (phaṇa), despite the claim below.
71
A doṇa is of uncertain measure.
72
vātetvā; EFEO pātetvā.
73
phaṇena.
74
thanato; EFEO janato. I assume the latter is in error for janako, father. Moreover, given
the confusion surrounding the graphemes t and g in the Khom script, it is quite possible that the
original EFEO reading was janago which, given the Thai alternation in the pronunciation between
g/k, might well have represented an original janako.

35
 The Asokaparinibbānakathā

thero rājānaṃ āgacchantaṃ disvā karuṇāvasena mahārāja nanu

pubbe dhammasoko ti nāmaṃ75 dhāresi | pubbe sucaritaṃ suṇāhī
ti vatvā gātham āha |
racchantaṃ patipannassa
kīlanto paṃsu dārako
paṃsudānānubhāvena
[du a] dhammaso <du a> ko ti visuto76
balacakkavattiladdho
jambūdipe mahāyaso
heṭṭhābhāgaṃ yojanaṭṭhānaṃ77
khuddhaṃ āṇāpavattikā
na sakkoti gaṇetuṃ va yasa78 koci mahiddhiko | soḷasāni yadāni
āhanti79 ca devatā nacce kiṇṇarā pakkhite māñe madhupakkhikā
khundharāsuvakā80 sabbe sattā yakkhā ca devatā tava
puñānubhāvena vasaṃ gacchanti |
attano caturāsīti sahasse ca vihāre thūpamaṇḍite puñaṃ anappakaṃ
katvā taṃ idāni kataṃ saraṃ | saṃghamittaṃ mahindañ81 ca pabbājesi
tav’ atrajaṃ sāsane tvañ ca pubbe vataṃ idāni kataṃ saraṃ |
yo mahindo tava putto
so ahaṃ āgato idha |
pitā mayhaṃ mahārāja |
putto sohaṃ tavatrajo |

75
EFEO nāma.
76
EFEO vissuto.
77
EFEO yojaṭṭhānaṃ.
78
EFEO vayasaṃ.
79
EFEO soḷasānighadhāni āharanti.
80
EFEO kiṇṇapakkhite hañe madhumakkhitā khandavāsuvakā.
81
EFEO mahinañ.

36
 The Asokaparinibbānakathā

making a hissing sound. Upon seeing the king coming, the elder,
out of compassion, said: ‘Surely, great king, you formerly bore the
name of Asoka,’ and then, after saying: ‘Please hear of your former
good conduct,’ uttered the following verse:
‘Whilst entering upon the edge of the carriage road, as a child
playing at making mud pies, you became renowned as Dhammasoka
through the majesty of that gift of mud.
‘Having become a powerful Cakravartin, you possessed great fame
in Jambudīpa; over an area of one yojana, in the section below,
angry, you exercised authority.
‘No one, not even one of great potency, is at all able to calculate
your fame;82 whilst devatās brought sixteen pitchers,83 dancing-
girls, kiṇṇarās and mountain parrots, their wings, methinks,
smeared with honey.84 All beings, yakkhas and devatās went under
your sway through the majesty of your merit.
‘You created eighty-four thousand vihāras, each adorned with
a thūpa—not trifling is that merit of yours; you should now
remember85 that which was done.
‘You had Saṅghamittā and Mahinda, your own born, go forth in
the Sāsana, and you should now remember that former observance.
‘This same I, who have now come here, am your son, Mahinda;
you, great king, are my father—I am your own-born son.

82
Reading EFEO va yasaṃ.
83
EFEO soḷasāni ghadhāni āharanti. Following EFEO and reading ghaṭāni for ghadhāni,
through confusion of the Khom characters ṭa and dha; and reading āharanti, assuming loss of ra
after the sequence of similarity between the Khom characters ā - ha - ra.
84
WPC nacce kiṇṇarā pakkhite māñe madhupakkhikā khundharāsuvakā; EFEO kiṇṇapakkhite
hañe madhumakkhitā khandavāsuvakā. Unclear, but proposing kandarā, ‘mountain’ for khandavā/
khundharā (k → kh, auditory confusion): kandarā suvakā = ‘mountain parrot’; and reading
pakkhike māñe madhu-makkhikā for WPC pakkhite māñe madhupakkhikā, ‘smeared with honey
on their wings’: pakkhike → pakkhite: (also see n.13, 40 for the change k → g → t) and makkhitā
→ pakkhikā (m → p and t → g, orthographic confusion)
85
saraṃ.

37
 The Asokaparinibbānakathā

pāpakammaṃ mahārāja
sāvajjaṃ satthārā desitaṃ |
kodhaṃ pakataṃ deva
kodhadosaṃ kāsitaṃ86
andhakāro87 ayaṃ loko
kodho dhumedhagocaro88
kodha [du b] nā pari <du b> muñcitvā
abbhāmutto89 va candimā |
andhakāro ayaṃ loko
kodho dumedhagocaro90
kodhanā parimuñcitvā
dukkhass’ antaṃ karissati |
rājā therassa dhammakathaṃ sutvā ayaṃ me putto mahindathero91
idh’ āgato ti assuparipuṇṇayano onasi so therassa pādamūle sīsaṃ
thapetvā bhusaṃ92 rodi |
thero taṃ assāmetvā93 mā bhāyi mahārāja tav’ atthāya mahārāja
idh’ āgato ’mhi | gaṇhāhi saraṇāni pañcasīlāni cā ti | rājā tathā akāsi
| thero raño tisaraṇāni pañcasīlāni ca datvā
jīvahetu94 pi deva anattikkamitabbaṃ mano padoso na95 kātabbo ti
ovādaṃ datvā cintesi ayaṃ rājā buddhasaraṇaṃ96 gato dhammaṃ
saraṇaṃ97 gato98 alam eva sugatiṃ gantun ti sakadānam99 eva gato |

86
EFEO kāsitaṃ.
87
EFEO andakāro.
88
EFEO dumedhakocaro.
89
EFEO abbhāputto.
90
EFEO dumedhakocaro.
91
EFEO mahindatthero.
92
EFEO bhūsaṃ.
93
EFEO assāsetvā.
94
EFEO jivihetu.
95
EFEO omits.
96
EFEO buddhaṃ saraṇaṃ.
97
EFEO dhammasaraṇaṃ.
98
EFEO adds saṃghaṃ saraṇa gato.
99
EFEO sakaṭṭhānam.

38
 The Asokaparinibbānakathā

‘The Teacher, great king, taught that evil deeds are blameworthy;
anger has been produced, your majesty—anger and hatred are
terrifying.100
‘This world is blind; anger is foolish pasture. You should rid yourself
of anger, as does the moon become free of the thunder-cloud.
‘This world is blind; anger is foolish pasture. The one who is
released from anger will make an end of suffering.’
Then, upon hearing the elder’s Dhamma-talk, the king bent down,
his eyes full of tears, realising: ‘This is my son, the elder Mahinda,
who has come here.’ He placed his head at the soles of the elder’s
feet, and wept bitterly.
The elder consoled101 him, saying: ‘Do not fear, great king; I have
come here for your sake, great king. You should take the refuges
and the five precepts.’ The king did as he said. The elder gave the
king the refuges and the five precepts, and exhorted him, saying:
‘These are not to be transgressed, even for the sake of your life,
my Lord: you should not make your heart one of anger,’ and then
thought: ‘This king has gone to the Buddha as refuge, has gone to
the Dhamma as refuge—this alone should be sufficient for him to
go to a happy destiny,’ and then went back to his own place.

100
Text kāsitaṃ. However, I suspect that the original reading may have been tāsitaṃ (which
I adopt), which was initially misread as gāsitam (orthographic confusion between t- and g-), and
then misprononunced as kāsitaṃ, by a reader dictating to a scribe (aural confusion).
101
Reading assāsetvā with EFEO for text’s assāmetvā.

39
 The Asokaparinibbānakathā

rājā gocaraṃ agahetvā parisuddhasīlaṃ102 rakkhitvā

sattadivasabbhantare yeva [dū a] kālaṃ katvā tāvatiṃsabhavane
<dū a> mahindiko103 devaputto104 hutvā dibbasampattiṃ
anubhavati |
aciraṃ vasitvā sagge katapuñonathodito105 pabbajjitvā106
nibbāyissaṃ iti cintesi | yoniso saggato cavitvā rājā laṅkādīpamhi
āgato anukkamena vayappatto kuṇḍalatisso pākaṭo |
vasitvā sagge katapuñena codito pabbajjitvā107 nibbāyissaṃ108 iti
pi cintesi devānaṃpiyatisso pi katvā puñaṃ anappakaṃ jahitvā
manussadehaṃ devaloke ajāyatha aciraṃ |
yoniso saggato cavitvā rājā laṅkādīpamhi āgato byaggho iti
ca nāmena laṅkādīpamhi pākato | kuṇḍalatisso byaggho ca
piyasahāyakā c’ ubho samānachandhā sapañā kusalesu samārattā109
jahitvāna sakagehaṃ pabbajjiṃsu buddhasāsane |
tesu ca kuṇḍalatisso
arañavāsiko ahu |
vasitvā lohapāsāde110
byaggho vācesi bhikkhūnaṃ111 |
ca <dū b> [dū b] te therā mahāpañā āsuṃ jinasāsane | tato tu
cirakālena mahātherassa tassa turogo uppajji | so rogaṃ oloketvā
na jivitaṃ sakaṃ āyukkhayaṃ112 ñatvā evaṃ adhiṭṭhahi | tadā
mayhaṃ nibbānakāl’ amhi

102
EFEO parisuddhasilaṃ.
103
EFEO omits.
104
EFEO devaputo.
105
EFEO katapuñena thodito.
106
EFEO pabbajitvā.
107
EFEO pabbajitvā.
108
EFEO nibbānayissaṃ.
109
EFEO kusalesu samārattā.
110
EFEO lohapāsāda.
111
EFEO bhikkūnaṃ.
112
EFEO āyukkhaya.

40
 The Asokaparinibbānakathā

The king, without adopting that pasture, guarded the precepts,

keeping them perfectly pure, and within an interval of a mere seven
days, finished his time and became the devaputta Mahindika in the
realm of the Thirty-three, experiencing heavenly excellence.
After dwelling but a short time in heaven, he was urged on by the
merit he had performed, thinking: ‘I should go forth and attain
nibbāna.’ Upon falling from the heavenly womb, the king came
to the island of Laṅkā, and in due course came of age, being well-
known as Kuṇḍalatissa.
Devānaṃpiyatissa also performed not trifling merit, abandoned
the human frame, and was born in the devaloka. After dwelling
but a short time in heaven, he was urged on by the merit he had
performed, also thinking: ‘I should go forth and attain nibbāna.’
Upon falling from the heavenly womb, the king came to the island
of Laṅkā, being well-known in the island of Laṅkā as Byaggha,
‘Tiger’. Both Kuṇḍalatissa and Byaggha became dear companions,
being of the same resolve, possessing insight and, having
undertaken113 things sound, abandoned their own home and went
forth in the Buddha’s Sāsana.
And, of these, Kuṇḍalatissa was a forest-dweller, whereas Byaggha
dwelled in the Lohapāsāda where he taught the monks.
And these elders were of great insight where the Sāsana of the
Conqueror was concerned; after a long time, however, a severe
illness arose for the great elder. He surveyed his illness, realised
that he had no life (left), that his own lifespan had terminated, and
then resolved as follows:
‘I have reached the time of my nibbāna.

113
Reading EFEO samāraddhā for text’s samārattā.

41
 The Asokaparinibbānakathā

jhāpanatthāya maṃ janā114

kūṭāgāre115 va thapetvā
pūjākammaṃ karissare
tadā mayhaṃ sarīro pi
tesaṃ hatthā vimuñcissa
kūṭāgārena116 sahito
abbhugantvāna gagalaṃ117
nabhasā118 rājagehaṃ119 va
gantvā rañābhivandito120
nivattitvā tappadesā
gantvāna nabhasā lahu
mahābyagghamha121 therassa
mahāyassava santikaṃ
mahārājanaparivāro122
tiṭṭhatu ambare tadā |
gato sahāyassa therassa kūṭāgārena123 me saha ākāsena nivattitvā
cittakamhi patiṭṭhatu124 | evaṃ katvā adhiṭṭhānaṃ nibbuto125
mahiddhiko | tato janā sabbe samāgamma samantato tassa sarīraṃ
pūjento sakaronto va sādhukaṃ126 mañjūse pakkhipitvā kūṭāgāre
va127 thāpayuṃ mahā [de a] tā parivārena nayiṃsu128

114
EFEO jarā.
115
EFEO kuṭāgāre.
116
EFEO kuṭāgārena.
117
EFEO bhagalaṃ.
118
EFEO nabhasa.
119
EFEO rājagahaṃ.
120
EFEO raño abhivandito.
121
EFEO mahābyagghavha therassa mahāyass’ eva.
122
EFEO mahājanaṃ parivāro.
123
EFEO kuṭāgārena.
124
EFEO tiṭṭhatu.
125
EFEO nibbato.
126
EFEO sāvadhukaṃ.
127
EFEO kuṭāgāre ’va.
128
EFEO parivārenayiṃsu; the last ola, <dū de>, in WPC is missing; hereafter the reading is
that of EFEO.

42
 The Asokaparinibbānakathā

‘Folk will place me in a pinnacled-house in order to cremate me,

and then perform an act of worship.
‘Then my body will be released from their hand; it will rise up into
the sky, together with the pinnacled-house.
‘It will go through the clouds to the king’s household and be
honoured by the king; then it will quickly return through the clouds
from that region;
‘And go into the presence of the great elder, the famous
Mahābyaggha; surrounded by the people,129 may it remain at that
point in the sky.
‘At that time, it should go, together with my pinnacled-house, to
my friend, the elder, return through the sky and become established
on the funeral pyre.’
After making that resolution in that way, the one of great potency
attained nibbāna. After that, folk from all around assembled,
worshipping and honouring his body and then placed it respectfully
in the casket, put this in the pinnacled-house, and then conducted
it with a great retinue.

129
Reading EFEO mahājana- for mahārājana-.

43
 The Asokaparinibbānakathā

cittakaṃ sampatto cittāsanne kuṭāgāro patiṭṭhahi asakontā va

cāletuṃ | sabbe janā samāgatā santaṭṭhā hatthato mutto kuṭāgāro
tadā pana abbhugantvā
gantvāna nabhasā kato
rājagahassa purato
patiṭṭhāsi nabhatalaṃ |
tato ca so rājā saddhātisso ti nāmako
nikkhamitvā sakā gehā
orodhaparivāritto
vanditvā pūjayitvāna
sabbadosaṃ khamāpayi |
tasmiṃ divase tato | so ca nivattitvā ākāsen’ eva laṅghi | so
mahābyagghatherassa vasanaṭṭhānam āgamā | rājā taṃ anubandhi
pacchato pacchato ca so |
tadā so byagghathero lohapāsāde bhikkhusaṃghassa dhammaṃ
vācesi | taṃ khaṇe kuṭāgāro nādure ākāse vaggiko ahu |
taṃ so byagghathero pi sutvā saddhaṃ mahantaṃ kim eso mahato
ghoso | āgato ko nu kho idha
iti vutte tuṃ taṃ theraṃ bhikkhū ārocayiṃsu te bhante tuyhaṃ
sahāyo ca kuṇḍalatisso ti vi [de b] suto thero | so disanto tuyhaṃ
santikam āgato | tasmiṃ kho mahato ghoso | mahārājā pi āgato |
tesaṃ vacanaṃ sutvā mahāthero mahiddhiko labhitvā
dhammasaṃvegaṃ evaṃ cintesi

44
 The Asokaparinibbānakathā

When it reached the funeral pyre, the pinnacled-house stood firm

on the funeral pyre platform, unable to move. All the folk who
had assembled were overjoyed. The pinnacled-house was released
from their hands.
Moreover, it then rose up, went through the sky, and then went and
stood firm at cloud-level in front of Rājagaha. Then the king named
Saddhātissa,
He emerged from his own household, surrounded by his harem, he
saluted it and worshipped it, and apologised for all his faults.
On that day, it returned from there, lurching through the sky. It
went to the dwelling-place of Mahābyaggha, the elder. The king
followed on behind.
At that time, Byaggha the elder, was teaching monks to recite the
Dhamma in the Lohapāsāda. At that moment, the pinnacled-house
was hovering130 nearby131 in the sky.
Byaggha the elder heard that great sound132 and then said: ‘What is
that great noise? Who can have come here?’
After he had spoken, the monks informed the elder, saying: ‘This,133
bhante, is your friend, the well-known elder Kuṇḍalatissa; he is
seen to have come into your presence. That great noise has to do
with this. The great king himself has come!’
Upon hearing what they had to say, the great elder, being of great
potency, gained Dhamma-shock and then thought as follows:

130
vaggiko, Skt vyagra.
131
nādure.
132
saddhaṃ mahantaṃ.
133
te.

45
 The Asokaparinibbānakathā

so tadā eso thero atite pi idāni vāci me sahadaḷhamitto ahu so so pi

idhāni parinibbuto mayhaṃ āyusaṃkhāro kidiso ti
passiya imasmiṃ divase yeva āyu khiṇan ti adassa | tato mahiddhiko
thero kuṭāgārādayo pamataṃ samānaṃ kaṃ yeva bhavantū ti
sabbe so aṭṭhānaṃ katvāna
abbhūgantvā nabhatalaṃ
kuṭāgāraṃ pavīsitvā
nibbato so mahiddhiko |
tato dvinnaṃ pi therānaṃ
nivatitvāna dve pi ca
kuṭāgārā cittasu
paṭiṭṭhahiṃsu tādeva
gato samuṭṭhahitvāna
aggikhandhā samantato
jhāpenti anavasesā
dhātumattakā |
sabbe devā manussā ca
ye tattha susamāgatā
sabbe te pūjayitvā
sakaṭṭhānaṃ nivattayiṃsū ti |
asokaparinibbāna |

46
 The Asokaparinibbānakathā

‘That one, this elder (come here), was then, in the past, and even
now, firm friends with me.134 Now that he has attained parinibbāna,
how many of the concomitants of my lifespan (remain)?’
And, upon looking, he saw that his lifespan would perish that same
day, whereupon the elder, as one of great potency, thought ‘When I
am dead, let (my) pinnacled-house become just the same (as it).’135
He, having created no opportunity for anyone, rose into the sky,
entered the pinnacled-house, whereupon the one of great potency
attained nibbāna.
And then the two pinnacled-houses of both elders returned and
those same136 became established on the funeral pyre.
Thereupon,137 masses of fire rose up on all sides and burned
(everything) without remainder, except for the relics.
All those, both devas and men, who had come together there, all
worshipped them, and then went back to their own place.
Asokaparinibbāna.

134
so tadā eso thero atīte pi idāni vāci me sahadaḷhamitto ahu (unclear).
135
kuṭagārādayo pamataṃ samānaṃ kaṃ yeva bhavantu is unclear, but makes some sense if
reading samānakaṃ yeva or samānaṃ taṃ yeva; see n. 13, 40, 100 above on the change t → g
→ k(h).
136
Assuming tādeva = tā-d-eva (where tā = te); or is tādeva a mistake for tāva-d-eva,
‘immediately’?
137
I assume gato is an error for tato.

47
 The Asokaparinibbānakathā

Appendix
Partial parallel account of the episode by Buddhaghosa in his Samantapāsādikā
(CSCD edition with relevant PTS page numbers in brackets)
[Be I 32] [Ee I 41] tena ho pana samayena bindusārassa rañño ekasataputtā
ahesuṃ. te sabbe asoko attanā saddhiṃ ekamātikaṃ tissakumāraṃ ṭhapetvā ghātesi.
ghātento ca cattāri vassāni anabhisittova rajjaṃ kāretvā catunnaṃ vassānaṃ
accayena tathāgatassa parinibbānato dvinnaṃ vassasatānaṃ upari aṭṭhārasame
vasse. sakalajambudīpe ekarajjābhisekaṃ pāpuṇi [Ee I 42]. abhisekānubhāvena
c’ assa imā rājiddhiyo āgatā—mahāpathaviyā heṭṭhā yojanappamāṇe āṇā
pavattati; tathā upari ākāse anotattadahato aṭṭhahi kājehi soḷasa pānīyaghaṭe
divase divase devatā āharanti, yato sāsane uppannasaddho hutvā aṭṭha ghaṭe
bhikkhusaṅghassa adāsi, dve ghaṭe saṭṭhimattānaṃ tipiṭakadharabhikkhūnaṃ,
dve ghaṭe aggamahesiyā asandhimittāya, cattāro ghaṭe attanā paribhuñji;
devatā eva himavante nāgalatādantakaṭṭhaṃ nāma atthi siniddhaṃ mudukaṃ
rasavantaṃ taṃ divase divase āharanti, yena rañño ca mahesiyā ca soḷasannañ
ca nāṭakitthisahassānaṃ saṭṭhimattānañ ca bhikkhusahassānaṃ devasikaṃ
dantaponakiccaṃ nippajjati. devasikam eva c’ assa devatā agadāmalakaṃ
agadaharītakaṃ suvaṇṇavaṇṇañca gandharasasampannaṃ ambapakkaṃ
āharanti. tathā chaddantadahato pañcavaṇṇa-nivāsana-pāvuraṇaṃ pītakavaṇṇa
hatthapucchanapaṭakaṃ dibbañ ca pānakaṃ āharanti. devasikam eva panassa
nhānagandhaṃ [Be I 33] anuvilepanagandhaṃ pārupanatthāya asuttamayikaṃ
sumanapupphapaṭaṃ mahārahañ ca añjanaṃ nāgabhavanato nāgarājāno
āharanti. chaddantadahe va [Ee I 43] uṭṭhitassa sālino nava vāhasahassāni divase
divase sukā āharanti. mūsikā nitthusakaṇe karonti, eko pi khaṇḍataṇḍulo na hoti,
rañño sabbaṭṭhānesu ayam eva taṇḍulo paribhogaṃ gacchati. madhumakkhikā
madhuṃ karonti. kammārasālāsu acchā kūṭaṃ paharanti. karavīkasakuṇā
āgantvā madhurassaraṃ vikūjantā rañño balikammaṃ karonti.
imāhi iddhīhi samannāgato rājā ekadivasaṃ suvaṇṇasaṅkhalikabandhanaṃ
pesetvā catunnaṃ buddhānaṃ adhigatarūpadassanaṃ kappāyukaṃ kāḷaṃ
nāma nāgarājānaṃ ānayitvā setacchattassa heṭṭhā mahārahe pallaṅke
nisīdāpetvā anekasatavaṇṇehi jalajathalajapupphehi suvaṇṇapupphehi
ca pūjaṃ katvā sabbālaṅkārapaṭimaṇḍitehi soḷasahi nāṭakitthisahassehi
samantato parikkhipitvā anantañāṇassa tāva me saddhamma-
varacakkavattino sammāsambuddhassa rūpaṃ imesaṃ akkhīnaṃ
āpāthaṃ karohī ti vatvā tena nimmitaṃ sakalasarīravippakiṇṇapuññappa
bhāva-nibbattāsītānubyañjanapaṭimaṇḍita-dvattiṃsamahāpurisalakkhaṇa

48
 The Asokaparinibbānakathā

sassirīkatāya vikasitakamaluppalapuṇḍarīkapaṭimaṇḍitam iva salilatalaṃ

tārāgaṇarasmijālavisada-vipphuritasobhāsamujjalitam iva gaganatalaṃ nī
lapītalohitādibhedavicitravaṇṇaraṃsi-vinaddhabyāmappabhāparikkhepa
vilāsitāya [Ee I 44] sañcāppabhānurāga-indadhanu-vijjulatāparikkhittam
iva kanakagirisikharaṃ nānāvirāgavimalaketumālā-samujjalitacāru-
matthakasobhaṃ nayanarasāyatanam iva brahmadevamanujanāgayakkhagaṇ
ānaṃ buddharūpaṃ passanto satta divasāni akkhipūjaṃ nāma akāsi.

Abbreviations
References to Pali texts follow the system adopted by the Critical Pali Dictionary.
Page references are to PTS editions, where available, otherwise to the Burmese
editions on the Chaṭṭhasaṅgāyana CD-ROM (Be), contained also in the Digital
Pali Reader (https://pali.sirimangalo.org).
CSCD Chaṭṭhasaṅgāyana CD-ROM
CPD Critical Pali Dictionary (Copenhagen)
DPPN Dictionary of Pali Proper Names
EFEO École française d’Extrême-Orient manuscript of the
Asokaparinibbānakathā (EFEO PALI 29)
PED Pali English Dictionary (Pali Text Society)
WPC Wat Phra Chetuphon (Wat Pho) manuscript Bangkok of the
Asokaparinibbānakathā (incomplete)

References
Cœdès, G. 1966. Catalogue des manuscrits en pāli, laotien et siamois provenant de
la Thaïlande. Copenhagen, Bibliothèque Royale de Copenhague (95); EFEO
DATA folder 108, file catal Wat Pho data, file 3.
Masefield, Peter. 2008. ‘Indochinese Pali’. Mahachulalongkorn Journal of Buddhist
Studies I: 1-9.
Skilling, Peter and Pakdeekham, Santi. 2002. Pāli Literature Transmitted in Central
Siam. Bangkok.
Strong, John S. 1989. The Legend of Asoka. Princeton.
Ver Eecke, Jacqueline. Le Dasavatthuppakaraṇa, édité et traduit. PEFEO CVIII,
Paris 1976: XVI-296 p.

49
 The Asokaparinibbānakathā

— 1980. Le Sīhaḷavatthuppakaraṇa, Texte pāli et traduction. PEFEO CXXIII, Paris

1980: XXXVI-158 p.
Ver Eecke-Filliozat, Jacqueline & Filliozat, Jean. 2003. Sahassavatthuppakaraṇaṃ,
Pāli text. Published by the Sangha Assembly of region III as a contribution to
the royal cremation ceremonies of Phra Thammarajanuwat (Kamon Kovido Pali
VI). Bangkok: Wat Thepsirin.

50
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Peter Masefield† and Nicolas Revire1

Abstract
This paper reconsiders the last meal of the Buddha from the little studied
perspective of ‘kammic fluff’ (kammapilotika). Although marginal in the
Nikāyas, this idea is more prominent in the commentarial accounts of
the Buddha’s death, and suggests that the Buddha’s final meal aided the
Buddha, rather than directly caused his death. Additionally, we examine
other evidence from some Theravāda traditions of mainland South East
Asia: modern mural paintings from Cambodia and Thailand which
indicate that the Buddha’s death possibly resulted from a complication of
a chronic peptic ulcer involving the vomiting of blood, and a little known
Pali text of ‘Indo-Chinese’ origin, which supports this interpretation, and
assumes that the Buddha’s final illness was caused by the remnants of
his former kamma.

1
This paper is part of a larger research project led by Nicolas Revire dealing with Pali and
vernacular hagiographies of the Buddha as depicted in the narrative texts, murals, reliefs and
sculptures of mainland South East Asia. The first paper in the series has been published as
Revire 2019. The author wishes to acknowledge support of the Center for Khmer Studies, the
École française d’Extrême-Orient, and the Thai Research Fund. We are also grateful to Dr Nithi
Nuangjamnong of Naresuan University who shared graciously the photos published here a Figs
1, 3–4 and 6. Final thanks are also due to Alex Wynne, the editor of this journal, for his essential
assistance with Pali sources and editorial rigor.

. 21(20): 51–82. ©21 Peter Masefield & Nicolas Revire

 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

The Buddha Siddhattha Gotama is recorded in the Mahāparinibbānasutta (and

the Udāna) as having, shortly prior to his final demise or parinibbāna, taken
his last meal in the house of Cunda Kammāraputta, i.e., ‘the smith’s son’, in
Pāvā (D II 126ff; Ud 81ff). It is said that the Buddha told Cunda that he, and he
alone, should be served with such sūkaramaddava as had been prepared, whilst
the monks accompanying him should be served any other foods, both hard
and soft, that had been prepared.2 Moreover, he told Cunda that any leftover
sūkaramaddava should be buried in a pit, since he beheld none in the cosmos—
humans and devas alike—other than the Tathāgata able to thoroughly digest it.3
Then, some time after the Buddha finished his meal, he was afflicted by a grating
affliction, whilst severe stomach pains, accompanied by blood and diarrhoea,4
that were potentially fatal,5 ensued. Nonetheless, he managed to suppress these
and make his way, shortly afterwards, on foot to Kusinārā,6 where he attained
final parinibbāna. This account has led many to conclude that it was Cunda’s
alms that had led to the Buddha’s death. According to Mettanando & von
Hinüber (2000: 106–107),
the onset of the Buddha’s illness was rapid. The disease started
while eating, so the Buddha assumed that there was something
wrong with this unfamiliar delicacy and he suggested to his host
that the food be buried … Soon the Buddha suffered severe stomach
pain and passed blood from his rectum.

2
D II 127: yaṃ te Cunda sūkaramaddavaṃ paṭiyattaṃ tena maṃ parivisa | yaṃ pan’ aññaṃ
khādanīyaṃ bhojanīyaṃ paṭiyattaṃ tena bhikkhusaṅghaṃ parivisa ||
3
D II 127: yaṃ te Cunda sūkaramaddavaṃ avasiṭṭhaṃ taṃ sobbhe nikhaṇāhi || nāhaṃ
taṃ Cunda passāmi sadevake loke samārake sabrahmake sassamaṇabrāhmaṇiyā pajāya
sadevamanussāya yassa taṃ paribhuttaṃ sammā pariṇāmaṃ gaccheyya aññatra Tathāgatassa ||
4
D II 127: kharo ābādho uppajji lohitapakkhandikā bālhā vedanā vattanti || It is worth
noting that this passage is syntactically ambiguous, and that the expressions kharo ābādho and
lohitapakkhandikā are quite rare in the Nikāyas. For the former, see Vin III 72, IV 70; for the
latter, see M I 316; the occurrence at Ja V 422 (no. 536) is paracanonical. The expression bālhā
vedanā only occurs in the accounts of the Buddha’s death.
5
māraṇantikā; potentially fatal, yes, but not necessarily so, as should be clear from what
follows. Defined, at Sv 546, as maraṇantaṃ maraṇasantikaṃ pāpanasamatthā and, at Ud-a 401,
as maraṇantā maraṇasamīpapāpanasamatthā, i.e., capable of causing one to reach death’s door.
6
Ud 82: atha kho Bhagavā āyasmantaṃ Ānandaṃ āmantesi | āyām’ Ānanda yena Kusinārā
ten’ upasaṅkamissāma ||

52
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

This short summary is quite misleading. The text of D II 127 tells us that even
before starting to eat, the Buddha told Cunda not to serve the sūkaramaddava
to anyone else, and then bury the remnants afterwards. After Cunda followed
the Buddha’s instructions, serving the food in the manner prescribed, the
Buddha told him to bury the remnants of the sūkaramaddava, which he did,
after which the Buddha delivered a Dhamma sermon. The text then states
explicitly that the Buddha got up from his seat and left (uṭṭhāyāsanā pakkāmi),
and only then became ill, an unspecified period of time after eating (bhagavato
… bhattaṃ bhuttāvissa kharo ābādho uppajji …). In other words, the disease
did not start precisely ‘while eating’ (although see below for commentarial
evidence to this effect).
As regards the illness, the text says that the Buddha suffered severe stomach
pain but does not explicitly state that he ‘passed blood from his rectum’. The
term used in this passage is lohitapakkhandikā, which could be taken as a
dvanda compound, viz., ‘blood and diarrhoea’, rather than ‘bloody diarrhoea’.
This point is important, for a dvanda interpretation of the compound would
suggest that the Buddha vomited blood, and if so his death could be ascribed
to a peptic or stomach ulcer. This was ruled out by Mettanando & von
Hinüber, but their claim that the Buddha ‘passed blood from his rectum’ rests
on the unwarranted assumption that the compound lohitapakkhandikā is to be
analysed as a kammadhāraya rather than dvanda. However, they point out that
‘for ulcers higher that the ligament of Treitz … when there is severe bleeding, it
would manifest as bloody vomiting, not a passing of blood through the rectum’
(2000: 107). The dvanda interpretation of the compound lohitapakkhandikā
thus raises the possibility that the Buddha died from a peptic ulcer; as we will
see, the notion that a spell of bloody vomiting (haematemesis) preceded the
Buddha’s death has been maintained in some Theravāda traditions of mainland
South East Asia.
What about the Buddha’s statement that only he can digest the
sūkaramaddava, and that its remnants should be buried? This part of the
narrative could indeed imply that the meal was regarded as dangerously harmful.
But if the sūkaramaddava was harmful, this would also wrongly suggest that
the immediate cause of the Buddha’s death was food poisoning. This has been
correctly ruled out by Mettanando & von Hinüber (2000: 107), based on the
account of the Buddha’s symptoms. Indeed, other aspects of the canonical and
commentarial accounts suggest that there was probably nothing wrong with the
meal itself. We should first note that the identity of the meal does not necessarily

53
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

suggest that it was harmful. In the Udāna Commentary (Ud-a 399ff), which is
rather more thorough than the account given in the Dīghanikāya Commentary
(Sv 516ff), Dhammapāla gives several possibilities as to the denotation of the
term sūkaramaddava:
It is said in the Great Commentary that sūkaramaddava is the
already available meat7 of the pig that is tender and succulent. Some,
however, say that sūkaramaddava is not pig’s meat (but rather)
bamboo shoots that pigs (sūkarehi) have trampled upon (maddita),
others that it is a mushroom that has come into being at a spot that
pigs (sūkarehi) have trampled upon (maddita), whilst still others
proclaim that sūkaramaddava is the name for a certain elixir.8
It is quite clear that, by the time of the commentarial period, knowledge as
to what sūkaramaddava may once have denoted had been lost.9 Nonetheless,

7
The word pavattamaṃsa recurs at Vin I 217 in the incident in which the female layfollower
Suppiyā instructs a servant to find same so that she might prepare meat-broth for a sick monk, such
servant, however, returning empty-handed, on account of the fact that it was an Uposatha day on which
animal slaughter was not permitted, as a result of which Suppiyā had to cut flesh from her own thigh for
the purpose. Sp 1094 explains pavattamaṃsa as ‘meat that is already dead’ (matass’ eva maṃsaṃ), in
accordance with which I.B. Horner renders same as ‘meat that is to hand’, adding the note ‘i.e., already
killed, and not to be killed on purpose for the monk’ (B Disc IV 296 n. 1). This also seems supported
by Sv-pṭ II 218, which states that sūkaramaddava is the meat of the wild boar (vanavarāhamaṃsa),
and that ‘meat that is already dead’ is implied at Sv 568 since Cunda, an ariyasāvaka and sotāpanna,
and the rest, in preparing the food for the Lord and the order of monks, did so blamelessly. Ñāṇamoli
(2001: 357), who takes sūkaramaddava as ‘hog’s mincemeat’, similarly renders pavattamaṃsa as
‘meat already on sale in a market’. Moreover, pavattamaṃsa is, presumably, to be distinguished from
āmakamaṃsa, raw or uncooked, meat, and which is not allowed (D I 5; M I 180); or else this is why
the commentaries explain paṭiyādāpetvā (had prepared) as pacāpetvā (had cooked).
8
Ud-a 399f: sūkaramaddavan ti sūkarassa mudusiniddhaṃ pavattamaṃsan ti mahā-
aṭṭhakathāyaṃ vuttaṃ || keci pana sūkaramaddavan ti na sūkaramaṃsaṃ | sūkarehi
madditavaṃsakalīro ti vadanti || aññe sūkarehi madditappadese jātaṃ ahichattakan ti || apare
pana sūkaramaddavaṃ nāma ekaṃ rasāyanan ti bhaṇiṃsu || Sv 568 gives the first and last
of these only, some editions adding in parentheses that it is a recipe for cooking soft-boiled
rice in the five products of the cow (eke bhaṇanti sūkaramaddavan ti pana mudu-odanassa
pañcagorasayūsapācanavidhānassa nām’ etaṃ | yathā gavapānaṃ nāma pākanāman).
9
See inter alia Wasson 1982, and Mettanando & von Hinüber 2000 who discuss the possible
nature of the sūkaramaddava-dish; contra, see Bareau 1968 who critically examines other parallel
passages in Sanskrit and Chinese where something called sūkaramaddava seems totally absent.
On this ground, Bareau concludes that the Pali sources discussing the last meal of the Buddha may
have been corrupt and of later elaboration.

54
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

sheer common sense suggests that the Great Commentary (Mahā-aṭṭhakathā),

no longer extant, is much more reliable in this regard, if only for the fact that
sūkaramaddava, possibly consisting of pig’s meat,10 was simply one of the huge
number of dishes that Cunda must have had prepared in his household and, we
may presume, from ingredients freely available in the local markets, in advance,
in anticipation of a visit by a ‘great’ and hungry Saṅgha.11
Although the precise identity of sūkaramaddava had seemingly been
forgotten well before the commentarial period, this need not mislead us into
thinking that there was anything inherently pernicious in ‘this mysterious
food’ (as An has it, 2005: 121 n. 5). And this remains true whether it were
pig’s meat, bamboo shoots, mushrooms or whatever.12 Indeed, the Buddha
declares that not only no blame should attach to Cunda, but also that, of all
meals received by the Lord, the two most meritoriously efficacious were that
given by Sujātā, prior to the night of his awakening, and that given by Cunda,
prior to his final extinction:
Of exactly the same fruition, of exactly the same ripening, are these
two almsfoods, being of greater fruition and of greater advantage
than other almsfoods in the extreme. What two? That almsfood
after consuming which the Tathāgata awakens to the unsurpassed
perfect awakening, and that almsfood after consuming which he
attains parinibbāna into that element of nibbāna that is without
remnant of substrate.13

10
It is a common interpretation in Thailand that sūkaramaddava consists of pig’s meat.
Modern Thai mural paintings depicting the life of the Buddha often represent a pig being cooked
and barbecued, or a wild boar being prepared and ready to be offered by Cunda to the Lord and
his fellow monks (Figures 1–3).
11
One that was ‘great’ by way of its greatness of good qualities and its greatness in number
(Ud-a 399: mahatā bhikkhusaṅghenā ti guṇamahattasaṅkhyāmahattehi mahatā).
12
It should be borne in mind, however, that pig’s meat and mushrooms—if this is indeed the
nature of that meal—are taboo in India, especially in the brahmin cast, on which, see Bareau 1968
and Wasson 1982.
13
D II 136f: dve’ me piṇḍapātā samasamaphalā samasamavipākā ativiya aññehi piṇḍapātehi
mahapphalatarā ca mahānisaṃsatarā ca || katame dve | yañ ca piṇḍapātaṃ paribhuñjitvā
Tathāgato anuttaraṃ sammāsambodhiṃ abhisambujjhati | yañ ca piṇḍapātaṃ paribhuñjitvā
anupādisesāya nibbānadhātuyā parinibbāyati || On the rather blurred distinction between the
terms nibbāna and parinibbāna, also involved in this passage, see Masefield 1979.

55
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

The Buddha also points out Cunda’s kammic benefits from offering the meal
as follows:
By Cunda has been heaped up a deed conducive to (long) life-span
… to (good) complexion … to happiness … to heaven … to fame
… to sovereignty.14
Perhaps the early Buddhist tradition had certain reasons to wish to absolve
Cunda of any blame; perhaps he and/or his family were important supporters
of the Saṅgha. But the account explicitly states that the Buddha was able to
digest the meal, and that he subsequently continued his journey on foot; the
Buddha was not, apparently, impaired or incapacitated as a result of the meal.15
The commentarial account continues in this vein. Whatever the precise nature
of the dish, Dhammapāla makes it clear that, although the affliction arose after
the Buddha had eaten the meal, it did not do so as a consequence of his having
partaken of that food. Instead, he claims that meal eased the pain brought on
through the recurrence of an illness that had originated, ten months previously,
in the hamlet of Beluva near Vesāli (but suppressed throughout the interval by
way of meditative attainment),16 thereby allowing him to complete the final leg
of his journey to Kusinārā where he would attain final parinibbāna. The verses
beginning ‘after eating Cunda’s meal’ were codified by the compilers of the
scriptures afterwards:
‘And along with the sūkaramaddava, to the one who had partaken
thereof’: there arose to the one who had partaken thereof, though
not with his having partaken thereof as its condition. For if (that
affliction) had arisen to him without his having partaken thereof, it
would have been far too grating; whereas, on account of his having

14
Ud 85: āyusaṃvattanikaṃ āyasmatā Cundena kammāraputtena kammaṃ upacitaṃ | vaṇṇa-
saṃvattanikaṃ āyasmatā Cundena kammāraputtena kammaṃ upacitaṃ | sukhasaṃvattanikaṃ
āyasmatā Cundena kammāraputtena kammaṃ upacitaṃ | saggasaṃvattanikaṃ āyasmatā Cundena
kammāraputtena kammaṃ upacitaṃ | yasasaṃvattanikaṃ āyasmatā Cundena kammāraputtena
kammaṃ upacitaṃ | ādhipateyyasaṃvattanikaṃ āyasmatā Cundena kammāraputtena kammaṃ
upacitan ti ||
15
It is therefore hard to credit Walshe’s dismissal of the claim that the sūkaramaddava the
Buddha ate could only be digested by the Tathāgata, as follows: ‘(or so we are told). The trouble
was, of course, that in fact even the Tathāgata failed to digest it!’ (1987: n. 418).
16
E.g., D II 99: atha kho bhagavato vassūpagatassa kharo ābādho uppajji | bālhā vedanā
vattanti māraṇantikā || tā sudaṃ bhagavā sato sampajāno adhivāsesi avihaññamāno ||

56
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

partaken of that succulent food, the pain became diminished, as a

result of which same he was able to continue on foot.17
Moreover, it is abundantly clear from Dhammapāla’s commentary that the
meal of sūkaramaddava was given by Cunda in good faith:
For it was this, they say, that Cunda, the smith’s son, gave out of a
desire to have the Teacher live for a long time, hoping, after hearing
that the Lord was to attain parinibbāna that same day: ‘Surely he
will remain a while longer once he has consumed this’.18
From the canonical account, supported by Dhammapāla’s interpretation, it
becomes evident that Cunda’s meal of sūkaramaddava in fact aided the Buddha and
should not be blamed for his death. If so, then what, we may ask, was it about the
dish that rendered it incapable of being thoroughly digested by anyone other than
the Tathāgata, such that any leftovers needed to be buried, and what was responsible
for the blood and diarrhoea that ensued? Let us again return to the commentaries.
According to the Udāna Commentary,19 the reason lies in the fact that
the devatās of the four great continents and lesser islands had infused the
sūkaramaddava with nutritive essence (ojas), thereby rendering it incapable of
being digested by anyone other than the Tathāgata.20 However, according to the
Milindapañha, this they did on every occasion the Tathāgata ate,21 thereby again
implying there was nothing special about Cunda’s meal per se, or at least prior
to its being offered.

17
Ud-a 401 (= D-a II 568): Cundassa bhattaṃ bhuñjitvā ti ādikā aparabhāge
dhammasaṅgāhakehi ṭhapitā gāthā || tattha bhuttassa ca sūkaramaddavenā ti bhuttassa udapādi
| na pana bhuttapaccayā || yadi hi abhuttassa uppajjissā atikharo abhavissā | siniddhabhojanaṃ
pana bhuttattā tanukā vedanā ahosi | ten’ eva padasā gantuṃ asakkhi ||
18
Ud-a 400: tañ hi Cundo kammāraputto ajja bhagavā parinibbāyissatī ti sutvā app’ eva nāma
naṃ paribhuñjitvā cirataraṃ tiṭṭheyyā ti satthu cirajīvitukamyatāya adāsī ti vadanti ||
19
Ud-a 400: tasmiṃ kira sūkaramaddave dvisahassadīpaparivāresu catūsu mahādīpesu
devatā ojaṃ pakkhipiṃsu | tasmā taṃ añño koci sammā jīrāpetuṃ na sakkoti ||
20
Or even by themselves, if Spk I 235f in a similar context is to be believed—see CD 447
n. 450 for a translation. See also Figures 1–2 where Sakka appears flying in the air, with his
typical green complexion, and infusing the pig’s meat with divine nutriments.
21
Mil 231: sabbakālaṃ mahārāja Tathāgate bhuñjamāne devatā dibbaṃ ojaṃ gahetvā
upatiṭṭhitvā uddhaṭuddhaṭe ālope ākiranti ||

57
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

The phenomenon of the dangers surrounding leftover food is outside the

scope of this paper. Nonetheless, we may note that, apart from the practice of
leaving uneaten alms for other bhikkhus (who, if they do not eat it, must throw
it away into water or a place without grass),22 there are a few other instances
in the Nikāyas where the Buddha tells others to bury leftovers. At S I 167ff ≠
Sn p. 15, for instance, we find the brahmin Sundarika Bhāradvāja, following
his performance of the Agni-oblation, going in search of some other brahmin
to whom he might offer the remnant (havyasesa) of that offering leftover in his
ladle.23 The commentary on this explains that he did so in the belief that, since the
oblation placed in the fire had been eaten by Mahābrahmā, he needed to offer the
remnant to another brahmin, if he were to please his forebears and successfully
find his way to the Brahmaloka.24 He therefore offers the remains to the Buddha
(S I 168f), whom he mistakes for a brahmin, but the latter refuses to accept
same, since he does not accept any food that has been chanted over by verses
(abhigītaṃ), adding the following, in much the same tone of the Cundasutta:
I do not behold anyone in this world with its devas, with its Māra,
with its Brahmā, with its generation of recluses and brahmins, with
its (generation of) devas and men, for whom that consumed could
become thoroughly digested, except for a Tathāgata or a sāvaka of
the Tathāgata.25
He then continues, saying that that brahmin should, instead, throw it away ‘in
a place where there is little grass, or immerse it in water devoid of living beings’,
such that, when he did so, it ‘hissed and seethed, and steamed and smoked, just
like a ploughshare, that had been heated all day, when plunged into water’.26

22
Vin I 157ff, I 352, II 216; M I 207, III 157.
23
S I 167: atha kho sundarikabhāradvājo brāhmaṇo aggiṃ juhitvā aggihuttaṃ paricaritvā
uṭṭhāyāsanā samantā catuddisā anuvilokesi | ko nu kho imaṃ habyasesaṃ bhuñjeyyā ti ||
24
Spk I 233: aggimhi tāva pakkhittapāyāso Mahābrahmunā bhutto | ayaṃ pana avaseso
atthi | taṃ yadi brahmuno mukhato jātassa brāhmaṇassa dadeyyaṃ | evaṃ me pitarā saha putto
pi santappito bhaveyya | suvisodhito c’ assa brahmalokagāmimaggo—Cp translation at CD 447
n. 447; also KS I 209 n. 5.
25
S I 168f: na khvāhan taṃ brāhmaṇa passāmi sadevake loke samārake sabrahmake
sassamaṇa-brāhmaṇiyā pajāya sadevamanussāya yasseso havyaseso bhutto sammā pariṇāmaṃ
gaccheyya aññatra brāhmaṇa Tathāgatassa vā Tathāgatasāvakassa vā ||
26
S I 169: atha kho Sundarikabhāradvājo brāhmaṇo taṃ havyasesaṃ appāṇake udake
opilāpesi || atha kho so havyaseso udake pakkhitto cicciṭāyati | ciṭiciṭāyati | sandhūpāyati |
sampadhūpāyati || Seyyathāpi nāma phālo divasasantatto udake pakkhitto cicciṭāyati | ciṭiciṭāyati

58
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Passing aside the question as to why Buddhaghosa should have thought

that an oblation offered to Agni had been consumed by Mahābrahmā,27 it is
nonetheless a fact that the late Vedic or Brahmanic milieu into which the Buddha
arose was one admitting of very strict rules on the use of leftover food, including
that gained by a brahmin student on his almsround. Hence, it is of interest to
find, in Āpastamba’s Dharmasūtra,28 that such a student should, inter alia, ‘after
he has eaten [food gained on his almsround] … not leave any food uneaten. If
he is unable to do so, he should bury the leftovers in the ground, [or] throw
them in the water’, the same text going on to list a hierarchy of individuals to
whom such food might then be offered, prior to its disposal. It surely cannot be a
coincidence that this is almost word for word the same as that put into the mouth
of the Buddha in the above quoted Saṃyutta passage (S I 167f).
The Buddha’s instruction to bury the remnants of his final meal thus reflects
Brahmanic ideas and practices about what to do with leftovers. If so, the
canonical texts and their commentaries are consistent in indicating that there
was nothing in the least unusual, let alone harmful, about the Buddha’s last
meal. So what, we may finally ask, really caused the Buddha’s death?
According to the Apadāna and its Commentary, as well as the Udāna
Commentary, various afflictions suffered by the Buddha were a consequence of
his ‘kammic-fluff’ (kammapilotika): the minor, leftover consequences of former
deeds already long since atoned for in hell and elsewhere.29 This concept is used
to explain various sufferings experienced by the Buddha during his final human
existence, including headaches,30 backache,31 cracked skin on his feet,32 the

| sandhūpāyati | sampadhūpāyati | evam eva so havyaseso udake pakkhitto cicciṭāyati | ciṭiciṭāyati

| sandhūpāyati | sampadhūpāyati ||
27
As C.A.F. Rhys Davids observes: ‘So obsolete apparently was Agni-worship become in
Buddhaghosa’s day, or even in that of his authorities, that he sees only the “Great Brahmā” as the
object of these rites’ (KS I 209 n. 4).
28
Olivelle 1999: 11. See also Deussen 1980: I, 148: ‘The residue (ucchiṣṭaṃ) of the offering…
is to be eaten only by a brāhmaṇa’, quoted in CD 447 n. 447.
29
On this concept, see Masefield 2010 and Anandajoti 2012: 10ff. A summary of the relevant
section of the Apadāna Commentary can also be found in the translation of Ud-a 263ff. Pilotika,
literally means ‘a small piece of cloth, a rag, a bandage’ (PED, sv.), that is, ‘the hanging thread(s)
at the end of a woven cloth’ (Sanjukta Gombrich, personal communication).
30
Ap 300, vv. 3363–64. See also Ud-a 265.
31
Ap 126: tasmā kadāci piṭṭhidukkhe uppanne sāriputtamoggallāne ito paṭṭhāya dhammaṃ
desethā ti vatvā sayaṃ sugatacīvaraṃ paññāpetvā sayati | kammapilotikaṃ nāma buddhamapi
na muñcati ||
32
Ap 300, v. 3362.

59
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

false accusations of Ciñcamāṇavikā and Sundarī,33 not to mention the various

machinations on the part of his major adversary, Devadatta.34 The Ap-a thus
gives the reason for the Buddha’s resurgence of the illness that first originated in
Beluva as being due to the time when he, as a doctor in a past life, had purged
the son of a wealthy banker.
atisāra:35 a purging of blood and diarrhoea (lohitapakkhandhikā).
In the past, the Bodhisatta made his living as a physician, after
apparently being reborn in the home of a householder. When he
was treating a particular banker’s son, who was afflicted36 with
illness, he prepared and administered a medicine, but owing to his
negligence on the day he was to be paid, he gave another medicine
which caused a purging with vomit (vamanavirecanaṃ). The banker
gave him a lot of money. As a result of the ripening of that kamma,
in rebirth after rebirth [the Bodhisatta] was afflicted by an illness
accompanied by blood and diarrhoea (lohitapakkhandikābādhena).
Moreover, in this, his last individual existence, at the time of his
parinibbāna, due to the meal of sūkaramaddava cooked by Cunda
the smith’s son, which had been infused with divine nutriments
by deities from the entire world-system, at the moment of eating
(bhuttakkhaṇe), there was a purging of blood and diarrhoea
(lohitapakkhandikā-virecanaṃ); the strength of a hundred
thousand crores of horses was expended. The Blessed One, going
to Kusinārā for his final nibbāna on the full moon of Visākhā,
sitting down in various places to drink water when he was thirsty,37

33
Ud-a 263, Ap 299 (vv. 3346, 3349, 3354).
34
Ap 300, vv. 3356–58.
35
This term in Pali is often translated as ‘dysentery’ (CPED; PED, svv.) which generally
manifests with the symptoms of ‘bloody diarrhoea’ but with no apparent vomiting. The CPD’s
definition as ‘dysentery’, however, is based on a single passage (Dhp-a I 182), which refers only
to ‘enteric (typhoid) fever’ (kucchiḍāhaṃ). In Sanskrit, atisāra literally denotes an excessive
‘discharge’ or ‘purging’ (SED, sv.), caused for instance by stomach or intestinal inflammation. It
could then, depending on context, refer to either ‘diarrhoea’—whereas the presence of blood is
not necessarily involved—or ‘(bloody) vomiting’. The latter interpretation thus opens again the
possibility that the Buddha died from a peptic ulcer provoking the vomiting of blood (on which
see also figures infra), and not from a dysentery, or a mesenteric infarction, mainly causing bloody
diarrhoea, as generally presented (e.g., Mettanando & von Hinüber 2000: 108–109).
36
vicchita: perhaps from the causative of Skt. vicch, ‘to press, bring into straits’.
37
For depictions of this motif in modern Khmer murals, see Figures 7 and 9.

60
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

reached Kusinārā with great difficulty and then passed into final
extinction just before dawn. Even the master of the triple world38
could not forsake this type of ‘kammic fluff’ (kammapilotika).39
This passage makes several noticeable points. First, it supports Mettanando &
von Hinüber’s claim (2000: 106) that the ‘disease started while eating’—but still,
nevertheless, goes against the account of the Mahāparinibbānasutta in this regard.
Apart from this, the account supports the notion that there was nothing intrinsically
wrong with the food: the statement that ‘the strength of a hundred thousand crores
of horses was expended’ emphasises the restorative effects of the meal, rather
than its adverse results. Finally, and most importantly, the principle of kammic
equivalence suggests that, since the Bodhisatta’s negligence in a former life caused
his patient to experience ‘a purging with vomit’, the transmitters of the story may
have understood lohitapakkhandhikā to involve vomiting blood.
Whether or not this is the case, this understanding of the story of the Buddha’s
death has been seemingly transmitted in some Pali and vernacular Buddhist
traditions of mainland South East Asia. This can be clearly seen today in specific
modern Khmer and Lao-Isan mural paintings from Cambodia and North East
Thailand illustrating the final sickness of the Buddha (Roveda & Sothon 2009:
164, 259; Brereton & Somroay 2010: 28–29). In these regions, the murals
invariably depict the Buddha vomiting, or about to vomit, blood (Figures 5–8,
10–11), and suggest an old and localised tradition in which lohitapakkhandikā
was regarded essentially as the purging of blood through the mouth, rather than
bloody diarrhoea expelled through the rectum.40

38
Reading lokattayasāmī pi for lokattayasāmim pi, assuming the character ī was misread as
anusvāra, and converted into -m for the purpose of sandhi before pi.
39
Ap-a 127: atisāro ti lohitapakkhandikā-virecanaṃ || atīte kira bodhisatto gahapatikule
nibbatto vejjakammena jīvikaṃ kappesi || so ekaṃ seṭṭhiputtaṃ rogena vicchitaṃ tikicchanto
bhesajjaṃ katvā tikicchitvā | tassa deyyadhammadāne pamādamāgamma aparaṃ osadhaṃ
datvā vamanavirecanaṃ akāsi | seṭṭhi bahudhanaṃ adāsi || so tena kammavipākena
nibbattanibbattabhave lohitapakkhandikābādhena vicchito ahosi || imasmim pi pacchimattabhāve
parinibbānasamaye cundena kammāraputtena pacitasūkaramaddavassa sakalacakkavāla-
devatāhi pakkhitta-dibbojena āhārena saha bhuttakkhaṇe lohitapakkhandikā-virecanaṃ ahosi
|| koṭisatasahassānaṃ hatthīnaṃ balaṃ khayam agamāsi || bhagavā visākhapuṇṇamāyaṃ
kusinārāyaṃ parinibbānatthāya gacchanto anekesu ṭhānesu nisīdanto pipāsito pānīyaṃ pivitvā
mahādukkhena kusināraṃ patvā paccūsasamaye parinibbāyi || kammapilotikaṃ evarūpaṃ
lokattayasāmim pi na vijahati ||
40
This tradition may have been pan-regional but, as far as we can ascertain, the visual and
narrative sources for Myanmar are lacking, and its current status for Laos is unknown. However,

61
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

While the modern interpretation of the episode of the last meal in Central
Thailand seems to remain ambiguous—the Thai Paṭhamasambodhi simply mentions
that the Buddha suffered from ‘diarrhoea and bleeding’ after having partaken of the
food offered by Cunda (trans. Paramanujitjinoros 2016: 547)—a Khmer vernacular
narrative text is illuminating in this regard since it gives a textual basis to the visual
evidence. The braḥ nibbān sūtr, only available in manuscript form mentions
‘vomiting’ directly in the Cunda episode. The crucial passage reads and translates as:
ព្រះ�ះអង្គគកក្អអកក្អួួ�តព្រះ�ះលោ�ហិិតស្រ�សៗ ចេ�ញមក
braḥ aṅg ka k’ak k’uat braḥ lohit sras 2 ceñ mak,
i.e., ‘The Lord then coughed and vomited out fresh blood’.41
In addition, the modern illustrations from Cambodia, often depicting Sakka,
Lord of the devas (devānaṃ inda), catching the blood vomited by the Buddha
after consuming Cunda’s alms (Figures 7–8, 10, 11a), might be compared with
a similar episode, recorded in the Dhammapada Commentary. In this episode,
Sakka is described as catching, and removing, on his head, the Buddha’s ‘blood
and diarrhoea’ (lohitapakkhandikā), following the onset of his sickness at
Beluva:42
Sakka permitted no other so much as to touch with his hand the
vessel which contained the excrement of the Teacher’s body,43

a mural painting from Phitsanulok province (Upper Central Thailand) which depicts the scene
is conveniently supplemented by the following caption: ลาก เลือด/lak lueat (to be understood as
*ราก เลือด/rak lueat in the standard dialect of Central Thailand), i.e., ‘vomiting blood’ (Figure 6).
Interestingly, the spelling of the first term substitutes the grapheme or letter < ร = r > for < ล = l >
and thus betrays a likely ‘provincial’ origin, probably of Laos descent, of the scribe and/or artist of
the murals. This may be explained historically by the fact that some Lao communities were deported
from their homeland in the 19th century to re-populate Phitsanulok and surrounding cities. On the
history of r’s disappearance from the modern Lao phonological system, see Davis 2015.
41
The passage is transliterated from MS FEMC 208-B.01.06.01.III.2, fascicle 1, folio ma 28
verso, line 2; it is held at Wat Phum Thmei, Kampong Cham province in Cambodia, and was
copied in 1948. We are very grateful to Trent Walker for bringing this Khmer vernacular and
unpublished reference to our attention, and for his translation of the cited passage.
42
Dhp-a III 269f: so Satthu sarīravalañjanabhājanaṃ aññassa hatthenā pi phusituṃ adatvā
sīse yeva ṭhapetvā nīharanto mukhasaṅkocanamattam pi na akāsi | gandhabhājanaṃ pariharanto
viya ahosi ||
43
The term sarīravalañjana is of obscure derivation. PED has ‘that which is spent or secreted,
i.e., outflow, fæces, excrement’, and ‘discharge from the body’ (valañja, sarīravalañja, svv.). In

62
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

but himself carried the vessel out on his own head. Moreover he
carried it out without the slightest contraction of the muscles of his
mouth,44 acting as though he were bearing about a vessel filled with
perfumes.45
Finally, another source which contests somewhat the common understanding
of the Buddha’s death comes from a Pali text of ‘Indo-Chinese’ origin.46 This
text has been published with a French translation some time ago by Ginette
Martini (1972).47 It is an extra canonical Jātaka composed in the so-called
mul script, possibly in the region now identified as contemporary Central
Thailand,48 and which reads as follows:
evam me sutaṃ || ekaṃ samayaṃ bhagavā bhoganagare viharanto
pāvacundassa piṇḍipātaṃ paribhuñjanto yathā hi amhākam
bhagavā cundassa gehe bhuñjitvā || taṃ divasaṃ yeva bhagavā
cundaṃ āmantesi | mam’ eva sukaramaduvamamsaṃ āhāram
sajjāhi tam āhāraṃ na aññesam bhikkhūnaṃ dehi sesāhāraṃ
nikkhāhī ti || taṃ sutvā Cando tathā akāsi || paribhuñjitamatte
tassa lohitaṃ paggharantaṃ || tasmiṃ khaṇe bhikkhūnaṃ taṃ

other words, the meaning is ambiguous, and the interpretation of Burlingame is likely to be based
here on the biased assumption that the Buddha had diarrhoea. However, just like with atisāra (see
note 35), a purging with vomit is equally possible in this context, and indeed supported by the
Khmer mural paintings.
44
The term mukhasaṅkocanamattam pi na akāsi might, perhaps, be better rendered as ‘without
so much as grimacing’.
45
BL III 79.
46
The nature of ‘Indo-Chinese’ Pali, with all its idiosyncrasies, has yet to receive the attention
it deserves from international Pali scholars. See, however, the preliminary grammatical surveys in
Martini 1936 and Terral 1956; also Masefield 2008 and 2009.
47
It may be worth pointing out that Ginette Terral, Ginette Terral-Martini, and Ginette Martini
are all one and same person, and wife of François Martini.
48
The mul script traditionally used for the notation of Pali is generally taken to indicate a text
of Cambodian (Khmer) origin, the khom script one of Central Thai origin, but shifting borders
over the centuries make it impossible to determine the provenance of any given text, especially
when it contains no information as to the year of its composition. At any rate, this Jātaka is found
in a manuscript once kept at the National Library of Bangkok, and is part of a longer text of the
ānisaṃsa genre titled Paṃsukuladānānisaṃsakathā. A cursory check in various Thai and Khmer
manuscript collections did not prove to be successful to find others variants of this text, although
much more research and editions (not to mention translations) is needed on this huge quantity of
still unpublished local Pali manuscripts from Thailand and its neighbouring countries.

63
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

disvā maṃsasaṃsaññaṃ ahosi || bhikkhu na saṃsayaṃ mama

vipākaṃ taṃ maṃsaṃ sukarassa yaṃ veraṃ mayā kataṃ vipākaṃ
patisevāmī ti || yadā ’haṃ bodhisattakāle daliddhakule nibbattitvā
|| pitā tassa kālaṃ akāsi || mātā pan’ assa vidhavā ahosi || tadā
bodhisatto araññaṃ pavīsitvā kaṭṭhaṃ tiṇṇañ ca āharitvā jivitaṃ
kappesi || tadā ayam sukaro yakkhajāto vessavaṇṇamahārājena
ānato || yakkho atikanto bārānasiyaṃ manusse piḷito ahosi || koci
manusso yakkhapiḷituṃ samattho nama n’ atthi || rājā nagare
bheriñ cāropetvā || tadā pan’ assa mātā puttassa balabhāvaṃ
ñatvā | Nārāyanassa balavā hoti rājānaṃ mama puttassa
balabhāvaṃ jānāpetvā rājā mama puttassa balam dhāressatī ti |
sahassakahāpaṇaṃ rājena dinnaṃ sabbam gahetvā bodhisattassa
āgatakāle vadati || bodhisatto mātuvacanaṃ anatikkanto
yakkhassa santikaṃ gantvā yakkhassa dubbalaṃ katvā māreti ||
yena kammavipākena pañcajātisate verā nāma avūpasamenti ||
(ed. Martini 1972: 254)
A new tentative49 translation is as follows:
So did I hear on one occasion that, whilst the Lord was dwelling in
the city of Bhoga, he would partake (of food gleaned) during his
almsround from Cunda of Pāva, such that our Lord would eat in the
household of Cunda.50 That same day, the Lord addressed Cunda,
saying: ‘You should prepare food consisting of sukaramaduva[sic]-

49
As Norman (2012: 38) once observed, ‘I discover each year that I know less and less [about
Pali philology], and increasingly find that I accept less and less of whatever I thought I understood
years ago … It is hard to be certain that anything is impossible in the field of Middle Indo-Aryan
studies’. If this be so, then how much more so in the case of ‘Indo-Chinese’ Pali, whose studies
are, at best, still in their infancy?
50
This would seem to be a basic misunderstanding. Although monks were generally expected
to gain their sustenance by walking on an uninterrupted almsround, when they would stand,
motionless and speechless, at the gate to some household, merely indicating their need of alms,
and without gesturing by altering the position of their bodies, nor breaking their silence in order
to attract attention (Ja III 162–168, no. 354), subsequently consuming any alms received upon
returning to their place of residence, the Buddha also allowed monks to accept an invitation for
a meal on the following day, as he himself frequently did, in the home of some lay supporter.
There is however, as far as can be determined, no record of a monk entering the home of a donor
in order to consume food just gained at the gate of that same household. See also Mil 229ff for a
long disquisition on the etiquette to be shown when on the almsround.

64
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

flesh for me alone;51 you should not give that food to the other
monks, but instead bury any leftovers’. Upon hearing this, Cunda
acted accordingly. No sooner had he consumed same than his blood
began flowing. That same moment, it occurred to the monks, upon
seeing this, that this must be due52 to the meat. (But the Buddha
said this:) ‘Monks, without doubt I am experiencing a kammic
ripening, my own kammic ripening, (due to) the meat of a pig
to whom I once showed enmity.53 At such time as I had, during
the time I was a Bodhisatta, come into being in a poor family, my
father finished his time,54 with my mother becoming his widow; I,
as a Bodhisatta, made my living by entering the forest and fetching
twigs and grass’. This pig had, at that time, been born a yakkha,
under orders55 of the Great King Vessavaṇṇa. That yakkha, in
transgressing (such orders), became an oppressor of the people in
Bārāṇasi. There was no man capable of restraining that yakkha.
The king had the drum paraded in the city. At that time, moreover,
his mother, aware of her son’s power, thought that after she had
apprised the king of her son’s power, telling him that he possessed
the power of Nārāyaṇa, he would reward her son’s power; she then
took the thousand kahāpaṇas the king had given her, informing the
Bodhisatta when he returned. The Bodhisatta, unable to go against
his mother’s wishes, went into the presence of the yakkha, rendered
him weak and then killed him. Through the ripening of that deed,
enmities have not been appeased over five hundred births.

51
It is, of course, a Vinaya offence for a monk to specify to a potential donor what food he
should be given (e.g., Suddhapācittiya 39 = Vin IV 88; Sekhiyā 37 = Vin IV 193).
52
maṃsasaṃsaññaṃ, possibly in error for maṃsasaññaṃ?
53
This is probably the best that can be done with what seems to be a rather clumsy sentence,
viz., bhikkhu na saṃsayaṃ mama vipākaṃ taṃ maṃsaṃ sukarassa yaṃ veraṃ mayā kataṃ
vipākaṃ patisevāmī ti. No doubt, other interpretations are possible. G. Martini (1972: 255), for
instance, refers here to the alleged noxiousness of the pig’s meat caused by the hatred of the
yakkha for the Bodhisatta arising in a former life as explained subsequently in the Jātaka.
54
pitā tassa kālaṃ akāsi; meaning, of course, that he died. But the time he ‘finished’ was the
kammic time that had given rise to that particular birth. Moreover, in the extended simile given
at Cp-a 97f, Dhammapāla likens saṃsāra to a prison, such that it seems quite legitimate to take
kālaṃ karoti as ‘to do time’.
55
Reading āṇatto for ānato.

65
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Even if this local Jātaka does not use the term ‘kammic fluff’
(kammapilotika), it assumes the concept by attributing the Buddha’s final
illness to the ripening of his former kamma. Admittedly, the passage does not
clearly specify through which channel the blood oozed after Cunda’s meal—
whether the rectum (bloody diarrhoea?) or the mouth (bloody vomiting?)—
and is open to interpretation. However, its use of the verb paggharati to
describe the flowing or dripping of blood, although not offering a decisive
interpretation of the compound lohitapakkhandikā, at least allows for the
possibility that blood flowed from the Buddha’s mouth. Indeed, the verb
paggharati is often employed in connection to the oozing or dripping of
blood in canonical sources. The same verb is also used at times to describe
the dripping of tears,56 which again suggests the possibility that later Pali
composers took it to describe the dripping or vomiting of blood from
the mouth, as already confirmed by Khmer and Lao-Isan artists in mural
paintings (see figures infra).
From the foregoing, and by way of concluding this paper, we are totally
rejecting the notion that the Buddha ate poisoned food. Indeed, how could the
Lord have deliberately accepted this meal consisting of sūkaramaddava should
he truly have been Omniscient, and should it really have been harmful for his
health as some authors claim? This would have been tantamount to committing
suicide proper,57 a negative act which should be avoided at all cost according to
the Pali Buddhist tradition (Wiltshire 1983).

56
E.g., S II 179: … yaṃ vā vo iminā dīghena addhunā sandhāvataṃ saṃsarataṃ
amanāpasampayogā manāpavippayogā kandantānaṃ rodantānaṃ assu passannaṃ paggharitaṃ …
57
It is a well-known fact that the Buddha deliberately decided at Vesāli, three months prior to this
episode at Pāvā, to enter into final parinibbāna, thus accepting the request of Māra (D II 104ff). The
impression given, therefore, is that the Buddha, at that particular point of time, was indeed determined
to die and hence, more or less, committed suicide. It is doubtful, however, that he really, and voluntarily,
decided to put an end to his own life. The fictional idea of the Buddha being able to stay on until the end
of the aeon (kappa), not realised however—or so we are told—because of Ānanda’s foolishness (D II
102–104), may possibly reflect a later anti-Ānanda faction among the early Buddhist lineage.

66
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

We, therefore, also contest the rendition of G. Martini’s French translation

of the pig’s meat as being ‘une chair nocive et vénéneuse’ in the first occurrence
of the above cited passage (1972: 253, 255; see also note 53 supra).58 Upon
reconsideration of all the above evidence related to the last meal, it thus appears
clear that it was not so much the nature of sūkaramaddava (pig’s meat?) that
directly caused the final illness of the Buddha. At best, the meal helped the
Buddha in his final hours to reach Kusinārā. At worst, the potentially fat and
heavy meal offered by Cunda may have triggered a resurgence of a past chronic
disorder (stomach or peptic ulcer?) leading to severe blood loss—apparently
manifested by black or bloody vomiting—and ultimately provoking his
death. This presumed chronic disease of the Buddha, which may have come
and gone over a period of many years, was simply due to his ‘kammic fluff’
or kammapilotika, that is, the leftover consequences of his former deeds as
illustrated in the above Pali commentaries and extra canonical Jātaka.

58
G. Martini, however, later correctly understands verā as hatred (‘haines’) and no longer as
noxiousness (‘nocivité’) as in the previous instance.

67
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Plates

Figure 1. Sakka, Lord of the devas, sprinkling the divine nutriments on the
Buddha’s last meal consisting of pig’s meat. Wat Arun Ratchawararam,
Bangkok, Thailand, repainted in the late 19th century
(Photo courtesy of Nithi Nuangjamnong, September 2017)

68
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 2. The Buddha’s last meal, consisting of pig’s meat, being prepared by
Cunda, and infused with divine nutriments by Sakka.
Wat Kasattrathirat Worawihan, Ayutthaya province, Thailand, c. 1879
(Photo courtesy of Nicolas Revire, June 2020)

69
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 3. Cunda and his attendants preparing the last meal, consisting of wild
boar’s meat (already dead?), for the Buddha and his retinue of monks.
Wat Photharam, Mahasarakham province, Thailand, early 20th century
(Photo courtesy of Nithi Nuangjamnong, March 2019)

70
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 4. The last illness showing the Buddha vomiting blood,

and the grief expressed in the faces of his followers. Wat Photharam,
Mahasarakham province, Thailand, early 20th century
(Photo courtesy of Nithi Nuangjamnong, March 2019)

71
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 5a. The last meal offered by Cunda and the subsequent illness of
the Buddha leading to his demise. Wat Ban Yang, Mahasarakham province,
Thailand, early 20th century (Photo courtesy of Nicolas Revire, January 2011)

72
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 5b. Detail of the Buddha, showing his stomach distress and vomiting
blood (Photo courtesy of Nicolas Revire, January 2011)

73
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 6. The Buddha ‘vomiting blood’ after the last meal

(the Thai caption to the viewer’s left clearly reads ลาก เลือด/lak lueat,
‘vomiting blood’, see n. 40), with Ānanda (Th. อานนท์/Anon) below him,
catching the purging. Wat Huai Kaeo, Phitsanulok province, Thailand,
mid-20th century (Photo courtesy of Nithi Nuangjamnong, April 2021)

74
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 7. The Buddha sitting down to drink water being fetched by Ānanda,
and Sakka getting ready to catch his vomit. Stung Treng province, Cambodia,
early 20th century (Photo courtesy of Nicolas Revire, July 2014)

75
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 8. The Buddha sitting down with Sakka getting ready to catch his
vomit. Wat Phnom Baset, Kandal province, Cambodia, mid-20th century
(Photo courtesy of Nicolas Revire, July 2014)

76
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 9. The Buddha sitting down to drink water being brought

by Ānanda, on his way to Kusinārā to reach final extinction.
Wat Prasat Andet, Kompong Thom province, Cambodia, late 20th century
(Photo courtesy of Nicolas Revire, July 2014)

77
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 10. Two panels depicting the meritorious offering of the last meal by
Cunda to the Buddha (left), and the Lord sitting down and about to vomit
blood in Sakka’s vessel (right). Wat Bakong, Siem Reap province, Cambodia,
early 21st century (Photo courtesy of Nicolas Revire, August 2018)

78
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 11a. The Buddha vomiting blood and attended by his retinue of monks,
with Sakka trying to catch the purging in his vessel. Angkor Wat
(modern pagoda), Siem Reap province, Cambodia, early 21st century
(Photo courtesy of Nicolas Revire, July 2018)

79
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Figure 11b. Detail of the blood dripping from the mouth of the Buddha
(Photo courtesy of Nicolas Revire, July 2018)

80
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Abbreviations
References to Pali texts follow the system adopted by the Critical Pali Dictionary.
Page references are to PTS editions, where available, otherwise to the Burmese
(Myanmar) editions on the Chaṭṭhasaṅgāyana CD-ROM (http://www.tipitaka.
org), contained also in the Digital Pali Reader (https://pali.sirimangalo.org).59
B Disc The Book of the Discipline, PTS 1949 onwards.
BL III Buddhist Legends, vol. III. Cambridge, Massachusetts 1921.
CD The Connected Discourses of the Buddha, PTS 2000.
CPD A Critical Pali Dictionary.
CPED Concise Pali-English Dictionnary.
KS The Book of the Kindred Sayings, PTS 1917 onwards.
PED Pali-English Dictionary.
PTS Pali Text Society.
SED Sanskrit English Dictionary.

References
An, Yang-gyu. 2005. The Buddha’s Last Days: Buddhagosa’s Commentary on the
Mahāparinibbāna Sutta. London.
Anandajoti, Bhikkhu. 2012. Pubbakammapilotika-buddhāpadānaṃ. Apadāna
39.10 and their Commentary in Visuddhajanavilāsiṇī [Accessible online:
https://www.ancient-buddhist-texts.net/Buddhist-Texts/K13-Apadana/
Pubbakammapilotikam.pdf].
Bareau, André. 1968. ‘La nourriture offerte au Buddha lors de son dernier repas’.
In Mélange d’indianisme à la mémoire de Louis Renou. Paris, 61–71.
Brereton, Bonnie Pacala & Somroay Yencheuy. 2010. Buddhist Murals of Northeast
Thailand: Reflections of the Isan Heartland. Chiang Mai.
Davis, Garry W. 2015. ‘The Story of Lao r: Filling in the Gaps’. The Journal of Lao
Studies 2: 97–109.
Deussen, Paul. 1980. Sixty Upaniṣads of the Veda. Delhi.

59
There would appear to be a few discrepancies between the edition of the Chaṭṭhasaṅgāyana
texts used in the Digital Pali Reader and their subsequent printed editions.

81
 On the Buddha’s ‘Kammic Fluff’: The Last Meal Revisited

Martini, François. 1936. ‘Dasa-bodhisatta-uddesa’. Bulletin de l’École française

d’Extrême-Orient 36: 287–413.
[Terral-]Martini, Ginette. 1956. ‘Samuddaghosajātaka’. Bulletin de l’École française
d’Extrême-Orient 48(1): 249–351.
———. 1972. ‘Un jātaka concernant le dernier repas de Buddha’. Bulletin de l’École
française d’Extrême-Orient 59: 251–256.
Masefield, Peter. 1979. ‘The Nibbāna-parinibbāna Controversy’. Religion 9(2):
215–230.
———. 2008. ‘Indochinese Pali’. Mahachulalongkorn Journal of Buddhist Studies
1: 1–9.
———. 2009. ‘A Brief Note on Methodology in the Transliteration of Indo-Chinese
and Other Pali Manuscripts’. Mahachulalongkorn Journal of Buddhist Studies
2: 11–14.
———. 2010. ‘Kammapilotika’. Mahachulalongkorn Journal of Buddhist Studies
3: 13–24.
Mettanando, Bhikkhu & von Hinüber, Oskar. 2000. ‘The Cause of the Buddha’s
Death’. Journal of the Pali Text Society 26: 105–117.
Ñāṇamoli, Bhikkhu. 2001. The Life of the Buddha. Onalaska.
Norman, K.R. 2012. ‘Pāli Philology and the Study of Buddhism’. The Buddhist
Forum. Vol. I, Seminar Papers 1987–1998. London.
Olivelle, Patrick. 1999. Dharmasūtras, the Law Codes of Ancient India. Oxford.
Paramanujitjinoros, Phra (trans.). 2016. Pathamasambodhi: The Life of the Lord
Buddha. Bangkok.
Revire, Nicolas. 2019. ‘The Birth of the Buddha at Angkor’. Journal of the Siam
Society 107(2): 63–90.
Roveda, Vittorio & Sothon Yem. 2009. Buddhist Painting in Cambodia. Bangkok.
Walshe, Maurice. 1987. Thus Have I Heard, the Long Discourses of the Buddha.
London.
Wasson, R. Gordon. 1982. ‘The Last Meal of the Buddha’. Journal of the American
Oriental Society 102(4): 591–603.
Wiltshire, Martin G. 1983. ‘The “Suicide” Problem in the Pali Canon’. The Journal
of the International Association of Buddhist Studies 6(2): 124–140.

82
 Did the Buddha address the monks in Māgadhī?

Ole Holten Pind†

‘Contrariwise, continued Tweedledee, if it was so, it might be, and if it
were so, it would be; but as it isn’t, it ain’t. That’s logic.’ Lewis Carroll,
Alice in Wonderland § 1.

§ 1. The purpose of this paper is to study the distribution of the two vocatives
bhikkhave and bhikkhavo in the Pāli canon, and to analyse the astute comments
on the issue by Aggavaṃsa, the eminent 12th century AD Burmese Pāli
scholar, who addressed it in a paragraph of his Pāli grammar, the Saddanīti.
Aggavaṃsa’s analysis of the evidence sheds light on their distribution in the Pāli
canon. Moreover, it raises some intriguing questions regarding the distribution
of bhikkhave and bhikkhavo in Burmese Pāli manuscripts, which deviates from
that of Sinhalese manuscripts. I have therefore found it necessary to re-examine
the question in the light of the evidence. I shall trace the textual background of
the readings that Aggavaṃsa’s analysis presupposes and draw the conclusion
that the distribution of the two vocatives reflects canonical prosody and has no
historical or regional implications for the occurrence of bhikkhave and bhikkhavo
in the Pāli canon. In fact, their occurrence is parallel to the distribution of the
two vocatives bhante and bhaddante.
§ 2. Pāli scholars have generally interpreted the vocative bhikkhave as an
“eastern” speech form or Māgadhism.1 This assumption, however, fails to

1
Cf. e.g. Geiger (1916, § 82.5). For an overview of the arguments for “Māgadhisms”
in the Pāli canon, cf. Bechert (1980: 24-34).

. 21(20): 83–106. ©21 Ole Holten Pind

 Did the buddha address the monks in Māgadhī?

address the obvious question why the compilers of the Pāli canon transmitted
it in a predominantly “western” Middle Indic (MI) dialect, but did not convert
this particular vocative into its alleged “western” cognate bhikkhavo. The use
of bhikkhave in the Pāli canon is assumed to represent a linguistic reflex of
popular usage that mirrors the monks’ recollection of how the Buddha used
to address them. This socio-linguistic explanation, however, does not apply to
the pervasive canonical usage of another alleged “Māgadhism,” the particle
seyyathā “(just) as, like” of which there are thousands of examples in canonical
prose. It would be irrational to maintain that the compilers of the Pāli canon used
seyyathā because it reflected, in their memory, the language of the Buddha or
popular usage as they evidently preferred to reproduce the speeches attributed to
the Buddha in a “western” MI linguistic idiom. This in itself raises the obvious
question why they would consistently utilise a particle that allegedly would
stem from an “eastern” MI dialect in a “western” MI linguistic context. The
only conclusion to draw from the evidence is that the early compilers of the Pāli
canon preferred to use seyyathā because they did not consider this particle as
dialectically incompatible with the canonical language.

§ 3. In spite of this, H. Lüders considered seyyathā as cogent linguistic

evidence of an original “eastern” Buddhist canon because he mentions it as an
instance of an “eastern” form in the first paragraph of his influential monograph
Beobachungen über die Sprache des buddhistischen Urkanons, followed by
bhikkhave in the second one.2 It is necessary, however, to show beyond doubt
that seyyathā and bhikkhave are dialectically “eastern” speech forms. There are
well over 26,000 instances of bhikkhave in the Pāli canon. This contrasts with
the highly restricted number of instances of bhikkhavo, of which there are only
about 165 examples, including a few instances of the nominative and accusative
plural that are identical to the vocative plural, thus constituting a mere fraction
of the total number of recorded examples of the two vocatives. This is significant
and underlines the linguistic markedness of bhikkhavo compared to bhikkhave.
Moreover, it raises an obvious question that no one has answered: why is the
use of bhikkhavo restricted to a mere fraction of the total number of instances
of bhikkhave?

2
Lüders (1954). Cf. Geiger (1916:§105.2) on seyyathā (“māgadhisierende” se).

84
 Did the buddha address the monks in Māgadhī?

§ 4. If we assume ex hypothesi that the Pāli recensionists in a limited number

of instances introduced the vocative bhikkhavo instead of bhikkhave, it is
necessary to address the corollary: did they substitute bhikkhavo for bhikkhave
randomly or systematically? The grammatical problem of the distribution of
bhikkhavo and bhikkhave in the Pāli canon attracted the attention of Aggavaṃsa,
who devoted a fairly long and interesting paragraph to it in the Saddanīti
190,6ff.3 The distribution of the two forms in the Pāli canon evidently puzzled
Aggavaṃsa. Since his attempt at explaining their distribution constitutes the first
and so far only analysis of the problem, it may well serve as a point of departure
for the following discussion. Addressing the use of bhikkhave and bhikkhavo,
Aggavaṃsa makes the following observations at Saddanīti 190,6-15:
tatra bhikkhave ti āmantaṇapadaṃ cuṇṇiyapadesv eva dissati
na gāthāsu, bhikkhavo ti pacattapadaṃ gāthāsu yeva dissati na
cuṇṇiyapadesv. api ca bhikkhave ti āmantaṇapadaṃ sāvakassa
bhikkhūnaṃ āmantaṇapāḷiyaṃ sandhivisaye yeva dissati na
asandhivisaye; Buddhassa pana bhikkhūnaṃ āmantaṇapāḷiyaṃ
sandhivisaye pi asandhivisaye pi dissati. bhikkhavo ti
āmantaṇapadaṃ Buddhassa bhikkhūnaṃ āmantaṇapāḷiyaṃ
gāthāsu ca dissati, cuṇṇiyapadesu ca sandhivisaye yeva dissati.
sāvakassa pana bhikkhūnaṃ āmantaṇapāḷiyaṃ na dissati, ayaṃ
dvinnaṃ viseso daṭṭhabbo.
“In the present context the vocative bhikkhave is observed
exclusively in canonical prose (cuṇṇiyapadesu eva), but not in
the verses (gāthāsu). The nominative bhikkhavo occurs in verses,
not in canonical prose. Moreover, the vocative bhikkhave only
occurs within the domain of sandhi, in the canon where a sāvaka
addresses the monks, but it never occurs outside the domain of
sandhi. However, in the canon where the Buddha addresses the
monks, it occurs both within and outside the domain of sandhi.
Moreover, the vocative bhikkhavo occurs both in the verses where
the Buddha addresses the monks, and exclusively in the domain of
sandhi in canonical prose. It never occurs, however, in canonical
prose where a sāvaka addresses the monks. This is how one should
regard the difference between the two forms.”

3
Allegedly completed in 1154 A.D.; cf. Mabel Haynes Bode (1909: 16).

85
 Did the buddha address the monks in Māgadhī?

In this important paragraph, Aggavaṃsa analyses the distribution of

bhikkhave and bhikkhavo in terms of:
a. Literary genre, i.e. if the two forms occur in prose
(cuṇṇiyapada) or in verse (gāthā);
b. Sandhi, i.e. if the particle ti follows bhikkhavo or bhikkhave;
and
c. The socio-religious status of the speaker, i.e. if it is bhagavat
or a sāvaka who addresses the monks.

§ 5. The crucial question is whether the distribution of the two forms in the Pāli
canon corroborates Aggavaṃsa’s observations. In order to decide whether they are
linguistically valid, it is necessary to address each of his statements individually. His
first claim that the vocative bhikkhave only occurs as vocative in canonical prose
and never in verse is true in that it describes a pervasive feature of the Pāli canon:
that bhikkhave never occurs in verse in contrast to bhikkhavo that only occurs in
verse4 and in sentence initial position in prose. Aggavaṃsa thus indicates that they
are contrastively distributed. The only recorded example of the use of bhikkhave
in verse is clearly a corruption. It occurs in pādas ab) of a śloka at Ap 470,20:
svākkhāto bhikkhave dhammo (– – – – | ˘ – – – |). The continuation of the verse:
dukkhantakaraṇāya vo caratha brahmacariyaṃ icc āha munisattamo, shows that
it is based on the well-known ordination formula that is recorded e.g. at Vin I 12,
37ff: etha bhikkhavo ti … svākkhāto dhammo. It is therefore evident that bhikkhave
is a corruption of bhikkhavo, which could easily have come about considering the
nature of the Sinhalese script in which the canon has been transmitted.

§ 6. The following claim that occurrences of the nominative plural bhikkhavo

are restricted to vippajahitvā canonical verses is also correct, except that
Aggavaṃsa understandably overlooked that this form also occurs in a single
prose passage at D III 34,13 foll: viparīto samaṇo Gotamo bhikkhavo ca. The
recorded occurrences of nominative plural bhikkhavo in verses comprise the
following 14 instances:

4
It is interesting in the present context that the only recorded example of Ardha-Māgadhī
vocative plural bhikkhavo also occurs in verse at Sūyag verse 157 (text ˚kkhuvo read ˚kkhavo ?);
Pischel (1900: §381, in fine).

86
 Did the buddha address the monks in Māgadhī?

1. sabbe c’ ime bhikkhavo sannisinnā, Sn 384;

2. bhikkhavo tisatā … tiṭṭhanti, Sn 5735 = 3. Th 841 = 4. M II 146
(not printed in Ee);
5. bhikkhavo samādahaṃsu, D II 254,1* = 6. S I 26,28*;
7. vajanti bhikkhavo, Th 21;
8. sotaṃ odhenti bhikkhavo, Th 1233;
9. vimuttā satta bhikkhavo, S I 35,6* = 10. S I 60,4*;
11. etaṃ rajaṃ bhikkhavo viharanti, Ja I 117,32*6
12. vippamuttā ca bhikkhavo, Ja IV 373,24*;
13. saṃviggā bhikkhavo, Ap 472,24;
14. ujjhāyanti … bhikkhavo, Ap 498,8.
In contrast to the limited number of instances of nominative plural bhikkhavo,
all other instances of nominative plural of bhikkhu in the Pāli canon invariably
read bhikkhū. In this regard, the language of the canon does not differ from other
MI instances of nominative plural of masculine u-stems.

§ 7. Aggavaṃsa does not quote examples of the accusative plural bhikkhavo, although
he evidently assumed their existence as it appears from the bhikkhu paradigm recorded
at Sadd 189,15ff. There are, in fact, eleven recorded instances of the accusative plural
in the Pāli canon.7 They occur mostly in verses (a) and rarely in prose (b):
a.
1. ath’ addasaṃ bhikkhavo, D II 272,24*;
2. so ’ham ete pajānāmi vimutte satta bhikkhavo (so read with
Sinhalese v.l. and Be; Ee °ve), S I 36,3* = 3. 60,27* (Ee °vo);
4. sakkaccaṃ ne upaṭṭhāsi bhikkhavo tatthavāsike, Ja VI
118,19*;

5
Qu. Sadd 190, 17 with yācanti for tiṭṭhanti, cf. Sn 566.
6
= Nidd I 505, 20* reading paṇḍitā for bhikkh°.
7
The earliest of the extant indigenous Pāli grammars, Kaccāyanabyākaraṇam, does not record
the nominative and accusative plural bhikkhavo, which indicates that the infrequent occurrence of
these forms in the canon went unnoticed by the early grammarians.

87
 Did the buddha address the monks in Māgadhī?

5. sā anadhivaraṃ upāgamiṃ pāsādike kusalarate ca

bhikkhavo namassituṃ, Vv 148;
6. sādhū ti sā paṭissutvā bhojayitāṭṭha bhikkhavo, Pv 159;
7. bhikkhavo paribhāsisaṃ, Pv 770 = 8. Pv 778;
9. addasaṃ sāsanakāre bhikkhavo, Ap 263,6;
10. oruddhe bhikkhavo disvā, Ap 599,18;

b.
11. viparītā ye bhagavantaṃ viparītato dahanti bhikkhavo ca,
D III 34,23.
It is thus evident that the distribution of the accusative plural bhikkhavo
follows the same pattern as that of nominative plural bhikkhavo, the majority of
the examples being found in verses, whereas only a single example is recorded
in prose. Apart from the limited number of accusative plural bhikkhavo, all other
instances of accusative plural are identical with nominative plural bhikkhū.
Thus, the use of nominative and accusative plural bhikkhavo for bhikkhū is
linguistically marked like the use of vocative plural bhikkhavo for bhikkhave.

§ 8. Aggavaṃsa’s subsequent statement that bhikkhave is found, where a sāvaka

addresses the monks and only within the domain of sandhi, is interesting because
it illustrates a feature that is primarily recorded in Burmese mss of the Pāli
canon. Aggavaṃsa quotes a formula that the compilers used, on the one hand,
as an introduction to suttantas in which an eminent sāvaka like Moggallāna or
Sāriputta addressed the monks and, on the other hand, as their way of addressing
the monks, when either one developed the Buddha’s talk for the sake of the
monks. In the latter instance it is introduced by the corresponding bhagavat
formula. As an example of the sāvaka formula, Aggavaṃsa quotes S II 274,7:
tatra kho āyasmā Sāriputto bhikkhū āmantesi: āvuso bhikkhave ti.

§ 9. Aggavaṃsa’s statement evidently presupposes a distinct Burmese reading

of the sāvaka formula because the recorded instances of the formula in the
Sinhalese version read with a few exceptions bhikkhavo8 ti. This difference

8
Cf. M I 95,20 (Mahāmoggallāno), A V 94,13 (Sāriputta), A V 155,29 (Mahāmoggallāna), A V
164,21 read bhikkhavo ti; similarly A V 41,29 (Mahācundo), 123,8 (Sāriputta), 157,23 (Mahācunda),

88
 Did the buddha address the monks in Māgadhī?

of reading raises the question of why only the Burmese tradition introduced
bhikkhave in the twenty-nine instances of the sāvaka formula,9 but not in any
of the 106 instances of the bhagavat formula,10 of which Aggavaṃsa quotes an
example at Sadd 190,25: tatra kho bhagavā bhikkhū āmantesi: bhikkhavo ti.

§ 10. The question is whether the Burmese reading is original or interpolated

from non-canonical sources. This necessitates an investigation of the background
of Aggavaṃsa’s analysis. The question is whether he records Burmese usage as
known to him or whether the Burmese readings reflect later sources. The latter
appears to be the case: Aggavaṃsa’s explanation of the sāvaka formula appears
to presuppose analyses found in Buddhaghosa’s aṭṭhakathās. Since they would
support the Burmese readings rather than the Sinhalese, they require careful
analysis. I shall quote them in the traditional order of the aṭṭhakathās:
1. āvuso bhikkhave ti sāvakānaṃ ālapanaṃ etaṃ. buddhā hi
parisaṃāmantayamānā: bhikkhave ti vadanti. sāvakā: satthāraṃ
uccaṭṭhāne ṭhapessāmā ti Satthu ālapanena anālapitvā: āvuso ti
ālapanti. Sv 1053,5 foll.11 on D III 272,5: °ve, v.l. °vo.
“āvuso bhikkhave ti is the vocative12 of the sāvakas. For the
Buddhas say ‘bhikkhave’ when addressing the congregation.

162,1 (Mahākassapa), and 315,2 (Sāriputta), for which the editor recorded the Burmese v.ll. bhikkhave.
9
Except the examples mentioned above, the following list records all remaining instances.
As will appear most of these occur in A. The Burmese readings are throughout bhikkhave, the
corresponding Sinhalese ones, however, are bhikkhavo; instances of bhikkhave apparently stem
from Burmese mss. The names of the respective sāvakas are quoted in brackets: M I 24,14:
Ee bhikkhavo (Sāriputta); 46,18: Ee bhikkhavo, om. āvuso (Sāriputta); 184,24: Ee bhikkhavo
(Sāriputta); III 249,2: Ee bhikkhavo (Sāriputta);—S II 274,8: Ee bhikkhave (Sāriputta); S III
105,6: Ee bhikkhavo (Mahānanda); IV 184,16: Ee °ve (Mahāmoggallāna); 263,2: Ee °vo, v.l.
°ve (Mahāmoggallāna);—A II 144,1: Ee °ve (Sāriputta); 156,36: Ee °vo (Mahānanda); 160,20:
Ee °vo (Sāriputta); III 186,14: Ee °vo (Sāriputta); 190,25: Ee °vo (Sāriputta); 314,18: Ee °vo
(Mahākaccāno); 355,4: Ee °vo (Mahācundo); IV 426,18: Ee °vo (Mahānanda); V 41,29: Ee °vo
(Mahācundo); 94,13: °vo (Sāriputta); 102,23: Ee °vo (Sāriputta); 123,8: Ee °vo v.l. °ve (Sāriputta);
126,8: Ee °vo (Sāriputta); 155,24: Ee °vo (Mahāmoggallāna); 157,33: Ee °vo (Mahācunda); 162,1:
Ee °vo v.l. °ve (Mahākassapa); 315,3: Ee °vo v.l. °ve (Sāriputta).
10
For occurrences that identify the place where the Buddha gave his talk to the monks, cf. § 28 below.
11
Cf. Sv-pṭ III 354,3: = sāvakānaṃ āmantanavasena ālapanasamudācāro, na kevalaṃ
bhikkhave ti, so pana buddhānaṃ ālapanaṃ.
12
ālapanaṃ is the term that denotes the voc. in the canon. It is used in the same sense by the
early Pāli grammarians, e.g., at Kacc §57.

89
 Did the buddha address the monks in Māgadhī?

The sāvakas, however, address them as ‘āvuso,’ avoiding

doing so by the teacher’s vocative, thinking: ‘we shall place
[our] teacher in a high position’.”
2. āvuso bhikkhave ti ettha pana buddhā bhagavanto sāvake ālapantā
bhikkhave ti ālapanti. sāvakā pana “buddhehi sadisā mā homā” ti
āvuso ti paṭhamaṃ vatvā pacchā bhikkhave ti bhaṇanti. buddhehi
ālapito bhikkhusaṅgho bhad(d)ante ti paṭivacanaṃ deti, sāvakehi
āvuso ti. (Ps I 100,7-12; ad M I 13,36: āvuso bhikkhavo sic!)
“āvuso bhikkhave ti. In this context, however, the blessed
Buddhas, when addressing the sāvakas, address them as
‘bhikkhave.’ The sāvakas, however, first address them as
‘āvuso,’ for fear they would be similar to the Buddhas, and
thereafter they say ‘bhikkhave.’ The congregation of monks
answers by saying ‘bhad(d)ante’13 when addressed by the
Buddhas, but ‘āvuso’ when addressed by the sāvakas.”
3. āvuso ti sāvakānaṃ ālapo. buddhā hi bhagavanto sāvake
ālapantā bhikkhave ti ālapanti. sāvakā pana buddhehi sadisā
mā homā ti āvuso ti paṭhamaṃ vatvā pacchā bhikkhavo (Ee;
-ve, Be) ti bhaṇanti. buddhehi ālapito bhikkhusaṅgho bhad(d)
ante (Ee; bhante, Be) ti paṭivacanaṃ deti, sāvakehi āvuso ti.
(Sāratthapakāsinī – Spk – II 233,4-9; ad S II 273,5: Ee ˚ve).
“āvuso is the vocative of the sāvakas. For the Blessed Buddhas,
when addressing the sāvakas, address them as ‘bhikkhave’ etc.
The sāvakas, however, first address them as ‘āvuso,’ for fear
they should be similar to the Buddhas, and thereafter they say
‘bhikkhavo.’ The congregation of monks answers by saying
‘bhad(d)ante’ when addressed by the Buddhas, but ‘āvuso’
when addressed by the sāvakas.”
4. āvuso ti hi avatvā bhikkhave (v.l. °vo) ti vacanaṃ
buddhālāpo nāma hoti. ayam pan’ āyasmā Dasabalena
samānaṃ ālāpaṃ na karissāmī ti Satthu gāravena sāvakālāpaṃ
karonto āvuso bhikkhave (v.l.°vo) ti āha. (Manorathapūraṇī –
Mp – II 127,18-22; ad A I 63,19: °vo sic!).

13
For the spelling of this term and its linguistic implications, cf. §20ff below.

90
 Did the buddha address the monks in Māgadhī?

“For by avoiding saying ‘āvuso’ the expression bhikkhave is

the vocative of the Buddhas. Moreover, when addressing the
sāvakas this honorable person (namely Sāriputta) says āvuso
bhikkhave out of respect for the teacher thinking ‘I shall not
use the same vocative as the honourable Dasabala.’ ”

§ 11. The four explanations are slightly divergent although those quoted
under 2. and 3. do not diverge from one another as do the mutually divergent
explanations quoted under 1. and 4. The underlying idea, however, is the same.
The aṭṭhakathās identify the following criteria for the use of bhikkhave as
opposed to that of āvuso bhikkhave:
1. The Buddhas exclusively address the monks as bhikkhave.
2. The sāvakas address them first as āvuso, subsequently adding
bhikkhave, because the use of the vocative bhikkhave without
further qualification is restricted to the Buddhas whose rank is
higher than that of a sāvaka like, for instance, Sāriputta.
Thus, the aṭṭhakathās explain the usage in terms of the socio-religious rank
of the person who addresses the monks. This is not surprising in itself. Indeed,
the modes of address recorded in the Pāli canon indicate the rank and social
status of the persons whose exchange of greetings and conversations the Pāli
canon reproduces in agreement with contemporary norms of cultured behavior.14

§ 12. It is not possible to identify with certainty the aṭṭhakathā on which

Aggavaṃsa based his explanation. It is likely that he relied on either
Sumaṅgalavilāsinī (henceforth Sv) or Papañcasūdanī (henceforth Ps) because
they are the only Buddhaghosa commentaries that do not record variants. The
readings °ve and °vo are recorded for the Dīgha passage that Sv addresses,
whereas the identical Majjhima passage as explained at Ps reads °vo. Spk
juxtaposes °ve and °vo, thus contradicting itself and Sv and Ps; and both °ve
and °vo are recorded in the exegesis at Manorathapūraṇī (henceforth Mp).
There is therefore no cogent reason to assume that the explanations of Sv and
Ps are trustworthy and reflect old usage. Aggavaṃsa’s conclusion that the mode

14
For a lucid analysis of the use of respect language in the context of social intercourse as
recorded in the Pāli canon, cf. Wangle (1962).

91
 Did the buddha address the monks in Māgadhī?

of address of any sāvaka when speaking to the monks is āvuso bhikkhave is

clearly not justified. It is evident that Burmese scribes substituted bhikkhave for
bhikkhavo as a linguistic device for distinguishing between the socio-religious
status of Buddha and sāvakas. The evidence would suggest their editorial
practice is based on extrapolation from a false reading occurring in the sāvaka
formula as recorded in the exegesis of the aṭṭhakathās.

§ 13. V. Trenckner thought that the reading bhikkhave at Ps I 100,7-12 was

incorrect and emended it to bhikkhavo in his annotated transcript of the Sinhalese
ms he used when editing the Majjhimanikāya. Whatever Trenckner’s motivation
for rejecting bhikkhave may have been, it is justified by the evidence: the Pāli
canon and the aṭṭhakathās use the vocative bhikkhavo in the bhagavat formula.
Elsewhere the aṭṭhakathās explain that the bhagavat addresses the monks as
bhikkhavo.15 Therefore, it is difficult to understand why their explanations of
the corresponding sāvaka formula portray the Buddhas as addressing them as
bhikkhave because this reduces the contrast between the two formulas to the
opposition between bhikkhave and āvuso bhikkhave, thus contradicting both the
Sinhalese and Burmese versions of the Pāli canon as well as the aṭṭhakathās.
Formally, it depends on the bhagavat formula that undoubtedly copies it.
Characteristically, it rarely introduces a suttanta, as it is the case with the
bhagavat formula:16 it usually occurs in the middle of a suttanta introduced
by the similar bhagavat formula. It is difficult to understand why this obvious
dependence on the bhagavat formula would not include the use of bhikkhavo
since the sāvaka’s use of āvuso and the monks’ answering the sāvaka by āvuso
in itself marks it as a sāvaka formula.

§ 14. It is therefore obvious that we should correct the reading bhikkhave to

bhikkhavo. The uniform distribution of bhikkhavo and bhikkhave in the bhagavat
and sāvaka formulas found in the Burmese tradition shows that at some point the
received readings that did not make the socio-linguistic distinction between the
modes of address of the bhagavat and the sāvakas were deliberately changed.
The Sinhalese tradition, however, remained faithful to the original ms. readings.
Since the aṭṭhakathās appear to share the same error, it may well be old.

15
Cf. e.g. Ps I 13,29: bhikkhavo ti āmantanākāradīpanaṁ (ad M I 1,6: bhikkhavo).
16
One example is found at D III 272,5; another at M I 46,18 (reading bhikkhavo ti om.
āvuso [sic]).

92
 Did the buddha address the monks in Māgadhī?

§ 15. The reason why Aggavaṃsa classifies the bhagavat or sāvaka formula as
belonging to the domain of sandhi is no doubt that whenever bhikkhave occurs
in the sāvaka formula it is invariably followed by ti. The same pattern also
characterises the examples of the use of bhikkhavo in canonical prose, when the
speaker, according to Aggavaṃsa’s analysis, is bhagavat. The classification is
based on the mere fact that bhikkhavo ti is analysed as °vo + iti > °vo + ti, the
elision of /i/ being considered a sandhi feature, cf. bhikkhavo ti ca sandhivasena
ikāralopo daṭṭhabbo bhikkhavo itī ti (Ps-pṭ I 51,22-23 = Spk-pṭ II 4,24-25 ≠ Mpṭ
I 45,8-9). The claim that bhikkhave when it is used by the Buddha occurs with or
without junctional features is puzzling. It is a well-known fact that the Buddha
in the canonical speeches constantly addresses the monks as bhikkhave, but
never initially, and iti never follows the vocative. Aggavaṃsa quotes as evidence
the following example, which he attributes to the Pāli, thus indicating that he
considers it as canonical: bhikkhū āmantesi: sotukām’ attha bhikkhave ti. The
attribution of this quotation to the Pāli canon turns out to be incorrect: the clause
occurs only in two of Buddhaghosa’s aṭṭhakathās (viz. Sv 676,5 = Spk I 71,23).
The introductory phrase bhikkhū āmantesi imitates well-known canonical usage
of the bhagavat formula in which the phrase bhikkhavo ti invariably follows it. It
is impossible to decide, however, if the reading bhikkhave in this particular case
is original or the result of a scribal mistake. Buddhaghosa and pre-Buddhaghosa
scholars were ignorant about what determines the distribution of bhikkhave and
bhikkhavo in the Pāli canon.

§ 16. The occurrence of the vocative bhikkhavo in canonical verses is restricted

to merely twenty-five instances, of which seven alone–about one third–occur
in the comparatively late Apadāna, evidently imitating the usage that the
Suttanipāta and the nikāyas had already established:
1. yaṃ evarūpaṃ janātha bhikkhavo gehanissitaṃ, Sn 280b;
2. suṇātha me bhikkhavo, Sn 385a;
3. nimmalā hotha bhikkhavo, Dhp 243d;
4. nibbanā hotha bhikkhavo, Dhp 283d;
5. vippamuñcetha bhikkhavo, Dhp 377d;
6. susīlā hotha bhikkhavo, D II 120,21*;

93
 Did the buddha address the monks in Māgadhī?

7–9. te vijānātha bhikkhavo, D II 256,3d*,11d* = D II 262,10*;

10. khemaṃ pattattha bhikkhavo, M I 227,13*;
11. araññe rukkhamūle vā, suññāgare vā bhikkhavo, S I
220,21b*;17
12. cha eva phassāyatanāni bhikkhavo, S IV 70,25*;
13. te rāgadose abhibhuyya bhikkhavo, S IV 71,23*;
14. mettaṃ cittaṃ vibhāvetvā satta vassāni bhikkhavo (Ee °ve v.l.
°vo), A IV 59,23*;
15. māraṃ sasenaṃ abhibhuyya bhikkhavo, It 41,2c*;
16. ahaṃ vo tena kālena ahosiṃ tattha bhikkhavo, Ja V 56,7*;
17. suṇotha bhikkhavo mayham yaṃ kammaṃ pakatam mayā, Ap
299,6;
18-20. evaṃ dhāretha bhikkhavo, Ap 464,10 = 498,18 = 569,13;
21. lābhīnaṃ Sīvalī aggo mama sissesu bhikkhavo, Ap 495,13;
22. samāgacchantu bhikkhavo, Ap 541,4;
23. suṇantu bhāsitam mayhaṃ bhikkhavo sugatorasā, Ap 541,6;
24. na buddho buddhanibbāne no padissati bhikkhavo, Ap 542,15;
25. evaṃ jānātha bhikkhavo, Ap 543,8.
In all of these examples, bhikkhavo occurs at the end of a pāda where
the metre requires – ˘ –. The question is why there are no examples of the
use of bhikkhave in canonical verses because from a purely metrical point
of view bhikkhave should be equivalent to bhikkhavo and thus suitable for
verse. However, in spite of the rythmical equivalence of the two vocatives,
bhikkhave is never found in verse—with the only exception of Ap 470,20
mentioned above § 5—a fact that requires a more sophisticated explanation
than the rather simplistic assumption that they have been introduced randomly.
The distribution of the vocative bhikkhavo would thus indicate that its use was
restricted to verses and to two types of suttanta introductions as well as to the

17
Qu. Sadd 190,24.

94
 Did the buddha address the monks in Māgadhī?

well-known formula for the initiation of monks. When the Milindapañha was
compiled, its author or authors were evidently aware that outside the domain
of the prose formulas the use of the vocative bhikkhavo was restricted to verses
because it occurs four times:
Mil 335,5: etaṃ pivatha bhikkhavo = Mil 335,24;
Mil 336,9: amataṃ ādetha bhikkhavo;
Mil 341,25: samiddhā hotha bhikkhavo.
Interestingly, these verses purport to reproduce the words of the Buddha.

§ 17. The pādas of S IV 70,25* and 71,23* are particularly illuminating

examples of the exclusion of the use of bhikkhave from verses. The text portrays
the Buddha addressing the monks in prose explaining: cha yime bhikkhave
phassāyātanā adantā aguttā arakkhitā saṃvuttā dukkhādhivāhā honti. The
content of this prose section corresponds exactly to pādas ab) of 70,25* in which
he is described addressing the monks as bhikkhavo: cha eva phassāyatanāni
bhikkhavo || asaṃvuto yattha dukkhaṃ nigacchati ||. The distribution of the
two forms is striking, and it contradicts the assumption that the occurrence of
bhikkhave in this suttanta represents a trace of its original language or of the
monk’s recollection of the Buddha’s diction because the introductory prose
section evidently imitates the verses whose content it paraphrases. In contrast
to the vocative bhikkhave of the prose introduction, the corresponding vocative
of the verse is bhikkhavo. Thus this particular example alone would indicate
that the opposition between the two forms is structural rather than historical;
the use of bhikkhavo is not a random substitute for bhikkhave but, rather, the
language of canonical verse excluded the use of bhikkhave and favoured that
of bhikkhavo.

§ 18. As noticed by Aggavaṃsa bhikkhavo occurs in prose passages where the

implicit or explicit speaker is bhagavat, in which case the particle ti follows
bhikkhavo. Thus, it constitutes an example of what he regards as the domain of
sandhi. These prose passages comprise:

95
 Did the buddha address the monks in Māgadhī?

1. Nine instances of the well-known ordination formula etha

bhikkhavo ti, etc.,18 that Aggavaṃsa apparently overlooked;
and
2. 130 instances of the bhagavat formula: tatra kho bhagavā
bhikkhū āmantesi.19 bhikkhavo ti, bhad(d)ante20 ti te bhikkhū
bhagavato paccassosuṃ,21 which the compilers used as an
introduction to suttantas whenever the bhagavat addressed the
monks before talking to them.
Thus, Aggavaṃsa’s survey of the distribution of bhikkhavo and bhikkhave
would seem to presuppose an exhaustive collection of examples from the
Pāli canon, although he apparently missed the solemn initiation formulas ehi
bhikkhū ti and etha bhikkhavo ti whose occurrence is restricted to the Vinaya.
However, he correctly concluded that the occurrence of the vocative bhikkhavo
is restricted to verses and the bhagavat and sāvaka formulas, although his
attempt at explaining either form in terms of junctional features of the Pāli has
no value.

§ 19. If the claim that the nominative, accusative and vocative plural
bhikkhavo in some cases has been substituted for bhikkhave were correct, it is
incomprehensible why the compilers of the Pāli canon would have introduced
bhikkhavo in a few verse passages and two introductory formulas and left
thousands of instances of bhikkhave untouched. It is also difficult to explain the
few instances of the nominative and accusative plural bhikkhavo, which one
would assume are introduced randomly too. However, sometimes bhikkhave
is introduced immediately after bhikkhavo as in the Mahāsatipaṭṭhānasuttanta
(Dīghanikāya II 290,2ff):

18
Cf. Vin I 12,37; 13,15; 19,30; 20,28; 33,10,26; 34,3; 43,4; the formula is often quoted by the
ct.s, e.g. at Mp I 152, satthā “etha bhikkhavo” ti hatthaṃ pasāresi; cf. e.g. Mp I 160,21; 202,20-
21; 206,18; 222,14-15; 302,5.
19
Its distribution in terms of the recorded occurrences in the various collections of the canon
is 1. D: 3; 2. M: 45; 3. S: 34; 4. A: 33; 5. Sn: 1 (Be bhaddante, cf. Sn 78 note 2.) = S I 188,25*
foll.; 6. Ud: 1; 7. Paṭis: 4.
20
The correct reading is bhaddante, v. infra.
21
Cf. D suttantas 22, 26, 30;—M suttantas 1-3, 5-7, 9-11, 16-20, 25, 28, 33, 34, 39, 45-47 49,
64, 65, 101-3, 106, 111-17, 120, 129-31, 137-39, 141, 148-49;—S I 155; 188; 216-18; 220; 222;
224ff; 231; 234-35; 237, 240; II 1; 3; 43; 80; 107; 118; 153; 178; 186-87; 190; 225; 267; III 66;
IV 1;—A III 1; 169; IV 1; 208; 216; 248; 302; 316; 320; 328; 351.

96
 Did the buddha address the monks in Māgadhī?

ekaṃ samayaṃ bhagavā kurūsu viharati. Kammāsadhammaṃ

nāma kurūnaṃ nigamo. tatra kho bhagavā bhikkhū āmaṇtesi:
bhikkhavo ti. bhad(d)ante ti te bhikkū bhagavato paccassosuṃ ...
ekayāno ayaṃ bhikkhave maggo.
It is thus evident that the distribution of the two vocatives is complementary:
bhikkhavo only occurs in verses and in initial position in prose, whereas
bhikkhave never occurs in such environments, and any valid analysis must
address the linguistic implications of their complementary distribution.

§ 20. The bhagavat formula is especially interesting because it makes it

possible to determine what distinguishes the syntactical constraints on the use
of bhikkhavo from those of bhikkhave with greater precision. Buddhaghosa
comments on the use of bhikkhavo in an interesting passage (Ps I 14,28 foll. =
Spk II 3,1 foll. = Mp I 18,18 foll.):
bhad(d)ante ti gāravavacanam etaṃ satthu (Spk °uno)
paṭivacanadānaṃ vā. api c’ ettha bhikkhavo ti vadamāno bhagavā
te bhikkhū ālapati bhad(d)ante ti vadamānā te bhagavantaṃ
paccālapanti. tathā bhikkhavo ti bhagavā ādimhi bhāsati
bhad(d)ante ti te paccābhāsanti, bhikkhavo ti paṭivacanaṃ
dāpeti bhad(d)ante ti paṭivadanaṃ denti:
“The word bhad(d)ante is an expression of respect for the Teacher
or (rather) it is an answer (to bhagavat’s greeting). Moreover,
when bhagavat addresses the monks by saying bhikkhavo, the
monks answer bhagavat by saying bhad(d)ante. Thus bhagavat at
first greets (the monks) as bhikkhavo and they answer the greeting
by the vocative bhad(d)ante. The vocative bhikkhavo invites (the
monks) to answer to the greeting, the vocative bhad(d)ante answers
the greeting.”22
The analysis is part of a comprehensive grammatical explanation of the meaning
and derivation of bhikkhavo. In contrast to the analysis of the sāvaka formula with
which it shares certain features, bhikkhavo is described as āmantaṇākāradīpanaṃ,

Cf. the less explicit interpretation at Spk I 29,12ff: bhikkhavo ti tesaṃ āmantanākāradīpanaṃ.
22

bhadante ti paṭivacanadānaṃ.

97
 Did the buddha address the monks in Māgadhī?

and the monks’ reply bhad(d)ante. In contrast to his analysis of the sāvaka formula,
Buddhaghosa interprets bhad(d)ante and bhikkhavo in terms of a respectful
exchange of greetings between bhagavat and the monks. bhagavat greets the
monks as bhikkhavo, and the monks answer respectfully bhad(d)ante.

§ 21. There is considerable uncertainty about the spelling of bhad(d)ante in this

formula: the readings of the manuscripts vacillate. They read bhadante or the
geminated form bhaddante (< Sanskrit bhadram + enclitic te),23 the latter often
being recorded as v.l. The anonymous compiler of the Pāli grammar attributed
to Kaccāyana was aware of the geminated and non-geminated spellings, and
stated in sutta 245: bhadantassa bhaddanta bhante, “bhaddanta and bhante are
substituted for bhadanta.” The reference to bhadanta is peculiar because this
vocative has become the norm in Buddhist Sanskrit literature.24 However, the
geminated vocatives bhaddanta (sg.) and bhaddantā (plural) commonly address
the monks in the Vinaya, and they evidently represent a thematised version of
bhaddante. Editors generally ignore the significance of the geminated form.
Trenckner, for instance, considered it an error,25 perhaps under the influence of
the spellings of the aṭṭhakathās, although the Sinhalese ms he used for his edition
of Majjhimanikāya invariably spells bhaddante.26 Even if one cannot exclude the
possibility that bhaddante was de-geminated, thus changing the syllabic quantity
from - - - to ˇ - -, there is no general tendency in the language of the Pāli canon
to reduce /-cc-/ to /-c-/. It is perhaps due to de-gemination of /-dd-/ between
vowels27 and writers may have introduced it at a time when they no longer
perceived the gemination as distinctive. In the oldest post-canonical literature
the non-geminated form bhadanta is recorded in Mil (cf. e.g. Mil 23,4: ~assa;
29,8,10: ~o), and, as already mentioned, the de-geminated vocative bhadanta
became the normative spelling in BHS and Buddhist Sanskrit literature.

§ 22. Burmese manuscripts maintain almost invariably the geminated form

in prose as well as in the derivatives ehibhaddantika and tiṭṭhabhaddantika in
contrast to the readings ehibhadantika and tiṭṭhabhadantika of the Sinhalese

23
Cf. Ai.Gr. III § 235 e).
24
Cf. BHSD s.v.
25
Cf. his preface to M Vol. I p. 2 line 3.
26
Trenckner drew attention to every example in his handwritten transcript of Majjhimanikāya.
27
Cf. L. Renou (1916: § 8).

98
 Did the buddha address the monks in Māgadhī?

tradition. The same contrast between the Sinhalese and Burmese readings recurs
in the analyses of the derivation of the two terms in the aṭṭhakathās attributed
to Buddhaghosa. Thus his explanation at Ps II 43,32-34 (ehi bhadante ti vutto
na etī ti na ehibhadantiko. tena hi tiṭṭha bhadante ti vutto na tiṭṭhatī ti na
tiṭṭhabhadantiko) is everywhere reproduced with the corresponding geminated
forms in the Burmese tradition. There is, therefore, no cogent reason for
assuming that the reading bhaddante does not reflect the original form.

§ 23. The interesting thing about this apparently insignificant question of spelling
is that the use of bhaddante is confined to verse, to the bhagavat formula, and
two prose passages in which the isi Kaṇha is respectfully addressed as follows:
sotthi bhad(d)ante hotu rañño, D I 93,13 fol., sotthi bhad(d)ante bhavissa rañño,
93,15 foll. In both cases, however, Be and the ṭīkās read bhaddante which, as
indicated above, must be the correct spelling. This is corroborated not only by
the derivative bhaddanta but also by the expression evaṃ bhaddantavā ti at D II
180,27 (cf. § 24), for which there are no recorded variants. Evidently, this term
is derived from the geminated form bhaddante. In verses the reading bhaddante
is invariably supported by metrical constraints (· · · · | – – – ||), e.g., paṭipadaṃ
vadehi bhaddante (Sn 921); samayo dāni bhaddante (D II 259,13*); abhidhāvatha
bhaddante (S I 209,14*); ahaṃ naṭo ’smi bhaddante (Ja II 169,5*); velaṃ karotha
bhaddante (Th 762); taṃ taṃ vadāmi bhaddante (Vv 697a); taṃ vo vadāmi
bhaddante (Ap 30,23); sabbaṃ harassu bhaddante (Ap 562,16); (– – – – | · ·
· ·) kiṃ bhaddante karitvāna (Th 721).28 In some cases the reading bhadante is
m.c., cf. aṅgārino dāni dumā bhadante (Th 527a, triṣṭubh).29 Thus, it is possible
to conclude that the original reading is bhaddante and that the degeminated form
bhadante is secondary and functionally equivalent to bhaddante.

§ 24. As noticed by Trenckner in 1879,30 bhaddante is the emphatic speech

form contrasting with the syncopated equivalent bhante. Whereas bhaddante
only occurs in verse and initially in a sentence in canonical prose, bhante, on
the other hand, never occurs in these environments, but invariably cliticizes on
the preceding syntactical word.31 It is interesting in the context of the use of

28
The reading bhadante is also possible in the pathyā cadence, cf. EV I: 205 n. ad Th 527.
29
Cf. EV I: 205 ad. loc. (v. EV I § 41).
30
Cf. Trenckner (1879: 69).
31
It is only in the linguistically hybrid Milindapañha and the aṭṭhakathās that we find bhante

99
 Did the buddha address the monks in Māgadhī?

emphatic and non-emphatic speech forms that the Pāli canon records a hyper-
emphatic form of bhaddante viz. bhaddantava derived from bhaddaṃ + tava,
genitive of the non-enclitic pronoun tvaṃ: evam bhaddantavā ti (D II 180,27
= 264,6 = 265,7 = 269,11 = M II 80,1,26 = S I 216,12,17,22ff).32 This form
contrasts with the correspondng non-emphatic usage evam bhante (e.g. at D II
81,16). The hyper-emphatic form is clearly a reflex of respectful language: the
gandhabba Pañcasikkha uses it as a respectful reply to the God Indra’s request.
It is highly likely that bhaddantavā ti imitates the use of pluti—protraction of
the last vowel of a vocative in sentence-final position. This usage indicates the
speaker’s respect for the addressee. Occasionally, it occurs in early Sanskrit
literature,33 and although it is not a pervasive feature of Pāli canonical discourse,
there are nonetheless a few examples in the canon.34

§ 25. The evidence thus shows that the occurrence of bha(d)dante and
bhikkhavo is restricted to verse and prose initially in a sentence, in contrast
to bhante and bhikkhave, which, as a rule, never occur in such environments.
Because of the syntactical parallelism between the use of bhaddante/bhante and
bhikkhavo/bhikkhave, it is possible to conclude that the use of bhikkhavo and
bhikkhave must be subject to the same syntactical constraints as bhaddante and
bhante: bhikkhavo representing the emphatic form corresponding to bhaddante
and bhikkhave, the non-emphatic one, corresponding to bhante. It is possible to
delimit the syntactical features that define the usage of emphatic bhikkhavo and
non-emphatic bhikkhave by focusing on the prose passages in which the two
terms occur since verse passages are subject to metrical constraints. Most of the
relatively few examples of bhikkhavo that occur in verse are characteristically
restricted to its occurrence in cadences and therefore do not convey information
about syntactical patterns such as word order. Pāli prose, however, displays a
marked difference in the syntactical properties of bhikkhavo and bhikkhave. A
systematic investigation of all instances of bhikkhave in the Dīghanikāya, for
instance, shows that it never occurs in initial position in a sentence and that
it cliticizes on a verb, nominal, pronoun, or particle. It is thus clear that the
distribution of bhikkhavo and bhikkhave parallels that of bhaddante and bhante.

in sentence initial position, cf. the recurrent phrase”bhante Nāgasena” at Mil 28,29 ff; Mp I 37,9;
126,19.
32
Ee w.r. bhaddanta vā at S loc. cit.
33
For pluti in Sanskrit literature, cf. AiGr. I §255-257.
34
For pluti in MI, cf. the examples mentioned in von Hinüber (2001: §311).

100
 Did the buddha address the monks in Māgadhī?

§ 26. The phrase that introduces the ordination formula etha bhikkhavo in the
Vinaya indicates that syntactically bhikkhavo does not cliticize on etha but has
the full force of the emphatic form, which in view of its syntactical properties as
constituting a syntactically independent utterance and its solemn enunciation is
hardly surprising. Pāṇini devotes a few sūtras in his Aṣṭādhyāyī to the description
of the use of pluti in connection with cultured exchange of greetings. If we take
into consideration that the Pāli canon imitates Sanskrit usage as recorded by
Pāṇini,35 one cannot exclude the possibility that the Pāli canon in the case of
exchange of greetings also reflects Sanskrit usage. At least in the case of the
ordination formula one might compare Pāṇini VIII 2:84: dūrād dhūte ca: “and
[the final vowel of a sentence becomes protracted (pluta) and acute (udātta)]
when used in calling [somebody] from a distance.” The distinction between
emphatic and non-emphatic vocatives in Pāli is analogous to the use of enclitic
and non-enclitic vocatives in Sanskrit, the non-enclitic form carrying the accent.36
There is no reason to believe that Pāli imitates the use of the Sanskrit accent.37
On the other hand, if the syntactical features of bhikkhavo and bhikkhave do
not imply the presence and absence of accent, respectively, as in Sanskrit, it
is difficult to explain the opposition between the two forms, unless we assume
that the opposition between bhikkhavo and bhikkhave imitates the opposition
between not enclitic and enclitic vocatives of Sanskrit.

§ 27. Since the Pāli canon represents the codification of an oral tradition, it is
natural to assume that the opposition between bhikkhavo and bhikkhave is a
reflex of the mode of recitation of the suttantas. The complementary distribution
of the two terms no doubt reflects the difference of enunciation of emphatic
and enclitic non-emphatic forms. It is thus understandable that bhikkhave never
occurs in initial position since the compilers of the Pāli canon used it as a non-
emphatic vocative. On this interpretation, it is understandable that the verse
pādas read bhikkhavo to the exclusion of bhikkhave. In the first place, it is
hardly likely that the authors of Pāli poetry would use a non-emphatic form of a
noun in a canonical verse text if a corresponding emphatic form were available,

35
Cf., e.g., the use of atthi nāma, kathaṃ hi nāma and yatra hi nāma in Pāli stereotypes; cf.
O.H. Pind, “Pāli Grammar and Grammarians from Buddhaghosa to Vajirabuddhi. A Survey”,
§ 12, in Buddhist Studies (Bukkyo Kenkyu) XXVI, 1997.
36
Ai.Gr. I § 248 b).
37
For the much debated question of accent in MI (in the Pāṇinian sense or any other sense),
cf. von Hinüber (2001: § 159).

101
 Did the buddha address the monks in Māgadhī?

because non-emphatic terms syntactically belong to prose. Thus, for instance,

the emphatic form bhaddante is, like bhikkhavo, primarily found in verses,
whereas bhante is not recorded except in comparatively late texts like Vv and
Pv, where its usage is clearly due to metrical constraints.38 Thus, the use of
bhikkhavo in Pāli poetry indirectly supports the interpretation of bhikkhave as a
non-emphatic version of bhikkhavo. The distribution of the two forms reflects
their syntactical properties.

§ 28. Let us assume for the sake of argument that the compilers added the
prose formulas containing bhikkhavo later in contrast to the common use of
bhikkhave in the sermons. This assumption is dubious because bhikkhavo
occurs in early strata of the canon like the Suttanipāta. Moreover, there are
examples of nominative accusative plural bhikkhavo in prose passages that
are not formulas. Assuming ex hypothesi that the bhagavat formula with
bhikkhavo was a default introduction to some suttantas, one has to explain
why the compilers introduced it immediately after the identification of the
place where the bhagavat gave a talk to the monks. In each instance, tatra
introduces the formula, referring anaphorically to the previously mentioned
place where the bhagavat or the sāvaka gave his talk. For instance, the
bhagavat and sāvaka formulas at D III 58,4-6 and 272,5-8, respectively,
start by identifying the places where bhagavat and Sāriputta addressed the
monks at Mātulā and Campā, respectively. Since the variety of places that
these suttantas identify in the introduction—Mātulā, for instance, only occurs
once as a place name in the canon—one must conclude that the formula
was not a kind of default introduction added at random to the preceding
identification of the place where the bhagavat or the sāvaka gave his talk.
Any other suggestion would be irrational. Although approximately two thirds
of the formulas identify the place as Sāvatthī, the topographical information
contained in other introductions shows that the use of the formula was not
restricted to talks given in Sāvatthī. For instance, the bhagavat propounded
the Mūlapariyāyasuttanta, M I 1, to the monks in Subhagavana at Ukkaṭṭhā
that is hardly ever mentioned in the canon.39 In every case where the bhagavat

38
Cf. idāni bhante, Vv 295 = 806; tuvaṃ ca bhante + 302 = 813; suṇohi bhante, Vv 650; ca
adadaṃ bhante, Vv 695, 793, 893, 1146, 1163, sace hi bhante, Vv 1188; Pv 98, 111, 133; naggo
kiso pabbajito ’si bhante, 246, 278, 335, 371, mātā pitā ca te bhante duggatā yamalokikā, Pv 412,
419ff, 564 ff.
39
I list below 55 examples of places mentioned in the context of the use of the bhagavat and

102
 Did the buddha address the monks in Māgadhī?

or sāvaka formula occurs, the addressees are the monks. Thus, the conclusion
is inevitable: the use of the formula is restricted to talks given to the monks.
It would seem odd that the tradition would keep the vocative bhikkhave in
the text itself but alter the introduction, as if the compilers of the canon were
ignorant of the prosodical distinction between bhikkhavo and bhikkhave.

§ 29. The evidence thus shows that the bhagavat fomula syntactically is an
integral part of the suttanta in which it occurs. The sāvaka version not only
introduces talks that emminent sāvakas gave to the monks, but occasionally the
compilers also introduced it in the middle of a suttanta, when describing how
the Buddha lets an emminent sāvaka take on the responsibility to develop his
own discourse. Thus, there is no cogent reason for assuming that the bhagavat
and sāvaka formulas are in any way later than other suttanta introductions. The
formulas as such are an indication of the text category to which the suttanta
they introduce belong: they record talks that the Buddha or eminent sāvakas
gave specifically to the monks at a well-known place, without any additional
information about the circumstances that caused the Buddha or the sāvaka to
address them. It is therefore understandable that the Majjhimanikāya, which
appears to be a text collection primarily meant for the use of monks, contains
a substantial number of examples of the bhagavat formula. Thus, the formal
features of the bhagavat formula are structural from a literary point of view,
contrasting with other types of canonical discourse directed to people other than
the monks.

§ 30. In the Vinaya, there is only one example of the solemn initiation formula
ehi bhikkhū ti. It occurs in the Mahāvagga narrative about the conversion of
Aññātakoṇḍañña who was the first convert. His story is related at Vin I 12,23-

sāvaka formulas, excluding those referring to Sāvatthī: 1. D III 58,3. 2. D III 272,3. 3. M I 1,4f. 4.
M I 95,7. 5. M I 225,2. 6. M I 281,2. 7. M II 214,2f. 8. M II 238,8. 9. M II 262,21. 10. M III 68,2.
11. M III 248,2. 12. S I 105,2. 13. S I 105,19. 14. S I 108,10. 15. S I 108,25. 16 S I 231,23. 17. S
II 107,7. 18. S II 153,20. 19 S II 185,7. 20. S II 263,13. 21. S V 144,12. 22. S V 227,12. 23. A I
111,33. 24. A I 274,2. 25. A I 276,11. 26. A I 291,22. 27. A II 1,5. 28. A II 24,29. 29. A II 79,9. 30. A
II 156,34. 31. A II 160,19. 32. A II 167,29f. 33 A III 169,10f. 34. A III 303,24f. 35. A III 355,2f. 36.
A IV 100,2f. 37. A IV 162,2. 38. A IV 208,18. 39 A IV 212,19f. 40. A IV 216,27. 41. A IV 317,23.
42. A IV 320,2. 43. A IV 402,23 (Ee om. formula of greeting, cf. loc. cit. no. 6). 44. A IV 414,25.
45. A IV 426,16. 46. A V 41,27. 47. A V 79,5. 48. A V 157,21. 49. A V 161,29. 50. A V 164,20. 51.
A V 315,2. 52. A V 316,11. 53. A V 326,21. 54. A V 354,24. 55. A V 357,16.

103
 Did the buddha address the monks in Māgadhī?

24. However, the use of the formula ehi bhikkhū ti is not confined to the
Vinaya, it is also attested in the narrative about the conversion of the robber
Aṅgulimāla recorded at M II 100,11*: tam “ehi bhikkhū” ti tadā avoca. Th 625
records a similar phrase: “ehi bhikkhū” ti maṃ āha.40 There is, therefore, no
reason to assume that it represents a recent addition to the canonical language
because it is also reflected in the way the Buddha is described as addressing
Māluṅkyāputta at M I 428,16-18: ehi tvaṃ Māluṅkyāputta mayi brahmacariyaṃ
cara ff., which may well have served as a literary model for the initiation
formula. The use of ehi underlines the solemnity of the utterance, which has
an analogous brahmanical counterpart in the solemn haviṣkṛt expression“ehi”
that is described as vācaṃ śantatamaṃ “the most solemn (form) of speech”41 at
Śatapathabrāhmaṇa I.1.4.2.42 The version in the plural “etha bhikkhavo ti” that
is used in situations describing the joint initiation of monks only occurs in the
Vinaya. The compilers most likely composed it on the analogy of the version
in the singular. The fact that it is limited to the Vinaya, however, is not a valid
reason for concluding that the formula “etha bhikkhavo” ti is a later addition
to the canon because the use of bhikkhavo ti is syntactically analogous to the
other canonical examples of its use initially in a sentence. In the case of both
formulas, one cannot exclude the possibility that bhikkhū ti and bhikkhavo ti
are instances of the use of pluti, as suggested in the case of bhaddantavā ti
quoted above. Consequently, the use of bhikkhavo ti in the ordination formula
has no chronological implications, its usage being intrinsic to the prosodical
structure of the language of the Pāli canon.

§ 31. The evidence thus justifies the conclusion that the use of bhikkhavo and
bhikkhave in the Pāli canon reflects contemporary verse and prose structures.
Thus, it is linguistically irrelevant to speculate whether bhikkhave historically
originated in another linguistic context than bhikkhavo: their usage presupposes
syntactical features and prosodies that are intrinsic to the language of the Pāli
canon. Therefore, the assumption that the occurrence of bhikkhave in Pācittiya

40
The fact that Th-a is claiming that the formula at Th 625 is due to the saṅgītikāras does
not indicate later usage because as already mentioned the use of ehi with the same intention also
occurs in Buddha’s talk to Māluṅkyāputta.
41
It represents the haviṣkṛt proper to a brāhmaṇa; cf. Śatapathabrāhmaṇa loc. cit. and
Āpastamba Śrautasūtra I.19.9: haviṣkṛd ehīti brāhmaṇasya.
42
Cf. loc. cit.: etad u ha vai vācaḥ śāntatamaṃ yad ehīti; Abhidh-k-trsl. III p. 61 no. 3: “on a
comparé Śatapatha, i.1,4,2.”

104
 Did the buddha address the monks in Māgadhī?

71 and Nissaggiya Pācittiya 10 of the Pātimokkha is a reflex of its early “eastern”

redaction is baseless, as the syntax of the two passages in which the vocative
occurs is similar to the general syntactical constraints on the use of bhikkhave
as shown above.43

§ 32. The failure to understand the prosodical function of bhikkhave and

bhikkhavo in the context of the language of the Pāli canon stems from addressing
linguistic problems ad hoc, without considering whether the distributional
pattern of the two forms would corroborate the suggested interpretations.
In the final analysis, the failure to suggest a plausible linguistic analysis of
the distributional pattern of bhikkhave and bhikkhavo is a matter of a flawed
methodology; and a flawed methodology leads invariably to wrong conclusions.
The present study shows, I believe, that it is necessary to study the language
of the Tipiṭaka as a language sui generis and not as a random patchwork of
borrowings from other linguistic environments, inter alia “eastern” ones. It
puts the immanent structure of the language of the Pāli canon at the centre
of analysis and illustrates the futility of addressing linguistic problems
atomistically, without consideration as to whether or not the suggested analyses
are compatible with the evidence recorded in the canon itself. To conclude in
Tweedledee’s sense of logic: if the vocative bhikkhave linguistically were a
Māgadhism it would be justified to infer that it is a vestige of another linguistic
environment, “but as it isn’t, so it ain’t.”

References
Bechert, H. 1980. ‘Allgemeine Bemerkungen zum Thema “Die Sprache der ältesten
buddhistischen Überlieferung”.’ In Die Sprache der ältesten buddhistischen
Überlieferung, ed. by H. Bechert. Abhandlungen der Akademie der
Wissenschaften in Göttingen. Phil.-hist. Klasse. Dritte Folge, Nr. 117. Göttingen.
Bode, Mabel Haynes. 1909. The Pali Literature of Burma. Royal Asiatic Society.
Dutt, Sukumar. 1962. Buddhist Monks and Monasteries of India. London.
Geiger, Wilhelm. 1916. Pāli Litteratur und Sprache (Grundriss der Indoarischen
Philologie und Altertumskunde. Bd. 1. H. 7). Strassburg.
v. Hinüber, Oskar. 2001. Das ältere Mittelindisch im Überblick. Wien, Österreichische
Akademie der Wissenschaften.

43
Cf. e.g. Dutt (1962: 68, note 6 with references).

105
 Did the buddha address the monks in Māgadhī?

Lüders, H. 1954. Beobachungen über die Sprache des buddhistischen Urkanons.

Abhandlungen der Deutschen Akademie der Wissenschaften zu Berlin. Klasse
für Sprachen, Literatur und Kunst. Jahrgang 1952 Nr. 10. Berlin.
Pind, O.H. 1997. ‘Pāli Grammar and Grammarians from Buddhaghosa to
Vajirabuddhi. A Survey’. Buddhist Studies (Bukkyo Kenkyu) XXVI.
Pischel, R. 1900. Grammatik der Prakrit-sprachen. Strassburg, Karl J. Trübner.
Renou, Louis. 1916. Grammaire Sanskrite Élémentaire. Paris, Librairie D’Amérique
et D’Orient.
Trenckner, V. 1879. Pāli Miscellany, Part I. London, Williams and Norgate.
Wangle, N. 1962. Society at the time of the Buddha. Bombay.

106
 Toward a Computational Analysis of the Pali Canon

Dan Zigmond

Abstract
This paper describes the results of applying computational text mining
to the Tipiṭaka, or Pali Canon, the canonical scripture of Therevāda
Buddhism. Individual volumes of the Tipiṭaka are divided into “clusters”
using purely computation tools, and in many cases these clusters appear
to match the rough scholarly consensus around the relative age of the
volumes. Texts are also summarized into “word clouds” based on relative
word frequency, and these also seem to reflect the underlying themes of
the texts. While these initial results are essentially confirmational rather
than novel, they suggest these approaches will be valuable additions to
the Pali scholar’s toolbox.

Computational text mining

Text mining can be defined as the process of extracting new information
from textual sources using computational means. Its practice goes back to the
invention of digital computers in the mid-twentieth century, and the explosion of
availability of both texts in electronic form and the computing capacity necessary
to analyze them have greatly accelerated progress in recent years. Although
much initial work was focused on commercial and government applications,
the past decade has seen increasing adoption of computational techniques in the
humanities (Jockers 2013).

. 21(20): 107–135. ©21 Dan Zigmond

 Toward a Computational Analysis of the Pali Canon

Thus far advances in text mining have typically been applied to English and
other modern (and primarily Western) languages, but this is starting to change.
Since 2007 researchers have convened regular international meetings on
Sanskrit Computational Linguistics (see, for example, Kulkarni and Dangarikar
2013). Along similar lines, the Classical Language Toolkit aims to make text
mining applicable to many ancient languages (Johnson et al. 2014). To the best
of our knowledge, however, there have been very few attempts to apply these
techniques systematically to the Pali Canon (e.g., Elwert et al. 2015).
In recent years a robust set of generalized tools have emerged to support
computational analysis, making application of these techniques to novel corpora
and languages more feasible. The work in this paper was carried out using the R
statistical programming language (R Core Team 2013), and the tidy (Wickham
2019), tidytext (Silge and Robinson 2016), factoextra (Kassambara and Mundt
2020), and wordcloud (Fellows 2018) packages.1

Particular challenges of the Pali Canon

The Tipiṭaka, or Pali Canon, is the canonical scripture of Therevāda Buddhism.
Purported to be the oral teachings of the historical Buddha (Sujato & Brahmali
2014), it is believed to have been first recorded in written form in what is now Sri
Lanka around the first century BCE. Although versions of these texts are preserved
in other languages, the oldest and most complete edition of the Tipiṭaka is recorded
in Pali, a Middle Indo-Aryan dialect whose name derives from the compound pāli-
bhāsa, “the language of the texts” (Geiger 2005, xxiii). In other words, the Pali
language and the Tipiṭaka are intimately linked: Pali is literally the language of the
Tipiṭaka. As Pali scholar Richard Gombrich put it succinctly: “For many Buddhists,
Pali occupies the kind of place that Arabic occupies for Muslims, Hebrew for
Jews, [and] Greek and/or Latin for various kinds of Christians” (Gombrich 2018).
Pali appears to have been originally a spoken language and has no fixed
written form. It has been traditionally written using the script of the various
Asian countries where Therevāda Buddhism has proliferated: in Khom and
Tham scripts in Thailand, Burmese in Burma, Sinhalese in Sri Lanka, etc. The
growth of the Vipassana meditation movement founded by S.N. Goenka in India
has led to a resurgence of Pali texts printed in Devanagari script. Beginning in

1
The novel tools and electronic texts used in this paper are freely available through the tipitaka
package (Zigmond 2020), and the source code for this package can be found at https://github.com/
dangerzig/tipitaka.

108
 Toward a Computational Analysis of the Pali Canon

the late 19th century, the Pali Text Society (PTS) pioneered the publication of
Pali texts in Roman script for Western scholars (and Western Buddhists) using
a system of diacritics similar to that typical for transliterated Sanskrit. This
Roman rendering is the written form of Pali used in this analysis.
For example, here is the first verse of the Dhammapada, perhaps the most
famous of the Pali scriptures and the first to be translated into a European
language, first in Roman-scripted Pali, then in two modern translations:
manopubbaṅgamā dhammā, manoseṭṭhā manomayā,
manasā ce paduṭṭhena bhāsati vā karoti vā,
tato naṃ dukkham anveti, cakkaṃ va vahato padaṃ. (Dhp 1)
Preceded by perception is mental states,
For them is perception supreme,
From perception that have sprung,
If, with perception polluted, one speaks or acts,
Then suffering follows,
As a wheel the draught ox’s foot. (Carter and Palihawadana 1987)
All experience is preceded by mind,
Led by mind,
Made by mind,
Speak or act with a corrupted mind,
And suffering follows
As the wagon wheel follows the hoof of the ox. (Fronsdal 2005)
The title itself of the Pali Canon, Tipiṭaka, can be translated as “consisting
of three baskets” and the Canon is composed of three distinct sets of scriptures:
• Vinaya Piṭaka, Basket of Discipline, describing the rules for the
monastic order.
• Sutta Piṭaka, Basket of Discourses, primarily recounting the
direct teachings of the Buddha (such as the verse quoted above).
• Abhidhamma Piṭaka, Basket of Special Teachings,2 summarizing
and systematizing the Buddha’s doctrines.

2
An alternative understanding of Abhidhamma Piṭaka would be the basket “about the
teachings.”

109
 Toward a Computational Analysis of the Pali Canon

Each of these is composed of several books, which in turn are often divided
into chapters and verses. The Sutta Piṭaka is the most widely studied and so its
divisions have particular significance. It contains four major collections of suttas
or discourses, plus a fifth collection of a wide variety of generally shorter material.
Table 1 shows these major divisions and the approximate length (in words) of
each. The total size of the Tipiṭaka is just under 2.7 million words, with the Suttas
alone totaling near 1.5 million. By way of comparison, the King James Version
(KJV) of the Christian Bible contains approximately 855,317 words (Project
Gutenberg 2020). Thus, in (very) rough terms, the Tipiṭaka (in Pali) is a bit more
than three times the length of the KJV (in English), while the Buddha’s discourses
alone (i.e., the Sutta Piṭaka) are a bit less than twice the length of the KJV.

Components of the Tipiṭaka Approximate length (in words)

Vinaya Piṭaka 414,887
Sutta Piṭaka 1,475,446
Dīgha Nikāya 142,313
Majjhima Nikāya 244,973
Saṃyutta Nikāya 264,973
Aṅguttara Nikāya 300,010
Khuddaka Nikāya 523,177
Abhidhamma Piṭaka 801,650
Tipiṭaka TOTAL 2,691,983

Table 1: Divisions of the Tipiṭaka and length in words

There are a number of challenges to working with this material using

computational tools. First there are several extant versions of the Canon. This
analysis was based on the digital edition of the Chaṭṭha Saṅgāyana Tipiṭaka
version 4.0 published by Goenka’s Vipassana Research Institute (hereafter
CST4; Vipassana Research Institute 2020). This edition originated at the so-
called ‘Sixth Buddhist Council’, held in Burma from 1954 to 1956. Originally
published after the Council meetings in Burmese script, the Vipassana Research
Institute in India began printing this edition in Devanagari and eventually Roman
(and several other) scripts in 1990 and later published the results electronically
as well.

110
 Toward a Computational Analysis of the Pali Canon

This CST4 edition differs somewhat from the more widely used Roman
edition published by the PTS in the UK, although no exhaustive catalog
of the inconsistencies appears to exist. While the PTS edition is available
electronically at the Göttingen Register of Electronic Texts in Indian Languages
(GRETIL: http://gretil.sub.uni-goettingen.de/gretil.html), the format used is
more cumbersome for computational analysis.
Beyond the occasional textual inconsistencies between these editions (which
tend to be minor), there is no comprehensive standard for organizing the Pali
Canon. To begin with, there are slight variations in which books are considered
canonical. For example, the Milindapañha and Peṭakopadesa are sometimes
included in the Khuddaka Nikāya, and sometimes not. (They are not included
in this analysis.)
Furthermore, even where the contents are agreed, the structure is sometimes
not. Some elements of the overall structure are canonical and universally
observed. For example, the previously discussed division of the Tipiṭaka into
three Piṭakas is well established, as is the division of the Sutta Piṭaka into five
Nikāyas (Webb 2011; von Hinüber 2015, 8). But beyond this, the different
editions do not always agree. The division of each individual Nikāya into
separate printed volumes is a publishing convenience and is fundamentally
arbitrary. Thus, for example, the PTS edition divides the Aṅguttara Nikāya, or
Numerical Discourses of the Buddha, into five volumes; the CST4 into only
four. Both the PTS and CST4 divide the Majjhima Nikāya, or Middle-Length
Discourses, into three volumes, but make the divisions at somewhat different
(though nearby) points in the text. This means the usual standard of reference,
by volume and page number, can be difficult to translate between editions.3
Finally, and perhaps most importantly, three characteristics of the Pali
language itself create computational challenges. First, most Pali words exist in
numerous declensions, generally based on number, gender, and case. Second,
consecutive words in Pali sentences can be combined through letter and syllable
elision in complex ways, forming what can appear to be novel words. Third,
Pali also makes substantial use of compounds. Taken together, this means that
individual words often appear in the Canon in a vast array of different forms.

3
This paper largely adopts the Chaṭṭha Saṅgāyana volume numbering as a natural consequence
of using an electronic version of the Chaṭṭha Saṅgāyana edition of the Canon. See the section
“Abbreviations” for a longer discussion of this.

111
 Toward a Computational Analysis of the Pali Canon

For example, there are 270 variations on the word bhikkhu (monk)4 if one
counts all words beginning with the base/stem bhikkh-. The 20 most frequent
such forms are shown in Table 2. These include declensions of bhikkhu such
as bhikkhave5 (vocative plural) and bhikkhū (plural), related words such as
bhikkhunī (nun), and compounds such as bhikkhusaṅghaṃ (congregation of
monks). Of these, fully 115 (about 42%) appear in the entire Canon only once.

1. bhikkhave 11. bhikkhūhi

2. bhikkhu 12. bhikkhunīnaṃ
3. bhikkhū 13. bhikkhussa
4. bhikkhuno 14. bhikkhuniṃ
5. bhikkhūnaṃ 15. bhikkhusaṅghaṃ
6. bhikkhunā 16. bhikkhavo
7. bhikkhuniyo 17. bhikkhusaṅghena
8. bhikkhuṃ 18. bhikkhusaṅgho
9. bhikkhunī 19. bhikkhūti6
10. bhikkhuniyā 20. bhikkhunīti6
Table 2: Most frequent words based on bhikkhu in the Pali Canon 6

4
All English definitions in this paper are from Buddhadatta (2014) unless otherwise noted.
Where Buddhadatta gives multiple definitions, I have generally taken the first few.
5
In fact, bhikkhave, the plural vocative case used in direct address, is the most common form
of bhikkhu and appears 2.6 times as often as the nominative case that one might expect to be most
common. This relatively obscure declension occurs so frequently in the Canon in conjunction
with this word because many of the Buddha’s discourses are directed toward a group of listening
monks, whom he addresses this way. (Geiger 2005)
6
Note that bhikkhūti and bhikkhunīti are not even single words; they are bhikkhu and bhikkhunī
with the quotation marker ti appended. This sort of issue is discussed in more detail in the section
“Limitations and future work” below.

112
 Toward a Computational Analysis of the Pali Canon

Altogether the various discourses of the Sutta Piṭaka contain 115,433 distinct
Pali words by our count. In comparison, the KJV contains only 13,306 words
in English. Thus while the Suttas are less than twice as long as the Bible, they
contain nearly nine times as many distinct words.
Because computational text mining typically depends on comparing word
frequencies across texts, having so many words, and so many with very low
frequencies, can pose a challenge. The most common word, ca (and; then; now),
appears 56,487 times; the 100th most common word, samannāgato (endowed
with; possessed of) appears only 2,508 times. The frequencies of all 100 most
common words are shown in Figure 1, demonstrating this precipitous decline.
The full lexicon of the Tipiṭaka follows the same frequency pattern we saw in
variations of the word bhikkhu: about 42% of all unique words in the Pali Canon
also occur only once. By way of comparison, only 31% of distinct words in the
KJV appear to occur just once. (The section “Future directions” below describes
some possible remediations to overcome this.)

113
 114
Count
10,000
20,000
30,000
40,000
50,000

ca
na
pe
vā
kho
ti
hoti
bhikkhave
dhammo
so
paccayo
uppajjati
paṭicca
dhammaṃ
taṃ
dhammā
tattha
te
evaṃ
pana
no
tassa
ekaṃ
bhikkhu
me
tīṇi
tesaṃ
bhagavā
dve
bhante
atha
ayaṃ
atthi
nti
hetupaccayā
tena
nava
yaṃ
dhammassa
yo
bhikkhū
ye
āmantā
cittaṃ
khandhā
yassa
rūpaṃ
viharati
pañca
āvuso
ceva
vuccati
āpatti
ka
kha
idaṃ
idha
nāma
ime
bhagavato
yattha
siyā
khandhe
saṃkhittaṃ
natthi
satta
yena
paccayā
āyasmā
yathā
iti
puggalo
vattabbe
hi

Figure 1: The 100 most common words in the Pali Canon

hetu
tato
tayo
etadavoca
tasmiṃ
ahaṃ
samaye
nu
uppajjissati
honti
bhagavantaṃ
yā
samayena
hetuyā
tvaṃ
ekamantaṃ
kiṃ
hevaṃ
khandhānaṃ
upasaṅkamitvā
vedanāya
maṃ
samannāgato
bho
jhānaṃ
bhikkhuno
Toward a Computational Analysis of the Pali Canon
 Toward a Computational Analysis of the Pali Canon

As shown in Figure 2, word frequency in the Pali Canon is inversely

proportional to word rank. This relationship roughly follows a classic power law,
as has been observed for many other language corpora (Zipf 1935). The main
divergence from a Zipf power function is that the Pali Canon does not have as
many high-frequency words; visually, the left side of the graph is flatter. Again, a
comparison to the KJV is instructive and is shown in light gray in Figure 2. Three
English words in the KJV exceed 2% frequency (the, and, and of), ranging from
7.5% to 4%, while no individual words in the Pali Canon are similarly common.
Overall, the KJV is more “head heavy,” meaning the most common words are
more common than they are in the Tipiṭaka, while the Tipiṭaka itself is more “tail
heavy,” meaning the least common words are more common there.

0.10000

0.00100
Frequency

0.00001

10 1 ,0 0 0 1 0 0 ,0 0 0
Rank

Figure 2: Word rank versus frequency across the Pali Canon (black, log-log
scale), and for the King James Version of the bible (light gray)

115
 Toward a Computational Analysis of the Pali Canon

Categorizing the Tipiṭaka through k-means clustering

We can further analyze the Canon using classical k-means clustering, one of the
oldest algorithms for computational categorization (see MacQueen 1967 and
Lloyd 1957). In the simplest terms, we compare the texts by using the relative
frequency of each unique word in each text.
More precisely, this approach reduces each of our texts to some number n
of quantitative features. The precise mechanism for transforming a text into
such features is discussed below, but one can then think of these features as
coordinates in an n-dimensional space. If our features are well chosen, then
texts with coordinates closer to each other in this space should be more similar
than texts further apart. In k-means clustering, we choose some number k of
clusters in which we wish to categorize our texts, then draw boundaries in this
n-dimensional space to create k distinct regions. Again, if our features are well-
chosen, the points within these boundaries will form clusters of similar texts.
Of course, reducing a complex text to some manageable set of meaningful
quantitative features is a difficult task. Some trivial approaches would obviously
not be particularly useful. For example, the length of a text might be a poor
choice because texts of the same or similar length do not necessarily have any
deeper linguistic connection. Paperback editions of The Da Vinci Code and A
Tale of Two Cities may both have 489 pages, but these two texts have little else
in common.
A more common and often successful approach is to use the frequency
of some number distinct words. In English, we might measure the relative
frequency of words like the, and, of, to, and that (the five most common words
in the KJV) as our features, such that the KJV would be represented as the five
quantities {0.075, 0.060, 0.040, 0.016, 0.015}. Another text similar to the KJV
would be presumed to have similar frequencies; in other words, it would occupy
a nearby position if plotted in five-dimensional space. It might seem surprising
that such mundane quantities could yield a useful analysis, but such analyses
have led to genuinely novel discoveries in other domains of literature and the
humanities (Jockers and Thalken 2020).
In this analysis of the Tipiṭaka, we will use the relative frequency of the 1,000
most common words in the Canon as our features. This number is admittedly
arbitrary, but the results appear similar across a wide range of thresholds. By
choosing the top 1,000, we are using words that appear more than 250 times
across the Canon. If we were to use the top 10,000 words, we would be including

116
 Toward a Computational Analysis of the Pali Canon

words that appear just 15 times, or less than once per volume.7 This might lead
to clusters determined by the presence or absence of a single word, or even a
single typographical error in our files.
Given the discussion above on the number of distinct words in the Pali
Canon, the top 1,000 may seem like a very small subset to use for our analysis;
it represents well under 1% of all distinct words. However, it is well established
that frequency variation among a very small number of words can often be
enough to identify authorship of past literary works (Jockers & Thalken 2020).
In fact, as we will discuss in our analysis of the Sutta Piṭaka, we can make
meaningful categorizations of canonical text based on the relative frequency of
far fewer than 1,000 words.
Our full methodology is roughly as follows:
1. The texts are read and separated into distinct words.
2. All numerals marking verses, pages, etc. are removed.
3. Each distinct word is counted, as well as the total words for
each volume.
4. The relative frequency is computed for each distinct word (i.e.,
the count of that word divided by the total words in a given
volume).
5. The 1,000 words with the highest average frequency across all
volumes of the Canon are selected as features.
6. The distance between each volume and every other volume is
calculated within this 1,000-dimensional space.
7. Boundaries are drawn to create two clusters within the space.

In effect, we are categorizing each volume of the Tipiṭaka based on the

relative statistical distribution of the 1,000 most common words. The underlying
hypothesis is that volumes with a more similar pattern of word usage are
intrinsically closer (i.e., more related) to each other than those with a more
dissimilar pattern of word usage.

7
As it happens, the 1,000th most frequent word is viññāṇassa (a declension of viññāṇa:
animation; consciousness). The 10,000th word is thīnaṃ (a declension of thīna: unwieldliness;
impalpability).

117
 Toward a Computational Analysis of the Pali Canon

For visual simplicity, this 1,000-dimensional space can be represented as

a simple two-dimensional chart, as shown in Figure 3. These two dimensions
are created by combining many of the underlying dimensions, with some loss
of information, in a long-established statistical process known as principal
component analysis. As shown on the axis labels of Figure 3, these two “principal
components” capture approximately 84.5% of the variation between our texts
in the full 1,000-dimensional space. (Volumes of the Tipiṭaka are shown in all
figures using the standard abbreviations from the PTS Pali-English Dictionary,
which are fully explained in Table 3 in the “Abbreviations” section below.)

A.I

S.V
A.III A.II
S.II
S.IV
2 S.III
A.IV
Vin.III
M.I Vin.IV It
Ud
M.III
Vin.II
D.II Vin.I Abh.III
Dim2 (6.5%)

M.II S.I
D.I

D.III
Abh.I
0

Nidd.I Abh.IV
Patis Abh.II
Nidd.II
Abh.V
Nett
Sn Vin.V Abh.VII
Vv Khp Abh.VI
Dhp Ap.I
Pv
-2
Thag Bv
Thig
Ja.II
Ja.I Ap.II
0 10 20 30
Dim1 (77.8%)

Figure 3: Cluster analysis of the Tipiṭaka

118
 Toward a Computational Analysis of the Pali Canon

We can see that the volumes of the Tipiṭaka shown here form two distinct
clusters. The first of these contains the Vinaya and Suttas (with one exception;
more on this below) while the second contains the Abhidhamma (again with
one exception). This division roughly follows scholarly opinion on the age
of the material; the Abhidhamma is considered the most recent of the three
baskets of the Tipiṭaka (von Hinüber 2015, 64). Thus we might consider the
left (blue) cluster to be our older texts and the right (red) cluster to be our
younger texts.
How then to explain the two exceptions to this otherwise clean separation
of our texts into older and younger clusters? The second volume of the
Abhidhamma, titled the Vibhaṅga and shown as Abh.II in our figure, is clustered
on the left, with our older texts, although the Abhidhamma is generally believed
to be younger. However, the Vibhaṅga is believed likely to be the oldest of the
Abhidhamma material, with some dating it to a similar period as the Vinaya
and Suttas (von Hinüber 2015, 69). It is thus not entirely surprising that our
algorithms might place it with the older material, which it likely matches in
linguistic style. Also note that Abh.II is about equidistant from its nearest volume
of the Abhidhamma (Abh.IV) as from the nearest volumes of the Suttas (Patis
and Nidd.I) and Vinaya (Vin.V). It may represent an intermediary between these
two periods of scripture.
This leaves the first volume of the Aṅguttara Nikāya, A.I in our figure, which is
shown in right/younger (red) cluster, despite being a volume of the original Suttas.
The Aṅguttara Nikāya, or “Numerical Discourses” (Bodhi 2012), is an
unusual collection. The volumes are organized according to number so that
we have the “book of ones,” “the book of twos,” etc. (von Hinüber 2015, 76).
The first volume, the Ekakanipāta, is the book of ones, containing discourses
referring to a single thing. For example:
Bhikkhus, I do not see even one other thing that when developed
leads to such great good as the mind. A developed mind leads to
great good. (A I 6; Bodhi 2012)
As this example demonstrates, many of these passages are extremely short.
Although there are some counter examples, most verses contain two sentences
with a total of a few dozen words. In this way, the book stands somewhat apart
from the other collections, and is not particularly similar to any of them – a
characteristic our computational analysis correctly highlights.

119
 Toward a Computational Analysis of the Pali Canon

In fact, the nearest neighbor of A.I in Figure 3 is A.II, which in the CST4
electronic edition contains the Aṅguttara Nikāya books of two, three, and four,
despite the fact that A.I is clustered with the Abhidhamma volumes. In some
sense, the grouping of A.I with the Abhidhamma may represent a limitation
of the standard clustering algorithms, which attempt to construct compact
polygons around the individual points. While A.I is closer to points in the left
(blue) cluster and one can imagine extending that cluster to include A.I, the
resulting polygon would be less compact, because A.I is somewhat further from
the center of the left polygon than from the center of the right.
Note that Abh.III represents a significant outlier from even the other volumes
of the Abhidhamma. (In fact, if we divide our volumes into three clusters, our
algorithm places Abh.III in a cluster of its own.) One explanation is that the Abh.III,
or the Dhātukathā, may be younger than the preceding volumes, and appears not to
have been recited at the first three Buddhist Councils at all (von Hinüber 2015, 69).
Figure 4 provides another view of these “distance” measures in a
hierarchical manner, using a cluster dendrogram to visualize the similarities and
dissimilarities among Tipiṭaka texts (Kassambara 2017). The y-axis represents
distance between the texts, so texts that are joined higher are less similar than
those joined lower. Color coding is used to cluster these texts into distance
groups. The seven “rainbow” texts on the left are all quite distant from the rest;
as in Figure 3, these include most of the Abhidhamma (with the exception of
Abh.II) as well as the first volume of the Aṅguttara (A.I). The remaining texts of
the Sutta and Vinaya Piṭaka form two broad clusters. On the far right, we have
most texts of the Khuddaka Nikāya (plus Abh.II); in the middle we have the
first four Nikāyas of the Sutta Piṭaka as well as the Vinaya Piṭaka. Once again,
this largely seems to reflect the scholarly consensus concerning the age of the
underlying texts, as will be discussed further in the next section.

120
 Toward a Computational Analysis of the Pali Canon

0.20

0.15

0.10
Distance

0.05

0.00
A.III

Vin.III

M.III
Vin.II

A.II
A.IV

Vin.IV
M.II
D.III
A.I

Vin.I

D.II

M.I
Abh.III

Abh.VII

Ap.II

Vin.V
It

Vv
Abh.VI

Abh.IV

S.III

D.I

Abh.II
Ap.I

Nidd.II
Bv
Abh.V
Abh.I

S.IV

S.II

Nidd.I
Ud

Thig

Khp
S.V
S.I

Cp

Pv

Nett

Dhp
Ja.II

Sn
Ja.I
Thag

Patis
Figure 4: Distance (i.e., dissimilarity) between Tipiṭaka texts

Categorizing volumes of the Sutta Piṭaka

If we confine our attention to the Sutta Piṭaka, we can apply the same
techniques and further divide these volumes into two more clusters, shown in
Figure 5 below.
The upper (blue) cluster contains the Dīgha, Majjhima, Saṃyutta, and
Aṅguttara Nikāyas, while the lower (red) cluster contains the many volumes
of the Khuddaka Nikāya (labeled according to their individual volume names,
as is customary) – with two exceptions. The Udāna (Ud) and Itivuttaka (It)
are clustered with the volumes of the other Nikāyas instead of the Khuddaka
Nikāya, where they are canonically placed. (More on this below.)
Here the clustering does not quite so closely mirror the scholarly consensus
on the age of the underlying material. The Khuddaka Nikāya or “minor texts”
represents something of a hodgepodge of “very heterogenous works” (von
Hinüber 2015, 41) that appear to have been collected later than the other Nikāyas.

121
 Toward a Computational Analysis of the Pali Canon

While some of these, such as the Dhammapada (Dhp) quoted earlier, are well-
known and well-loved among Buddhists, they are generally quite distinct from
the other Sutta collections. Von Hinüber (2015, 45) goes so far as to say that
many Dhp verses “have hardly any relation to Buddhism.” It therefore seems
sensible that these texts can be linguistically distinguished from the first four
Nikāyas.
While the Udāna (Ud) and Itivuttaka (It) are among the oldest elements
of the Khuddaka Nikāya, others of likely similar age include the previously
mentioned Dhp and the Suttanipāta (Sn). This suggests our algorithm is not
clustering these texts by age per se, or at least not by age alone. On the other
hand, Ud and It generally take the form of suttas or discourses, whereas many
of the other Khuddaka Nikāya texts do not. In some cases, material from Ud
and It is also found elsewhere in the Canon, creating inherent similarities. The
placing of these texts in the upper (older) cluster may result from these textual
and stylistic elements rather than, or in addition to, age.
Sn may create particular challenges for algorithms of this sort, because it is
itself a collection of diverse texts of varying ages (Norman 2010, xxxi–xxxiii).
The proximity of Sn to the Niddesas (Nidd.I and Nidd.II) in Figure 5 is likely
due to the latter texts being commentaries on sections of the former. Nidd.I and
Nidd.II are clearly much later than Sn, so the relationship uncovered here is not
chronological but perhaps simple concordance.
Note that once again, A.I is very much an outlier with respect to the other
Suttas. Although it falls within the upper (older) cluster, it is not particularly
close to any other volume there. It is closest to It, which is also organized
numerically. In fact, if we group into three clusters using the same algorithm,
A.I ends up in a separate cluster of its own.

122
 Toward a Computational Analysis of the Pali Canon

5.0
A.III
S.V
S.IV
A.IV S.II A.II
M.I
S.III
M.III
D.II
2.5
M.II
D.I Ud
S.I
D.III It A.I
Dim2 (31%)

0.0

Nidd.I
Nidd.II

Patis
Sn Nett

-2.5
Dhp
Ja.I Pv
Thag Khp
Ja.II Vv
Thig Cp
Ap.I
Bv
Ap.II
0 5 10 15
Dim1 (44.7%)

Figure 5: Cluster analysis of the Sutta Piṭaka

As in the analysis of the full Tipiṭaka above, this clustering was based on the
relative frequency of the 1,000 most common words in the canon. As it turns
out, the suttas can be similarly categorized using a much smaller set.

123
 Toward a Computational Analysis of the Pali Canon

Figure 6 shows a clustering of the Sutta Piṭaka based only on the 13 most
common Pali words, which represent all the words with an average frequency of
at least 0.5% across the Canon.8 Although the shape of the clusters is inevitably
different, the results are exactly the same as the 1,000-word clustering. We are
able to distinguish the predominantly older and younger suttas based only on their
use of the following words: ca, na, kho, vā, ti, bhikkhave, hoti, pe, te, so, dhammā,
taṃ, and me, most of which are simple grammatical particles and the like.9

A.I

10

5 Vv
Dim2 (32.7%)

Bv
Khp
Nidd.I Ap.I Ap.II
Cp Thig
Cp
Dhp
Pv Ja.II
It Nidd.II
Thag
A.II Patis
0
Nett
Sn
S.V Ud
A.III S.I
S.II A.IV D.III
D.I
S.III D.II M.II
S.IV M.III
-5 M.I
-15 -10 -5 0 5
Dim1 (49.9%)

Figure 6: Cluster analysis of the Sutta Piṭaka based on the top 13 words

8
Note that only 5 words have an average frequency of 1%, a further testament to the great
linguistic variety of the canon. In contract, the KJV has 11 words with a frequency of at least 1%,
and 29 with a frequency of at least 0.5%.
9
As in the previous two analyses, A.I is a substantial outlier; in fact, even more distant from
all other texts. This remains somewhat of a mystery.

124
 Toward a Computational Analysis of the Pali Canon

We can again view this clustering hierarchically, as shown in Figure 7. We see

the same broad grouping of texts into older (right) and younger (left) clusters,
with A.I standing out as most distant within its cluster.

0.15

0.10
Distance

0.05

0.00
M.III

A.III
D.III

M.II

A.II

A.IV
A.I

M.I

D.II
Ap.II
Vv

It
Nidd.II

Ap.I

D.I

S.III
Bv
Nidd.I

S.IV

S.II
Thig

Ud
Khp

S.I

S.V
Pv
Nett

Dhp
Cp
Ja.II
Sn

Ja.I
Thag
Patis

Figure 7: Distance between Sutta Piṭaka texts

Summarizing the Sutta Piṭaka

Word frequency can also provide clues to the core meaning of different
volumes of the Canon. A common way to illustrate this is with “word clouds,”
graphical arrangements of individual words where the size of each word is
in proportion to its frequency. In order to focus on the words with the most
semantic content, a set of very common “stop words” are first removed from
the corpus (Lewis et al. 2004). For example, in English, words such as a, and,
and the are prototypical stop words: common in virtually every text and so not
a very useful guide to meaning.10

10
Stop words are not removed prior to the earlier clustering analysis because the relative

125
 Toward a Computational Analysis of the Pali Canon

As far as we know, no definitive set of stop words has been defined for
Pali, so a tentative set was created for this analysis. This was derived by
combining the words labeled as “indeclinable” or “participle” in the PTS Pali-
English Dictionary (PTS 1925)11 plus the most common Pali pronouns (Geiger
2005, 98–109). The full list of 245 words included is shown in Table 4 in the
Appendix.
Figure 8 shows such a word cloud for the Therīgāthā, or Poems of the
Early Buddhist Nuns (left), and the Theragāthā, or Poems of the Elder Monks
(right). As might be expected, the most prominent word in the Therīgāthā is
therī (senior nun), while the most prominent word in the Theragāthā is thero
(senior monk).

dhammaṃ

thero
sāsana
sumedhā buddhassa
dhammamadesesi
vijjā

niṭṭhito pāde samaṇā yadā

anaññathā

disvā nibbutā bhaṇati satthā

nti rattiyā
tattha sāsanaṃ disv ā niṭṭhito pañca
evaṃ nti
dukkhaṃ

ahu loke etaṃ tadā

sabbattha

tisso anāsavā
piyā sace yathā kataṃ sātatiko

hoti
sabbe dāni
natthi tathā

sukhaṃ
bhikkhu
tassāhaṃ
vacanaṃ
maggaṃ
dāni
kataṃ jhāyati tisso sato evaṃ
tato kāmā dhammaṃ tiṭṭhati
citta ṃ

sobhare tattha
buddhassa cittaṃ

therī
jarāya dukkhassa yāme
gāthāyo tato
vijjā

tatruddānaṃ
sabbe anuppattā pe
sutvā

citta
yathā passa
punappunaṃ
saccavādivacanaṃ

Figure 8: Word cloud for the Therīgāthā (left) and Theragāthā (right)

frequency of such words can be very useful in dating and identifying authorship. However, these
differences are too subtle to show up in word cloud images and tell us little about meaning. By
way of example, my own use of words like a, and, and the might help establish that I am the
author of this paper or perhaps even assist in dating this paper based on the prevailing usage of
such terms, but would reveal very little else about the paper’s content.
11
These were collected using the online interface to the PED available through the University
of Chicago’s Digital Dictionaries of South Asia project, at https://dsalsrv04.uchicago.edu/
dictionaries/pali/.

126
 Toward a Computational Analysis of the Pali Canon

These word clouds are potentially more interesting when applied to smaller
sections of the Canon, which are likely to be more focused in meaning. Figure
9 shows such a cloud for the Mahāsatipaṭṭhāna Sutta (M I 56) on the left,
the ‘(Great) Discourse on Mindfulness Meditation’ (again with stop words
removed). Here we see words like pajānātiī (knows clearly), viharati (lives;
abides), and loke (declension of loka, the world) emphasized, which are
central to the meaning of the sutta. On the right we see a word cloud for the
full first volume Aṅguttara Nikāya (A I), which covers a wide-ranging set of
themes. The only substantive word that stands out is bhikkhave, indicating that
these disparate discourses were addressed to monks but had no other obvious
common thread.

bhikkhave anuppannā
bhāveti
viharati
uppajjanti
vijjati anatthāya
evaṃ akusalā saṃvattati

bhikkhave
sātarūpaṃ
yathā
panassa
sati ajjhattaṃ
uppannassa
dukkhaṃ nivisamānā
kāye

vedanā ettha
etthesā
aya ṃ

kāyānupassī ta ṇhā
tañca

evampi hoti ṭhānaṃ netaṃ pe amataṃ eteva

vuccati uppādo
dhammesu uppajjamānā aññaṃ dīpenti
mahato cittaṃ dhammā
atthi
aññena

dhammānaṃ

loke
nirujjhati pe piyarūpaṃ bhikkhū
pahīyati sattā ekadhammampi atthāya
cittanti vedanaṃ
hoti yadidaṃ ṭhānametaṃ
appakā
cittaṃ uppajjati ceva saṃvattatī
saṃyojanassa yathayidaṃ kusalā

pajānāti
anuppannassa
nāhaṃ evamevaṃhetu
bhikkhu paccājāyanti
uppannā bahutarā
uppajjati
dhammānupassī samanupassāmi

Figure 9: Word cloud for the Mahāsatipaṭṭhāna Sutta (left)

and Aṅguttara Nikāya I (right)

127
 Toward a Computational Analysis of the Pali Canon

Limitations and future directions

There are advantages and disadvantages to using the exact Pali syntax found in
the Canon as the basis for our analysis. By way of analogy, the English words
monk and monks are obviously distinct, and different authors may vary in the
relative frequency of each. On the other hand, something is clearly lost if we
treat the two as entirely unrelated, with no more connection than that between
monk and mouse. Yet that is exactly what we are doing when we treat bhikkhu
and bhikkhū as entirely distinct words.
One approach to overcoming this limitation would be to convert words to
their Pali bases or stems (in this case, bhikkh). One could then use the base/
stem frequencies as features, either replacing or augmenting the exact word
frequencies. This would also avoid the issue noted above, where, for example,
bhikkhūti is treated as a single word when it is, in fact, a concatenation of the
two words bhikkhu and ti. However, developing an accurate stemming algorithm
will be a substantial undertaking. Some progress has been made by others (see,
for example, Basapur 2019, Elwert 2015, and Alfter 2014), but no complete
algorithm appears yet publicly available. This is important work to undertake.
Our tentative list of stop words is also unsatisfactory. It was created in a
somewhat manual process that may have included errors. It is also possible that,
for example, all adverbs should be added to this list. Another approach would
be to add all very common words, regardless of grammatical function, although
this would result in meaningful words like bhikkhave and dhammo (doctrine;
nature; truth) being excluded. It will likely take a good deal of trial and error, as
well as a healthy dose of human judgment, to arrive at a definitive set. Our initial
list is at best a good starting place for a much longer effort.12
The analysis described thus far has been at the “macro” scale of entire volumes.
While this is interesting, it is also limiting, and in some cases arbitrary. In the
future we would like to descend to the more “micro” levels of individual suttas
and verses. The tools we have now discard demarcation of particular verses as
well as word ordering within the volumes. In order to facilitate microanalysis at
the verse and word level, all such material would need to be preserved.
More advanced techniques of text mining and natural-language processing
could also be applied. Topic modeling—a machine learning technique for
clustering and summarizing texts—is one broad example; extraction of n-grams,

12
Elwert (2015) alludes to a set of stop words used in that work, but it does not appear to have
been published.

128
 Toward a Computational Analysis of the Pali Canon

or key phrases, is another. However, this would depend on the sort of further
advances in stemming discussed above to be truly useful.
Finally, several purely technical challenges remain. The inconsistent volume
numbering between the CST4 and PTS editions is an annoyance, and the solution
arrived at here, sitting in between the two, is a poor one. In the future we will
edit the underlying files to match the PTS numbering for consistency with other
scholarly material. This is a laborious process of careful editing and was deemed
too much to attempt right now.

Tentative conclusions
The analyses described here have been largely confirmational; they do not
yet bring new knowledge to the study of the Pali Canon. While the apparent
separation of volumes into groups of newer and older texts generally matches
scholarly consensus, the discrepancies appear to be artifacts of the algorithms
rather than novel discoveries. Other analyses help us visualize the relationships
between the texts and some of their central themes but are not yet revealing
previously undiscovered truths.
Nevertheless this style of macroanalysis shows promising potential. As
these methods are refined, they may be helpful in dating noncanonical and
paracanonical texts and tracing the overall evolution of the Canon. As we expand
these techniques to the level of individual suttas and verses, we may gain still
further insight into the authorship of these various component texts.
This first analysis only scratches at the surface of these ancient scriptures,
showing that modern computational tools can be applied. The tools developed
have been released publicly so that other scholars may continue analysis in a
similar vein. We hope this is the beginning of the application of these tools to
the Tipiṭaka and not the end.

129
 Toward a Computational Analysis of the Pali Canon

Abbreviations
The text and figures above generally follow the standard of the Pali-English
Dictionary (Pali Text Society 1925) but are shown in Table 3 for clarity. Note
that discrepancies between the PTS and CST4 editions make volume numbering
difficult. It has been handled here (admittedly somewhat inconsistently) as
follows:
• Volume numbering within the Vinaya Piṭaka has been adjusted
to match the PTS order.13
• Volume numbering within the Abhidhamma Piṭaka is consistent
between the two editions and is unchanged.
• Volume numbering within the Dīgha Nikāya is also consistent
between the two.
• Volume division and numbering within the Majjhima Nikāya,
Saṃyutta Nikāya, and Aṅguttara Nikāya is inconsistent and has
been left according to the CST4.
• Volumes of the Khuddaka Nikāya are listed under their separate
titles rather than by number, as is the norm for these works.
The inconsistent volume numbering for the Majjhima Nikāya, Saṃyutta
Nikāya, and Aṅguttara Nikāya is unfortunate. Reconstructing the CST4
electronic files to follow the PTS numbering would have been possible but quite
laborious and so was not attempted at this time.

Vin.I – Vin.V Vinaya Piṭaka volumes I – V

D.I – D.III Dīgha Nikāya volumes I – III
M.I – III Majjhima Nikāya volumes I – III
S.I – V Saṃyutta Nikāya volumes I – V
A.I – A.IV Aṅguttara Nikāya volumes I – IV
Khp Khuddakapāṭha
Dhp Dhammapada
Ud Udāna

13
The CST4 numbering for what PTS labels volumes I through V would be III, IV, I, II, V.

130
 Toward a Computational Analysis of the Pali Canon

It Itivuttaka
Sn Suttanipāta
Vv Vimānavatthu
Pv Petavatthu
Thag Theragāthā
Thig Therīgāthā
Ap.I Therāpadāna
Ap.II Therīapadāna
Bv Buddhavaṃsa
Cp Cariyāpiṭaka
Ja.I – J.II Jātaka volumes I – II
Nidd.I Mahāniddesa
Nidd.II Cūḷaniddesa
Patis Paṭisambhidāmagga
Nett Nettippakaraṇa
Abh.I – Abh.VII Abhidhamma Piṭaka volumes I – VII

Table 3: Abbreviations for Tipiṭaka volumes used in figures

Acknowledgements
The texts used in this analysis were generously provided by Frank Snow at
Tipitaka.org. Boris Veytsman and Gully Burns provided valuable feedback on
an early draft, as did an anonymous peer reviewer, who also corrected some
of my misconceptions on the Khuddaka Nikāya in general and Suttanipāta
in particular. Alexander Wynne corrected several more errors in his thorough
editing of the manuscript, which I greatly appreciated. I was introduced to the
techniques of computational text mining at the Chan Zuckerberg Initiative in
2019–2020 under the tutelage of the superb team of research scientists working
there, including Drs. Veytsman and Burns, as well as Ana-Maria Istrate, Sunil
Mohan, and Ivana Williams. This work would also not have been possible
without the software packages cited in the text and the great labor that has gone
into developing and maintaining them, for which I am very grateful.

131
 Toward a Computational Analysis of the Pali Canon

Bibliography
Alfter, D. (2014). “Morphological analyzer and generator for Pali.” Bachelor thesis,
University of Trier, Department of Digital Humanities and Computational
Linguistics. https://arxiv.org/pdf/1510.01570.pdf
Basapur, S., V, Shivani, and Nair, S. S. (2019). “Pāli Sandhi – A Computational
Approach.” Proceedings of the 6th International Sanskrit Computational
Linguistics Symposium.
Bodhi, Bhikkhu (trans). (2012). The Numerical Discourses of the Buddha. Wisdom
Publications.
Buddhadatta, A.P. (2014). Concise Pali-English Dictionary. Motilal Banarsidass
Publishers.
Carter, J.C., and Palihawadana, M. (1987). The Dhammapada. Oxford University
Press.
Elwert, F., Sellmer, S., Wortmann S., Pachurka, M, Knauth, J., and Alfter, D. (2015)
“Toiling with the Pāli Canon.” Proceedings of the Workshop on Corpus-Based
Research in the Humanities. Warsaw, Poland.
Fellows, I. (2018). wordcloud: Word Clouds. R package version 2.6.
https://CRAN.R-project.org/package=wordcloud.
Fronsdal, G. (2005) The Dhammapada. Shambhala Publications.
Geiger, Wilhelm. (2005) A Pāli Grammar. Pali Text Society.
Gombrich, Richard (2018) Buddhism and Pali. Mud Pie Books.
von Hinüber, Oskar. (2015). A Handbook of Pali Literature. Munshiram Manoharlal
Publishers.
Jockers, M. L. (2013). Macroanalysis: Digital Methods & Literary History.
University of Illinois Press.
Jockers, K.L and Thalken, R. (2020). Text Analysis with R: For Students of Literature.
Springer Nature Switzerland.
Johnson, K. P., Burns, P., Stewart, J and Cook, T. (2014). CLTK: The Classical
Language Toolkit. https://github.com/cltk/cltk
Kassambara, A. (2017). Practical Guide to Cluster Analysis in R: Unsupervised
Machine Learning. STHDA.
Kassambara, A. and Mundt, F. (2020). “factoextra: Extract and Visualize the Results of
Multivariate Data Analyses.” https://CRAN.R-project.org/package=factoextra

132
 Toward a Computational Analysis of the Pali Canon

Kulkarni, M. and Dangarikar, C. (2013). Recent Researches in Sanskrit

Computational Linguistics: Fifth International Symposium IIT Mumbai, India,
January 2013 Proceedings. D.K. Print World.
Lewis, D. D., Yang, Y, Rose, T. G., and Li, F. (2004). “RCV1: A New Benchmark
Collection for Text Categorization Research.” J. Mach. Learn. Res. 5, 361–397.
Lloyd, S. P. (1957, 1982). Least squares quantization in PCM. Technical Note,
Bell Laboratories. Published in 1982 in IEEE Transactions on Information
Theory, 28, 128–137.
MacQueen, J. (1967). Some methods for classification and analysis of multivariate
observations. In Proceedings of the Fifth Berkeley Symposium on Mathematical
Statistics and Probability, eds L. M. Le Cam & J. Neyman, 1, pp. 281–297.
Berkeley, CA: University of California Press.
Norman, K. R. (2001). The Group of Discourses. Pali Text Society.
Pali Text Society. (1925). The Pali Text Society’s Pali-English Dictionary. Pali Text
Society.
Project Gutenberg (2020). The King James Bible. http://www.gutenberg.org/
ebooks/10900.
R Core Team (2013). R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-
project.org/
Silge, J., and Robinson, D. (2016). “tidytext: Text Mining and Analysis Using Tidy
Data Principles in R.” JOSS, 1(3). doi: 10.21105/joss.00037, http://dx.doi.
org/10.21105/joss.00037.
Sujato, Bhikkhu, and Brahmali, Bhikkhu. (2014). The Authenticity of the Early
Buddhist Texts. Supplement to Journal of the Oxford Centre for Buddhist
Studies, Volume 5 (https://ocbs.org/journal-supplements/)
Vipassana Research Institute (2020). Chaṭṭha Saṅgāyana Tipiṭaka Version 4.0.
http://tipitaka.org/
Webb, Russell (ed.). (2011). An Analysis of the Pali Canon. Buddhist
Publication Society.
Wickham et al., (2019). “Welcome to the tidyverse.” Journal of Open Source
Software, 4(43), 1686, https://doi.org/10.21105/joss.01686
Zigmond, D. (2020). “tipitaka: Data and Tools for Analyzing the Pali Canon.” R
package version 0.1.1. https://CRAN.R-project.org/package=tipitaka
Zipf, G. K. (1935) The Psychobiology of Language. Houghton-Mifflin.

133
 Toward a Computational Analysis of the Pali Canon

Appendix
ati kati tassaṃ paricca re
atīva kadā tassā pariññā labbhā
atha kamhi taṃ pariyādāya lesa
atho kayaṃ tā pātur va
adu kasmā tāni pi vaka
anu kasmiṃ tāya puna vata
anti kassa tāyaṃ purā vā
anto kassaṃ tāyo pure vāhasā
api kassā tāsaṃ ba vi
abhito kaṃ tāsānaṃ byā/vyā vinā
ambho kā tāsu bha vinidhāya
amma kāni tāhi bhaṇe viparakkamma
amhākaṃ kāya ti bho viya
amhe kāyo tu maññe vivicca
amhesu kāsaṃ tuṇhī mama visuṃ
are kāsānaṃ tumhaṃ mamaṃ vīsati
alaṃ kāsu tumhākam mayā ve
alālā kāhi tumhākaṃ mayi vo
assu kiṇi tumhe mayhaṃ sakkā
aha kim tumhesu maṃ samma
ahaṃ kimhi tumhehi mā sammā saha
ahe kismā tuyhaṃ murumurā sā
aho kismiṃ tuvaṃ me sāgataṃ
ā kiṃ te yagghe su
ādu ke tena yadi suṭṭhu
āma kena tesaṃ yamhā sudaṃ
ārabbha kesaṃ tesānaṃ yamhi suru
ārā kesānaṃ tesu yasaṃ sū
āsajja kesu tehi yasānaṃ so
āsu kehi tvayā yasmā soḷasa
iti ko tvayi yassa ha
ito kvaṇ tvaṃ yassaṃ hañci
ittha khalu dabhakkaṃ yassā han
itthaṃ kho diṭṭhā yaṃ handa
ida ca dhi yā hambho

134
 Toward a Computational Analysis of the Pali Canon

idāni cana na yāni have

idha ci nanu yāya haṃ
iṅgha ce nānā yāyaṃ haṃsi
iva codanā nāma yāyo hā
iha jātu nu yāvatā hi
uda taggha nūna yāsaṃ hiṅkāra
udāhu tamhā neva yāsu huṃ
uddissa tamhi no yāhi he
uddhaṃ tayā paññāya ye heṭṭhā
upanidhāya tayi paṭi yena
upari tava paṭikacca yeva
upasagga tavaṃ paṭṭhāya yesu
ubbhaṃ tasmā pati yehi
ūhacca tasmiṃ pada yo
kacci tassa pana ruṇ

Table 4: Tentative “stop words” for Pali

135
 Nissāraṇīya:
A Codified Term Updating the Development of the Pāli Vinaya, Part I

Juo-Hsüeh Shih

Abstract
Nissāraṇīya is a term added to conclude the saṅghādisesa rules for nuns
only in the Pāli Vinaya. It refers to a temporary expulsion of the guilty
nun, yet this is beyond the penalty prescribed. A comparative study of the
relevant passages in the other Vinayas attests to the controversy hinted
at in the Sp. The Pāli Vinaya is alone in asserting the expulsion of the
nun, whereas the other traditions are concerned with the nun’s release
from her offence. The key to such controversy lies in orthographical
variation: nissāraṇīya vs niḥsaraṇīya. Our study points to the assumption
that the Vinaya may have borrowed a term from the Suttas to supplement
the offence name saṅghādisesa. It was nissaraṇīyaṃ saṅghādisesaṃ
in the Pātimokkha, which is confirmed by internal evidence from the
Sp. Nissaraṇīya was later replaced by nissāraṇīya and its meaning and
reference underwent a dramatic change. Moreover, nissāraṇīya then
found its way back into the Suttas in which there is some confusion
between nissaraṇīya and nissāraṇīya.

. 21(20): 136–162. ©21 Juo-Hsüeh Shih

 Nissāraṇīya

Preliminaries
In the Pāli Vinaya nissāraṇīya appears solely in the Saṅghādisesa chapter
of the Bhikkhunī Vinaya with one exception in the Parivāra: nissāraṇīyaṃ
paññattaṃ, ‘enactment of expulsion’. Its variant forms nissāraṇā and nissare
make a couple of appearances in the Mahāvagga and Parivāra, respectively.
When a monk violates a saṅghādisesa rule, he is said to have committed an
offence of saṅghādisesa, ‘an offence entailing legal acts of the Saṅgha (for its
removal)’, whereas in the case of nuns, the offence now has an additional term
qualifying it: nissāraṇīyaṃ saṅghādisesaṃ.
What does nissāraṇīya mean in this context? As we will see, in the canonical
commentary (padabhājana, ‘Analysis of Words’ = AW), nissāraṇīyaṃ is
glossed as saṅghamhā nissāriyati, ‘she is made to leave the Saṅgha’. The post-
canonical commentaries, particularly the Samantapāsādikā (Sp), reinforces this
position by making the canonical gloss even clearer. With regard to whether
there is indeed something extra, there are opinions pro and con. Those who
believe that nissāraṇīya denotes something extra agree upon the temporality of
such expulsion, besides this, however, nothing about how, when and where to
put this into practice is found in any Vinaya literature.
The term nissāraṇīya gives rise to different interpretations, probably because
the term is new (i.e. absent from the Bhikkhu Pātimokkha = BhuPām) and the
penalty of expulsion is beyond the scope of the mending procedures for an
offence of saṅghādisesa. Without reliable clues, the meaning and reference of
nissāraṇīya remain arguable and the problem whether nissāraṇīya denotes an
extra punishment remains pending.
Despite all these ambiguities and uncertainties, we must not overlook what the
Pāli Vinaya has ever said. According to the canonical texts and commentaries, there
can be no doubt that for the Pāli tradition nissāraṇīya denotes the expulsion of the
nun guilty of a saṅghādisesa offence. This is the starting point for our investigation.
Section I discusses whether nissāraṇīya denotes an extra punishment
or involves nothing extra. A brief summary of the penalty for an offence of
saṅghādisesa will first be presented to show the procedures required of the
offenders to escape from their offences. All the procedures take place within the
monastery. Nowhere in the texts is there ever an indication that an offender is to
be expelled from the Saṅgha. In this respect the Pāli Vinaya is very limited. It is
therefore necessary to collate the other Vinayas to advance our understanding of
the saṅghādisesa offence for nuns.

137
 Nissāraṇīya

The Sp’s commentary deserves special attention in that it hints at an

existing controversy over how the additional term qualifying saṅghādisesa was
understood. To verify such controversy, a comparison of the closing phrases of
the saṅghādisesa rules in the other Vinayas will be provided. This comparison
demonstrates that none of them suggest anything extra. More importantly,
the other texts read niḥsaraṇīya, derived from a normal stem, whose Pāli
correspondence should be nissaraṇīya, not nissāraṇīya as is seen in the present
Pāli Bhikkhunī Vinaya (= BhīVin).
Section II draws the attention back to the Pāli Vinaya to explore further
the issue in question. As mentioned above, other traditions read niḥsaraṇīya
whereas the Pāli Vinaya has nissāraṇīya. Is the additional term from a normal
stem: niḥsaraṇīya/nissaraṇīya, or is it from a causative stem: nissāraṇīya?
Which is the right word, or are they just interchangeable? While a difference of
terms may not really matter, their interpretation does, especially when the two
differing forms of the same word may lead to completely opposite results.
The evidence of the Sp is important in broadening the scope of our investigation
and deepening our perspective. It was nissaraṇīya in the Pātimokkha, in
agreement with the reading niḥsaraṇīya in the other Vinayas available to us.
If this is the case, we may infer that once there was full agreement among the
various traditions on what the additional term niḥsaraṇīya/nissaraṇīya refers to.
It is interesting to ask why and when an additional term was added to qualify
the offence title saṅghādisesa. Section III turns to the Suttas in search of the
possible origin of the term added. While nissaraṇīya was a later addition in the
Vinaya, this term and the variant form nissaraṇā are relatively well attested in
the Suttas, with a meaning and reference that fits the context of the saṅghādisesa
offence perfectly. A comparative study also shows that in the Skt texts of both
the Sūtras and Vinaya, niḥsaraṇīya literally means ‘going out’, hence escape or
leading to freedom in a soteriological sense. There is no disagreement on the
usage of niḥsaraṇīya/nissaraṇīya among the Pāli Suttas, Skt Sūtras and Vinaya.
The investigation of Section III leads to the assumption that the Pāli Vinaya
may have borrowed the term nissaraṇīya from the Suttas, yet as can be seen,
the term now is nissāraṇīya. Section IV looks further into the phenomenon of
the nissaraṇīya/nissāraṇīya confusion in the Pāli Suttas. If the Pāli Vinaya had
retained the original form of nissaraṇīya, the causative derivative nissāraṇīya
could not have come into being. The Pāli Vinaya is the only source for
nissāraṇīya, thus it must have contributed to the above-mentioned confusion.

138
 Nissāraṇīya

I. Nissāraṇīya: Nothing Extra or Something Extra?

To judge whether or not nissāraṇīya indicates an extra punishment, it
is necessary to see what the penalty for an offence of saṅghādisesa is.
Saṅghādisesa is an offence next to Defeat (Pārājika) in gravity, and can be
amended. A monk violating a saṅghādisesa rule is subject to the penalty of
parivāsa, literally ‘living apart’ (or ‘probation’), if he conceals his offence,
which is followed by a six nights’ penance (mānatta). Since a pārivāsika
monk (a monk undergoing parivāsa) is not allowed to share a dwelling
under the same roof with regular monks, he has to live alone in separate
quarters. A guilty monk undertakes six nights’ mānatta straightforwardly
without parivāsa if he does not conceal his offence.
For nuns there is no parivāsa: because staying alone is not befitting for
a nun, she is exempted from this.1 But the duration of mānatta is, for nuns,
extended to fourteen days. On the completion of mānatta, an offender is
entitled to request the Saṅgha for rehabilitation (abbhāna), through which
one becomes purified of guilt and is re-admitted into the Saṅgha as a
regular member.
There is no substantial difference between the penalties for parivāsa and
mānatta except that the duration of the latter is fixed, whereas that of the former
varies according to the length of concealment. The Cv enumerates ninety-four
observances for parivāsa, but they also apply to mānatta with some variations
(Vin II 31ff).
In the first place, a parivāsika monk has to report his case to any incoming
monks or to the monks he visits. He is to inform the Saṅgha of his status on
the occasions of Pātimokkha-recitation (Uposatha) and Invitation (Pavāraṇā).
There is no way of escaping; in the case of illness, he must have someone report
on behalf of him.2 A mānattacārika monk (a monk undergoing mānatta) needs
to report his status to the Saṅgha on a daily basis.3 In the case of a nun, she has
to make daily report to both Saṅghas.

1
KKh 166,25-26 has it that if a nun conceals her offence against a saṅghadisesa rule, she
is guilty of a wrong-doing even though there is no parivāsa for her. (bhikkhuṇiyā hi āpattiṃ
chādentiyāpi parivāso nāma n’atthi chādanappaccayāpi pana dukkaṭaṃ āpajjati.) This verdict is
not found in the canonical commentary, nor in the Sp.
2
Vin II 32,19-22.
3
Vin II 35,26-30: mānattacārikena bhikkhave bhikkhunā āgantukena ārocetabbaṃ, āgantukassa
ārocetabbaṃ uposathe ārocetabbaṃ, pavāraṇāya ārocetabbaṃ devasikam ārocetabbaṃ. Sace
gilāno hote dūtena pi ārocetabbaṃ.

139
 Nissāraṇīya

Secondly, a parivāsika monk may visit monks belonging to his own

community if he can reach there on the same day.4 He must not go about or
visit monks belonging to another community (nānāsaṃvāsakā) without being
accompanied by a regular monk unless in an emergency.5 The same applies to a
mānattacārika monk except he is allowed to go out with the Saṅgha.6
In addition to the above regulations, one reads among the ninety-four observances
that a mānattacārika monk should not live away from regular monks, or stay alone in
the forest (in order to avoid the embarrassment of being deprived of many privileges
granted to a regular monk), or avoid reporting his present status to the monks he
meets,7 or stay under the same roof as a regular monk, whether it be a residence or
not.8 A failure to observe these restrictions will incur a “break” (ratticchedā).9 The
above restrictions are to be observed also by a guilty nun. She is not allowed to dwell
under the same roof with regular nuns, and is not allowed to live alone or away from
nuns belonging to the same community. She is therefore supposed to live with a
companion assigned by the Saṅgha in a separate quarter within the nunnery.10 The
foregoing discussion shows that the penalties for an offence against saṅghādisesa
do not involve the culprit’s expulsion from the Saṅgha.11
As regards the placement of the term nissāraṇīya before saṅghādisesa, the
AW offers the following terse gloss on Saṅgh 1 (N):
Nissāraṇīyan ti saṅghamhā nissāriyati. ‘Expulsion means she is
made to leave the Saṅgha.’ (Vin IV 225,7)

4
Vin II 33,5-12.
5
Vin II 32,22-33,5.
6
Vin II 35,32-36,7. Mostly the same as the above quotation, but read mānattacārikena for
pārivāsikena and aññatra saṅghena for aññatra pakatattena. Here Saṅgha means, according to
the Sp (1170,21-23), a chapter of four or more monks.
7
Vin II 32,17: na āraññakaṅgaṃ samāditabbaṃ. Sp 1164,21-23 glosses: na āraññakaṅgan ti
āgatāgatānaṃ ārocetuṃ harāyamānena araññikadhitaṅgaṃ na samādātabbaṃ.
8
Vin II 33,12-15: na bhikkhave pārivāsikena bhikkhunā pakatattena bhikkhunā saddhiṃ ekacchanne
āvāse vatthabbaṃ, na ekacchanne anāvāse vatthabbaṃ, na ekacchanne āvāse vā anāvāse vā vatthabbaṃ.
9
Vin II 36, 21-24:Cattāro kho Upāli mānattacārikassa bhikkhuno ratticchedā: sahavāso
vippavāso anārocanā une gaṇe caraṇan ti. (There are, Upāli, four kinds of break: living under
the same roof as a regular monk, living away from [the regular monks], failing to report [daily his
case to the Saṅgha], and going about in less than a group.)
10
This is confirmed by the tenth chapter of the Cv (Vin II 279,22-24) in which another nun was
assigned as a companion (dutiyā) to a nun who had to undergo mānatta.
11
For a detailed discussion on the technical aspects of the saṅghādisesa penalties, cf. Nolot
1996, SVTT III, pp. 116-136.

140
 Nissāraṇīya

The translation here is based on the traditional position of the Pāli Vinaya,
yet this gloss could be interpreted differently. Nissāraṇīya is a gerundive of
nissāreti, derived from the causative stem of niḥ-√sṛ, which means to go out,
depart, or withdraw. The passive niḥsāriyati/nissāriyati means “being caused to
go out, turn out”, and hence to be removed or expelled.
One may take either the offence or the guilty nun as the subject of nissāriyati.
In the case of the former, it means “the offence is removed (Literally: made to go
out) from the Saṅgha. The gloss at Saṅgh 9 for nuns (Vin IV 240,21: nissāraṇīyan
ti saṅghamhā nissāriyati) could be read in this way. The introductory story to this
rule recounts that some nuns lived in close association (bhikkhuniyo saṃsaṭṭhā
viharanti). While the guilty nuns are in the plural, nissāriyati remains singular.
If this is the correct reading, the offence must be the subject of nissāriyati and
hence nissāraṇīyaṃ saṅghādisesaṃ means: an offence entailing legal acts of the
Saṅgha, through which the offence is removed (Literally: ‘the offence is made
to go out from the Saṅgha’). It is worthy of note that in the commentarial texts
the offence is always the referent (grammatical subject), although the exposition
ends up with the nun (grammatical object → logical subject) being expelled.
One may, however, argue that the singular nissāriyati could be merely a
formalistic error, a certain expression being repeated automatically. Should this
be the case, one could read nissāriyanti instead of nissāriyati. Oldenberg held
this opinion,12 yet no manuscript evidence is adduced. If, however, we take the
offence as the subject of nissāriyati, the additional expression then adds no new
idea; on the contrary, it makes clear the final result of the amending proceedings.
It may thus serve as a supplement to the term saṅghādisesa, whose meaning
is not self-explanatory. If we take the guilty nun to be the subject, as the Pāli
commentaries have done, an immediate difficulty comes up. Following the gloss
on nissāraṇīya is that on saṅghādisesa:
An offence entailing legal acts of the Saṅgha means: on account
of her offence the Saṅgha inflicts the mānatta penalty13 [upon her],
draws [her] back to the beginning,14 and rehabilitates [her]. These

12
Horner agrees with Oldenberg, Cf. BD III xxxvi.
13
A summary of this penalty dealt with in the Cv is given in Nolot 1996, “SVTT II” pp. 116ff.
14
Mūlāya paṭikassati. If the offending nun commits another offence of the same category while
undergoing the penalty of mānatta (six nights’ duration for monks but a fortnight’s duration for
nuns), she then has to retake the penalty from the beginning. For more information, cf. the third
chapter of the Cv (Vin II 44ff)

141
 Nissāraṇīya

things are carried out neither by several nuns nor by one single nun,
it is therefore called an offence entailing legal acts of the Saṅgha.
Legal act is indeed the name of this class of offence, thus it is called
an offence entailing legal acts of the Saṅgha.15
This exposition is formulated on the model of that for monks, and our
foregoing discussion on the penalty for an offence of saṅghādisesa has shown
that all the mending procedures are carried out within the compound of the
Saṅgha, which does not involve expelling the culprit out of the monastery.
As the gloss on nissāraṇīya comes first indicating expulsion of the guilty
nun from the Saṅgha, it is strange that what immediately follows suggests no
expulsion at all.
I.B. Horner takes the guilty nun as the subject and remarks: “Nissāraṇīya,
involving being sent away, adds nothing to the saṅghādisesa penalty incurred
by a nun, and hence makes no difference in the penalty imposed on monks and
nuns for having committed such an offence. Only the words, as found in each
‘rule’ of the Nuns’ Saṅghādisesas, is extra.”16 (BD III xxxvii)
By “being sent away” Horner means a temporary exclusion (BD III
xxxvi), but it is not clear what that exactly refers to and how it will be put
into practice; this disagrees with the gloss on saṅghādisesa that immediately
follows. Nevertheless, later on she shifted her position. In rendering the phrase
saṅghamhā dasa nissare, she takes the offence to be the subject: ‘ten are to be
escaped from by means of the Order’ (BD III xxxvi). We shall come back soon
to this subject (see below p. 147). In commenting on the term nissāraṇīya, the
Sp writes:
Expulsion means it (her offence) causes the nun to be expelled from
the Saṅgha. But in the AW, to expound this meaning, ‘expulsion’ is
explained as ‘she is expelled from the Saṅgha’. Here the meaning
should be understood in this way: the offence, having committed
which the nun is expelled from the Saṅgha, that is to be removed. It

15
Vin IV 225,8-12: Saṅghādisesan ti saṅgho ‘va tassā āpattiyā mānattaṃ deti
mūlāya paṭikassati abbheti na sambahūla na ekā bhikkhunī vuccati saṅghādiseso ti.
tass’ eva āpattinikāyassa nāma kammaṃ adhivacanaṃ tena pi vuccati saṅghādiseso ti.
16
Édith Nolot concludes her detailed discussion on nissāraṇā/nissāraṇīya with two hypotheses
posited by others reflecting opinions pro and con: nothing extra vs something extra. Nolot 1999,
“SVTT V”, pp. 54-55.

142
 Nissāraṇīya

is indeed not that very offence which is removed from the Saṅgha
[by anyone], but it is the nun who is expelled from the Saṅgha
because of that offence. Therefore “expulsion” means that [offence]
causes (her) to be removed.17
The first commentary on the Sp, the Vajrabuddhi-ṭīkā, is silent on this subject,
perhaps because the Sp has made what nissāraṇīya refers to clear enough. In
commenting on the phrase bhikkhuniṃ saṅghato nissāretī, the Kaṅkhāvitaraṇī-
purāṇaṭīkā reinforces the standpoint that the offence, as the agent, is indeed the
cause for nissāraṇā, and so it reads nissāraṇīyo to explain the reason why the
nun is expelled: her offence causes her to be expelled.18
The following information can be extracted from the Sp’s commentary: 1. There
exists a controversy over what is to be removed from the Saṅgha: the guilty nun or
the offence committed; 2. The opponents consider the offence as the referent; 3. The
Sp also takes the offence as the referent but explains that the offence causes the nun
to be expelled.19 Note that a grammatical concern is involved here, and that there is a
consensus among the Pāli and other Vinayas that the added word refers to the offence.
In view of the controversy, one would expect the Sp to have an opinion
on the referent different from its rivals’. Surprisingly it was not the case. The
logic of the Sp’s interpretation precludes the possibility of the nun as the
agent. The added word in the Pāli reads nissāraṇīya, a causative derivative.
If the nun is taken as the referent, nissāraṇīya would mean: [A] The nun
(grammatical subject) causes the offence (grammatical object → logical
subject) to be removed. This will happen after the nun has undergone required
amends. But if the offence is taken as the referent, the interpretation in the
passive voice will lead to a result which the Sp desires: [B] The offence
(grammatical subject) causes the nun (grammatical object → logical subject)
to be expelled.

17
Sp 908,5-11: bhikkhuniṃ saṅghato nissāretī ti nissāraṇīyo, taṃ nissāraṇīyaṃ padabhājane
pana adhipāyyamattaṃ dasettuṃ saṅghamhā nissārīyatī ti vuttaṃ. tattha yaṃ āpannā bhikkhunī
saṅghato nissārīyati so nissāraṇīyo ti eveṃ attho daṭṭhabbo, na hi so eva dhammo saṅghamhā
nissārīyati, tena pana dhammena bhikkhunī nissārīyati. tasmā so nissāretī ti nissāranīyo.
18
Bhikkhuniṃ saṅghato nissāretī ti āpannaṃ bhikkhuniṃ bhikkhunisaṅghamhā nissāretī.
Hetumhi cāyaṃ kattuvohāro “nissāraṇahetubhūtadhammo ‘nissāraṇīyo’ ti vutto” katvā. (Chaṭṭha
Saṅgāyana online edition)
19
Édith Nolot argues that in the Pāli texts, nissāraṇa/nissāraṇīya refer “exclusively to persons,
not to objects.” (Nolot, “SVTT IV-X”, p. 52). Noting “that [offence] causes to expel [her]”, the Sp’
exposition shows that the guilty nun is an indirect referent of nissāraṇīya.

143
 Nissāraṇīya

Grammatically speaking, both are all right, but in terms of connotation, they
make a great difference. [B] suggests that the expulsion must take place in the
beginning, in that the prescribed mending procedures conclude with the nun’s
being reinstated. This may explain why most of the discussants on the nissāraṇīya
problems focus their attention on the mānatta penalty, taking nissāraṇīya as
referring to some sort of “dismissal” or “isolation” during the period of undergoing
mānatta.20 Nevertheless, our subsequent discussion will demonstrate that
nissāraṇīya can be something really extra to the traditional set of the saṅghādisesa
proceedings. As [A] is not favored by the Pāli tradition, the Sp, in support of
[B], must comply with the traditional view on the referent. Despite the dictional
variation, the core of the controversy seems to be a matter of interpretation, yet to
reach a desired interpretation, a corresponding wording has a role to play.
A comparison of the Saṅghadisesa chapters of other extant Vinayas finds
the Sp to have stood firm but alone, arguing against almost all traditions. Three
Vinayas, the Mahāsaṅghīka (Mā), Chinese and Tibetan Mūlasarvāstivādin
(= CMū and TMū, respectively), contain no additional expression qualifying
the term saṅghādisesa. It is evident that there exists no expulsion of the
guilty nun in these traditions.
In the Mahīśasaka (= Mī) and Sarvāstivāda (= Sa) Vinayas, one does read an
additional expression.
Mī, T22[1421]79a16-17: 是比丘尼初犯僧伽婆尸沙，可悔
過。This nun commits a first-offence saṅghāvaśeṣa, [which is] a
repentable fault.21
Sa, T23[1435]b4: 是法初犯僧伽婆尸沙，可悔過。This is a first-
offence saṅghāvaśeṣa, [which is] a repentable fault.
Waldschmidt in his work Bruchstücke des Bhikṣuṇī-Prātimokṣa der
Sarvāstivādins reconstructs niḥsaraṇīya for the Chinese rendering “a repentable
fault”. His reconstruction is in fact corroborated by a tiny bit of evidence
from a Sanskrit text of the Bhikṣuṇī Prātimokṣa. It is a four-line fragment
of Saṅghavāśesa 8-9, and the third line reads: “[D]harmaḥ Pratthamāpattiḥ
saṅghavāśesa niḥsa…”22

20
Nolot, “SVTT IV-X”, pp. 54-55.
21
Ann Heirmann renders 可悔過 as “it has to be confessed”. Heirmann 2002, Part II, p. 388, fn. 10.
22
Finot, PrMoSū(Sa), p. 549.

144
 Nissāraṇīya

The ending phrase of the corresponding rules of the Dharmaguptaka (= Dha)

is the closest version to that of the Pāli.
Dha, T22[1428]1032a10f: [是]比丘尼犯初法，應捨，僧伽婆尸
沙。[T]his nun commits a first-offence, [which/who] should be
abandoned/removed, and this entails legal acts of the Saṅgha.
There is some ambiguity in the Chinese rendering, that is to say, the referent
of the additional term could be the guilty nun herself or the offence. I shall deal
with this problem in the following discussion. (See below pp. 148-149)
It is obvious that in the Mā, CMū, TMū, Mī and Sa, no expulsion is indicated
by the additional expression, and the word order in the Mī and Sa is different:
niḥsaraṇīya comes after saṅghavāśesa. As is shown above, the Sp’s commentary
holds a strong position against an opposite opinion: it is the offence that is to
be removed from the Saṅgha. This is exactly what is spelt out in the Bhikṣuṇī
Vinaya of the Mahāsāṅghika-Lokottaravādin (= BhīVin(Mā-L)).
Bhī(Mā-L) 103,5-7: ayaṃ dharmo prathamāpattiko saṃghātiśeso
upādiśeṣo saṃgho saṃghaṃ evādhipati kṛtyā niḥsaraṇīyo.
‘This is a first-offence saṃghātiśeso.23 Having remainder in the
Saṅgha; with the Saṅgha having acted as an authority, this offence
should be removed.’
With dharmo (the offence) as the subject and qualified by niḥsaraṇīyo, the
statement, formulated in this way, leaves no room for any other interpretation.
But this does not necessarily mean the Mā-L also accepted the additional term
niḥsaraṇīya. This term together with its explanation as quoted above appears only
in the first and the last rule; this presumably means the explanatory phrase is to
be carried throughout the entire chapter. The text as we have it now is a complete
version of the BhīPām, and it would seem that when the text was compiled,
the redactor(s) must have been aware of the controversy over nissāraṇīya/
niḥsaraṇīya. Being a sub-sect of the Mā, which added no term to saṅghātiśeso,
the Mā-L may have interpolated the term niḥsaraṇīya with a precise explanation.
This can be seen as an opponent’s response to the Pāli tradition.

23
In fn. 4 (BhīVin(Mā-L) , 103), Roth suggests a translation of upādiśeso saṃgho: “groups of
offences (saṃgha) which is the supplement (śesa) to the first group (upa+ādi) [the group of the
Pārājika-offences]”.

145
 Nissāraṇīya

The Pāli Vinaya insists that nissāraṇīyaṃ saṅghādisesaṃ means the

guilty nun should be expelled from the Saṅgha, whereas in the Mā-L Vinaya,
niḥsaraṇīyo means the offence should be removed from the Saṅgha. The Vinayas
having an additional expression are all the Sthāvira-affiliated schools, this fact
betrays the possibility that the additional expression most likely originated from
the Sthāvira side with the wording niḥsaraṇīya/nissaraṇīya. Nissaraṇīya had
later been replaced by nissāraṇīya in the Pāli and its reference had undergone
dramatic change from removing the offence to expelling the guilty nun. It cannot
be more evident that the Mā-L took a position opposite to the Pāli.
Now we notice that instead of nissāraṇīya, what the other Vinayas read
is a different word niḥsaraṇīya. As is mentioned previously, the variant form
of this word must have contributed to the controversy. The Pāli Vinaya alone
reads nissāraṇīya, insisting that the guilty nun should be expelled from the
Saṅgha. Given the expulsion is temporary, it is something extra to the traditional
prescription. For the other Vinayas which reads niḥsaraṇīya, there is nothing
extra in that this term is supplementary. Why is there such controversy? Which
is the right word in the Pāli Vinaya, nissaraṇīya or nissāraṇīya? Now we turn
to this very issue.

II. Nissaraṇīya or Nissāraṇīya

To remove the offence or to expel the nun? It indeed was the addition of a new term
that had given rise to such controversy. Why and when was the term added? As
is understood, those rules for nuns were originally embedded in the BhuVin, and
from which the rules for nuns were latter extracted to form an independent BhiVin.
It was presumably around this time that a new term may have been attached to
qualify the word saṅghādisesa, presumably for the purpose of refining the text. The
BhuPāms of the various traditions agree to a great extent in terms of the numbers
and contents of the rules and their sequential order. On the contrary, the BhīPāms
vary to a great extent in every aspect, which may suggest a poor textual transmission
or free composition (?) during the sectarian period. Moreover, from the fact that not
all the Vinayas (i.e. Mā, CMū and TMū) have a term added to saṅghādisesa, one
may infer that sometime in early history of the BhīPāms a certain school innovated
to add a new term; some schools followed but some did not.
All traditions that have an additional expression reads niḥsaraṇīya, but it
is nissāraṇīya in the Pāli text alone. It is interesting to further explore whether
nissāraṇīya is the original form of the additional term in the Pāli Vinaya.

146
 Nissāraṇīya

Nissāraṇīya/nissāraṇā appears only in some contexts concerning penalties.

Throughout the Pāli Vinaya one reads the form derived from a causative stem
consistently with one exception. In the eighth chapter (Gāthāsaṃgaṇika:
Collection of Stanzas) of the Parivāra (Vin V 144ff), stanzas 10-30 form a
passage relating how many Pātimokkha rules are shared or not shared by nuns
and so on. When it comes to the training rules peculiar to nuns, the text says:
pārājikāni cattāri saṅghamhā dasa nissare: ‘There are four Defeats and ten
should go out from the Saṅgha’ (Vin V 147,23).
Nissare is an optative form of niḥ-√sṛ derived from the normal stem, hence
“ten should go out” here must refer to the ten offences committed, which should
be removed by means of the legal acts of the Saṅgha. In glossing this stanza,
the Sp writes:
Ten should go out from the Saṅgha: in the Vibhaṅga it is said that
[she] is made to leave the Saṅgha, but in the Pātimokkha ten have
come down to us with the wording: nissaraṇīyaṃ saṅghādisesaṃ.24
The Sp here spells out that although in the Vibhaṅga it is interpreted in
the sense of expulsion (nissāraṇīya), in the Pātimokkha it is nissaraṇīyaṃ
saṅghādisesaṃ, which means ‘an offence entailing legal acts of the Saṅgha,
through which the offence should be escaped’.
It should be noted that nissaraṇīyaṃ saṅghādisesaṃ in this context records
no variant reading, and this means the Sinhalese and Burmese manuscripts
and printed texts consulted by the PTS editors fully agree upon the form of
nissaraṇīyaṃ. Moreover, from a syntactic point of view, nissaraṇīyaṃ should
be the correct wording. Should this be the case, nissāraṇīyaṃ as we have it
now in the BhīPām must be a later change, and there must be reason(s) for
such change.
As is pointed out above (see above pp. 140-142), the Pāli canonical
commentary (Vin IV 225,7; 240,21; 225,8-12) glosses nissāranīya first and then
saṅghādisesa. Amending a saṅghādisesa offence involves no expulsion, nor are
the proceedings carried out outside the compound of the monastery. It would
be odd to say of expelling the nun first but in what follows no expulsion is
indicated, unless the gloss (nissāraṇīyan ti saṅghamhā nissāriyati) was inserted
here later at a certain time when the added word nissaraṇīya has been changed

24
Sp 1350,4-6: Saṅghamhā dasa nissare ti saṅghamhā nissāriyatī ti evaṃ Vibhaṅge vuttā
mātikāyaṃ pana nissaraṇīyaṃ saṅghādisesan ti evam āgatāni dasa.

147
 Nissāraṇīya

into nissāraṇīya, which points to the culprit’s expulsion outside of the normal
procedures.
None of the Vinaya texts that contain a new expression offers a gloss on
it. It would be no wonder if this new term brings no new idea beyond what
is connoted in the title saṅghādisesa. The text of the Dha is important in this
context. The Pāli and Dha Vinayas resemble each other in many ways, and we
find a remarkable similarity of wording and phrasing in the closing part of the
saṅghādisesa rules.25
Pāli, Vin IV 224,27-8: ayaṃ bhikkhunī paṭhamāpattikaṃ dhammaṃ
āpannā nissāraṇīyaṃ saṅghādisesan ti.
[T]his nun commits a first-offence, which entails legal acts of the
Saṅgha involving expulsion.
Dha, T22[1428]1032a10f: [是]比丘尼犯初法，應捨，僧伽婆尸
沙。
[T]his nun commits a first-offence, [which/who] should be
abandoned/removed, and this entails legal acts of the Saṅgha.
In the case of the Dha, with the referent unindicated, there is some ambiguity
in the expression 應捨 (‘should be abandoned/removed’). What should be
abandoned/removed? The guilty nun or the offence committed?
An examination of the Dha’s renderings for technical terms indicates
that banishment is rendered as 擯 (T22[1428]891a6), and expulsion as 驅出
(T22[1428]889a10). If expulsion of the guilty nun is meant here, we would not
expect to read the character 捨 (‘abandoned/removed’). In view of the Dha’s
terminology, it seems plausible to take the offence as the referent.26
As can be seen, the passages of the Pāli and Dha are word-for-word verbatim.
Because of such resemblance, one would expect to read in the Dha the same
gloss on nissāraṇīya as is seen in the Pāli text. But the Dha, like the Mī and

25
For a detailed discussion on the concluding phrases of the saṅghādisesa rules in the various
Vinayas, cf. Shih 2003, pp. 213-218.
26
Ann Heirmann’s translation reads: the bhikṣuṇī violates an immediate rule, a saṃghāvaśeṣa, that
has to be given up. She has a subsequent discussion on this expression, quoting the corresponding phrases
from the other Vinaya recensions. Heirmann, 2002, Part II, 388-389, fn. 10. In an article (Heirmann
2003, p. 17) she further points out that “In the Dharmaguptakavinaya, the character 捨 is never used
when one excludes (滅、擯）or suspends (舉) a bhikṣuṇī, but is used when one gives up bad behavior.”

148
 Nissāraṇīya

Sa, offers no explanation of the new term. A newly added term requires no
exposition only when its meaning is already known or is readily understandable.
The reason why the Pāli inserted an explanation is probably that the new term
had been changed from nissaraṇīya to nissāraṇīya, and the latter indicated a
new institution.
It is not surprising to read an additional term in the Pātimokkha, as many instances
have demonstrated where the Pāli BhiVin seems to be more advanced in wording
and phrasing, compared to its Bhikkhu counterpart.27 The additional nissaraṇīya
supplements saṅghādisesa in its meaning, and the new phrase nissaraṇīyaṃ
saṅghādisesaṃ would therefore mean: an offence entailing legal acts of the Saṅgha,
through which the offence should be removed or the guilty nun should be released
from her offence. The latter is exactly what the Chinese commentarial text, the 毗尼
母經 Pi-ni-mu Jing (Vinayamātṛka-sūtra), says in explaining how the Saṅgha helps
the offenders remove their offences by means of parivāsa (lit. living apart), mānatta,
and then reinstatement: Having been reinstated, an offender becomes pure and is
“released from the offence” (於所犯處得解脫T24[1463]842c27).

III. Where may have Nissaraṇīya Come from?

The internal evidence studied above has shown that originally it was nissaraṇīya
in the Pātimokkha, and through comparison, the external evidence demonstrates
that it is niḥsaraṇīya in some of the other Vinayas. It becomes clear why
nissaraṇīya is employed here when we look it up in the Suttas, where the term
nissaraṇīya/nissaraṇa occurs in a specific soteriological context. The Itivuttaka
records a discourse on tisso nissaraṇīyā dhātuyo:
There are, monks, these three elements that should be escaped.
Which three? This is the escape from sensuous desires, that is,
renunciation. This is the escape from forms, that is, formlessness,
while cessation is the escape from whatever has come into being,
conditioned, and dependently arisen. These, monks, are the three
elements that should be escaped.28

27
Cf. Shih 2000, p. 24.
28
Iti 61,2-7: tisso imā bhikkave nissaraṇīyā dhātuyo. katamā tisso? kāmānam etaṃ nissaraṇaṃ
yad idam nekkhamaṃ. rūpāṇam etaṃ nissaraṇaṃ yad idam ārupaṃ. yaṃ kho pana kiñci bhūtaṃ
saṅkhataṃ paṭiccasamuppannaṃ nirodho tassa nissaraṇaṃ. imā kho bhikkave tisso nissaraṇīyā
dhātuyo ti.

149
 Nissāraṇīya

In the commentary, nissaraṇa is glossed as “departing” (Iti-a 42,5: nissaraṇan

ti apagamo.), and nissaraṇīya as “connecting with nissaraṇa” (Iti-a 40,24:
nissaraṇīyā ti nissaraṇapaṭisaṃyuttā). The canonical text read nissaraṇīya with
a variant reading of nissāraṇīya recorded. The commentary, however, reads
nissāraṇīya. It is evident that there is some confusion between nissaraṇīya and
nissāraṇīya in the Pāli texts.
The idea of nissaraṇa is well attested in the Suttas, particularly the
Saṃyutta-Nikāya. In the context when the Blessed One recounts his pre-
enlightenment practice, one reads a triple expression of “gratification, danger,
escape” (assāda, ādīnava, nissaraṇa). The triad is a series of subjects to
work on, which leads to final liberation. This formula of praxis is applied
to the four elements, five aggregates, six internal sense bases, and so on.
Whatever the subject may be, nissaraṇa of that subject means the “removal
and abandonment” of it. For instance:
The pleasure and joy that arise in dependence on form, this is the
gratification in form. That form is impermanent, unsatisfactory,
and subject to change; this is the danger in form. The removal
and abandonment of desire and lust for form; this is the escape
from form.29
A similar account with a slightly differing wording recurs in the beginning
of the Sambhodhi-vagga in the AN. Here the triad of “gratification, danger,
escape” is expressed with loke added: “What is the satisfaction in the world?
What is the danger, and what is the escape?” (AN I 258,25-26: ko nu kho loke
assādo ko ādīnavo kiṃ nissaraṇan ti).
Still, another instance shows a differing usage of nissaraṇa with the
ablative: “If there were no escaping from the world, beings in this world
could not escape. But as there is in the world escaping, that is why beings do
escape therefrom.”30

29
SN III 28,2-6: yaṃ kho rūpam paṭicca uppajjati sukhaṃ somanasaṃ ayaṃ rūpassa assādo.
yaṃ rūpam aniccaṃ dukkham vipariṇāmadhammaṃ ayaṃ rūpassa ādīnavo. yo rūpasmiṃ
chandarāgavinayo chandarāgapahānaṃ idaṃ rūpassa nissaraṇaṃ.
30
AN III 260,6-8: no ce taṃ bhikkhave lokamhā nissaraṇaṃ abhavissa na-y-idaṃ sattā loke
nissareyyuṃ, yasmā ca kho bhikkhave atthi loke nissaraṇaṃ tasmā sattā lokamhā nissaranti.

150
 Nissāraṇīya

The terms pañca niḥsaraṇīya dhātavaḥ and ṣaḍ niḥsaraṇīya dhātavaḥ appear
in both the Skt Saṅgīti-sūtra and Daśottara-sūtra.31 Their corresponding sūtras
in the Chinese Dīrgha-Āgama read 五出要界 (five factors leading to freedom
from bondage; T1[1] 51b27) and 六出要界 (six factors leading to freedom from
bondage; T1[1]52a9) respectively. The character 要 (yao) means bondage or
debarring, and 出要 (chu-yao; freedom from the bondage) is the rendering for
nissaraṇa.
In the Udumbarika-Sīhanāda-Suttanta (DN III 43,29; 46,28), we read a pair
of contrasting expression nissaraṇa-pañño (knowing the means of escaping)
and anissraṇa-pañño (not knowing the means of escaping). The latter occurs
in one of the contexts in which the Blessed One presents his insight into the
possible subsequential defilements (upakkilesa) resulting from ascetic praxis
(tapa). This is one of the subsequential defilements:
Moreover, Nigrodha, an ascetic who undertakes a course of
austerity makes distinctions about foods: “This pleases me; this
does not please me.32” Because he rejects with desire whatever
is not pleasing to him, and whatever pleases him, being bound to
it, infatuated, going too far, blind to the disadvantage (in doing
so), not knowing the means of escaping, he enjoys it… etc. This,
Nigrodha, also becomes a [kind of] subsequential defilement.33
In a soteriological context, “knowing the means of escaping” (nissaraṇa-
pañño) usually means to get rid of one’s desires, which keep one going round
the samsāra world. The way out of samsāra is doubtless to “escape” from those
desires. In the Saṅghīti-suttanta, one reads a passage on pañca nissāraṇīyā
dhātuyo. For example, the first nissaraṇa reads:
Herein, friends, when a monk is contemplating sensuous desires,
his heart does not spring forward to them, nor does he feel satisfied
with them, dwell on or become attached to them. However, when he
is contemplating renunciation of them, his heart springs forward to,

31
Karashima 2014, p. 208.
32
Sv III 837,8: khamatῑ ti ruccati. na khamatῑ ti na vuccati (misprint for ruccati).
33
DN III 43,25-31: puna ca paraṃ Nigrodha tapassῑ tapaṃ samādiyati, bhojanesu vodāsaṃ
āpajjati – “Idaṃ me khamati, idam me na-kkhamatīti.” so yaṃ hi kho ‘ssa na kkhamati taṃ
sāpekho pajahati, yaṃ pan’ assa khamati taṃ gathito mucchito ajjhāpanno anādīnavadassāvī
anissaraṇa-pañño paribhuñjati … pe … ayam pi kho Nigrodha upakkileso hoti.

151
 Nissāraṇīya

is satisfied with, dwells on and is inclined to renunciation. His heart

is blissful, well developed, well lifted up, well freed and detached
from sensuous desires. He is released from those intoxicants,
distress, and fever [of passion], which arise in consequence of
sensuous desires. He does not experience that kind of feeling. This
is called the escape from sensuous desires.34
Some variant readings are recorded and the spelling in both the Burmese
manuscript and printed edition read nissaraṇīyā and nissaraṇiyā, respectively.35
Later in the same Sutta, another set of cha nissāraṇīyā dhātuyo is given with
more variants recorded. The texts seem to be struggling between the two forms,
whereas the Burmese versions are more consistent in the form derived from
the normal stem.36 One reads within the paragraph of exposition a stereotyped
sentence with variation of key words. To give just one example: ‘Because, my
friend, it is the escape from malevolence that is [called] emancipation of heart
through benevolence’ (DN III 248,10-11: nissaraṇaṃ h’etaṃ āvuso vyāpādassa,
yadidaṃ mettā ceto-vimutti.)
Now I shall draw attention to the term nissāraṇīyā in the Saṅghīti-suttanta.
As nissaraṇa is consistently used in the context meaning “escaping”, it would
seem etymologically correct to emend pañca nissāraṇīyā dhātuyo (DN III
239,18) to pañca nissaraṇīyā dhātuyo. In fact, the PTS text has been changed
intentionally. In this very passage, the first edition (1911) of the DN reads
nissāraṇīyā, but the 1960 reprint reads nissaraṇīyā instead.37 It is discernable
that the word nissaraṇīyā in the reprint is “in a slightly different typeface from
the other words”, and hence this suggests a conscious decision to change the
text.38 It is interesting to note that in the Skt Saṅgīti-sūtra and Daśottara-

34
DN III 239,18-240,4: idh’ āvuso bhikkhuno kāme manasikaroto kāmesu cittam na pakkandati
nappasīdati na santiṭṭhati na vimuccati, nekkhammaṃ kho pan’ assa manasikaroto nekkhamme
cittaṃ pakkhandati pasādati santiṭṭhati vimuccati, tassa taṃ cittam sugataṃ subhāvitaṃ
suvuṭṭhitaṃ suvimuttaṃ visaṃyuttaṃ kāmehi, ye ca kāmapaccayā uppajjanti āsavā vighātā
pariḷāhā, mutto so tehi, na so taṃ vedanaṃ vedeti, idam akkhātaṃ kāmānaṃ nissaraṇaṃ.
35
DN III 239, fn. 8.
36
DN III 247, fn. 9.
37
DN III 239,18 and 247,21 (the editions of 1960 and 1992, perhaps since 1960 onwards) show
the same alteration, but elsewhere (p. 275,13) nissāraṇīyā remains the same and p. 278,21 reads
nissāraṇīya (misprint for nissāraṇīyā). Perhaps the person who was responsible for this change
was unaware of the other occurrences in these places and therefore did not repeat changing.
38
I am indebted to Professor K.R. Norman for pointing out, through the observation of the

152
 Nissāraṇīya

sūtra, one reads consistently pañca niḥsaranīyā dhātavaḥ and ṣaḍ niḥsaranīyā
dhātavaḥ, respectively.
In the Pāli Suttas, the usage of nissaraṇa/nissaraṇīya focuses specifically
on the issue of freedom or release from negative or undesired elements, which
is one of the factors leading to final liberation. In the Skt and Chinese Vinaya
texts, nissaraṇa is used in the same sense as that in the Pāli Suttas, although
articulating a more specific disciplinary concept of release from a monastic
offence. The following are some citations of the stock phrase from the Skt and
Chinese texts in contrast with the Pāli Cv:
Cv Vin II 15,12-13: sammā vattāmi, lomaṃ pātemi, netthāraṃ
vattāmi, pabbājanīyassa kammassa paṭippassaddhiṃ yācāmī ti.
‘I am comporting myself properly; I am subdued, and I am
proceeding towards release [from the offence]. So now I request a
revocation of the legal act of banishment.’
BhīVin (Mā-L) 164,3-4: sā vartaṃ vartayati, lomaṃ pātayati,
niḥsaraṇaṃ pravartayati.
‘She comports herself properly; she is subdued; she proceeds
towards release [from her offence].’
MSV (Pāṇḍ-v & #167; 1.12): utkacaprakacāḥ saṃghe roma
pātayanti niḥsaraṇaṃ pravartayanti sāmīcīm upadarśayanty
antaḥsīmāyāṃ sthitvā osāraṇāṃ yācante.
‘They are in full-blown awe, they are subdued towards the Saṅgha;
they proceed towards release [from the offence]; they pay homage;
staying within the bounded area, they request for reinstatement.’
BhīKavā (28b1): saṃghe roma pātayantan niḥsaraṇaṃ
pravartayantaṃ sāmīcīm upadarśayantaṃ antaḥsīmāyāṃ sthitam
osāraṇāṃ yācantam.

different typeface of words in the text, the deliberate changing of the text.

153
 Nissāraṇīya

They are subdued towards the Saṅgha; they proceed towards

release [from the offence]; they pay homage; staying within the
bounded area, they request for rehabilitation.’39
CMū-Kavā (T24[1453]487a12-14): 極現恭勤，於僧伽處不生輕
慢，希求拔濟，恆申敬禮，界內而住，請乞收攝法。
‘[He] displays extreme respect and sincerity; [he] yields no
irreverence towards the Saṅgha; [he] wishes for rescue; [he]
constantly pays homage; [he] stays within the bounded area and
requests for reinstatement.’
The Chinese text reads “wishing for rescue” (希求拔濟), identical in
meaning with niḥsaraṇaṃ pravartayati (proceeding towards release [from
the offence]). Netthāraṃ in the Cv in fact conveys the same meaning but the
wording is different. As can be seen, there is full agreement on the meaning
of niḥsaraṇa/nissaraṇa between the Suttas and Vinaya in different languages
except for the Pāli Vinaya.
The above discussion has shown that in the Pāli Suttas the application of
nissaraṇa prevails, with the non-person as the referent, e.g. kāmānaṃ nissaraṇaṃ
(escape from sensuous desires), nissaraṇaṃ ... vyāpādassa (escape from
malevolence), loke nissaraṇaṃ (escape from the world), lokamhā nissaraṇaṃ
([such thing as] escape from the world). In the case where the subject is the
person, the verb is derived from the normal stem and the sentence is construed
with the active voice: na-y-idaṃ sattā loke nissareyyuṃ (‘Beings in this world
could not escape’); atthi loke nissaraṇaṃ tasmā sattā lokamhā nissaranti
(‘There is in the world escaping, that is why beings do escape therefrom’).
In this specific context, no passive structure with causative derivations is
seen. Confusion between the normal/causative derivations occurs only in the
particular expression nissaraṇīya/nissāraṇīya. Such confusion is highly likely
to have resulted from the introduction of nissāraṇīya in the Vinaya into the
Sutta. (See discussion below in Section IV)

39
This text was first edited by C.M. Ridding and L. de la Vallée Poussin in “A Fragment
of the Sanskrit Vinaya: Bhikṣunikarmavacana”, who regarded it as a work of the Sarvāstivāda,
yet M. Schmidt has re-identified it as belonging to the Mūlasarvāstivāda (M. Schmidt, “Zur
Schulzugehörigkeit einer nepalesischen Handschrift von Bhikṣuṇī-Karmavācanā”, SWTF
Beiheft 5. This quotation is taken from GRETIL.

154
 Nissāraṇīya

When, in the Pāli tradition, a separate BhīPām was being compiled, the
redactor(s) had the concern with refining the text by coining new or special
terms, for instance, four technical designations are assigned to the offenders
against the four Defeats peculiar to nuns. They are “above the knee-caps”
(ubbhajānumaṇḍalikā), “a fault-concealer” (vajjapaṭicchādikā), “a follower of
the suspended one” (ukkhittānuvattikā), and “an offender by the eight conditions”
(aṭṭhavatthukā). Such technical designations are not found in any of the other
Vinaya tradition except for the Dha.
Special treatment has given to the first grave offence (i.e. Defeat) and would
it not be natural to try to make the second class (i.e. Saṅghadisesa) more
comprehensible? In view of the meaning and reference of nissaraṇa/nissaraṇīya
in the Suttas, the Saṅghadisesa chapter would seem to be just the right place
for this word and its variant nissaraṇīya. It is therefore plausible to assume
that the Vinaya redactor(s) may have borrowed from the Suttas nissaraṇīya to
supplement saṅghadisesa in its meaning and reference.

IV. The Confusion between Nissaraṇīya and Nissāraṇīya in the

Pāli Suttas
We assume that the Suttas were highly likely to be the source of nissaraṇīya,
yet the present BhīPām reads nissāraṇīya. This suggests a later change of the
wording to suit an intended new institution. In the Vinaya, there is no problem
at all in that the original form nissaraṇīya, once being changed, has ceased to
appear throughout the canonical Vinaya. Nevertheless, the variation between
either nissaraṇīya or nissāraṇīya occurs unexpectedly in the Suttas. Where
nissaraṇīya is read, nissāraṇīya is usually recorded as a variant, and vice versa.
It is not possible to date the confusion, nor are we able to tell how it actually
happened. There is, however, a clue found in the Atthavasa-vagga of the AN (I
99,13-14), where one finds the phrase osāraṇīyaṃ paññattaṃ (‘Enactment of
reinstatement’) paired with nissāraṇīyaṃ paññattaṃ (‘Enactment of expulsion’).
This is the only occurrence, and in the Chinese Ekottarika-Āgama one finds
no Sūtra parallel to the Pāli Suttanta where the above dyad appears. It would
seem that once nissaraṇīya had been borrowed by the Vinaya, and having been
changed, it had found its way back into the Suttas. This two-way borrowing
may have interfered with the consistency in wording and the transmitters’
perceptions of what the correct reading should be.

155
 Nissāraṇīya

Apart from the Pāli tradition, we have not yet found any case of nissāraṇīya
in any other texts, be them in the Sūtras or Vinaya. We therefore infer that
nissaraṇīya must be the original form in the Pāli Suttas, and later it became
confused with the causative form nissāraṇīya in the Vinaya. The discussion in
Section III has exemplified some cases of such confusion. This may have resulted
in a conscious change of the word according to what one deems to be correct
(see above pp. 152). These changes were then replicated in the commentarial
tradition.
When commenting on the expression pañca nissāraṇīyā dhātuyo, the
post-canonical commentary, the Sumaṅgala-vilāsinī, writes:
Sv III 1031,31: Nissāraṇīyā ti nissaṭā visaññuttā.
‘Nissāraṇīyā means gone out, detached from.’
Later in the same text we read another gloss on cha nissāraṇīyā
dhātuyo:
Sv III 1036,13: nissāraṇīyā dhātuyo ti nissaṭā dhātuyo va.
‘Elements of escape means just elements which have gone out.’
As the above quotations show, although the text adopts the reading of
nissāraṇīya, it explains in the sense of nissaraṇīyā. Nevertheless, the sub-
commentary has corrected nissāraṇīya back to nissaraṇīyā:
After shortening [ā], the exposition should read: Escape means
they go out. Because this word -aṇīya is used of the agent, just like
niyyāniyā (leading to salvation). Therefore “gone out” is said. But
from what have they gone out? From their respective opposites.40

40
Sv-ṭ III 324,21-24: nissarantī ti nissaraṇīyā ti vattabbe rasaṃ katvā niddeso. kattari h’ esa
anīya-saddo yathā niyyāniyā ti. ten’ āha nissaṭā ti. kuto pana nissaṭā ti? yathā sakaṃ paṭipakkhato.
(For rasaṃ, the text has dīghaṃ, read with the v.l.)

156
 Nissāraṇīya

Another passage, which comments on the expression kāmānaṃ nissaraṇaṃ

(escape of the sensuous desires), further explains:
Escape means they go out from there. What go out? Sensuous
desires. And having so construed it, the genitive case fits well in
the sense of agent: “of sensuous desires”.41
The explanation offered here suggests the change by the commentator is
based on the traditional sentence structure with non-person as the agent in the
active voice. It is only in the Vinaya that a causative derivative nissāraṇīya is
applied. Although the Sp also takes the offence as the agent but its object (the
guilty nun) is construed in the passive.
Semantic divergency is what distinguishes nissaraṇīya from nissāraṇīya.
They are not interchangeable. It may cause considerable loss to change the text
without a firm grasp of the divergent denotations between these two forms. Now
we have a case of this. As has been pointed out in Section II (above, p. 147),
the Sp states that “[I]n the Vibhaṅga it is said that [she] is made to leave the
Saṅgha, but in the Pātimokkha ten have come down to us with the wording:
nissaraṇīyaṃ saṅghādisesaṃ (mātikāyaṃ pana nissaraṇīyaṃ saṅghādisesan
ti evam āgatāni dasa). Nevertheless, the Chaṭṭhasaṅgāyana electronic
edition (https://tipitaka.sutta.org) has changed nissaraṇīyaṃ to nissāraṇīyaṃ
(mātikāyaṃ pana “nissāraṇīyaṃ saṅghādisesa”nti evaṃ āgatā dasa). It is
thanks to the PTS edition, which has preserved the text as it was, this valuable
evidence of nissaraṇīyaṃ saṅghādisesaṃ would otherwise never see the light
of day!

41
Sv-ṭ III 325,16-18: nissaranti tato ti nissaraṇaṃ. ke nissaranti? kāmā. evañ ca katvā kāmānan
ti kattari sāmivaccanaṃ suṭṭhu yujjati.

157
 Nissāraṇīya

Concluding Remarks
The customary mending procedures for an offence of saṅghādisesa are explained
in detail in the Vinaya literature, which involve no expulsion of the culprit out of
the monastery. It is therefore bewildering when a new term nissāraṇīya appears
in the saṅghādisesa rules for nuns, possibly denoting expulsion, but without
explanation except for a gloss terse enough for differing interpretations.
The commentarial literature has unambiguously confirmed the nuns’
expulsion rather than their release from the offence. However, a hint at an
existing controversy over such polemic views can be discerned in the Sp. This
hint proves to be true. A comparative examination of the related passages
in the other Vinayas demonstrates that the other traditions either contain no
additional expression to the offence name saṅghādisesa, or that the additional
expression was a supplement. Thus it is evident that the Pāli Vinaya is unique
in applying the causative form of nissāraṇīya and asserting its denotation of
the nun’s expulsion.
The four Vinayas (Pāli, Dha, Mī, and Sa) containing an extra expression
attached to the term saṅghādisesa belong to the schools affiliated with the
Sthavira. This suggests that such an addition may have taken place early in the
sectarian period, originating in the Sthavira tradition.
The new term added was originally nissaraṇīya, a term which makes sense
in the Suttas within a specific soteriological context, and which fits the nature
of the saṅghādisesa rules. It seems plausible to assume that the Vinaya may
have borrowed this term to make explicit this class of offence as remediable.
The evidence from the Parivāra and the Sp’s commentary proves that in the
Pātimokkha the wording was originally nissaraṇīyaṃ saṅghādisesaṃ “an
offence entailing legal acts of the Saṅgha leading to the removal of the offence”.
This use of nissaraṇīya is not isolated. The relevant passages in the Sūtras
and Vinayas of the other schools available to us read niḥsaraṇīya consistently;
only the Pāli Vinaya reads nissāraṇīya. What is interesting is that in the Pāli
tradition nissāraṇīya also appears in the Suttas. The discussion in Section IV
has shown that the confusion between the two forms interfered not only with
the consistency in wording but also the readers’ judgement of what is correct,
based on which changes (either by the ancient transmitters or the PTS editors,
see above pp. 150, 152, fn. 38, 155-157) had in fact been made. In some cases
changes may have recovered the correct form, but in other cases it may have
caused the loss of valuable textual evidence.

158
 Nissāraṇīya

As to whether there exists an extra punishment for nuns offending against a

saṅghādisesa rule, Part I of this study has demonstrated that the key lies in the
difference between nissaraṇīya and nissāraṇīya, and that the former was later
on replaced by the latter. The most curious is why and how such replacement
may have taken place. This is the issue to be addressed in Part II.

Abbreviations and References

All Pāli texts refer to the editions of the Pali Text Society.

Abbreviations
AN Aṅguttara-Nikāya
AW Analysis of Words (= Vinaya padabhājana)
BD The Book of the Discipline (Horner, 1938-1966)
BhīKavā Bhikṣuṇī-Karmavācanā
BhīPr Bruchstücke des Bhikṣuṇī-Prātimokṣa der Sarvāstivādins.
(Waldschmidt, 1979)
BhīPām Bhikkhunī Pātimokkha
BhīVibh Bhikkhunī Vibhaṅga/Bhikṣuṇī Vibhaṅga
BhīVin Bhikkhunī Vinaya/Bhikṣuṇī Vinaya
BhīVin(Mā-L) Mahāsāṃghika-Lokottaravādin Bhikṣuṇī-Vinaya (Roth, 1970).
BhuPām Bhikkhu Pātimokkha
CMū-Kavā The One Hundred and One Karmavācanā of the
Mūlasarvāstivādin Vinaya (根本說一切有部百一羯磨), T24,
No. 1453.
DN Dīgha-Nikāya
Dha Dharmaguptaka-Vinaya, T22, No. 1428.
Dutt Gilgit Manuscripts, Vol. III (1943).
Kkh Kaṅkhāvitaraṇī
Mā Mahāsāṅghika-Vinaya, T22, No. 1425.
Mā-L Mahāsāṅghika-Lokottaravādin-Vinaya
Mi Mahīśāsaka-Vinaya, T22, No. 1421.
Mp Monoratha-pūraṇī

159
 Nissāraṇīya

MSV Mūlasarvāstivādin-Vinayavastu
Mū Mūlasarvāstivādin-Vinaya, T23-24, Nos. 1442-1451.
Mv Mahāvagga (Vinaya-Piṭaka)
Pāṇḍ-v Pāṇḍulohitakavastu
PrMoSū (Sa) Le Prātimokṣasūtra des Sarvāstivādins (Finot and Huber)
Sa Sarvāstivāda-Vinaya, T23, No. 1435.
Saṅgh Saṅghādisesa
Saṅgh (N) Saṅghādisesa rules for nuns.
Skt Sanskrit
Sp Samantapāsādikā
Sv Sumaṅgalavilāsinī
Sv-ṭ Dīghanikāya-aṭṭhakathā-ṭīkā
SVTT I-III Édith Nolot (1996)
SVTT IV-X Édith Nolot (1999)
T Taishō Shinshū Daizōkyō 大正新修大藏經
Vin Vinaya-Piṭaka
Vmv-ṭ Vimativinodanī-ṭīkā

References
Dutt, N. (ed.). 1943. Gilgit Manuscripts, Vol. III: Mūla-sarvāstivādavinayavastu.
Srinagar.
Finot, Louis and Huber, É. 1913. “Le Prātimokṣasūtra des Sarvāstivādins,” texte
Sanskrit par M. Louis Finot, avec la version chinoise de Kumārajīva traduit en
Français par M. Édouard Huber”. Journal Asiatique, Novembre-Décember.
Hardy, E (ed.). 1885- 1900. Aṅguttara-Nikāya, 5 Vols. London: Pali Text Society.
Heirmann, Ann. 2002. Rules for Nuns according to the Dharmaguptakavinaya.
Delhi: Motilal Banarsidass.
———. 2003. “A Lexicographical Research: Technical Terms of Vinaya Texts” (
漢語律本名相詞彙研究). Universal Gate Buddhist Journal, Issue 18, Offprint
Edition.
Horner, I. B. 1938-1966. The Book of the Discipline (Vinaya-piṭaka), 6 Vols. Sacred
Books of the Buddhists 10, 11, 13, 14, 20, 25. London.

160
 Nissāraṇīya

Karashima, Seishi. 2014. “The Sarvāstivādins’ “encroachment” into the Chinese

Translation of the Daśottara-sūtra in the Dīrgha-āgama of the Dharmaguptakas”.
Research on the Dīrgha-āgama, ed. Dhammadinnā. Taipei: Dharma Drum
Publishing Corporation, pp. 197-235.
Kassapa, Coḷiya. 1960. Vimativinodanī-ṭīkā, 2 Vols. Rankun (ChS).
Maskell, Dorothy (ed.). 1956. Kaṅkhāvitaraṇī nāma mātikaṭṭhakathā. London: Pali
Text Society.
Nolot, Édith. 1996. “Studies in Vinaya Technical Terms I-III”. Journal of the Pāli
Text Society, Vol. XXII, pp. 73-150.
———. “Studies in Vinaya Technical Terms IV-X”. Journal of the Pāli Text Society,
Vol. XXV, 1999, pp. 1-111.
Oldenberg, Hermann (ed.). 1879-1883. Vinaya-Piṭaka, 5 Vols. London : Pali
Text Society.
Rhys Davids, T. W., Carpenter, J. E., Stede, W. (eds.). 1886-1932.
Sumaṅgalavilāsinī, Dīghanikāya-aṭṭhakathā, 3 Vols. London: Pali Text
Society.
Rhys Davids, T. W. and Carpenter, J. E. (eds.). 1890-1911. Dīgha-Nikāya, 3 Vols.
London: Pali Text Society.
Roth, Gustav (ed.). 1970. Bhikṣuṇī-Vinaya, including Bhikṣuṇī-Prakīrṇaka and a
summary of the Bhikṣu-Prakīrṇaka of the Ārya-Mahāsāṃghika-Lokottaravādin.
Tibetan Sanskrit Work Series, 12. Patna.
Schmidt, Michael. 1993. “Bhikṣuṇī Karmavācanā: Die Handschrift Sansk. c. 25 (R)
der Bodleian Library Oxford”. In Studien zur Indologie und Buddhismuskunde.
Festgabe des Seminars für Indologie und Buddhismuskunde für Professor Dr.
Heinz Bechert, zum 60. Geburtstag am 26. Juni 1992. Herausgegeben von
Reinhold Grünendahl, Jens-Uwe Hartmann und Petra Kieffer-Pülz. Bonn, pp.
239-288.
Shih, Juo-hsüeh. 2000. Controversies over Buddhist Nuns. Oxford: The Pali Text
Society.
———. 2003. “What should be Relinquished: Problems in the Saṅghādisesa Chapter
of the Pāli Bhikkhunī Vinaya”. Journal of Humanities and Social Sciences of
Tzu Chi University, No. 2, pp. 199-236.
de Silva, Lily (ed.). 1970. Dīghanikāya-aṭṭhakathā-ṭīkā, Līnatthavaṇṇanā, 3 Vols.
London: Pali Text Society.
Takakusu, J., Nagai, M. (eds). 1924-47. Samantapāsādikā, Vinaya-aṭṭhakathā, 7
Vols. London: Pali Text Society.

161
 Nissāraṇīya

Takakusu, J., Watanabe, K. (eds.) 1924-1935. Taishō Shinshū Daizōkyō 大正新修

大藏經. Tokyō.
Waldschmidt, Ernst. 1979. Bruchstücke des Bhikṣuṇī-Prātimokṣa der Sarvāstivādins.
Kleinere Sanskrit-Texte, 3. Wiesbaden: Franz Steiner. First Published 1926, Leipzig.
Yamagiwa, Nobuyuki. 2001. Das Pāṇḍulohitakavastu, Über die verschiedenen
Verfahrensweisen der Bestrafung in der buddhistischen Gemeinde, Neuausgabe
der Sanskrit-Handschrift aus Gilgit, tibetischer Text und deutsche Übersetzung.
Indica et Tibetica 41, Marburg.

162
 Book reviews

Wisdom as a Way of Life. Theravāda Buddhism Reimagined.

By Steven Collins. Edited by Justin McDaniel; Preface by Dan
Arnold; Afterword by Charles Hallisey. 2020. Columbia University
Press. ISBN: 9780231197205 (cloth), 9780231197212 (paperback),
9780231552042 (ebook)

Reviewed by Alexander Wynne

Steven Collins’ final book, published posthumously, consists of two parts by

the author and three sections supplied by his former friends and colleagues
(Dan Arnold, Justin McDaniel and Charles Hallisey). McDaniel saw the book
through to publication, after receiving a draft before Collins’ death in February
2018. To make the book workable, McDaniel tells us (p.xxiii) he had to cut three
sections from the original manuscript: a preface, an introduction and a chapter,
although lengthy sections from these are cited in McDaniel’s own introduction.
As the title suggests, the book reimagines the meaning of wisdom in
Theravāda Buddhism. Part One, ‘Wisdom’, focuses on the Jātakas, which
Collins opposes to the ‘texts of systematic thought (“doctrine”)’, and claims are
‘the heart and humanity of the Pali tradition’ (p.2). He argues that the Jātakas
exemplify the importance of living correctly and well, providing a person with
the ‘capacity of judging rightly in matters relating to life and conduct’ (p.9). Part
Two, ‘Practices of Self’, focuses on spiritual practice and although conceptually
simpler, is more difficult to describe. According to McDaniel, Collins
… firmly believed that Theravāda Buddhists had something
important to offer intellectually to the project that modern
historians and philosophers like Pierre Hadot, Michel Foucault,
and Derek Parfit spent their lives investigating. He saw these
Western philosophers as attempting to articulate what Theravāda
Buddhists had been arguing for over two millennia – that the study

163
 Book reviews

of philosophy and ethics is largely “practices of the self,” and

therefore has to involve both textual study and an ascetic lifestyle.
(pp.xxx-xxxi)
Part Two of the book is also an ‘attempt [at comparison]’ (p.xxxi) between
Theravāda practice and Pierre Hadot’s writings on ‘spiritual exercises’ or
‘philosophy as a way of life’, and Michel Foucault’s ‘practices/technologies
of self’ (p.87). According to Collins (p.xxxiii), ‘practices of self’ include the
‘entire process of acculturation’, which ‘in all societies, all civilizations, all
cultures, is the cultivation of a certain kind of self, a certain kind of subject of
experience’:
This learning of specific forms of physical and mental self-control,
this askēsis, from childhood on, and the introjection of culturally
specific ideals, is part of what constitutes sanity in any given
social context . . . the introjection and performance of certain basic
components of human sociality (so-called Morality) can be seen
as a kind of wisdom, promulgated at length in Buddhist texts.
(p.xxxiv)
The argument seems to be that Theravāda Buddhism is not really or mostly
about the meditative quest for Nirvana. It is rather a civilisational project, in
which Jātaka type narratives are a guide to practical wisdom in everyday life,
and in which ‘practices of self’ are part of a process of acculturation, a sort of
wisdom leading to the creation of certain types of human subject. This would all
amount to a bold reimagination of the Theravāda tradition, one which McDaniel
would be correct to call ‘revolutionary’ (p.xxiv). Unfortunately, however,
Wisdom as a Way of Life does not fulfil Collins’ lofty ambitions.

1. In Part Two, ‘Practices of Self’, Collins struggles to say what he is trying to

achieve. He begins by mentioning two intentions (p.85), and yet the pages that
follow do not state what the first intention actually is. Collins does, however,
state that practices of self are
demographically tiny but civilizationally of great importance …
the texts in which they are described and prescribed remain of great
importance to the intellectual history of the Pali tradition. (p.86)

164
 Book reviews

This suggests that the first aim of Part Two is to show that practices of self form
part of elite Buddhist practice, or are derived from it, and so are a minority concern
in any Theravāda society. The same is suggested when Collins wonders whether
practices of self constitute ‘an elite regimen of truth’ (p.154). And yet a development
from the ‘worldly wisdom’ focus of Part One, to an ‘elite regimen of truth’ in Part
Two, is never made clear. The second aim of Part Two is more clearly stated:
The second reason for writing this chapter, indeed for writing the
entire book, is to provide some comparative material to the work of
Pierre Hadot on – to use the standard slogans – “spiritual exercises”
and “philosophy as a way of life,” and to that of Michel Foucault
on “practices/technologies of self” and “subjectivity of truth”. (p.87)
This is different from the professed ‘attempt [at comparison]’ with Hadot/
Foucault stated in the introduction (p.xxxi). It would seem that Collins never
finally conceptualised what the purpose of Part Two should be. Indeed, although
he goes on state that he wishes to correct Hadot’s/Foucault’s ‘lack of attention
to the social and institutional contexts of the ideas they were writing about’
(p.87), he admits that he has ‘provided no serious empirical study’ (p.87). There
is no serious institutional study either. Collins describes his personal experience
of Pali chanting at Wat Suthat, Bangkok (section 2.4.1), but this is neither
an empirical study nor an analysis of institutions. It is purely descriptive and
unremarkable; chanting occurs in Buddhist temples, as everybody knows. The
point that devotional ritual requires some degree of ‘training’ and ‘concentration’
(p.110-11) is a simple observation, and not part of any apparent argument. What
are the implications of this practice for acculturation, and in what ways do social
and institutional contexts affect it? Collins does not try to explain.
The same is true of other aspects of Theravāda practice covered in Part Two.
These sections read as a descriptive overview of spiritual practice rather than a
sociological or civilisational analysis. Collins neither shows how the study of
philosophy and ethics is part of ‘practices of the self’ (pp.xxx-xxxi), nor explains
how certain kinds of ‘selves’ are cultivated in the Theravāda context, ‘from
childhood on’ (p.xxxiv). There is no meaningful comparison with the ‘Spiritual
Exercises’ of Hadot or ‘spirituality’ of Foucault, and no exploration of how
sociological or institutional contexts aid acculturation in Theravāda societies.
Rather strangely, Collins does not explain why the expression ‘practices of self’
is any better than ‘Buddhist meditation’ or ‘Buddhist spiritual practices’.

165
 Book reviews

2. Just as Part Two fails to explain how Theravāda acculturation is a form

of wisdom, informed by elite spiritual ideals, so too Part One fails to show
how Buddhist narratives are a source of worldly wisdom. The discussion
here is constructed around a simple dichotomy between the ‘quotidian’ and
‘supererogatory’ (p.2). Whereas the former consists of ‘forms of wisdom and
wisdom seeking that are matters of practice’ (p.2), the latter are expressions of
absolute Buddhist values which need not be practised by everyone: celibacy,
asceticism, meditation and so on.
The distinction between the ‘quotidian’ and ‘supererogatory’ corresponds
to the categories ‘dhamma 1’ and ‘dhamma 2’ (p.7), first made by Collins in
Nirvana and Other Buddhist Felicities (1998). Dhamma 1 is defined as ‘an ethics
of reciprocity in which the assessment of violence is context dependent and
negotiable’ (p.7); Collins also states that it consists of ‘the kinds of good moral
character other than a basic civilizational necessity’ (p.10), which differ from
the standard lists of Buddhist virtue (i.e. sīla). In this sense dhamma 1 of the
Jātakas is similar to Aśoka’s Dhamma, which Collins claims is not ‘specifically
Buddhist’ (p.11).
Collins also mentions that the distinction between the quotidian and
supererogatory is ‘implicitly between forms of wisdom and wisdom seeking
that are matters of practice, or might be, or are textual tropes’ (p.2), the latter
being ‘ideals and aspirations that will be matters of practice in actual life
only for a minuscule proportion of any population in Theravāda civilization’
(p.2). Since on the same page he contrasts the Jātakas, ‘stories about wisdom’,
with ‘the ‘simple four truths, five aggregates and the rest’ (p.2), it seems that
supererogatory wisdom consists of orthodox Buddhist doctrine, and the spiritual
practices which lead to it. Of course dhamma 2, ‘exemplified and promoted in the
Discourses (Suttas) and Monastic Rules (Vinaya) texts’ (p.8), and which mostly
consists in ‘simply the living of a celibate monastic life, itself supererogatory’
(p.10), cannot be mentioned in the Jātakas, which are tales about a mythic time
before the Buddha. Nevertheless, Collins points out that asceticism is mentioned
in relation to Paccekabuddhas (pp.17-18, 79-80).
The analytical structure of Part One is therefore rather rich and dense. The
overall purpose is to portray the Jātakas as a sort of ‘wisdom literature’ (p.12),
and an important part of court culture in South Asia and Theravāda kingdoms
(pp.19-20). However, the various arguments advanced in support of these
claims are not convincing. The Jātakas do not distinguish quotidian wisdom/

166
 Book reviews

dhamma 1 from supererogatory wisdom/dhamma 2, and then advise the former.

Indeed, the dichotomy between dhammas 1 and 2 is misleading; it overlooks the
universalist agenda of the Jātakas, and so fails to notice what was the original
Buddhist project in civilisation.

2a. Value Conflicts

In support of the idea that the Jātakas teach a quotidian dhamma 1, Collins
claims that one of their most important themes is value conflicts:
Human life, apart from systems of specialist askēsis, contains
irresolvable value conflicts. The Birth Stories can be enjoyed and
admired by everyone, for many different reasons, without being
subjected to the classificatory categories of, for example, “the
eightfold path,” “conditioned co-origination,” and still less – since
in The Birth Stories it does not occur – nibbāna (nirvana). They
express many of the aspirations of Theravāda civilization, and
thence of its intellectual history. (p.3)
So there are stories that offer examples of quotidian wisdom, some
of which confound and challenge the reader/audience by offering
perhaps irresolvable ethical and value dilemmas. (p.35)
[S]ome of the stories challenge the most revered of Buddhist
values, just as does the great Vessantara, which has even greater in-
your-faceness, if I may put it that way, with no convincing solution,
or indeed no solution at all, to its central moral dilemma. (p.75).
The conflict between kingship and renunciation is of course explored with
great skill in the Vessantara Jātaka (Ja 547). The same is true of the Temiya
Jātaka (Ja 538), in which the Bodhisatta feigns being deaf, mute and crippled to
avoid accruing the bad karma of kingship (p.37ff). And yet it is important to note
that other Jātakas resolve the conflict between kingship and renunciation. Both
the Mahā-Janaka and Nimi Jātakas (Ja 539, 541), for example, navigate this
conflict through a temporal sequence: the king rules justly first and renounces
second, ‘after fulfilling his duties as ruler’ (Appleton and Shaw, 2015: 54). This
ideological solution proved useful later on in Indian civilisation, when it was
utilised within Brahmanism in the sequential form of the four āśramas (Olivelle,
1993: 117ff).

167
 Book reviews

Even if the dilemma between kingship and renunciation is an important

theme of the Jātakas, the other stories Collins cites in this context have nothing
to do with values: when the monkey deceives and so evades the crocodile (Ja
208, pp.28-29), when the mouse kills the jackal (Ja 128, p.34), or when the crane
eats some fish, but is then killed by an avenging crab (Ja 38, p.35), the Jātakas
deal with conflicts of interest rather than conflicts of value. Their general point
is that individual conflict is endemic in human life, a fairly obvious Buddhist
extension of the principle of dukkha that is not accompanied by words of advice.
Collins further claims (p.35) that value conflicts emerge from the redaction of
Jātakas in collections:
I want to stress that [The Birth Stories], like proverbs, fables,
and other genres, what I will call in 1.4 “wisdom literature” as
a cross-civilizational phenomenon, almost always were and are
redacted in collections. This means that as well as their internal
nature, which may and often does itself contain problematizations
and conflicts of values, a collection as a whole clearly does this.
Yes, perhaps individual stories, especially the long ones, were
read or heard separately, but they would necessarily have been
read or heard as coming from a collection, many or most of
which the readership/audience would have been familiar with and
remembered.’ (pp.12-13)
Collins does not elaborate on this claim, despite saying (p.21), with
reference to collections, that he will ‘return to the issue of diversity and unity
in the conclusion (1.4.1).’ Elsewhere, he puts the task off into the future: ‘The
kind of comparative project I am envisaging would compare and contrast the
kinds of collection and context in which such wisdom literature circulated –
both literary form and narrative content’ (p.56). Perhaps a redactional analysis
of the Jātakas, supplemented by anthropological and historical work, will
one day show that Jātakas were transmitted, recited or performed in groups to
highlight value conflicts. But Collins does not provide any reason to believe
that this might be the case. And since he provides no evidence for value
conflicts, apart from that between kingship and renunciation, his comparative
project would appear dubious.

168
 Book reviews

2b. Kingship
With regard to kingship, Collins claims that the Jātakas sometimes adopt a more
pragmatic approach which exemplifies dhamma 1:
Buddhist advice to kings in dhamma 1 tells them to not to pass
judgment in haste or anger, but appropriately, such that the
punishment fits the crime. (p.7)
In connection with this Collins notes that in ‘a number of places in Pali an
executioner’s block, gaṇḍikā, is, astonishingly, prefixed with dhamma-, so that the
compound is perhaps best translated here as “block of justice”.’ (p.8). Although
Collins does not return to this topic – despite claiming ‘I shall tell one of these
stories below’ (p.8) – his previous book, Nirvana and other Buddhist Felicities
(1998: 459), refers to the use of the dhamma-gaṇḍikā in the Janasandha Jātaka
(Ja 468). Collins there claims that this story ‘is a striking example of how
different are the meanings of the word Dhamma in Mode 1 and in Mode 2.’
This is not an accurate interpretation of the Janasandha Jātaka, however, in
which Prince Janasandha destroys the executioner’s block (Ja IV.176: dhamma-
gaṇḍikaṃ bhedāpetvā). Since the Bodhisatta is here an agent of non-violence
opposed to capital punishment, it would seem that this Jātaka promotes the
triumph of absolute Buddhist values (dhamma 2) over the norms of Indian
kingship (dhamma 1). The same is true of the Maṇicora Jātaka (Ja 194), in
which a wicked king wishes to behead the Bodhisatta, but is himself beheaded
through the intervention of the god Sakka. The story does not advise the use
of the dhamma-gaṇḍikā, but shows that it is used against the evil-doer; the
principle of direct karmic retribution is implied.
These two stories undermine Collins’ claims about dhamma 1. But Collins
also points out that the principle of reciprocity, a subtler form of dhamma 1 not
confined to kingship, is also mentioned in the Jātakas:
Justified violence is, of course, explicit in all the stories where a king
hands out justice. The ethical and narrative principle of reciprocity,
central to dhamma 1, requires it, since crime is inevitable in the
quotidian world. (p.34)

169
 Book reviews

Although Wisdom as a Way of Life does not elaborate the principle of

reciprocity, the same argument can be found in Nirvana and other Buddhist
Felicities. Once again, however, the Jātakas which Collins believes advise
the principle of reciprocity do nothing of the sort. Collins claims (1998: 451)
that in the Puṭabhatta and Godha Jātakas (Ja 223, 333) the Bodhisatta teaches
reciprocity: to ‘respect someone who respects (you), share with someone who
shares with you; do a favour for someone who returns it’ etc. These stanzas
on reciprocity certainly exist in Ja 223/333 (Ja II.205, III.108) but they are
not given as advice. They are instead words of warning, which the Bodhisatta
uses to shock a selfish king into seeing the error of his ways. The principle of
reciprocity is merely a stratagem, a sort of ‘skill in means’, in other words, used
to elicit core Buddhist values of pity and compassion, as befits the occasion.

2c. Aśokan Dhamma

The principle of reciprocity is also mentioned in Collins’ discussion of the Sāma
Jātaka (Ja 540), ‘a fairy story that expresses and celebrates an important virtue
of dhamma 1, caring for parents’ (p.34). Parental care is frequently mentioned in
Jātakas which urge individuals (often kings) to ‘practise Dhamma’ (dhammaṃ
cara) in the sense of acting righteously towards such groups as one’s parents,
children and wives, ascetics and Brahmins, towns and countries, friends and
associates, elephants and army, villages and towns, kingdoms and countries,
birds and beasts and so on.1 These teachings obviously resemble Aśoka’s
Dhamma, but this does not mean that they are not ‘specifically Buddhist’.
The injunctions to ‘practise Dhamma’ are nothing more than elaborations
of the Sigālovāda Sutta (DN 31). This foundational ‘skill in means’ discourse
generalises Buddhist values into a non-denominational form, using a set of
categories similar to those used in the Jātakas (parents, children, wives etc.).
Although the layman Sigāla continues to carry out his ritual acts as before, by
following the Buddha’s teachings on how to do it properly he acts in accord with
Buddhist values.2 The same is true of the Jātaka advice to ‘practise dhamma’: it
is a way of behaving like a Buddhist without knowing it; when applied to rulers
it becomes almost a charter for Buddhist kingship.

1
E.g. Ja I.152, Ja IV.421.
2
Gombrich (2006: 81): ‘the Buddha constantly slips new ethical wine into the old brahminical
bottles: pretending to interpret traditional ritual, he in fact abolishes it.’

170
 Book reviews

The tendency to read Aśoka’s Dhamma as non-Buddhist just because it does

not mention Buddhist doctrine (the Four Truths and so on) is fundamentally
mistaken. Aśoka’s edicts are saturated with Buddhist references,3 which is
hardly surprising given his claim to have received instruction from the Buddhist
Sangha, and even his listing of a few favourite texts.4 When Aśoka advises his
officials to ‘practise the middle’ (majhaṃ paṭipādayema),5 what else could it be
but a political extension of the Middle Way?
It is unrealistic to suppose that Aśoka formulated ‘his’ Dhamma all by himself.
A more plausible interpretation is that Aśoka’s Dhamma was an elaboration of
the approach to Buddhist values found in the Jātakas, itself a reworking of the
Sigālovāda Sutta.6 Thus the Jātaka advice to ‘practise dhamma’ with regard to
parents etc. is not an example of dhamma 1, but was an extension of Buddhist
‘skill in means’ which came to be applied by Aśoka across India. Both the
Dhamma of the Jātakas and Aśokan Dhamma are quintessentially Buddhist, the
former providing an ideology for kingship realised by the latter.

2d. Paccekabuddhas
Collins claims that Paccekabuddhas teach ‘what is right in everyday human
life … In The Birth Stories pacceka buddhas do teach dhamma. However,
this is dhamma 1’ (p.17). While it is true that Paccekabuddhas do not teach
the Eightfold Path or the Four Truths – a concern of Buddhas alone – in all
other respects Paccekabuddhas are connected to ‘supererogatory’ rather than
‘quotidian’ values. As Appleton (2018: 4-5) has pointed out, these include the
benefits of renunciation, the importance of dispassion, the necessity of controlling
the sense faculties, the avoidance of attachment to sensual pleasure and so on.
Collins provides no evidence to support his claim that Paccekabuddhas were a
means of introducing non-Buddhist values into the Jātakas.

3
See Sujato & Brahmali (2015: 103ff) and Wynne (2015: 103-04) on the psychological aspect
of Aśoka’s edicts.
4
See MRE 3 (Bairāṭ); Sujato & Brahmali (2015: 105).
5
RE 13, Kalinga; Sujato and Brahmali (2015: 104).
6
See Gombrich (2006: 131): ‘Some scholars have questioned Asoka’s Buddhism on the
grounds that he never mentions nibbāna or other key concepts of Buddhist soteriology. Our
description of Buddhist lay religiosity, both in the Canon and after, proves that this objection
is foolish.’

171
 Book reviews

2e. A False Dichotomy and Buddhist Universalism

In his preface, Dan Arnold notes that ‘in Wisdom as a Way of Life, the basic
distinction between systematic and narrative thought remains central’ (p.xiii).
This distinction is indeed a major underlying feature of Part One. It forms the
basis of the binary opposition between the Suttas and Vinaya, on the one hand,
and the Jātakas on the other; whereas the former are the source of transcendent
aspirations expressed in systematic form (dhamma 2), the latter are narratives
without a Buddha, and exemplify a different set of values (dhamma 1):
Paññā in quotidian dhamma 1 is skill is some particular domain. In
The Birth Stories there is no Buddha, so naturally no dhamma 2 of
the kind exemplified and promoted in the Discourses (Suttas) and
Monastic Rules (Vinaya) texts, although the motif of renunciatory
askēsis certainly is. (p.8)
Collins draws a sharp distinction between the Jātakas and the Buddha/
canonical teaching. But this is a false dichotomy. The Pali Suttas are concerned
with much more than transcendent or ascetic values, and are also mostly expressed
in a narrative form; they frequently narrate tales of piety, faith, devotion, stream-
entry and so on. The Jātakas extend this Suttanta style of spiritual teaching, and
are not typologically or didactically distinct from it.
A good example of this is the ‘ten dhammas of a king’ (dasa rājadhamma):
giving, virtue, liberality, honesty, gentleness, austerity, non-anger, non-violence,
forbearance and concord.7 Collins states (p.33) that this list is a ‘very common
motif throughout the Birth Stories’. But he fails to see that it is a broadly
ascending set of virtues which bridges lay and ascetic ideals, and so brings
absolute Buddhist values into the domain of kingship. It could perhaps be
regarded as an example of what Max Weber called ‘inner-worldly asceticism’,
and is typical of the universal nature of early Buddhist teaching.
The dichotomy between dhamma 1 and dhamma 2 is false, therefore,
and obscures the fact that the Jātakas merely extend the style and content of
Suttanta teaching. A good example of this is the Kukkura Jātaka (Ja 22), a tale
in which the Bodhisatta is reborn as leader of a pack of stray dogs. When the
stray dogs get set up by the palace dogs for a crime they did not commit, the
king issues an order to have them killed. The Bodhisatta-dog then sneaks into

7
E.g. Ja III.274, III.320, III.412, V.378 etc.

172
 Book reviews

the palace to fix the conflict of interest (not values), but initially hides under
the throne, just like a scared animal, before emerging to teach the king.
This Jātaka is a parable showing that conflicts of interest are inevitable, but
can be difficult to understand and so often result in poor judgments and bad policy.
But there is no worldly wisdom for aspiring rulers. Instead, the Bodhisatta-dog
advises the king to ‘practise dhamma’ with regard to one’s parents etc., before
establishing him in virtue (sīla), a ubiquitous feature of the Jātakas. The king
thereupon grants safety to all creatures (Ja I.178: sabbasattānaṃ abhayaṃ
datvā), makes merit for the rest of his life and on death ascends to heaven. This
Jātaka thus teaches a sort of moral spirituality that harmonises with Buddhist
cosmology; like the Jātakas in general, it is standard Buddhism in all but name.
The Kukkura Jātaka is a good guide to the nature of Dhamma in the Jātakas.
Collins’ claim (p.55) that these tales belong to ‘wisdom literature as a cross-
civilizational category’ is simply a mistake, and a very strange one at that.
Historical studies have shown that rulers used the Jātakas for ideological rather
than practical purposes. Thai monarchs, going back as far as King Lithai in
the Sukothai period (c.1361 AD), valued the Jātakas in so far as they allowed
kings to portray themselves in the image of the Bodhisatta, and so promote an
ideal of royal authority and charisma based the Bodhisatta’s accumulation of
‘spiritual perfections’ (pāramī).8 No doubt Buddhist monks provided some of
the statecraft and worldly wisdom which kings needed, in the forms of Nīti and
Dhamma-sattha texts. The Jātakas served a different end, one that was more
ideological and specifically Buddhist.

3. Collins’ claims about dhamma 1, which amount to a misunderstanding of the

Jātakas, are difficult to understand. But a couple of digressions in Wisdom as a
Way of Life suggest that the problem stems from a faulty text-critical method.

3a. The Middle Way

In section 2.4.2, Collins discusses the First Sermon with reference to an article
by Oliver Freiberger (2006). Freiberger argues that the ‘middle way’ of this
sermon really deals with two early Buddhist tendencies: extreme asceticism
versus monastic laxity (2006: 250-51). Collins uses this analysis to ‘set right
what is an extraordinary mistake made in so many secondary sources’ (p.115).

8
Jory 2016, particularly the section entitled ‘The Doctrine of Perfections (barami)’ in chapter 2.

173
 Book reviews

This mistake is apparently to understand the Middle Way as a path ‘between the
life of a householder, given over to sense pleasures, and that of extreme self-
mortificatory asceticism’ (pp.115-16). As Collins points out, the Middle Way
is advice for renouncers (pabbajita); the recipients of the teaching are ascetics.
Collins does not state which secondary sources have misunderstood this
rather obvious point. But it leads to a very strange mistake of his own. In the First
Sermon, the adjective gammo, ‘belonging to the village (life), common, vulgar’
(DOP s.v.), is used to describe ‘sensual indulgence’ (kāmasukhallikānuyogo).
The commentary then interprets gammo as gāmavāsīnaṃ santako (Spk III.297),
‘the property of village dwellers’. But Collins somehow believes that gammo
qualifies ‘renouncers’ rather than ‘sensual indulgence’: ‘Santako (“the property
of”) is satirical: these renouncers haven’t renounced, they are owned by the
villagers they depend on so closely for a living’ (p.117). This is a bizarre error,
which leads to the following claim:
As Freiberger suggests persuasively, given that this is something
specifically not to be followed by renouncers, it must refer to some
kind of asceticism that the Buddha is saying should be avoided.
Most likely this is a familiar South Asian stereotype: scruffy
layabouts who live close to villages for the sake of an easy life and
a free lunch… (p.117)
This is an unfortunate misreading of Freiberger’s argument, which
understands the First Sermon as a warning against non-institutional modes of
asceticism, and monasticism which strays too close to household comforts.
Collins’ interpretation of the First Sermon in terms of modern Indian layabouts
is a peculiar piece of Orientalism.

3b. The ‘Rhinoceros Horn’ Sutta

Another peculiarity occurs in Collins’ treatment of the Khagga-visāṇa Sutta (pp.123-
24). This verse text, third poem of the Suttanipāta, famously likens the wandering
bhikkhu to the Indian rhino, in the refrain ‘one should wander alone, like a horned
rhinoceros (or ‘like the horn of a rhinoceros’)’.9 Collins interprets as follows:
The “One Horn of the Rhinoceros” poem certainly seems to
recommend in many verses that “he [the monk] should live his life

9
Sn pp.6-12: eko care khagga-visāṇa-kappo.

174
 Book reviews

alone, like the one horn of a rhinoceros.” The verb is carati, which
almost all translators take, literally and naïvely, as “wander,” which
is only one of its meanings. I discussed this verb in the previous
section: cariyā is one’s way of being, one’s way of life. “Wander”
suggests that the idea is that the monk moves around, but in fact
it refers to a monk’s psychological way of life, his inner mode of
being, not his behavior in the outer world. (p.123)
Collins reads the verb carati according to the use of the noun cariyā in the
Visuddhimagga (Collins, p.121), where it means something like ‘mode of being’.
And yet the poem betrays no trace of settled monasticism, let alone an urban
context, and is not obviously addressed ‘to monks who live in busy, bustling city
monasteries’ (Collins, p.124). It instead offers quite literal injunctions to ‘resort
to remote lodgings, and live/wander alone like a rhinoceros horn’.10 What would
it mean to ‘live’ or ‘behave’ like a rhino in a monastery anyway? The verb carati
must have the sense of ‘wander’, the only thing about a rhino’s lifestyle that a
Buddhist bhikkhu could conceivably do.
Collins also overlooks ancient Buddhist interpretations of the poem. As Norman
has pointed out (1992: 144), the poem was a source of the oldest Pali commentary,
the Niddesa, which is so old that it is included in the Pali canon.11 This antiquity
is confirmed by the existence of another recension of the poem in the Mahāvastu
of the Mahāsāṃghika-Lokottaravādins.12 Both of these ancient interpretations
attribute the poem’s celebration of anti-monastic wandering to pre-Buddhist
Paccekabuddhas. Should we understand that both wings of the ancient Buddhist
tradition, Theravāda and Mahāsāṃghika alike, took the verb carati ‘literally and
naïvely’? Or is it more likely that both inherited a way of interpreting an awkward
text from pre-monastic times? No doubt Theravāda monks in bustling monastic
centres have long drawn inspiration from the poem. But this has nothing to do with
its original meaning, which Collins was unable or unwilling to see.

4. Synchronism and the ‘Pali Imaginaire’

In his consideration of the Khagga-visāṇa Sutta, Collins ignores facts about the
text’s antiquity and ancient interpretation, choosing instead to synthesise its use

10
Sn v.72 (p.12): sevetha pantāni senāsanāni, eko care khaggavisāṇakappo.
11
For the interpretation of the Niddesa, see Bodhi (2017: 420ff).
12
For the interpretation of the Mahāvastu, see Senart (1882: 359) and Jones (1949: 305).

175
 Book reviews

of the verb carati with the noun cariyā from the Visuddhimagga, a text nearly
1000 years younger. The synchronic approach to Pali texts has its uses. In Selfless
Persons (1983) it resulted in a more sophisticated understanding of Buddhist
doctrine; in Nirvana and other Buddhist Felicities, Collins conceptualised
textual synchronism in terms of the ‘Pali imaginaire’, which resulted in original
and useful ways of considering Buddhist values. But this method is not always
appropriate; sometimes it is unhelpful and misleading to think of Buddhist texts
in terms of the Pali imaginaire, which consists of
… any and every text written (or translated into) Pali. I think it is
a matter of empirical fact that, as far as the grand issues of life,
death, suffering, and nirvana are concerned, all texts in Pali show
a remarkable consistency, and can be treated as a single whole.
(2010: 4–5)
We have seen that a synchronic approach fails when applied to canonical
texts such as the Khagga-visāṇa Sutta. The same is true of the Jātakas. Instead
of regarding this collection as a northern Indian composition stemming from
pre-Aśokan times, Collins follows the interpretation found in Nirvana and other
Buddhist Felicities, which discusses the Jātakas alongside medieval Nīti texts
(manuals for Buddhist kingship) to form an overall impression of Theravāda
advice to kings. In Nirvana and other Buddhist Felicities this makes some sort
of sense, since Buddhists have used the Pali canon in all sorts of ways, including
political instruction, and Nīti texts draw upon the Jātakas.13 But in Wisdom as
a Way of Life, when the Jātakas are the focus and Nīti texts have faded away
into the background, the use of the Pali imaginaire involves abstracting the
Jātakas from their historical context and understanding them almost as medieval
manuals for kingship.
This misapplication of the Pali imaginaire can only be regarded as a form
of hermeneutical extremism. In Nirvana and other Buddhist Felicities (1998:
xx), Collins recognised the historical difference between canonical Pali texts
(‘c. fourth-third C. BC’) and their commentaries (‘fifth-sixth C. AD, some
perhaps later’). But Wisdom as a Way of Life rejects these necessary historical
foundations. In an introductory section entitled ‘On early Buddhism and
Buddhaghosa’s Fantasy’ (p.l), Collins writes as follows:

13
v. Hinüber (1996: 195).

176
 Book reviews

What did Buddhaghosa do, apart from writing commentaries and

The Path of Purification? He created, or better put together, no
doubt at least from some earlier materials, a make-believe world
of the time of the Buddha, when the Great Man walked the earth
and Enlightenment was readily available, sometimes after a single
sermon, sometimes even after he uttered a single telling phrase. Bliss
was it in that dawn to be alive. Although some pre-Buddhaghosan
textual sources in languages other than Pali do exist – all of them
from the first five centuries AD – almost all modern scholarly
accounts of Early Buddhism, with only a very few exceptions,
rely on the Pali Canon (usually translations of it, of course). I call
this “Buddhaghosa’s fantasy” not because I wish to criticize it or
be supercilious about it, but simply as a phrase depicting the Pali
Canon as a roseate textual world of the imagination collected and
constructed by Buddhaghosa, as “The Early Days” … (p.lv)
The notion that Buddhaghosa ‘collected and constructed’ the Pali canon, and
so created ‘a make-believe world of the time of the Buddha’, is simply wrong.14
But it can be discounted as a strange bias, for as Justin McDaniels notes, Collins’
attacks on the study of early Buddhism were virtually ‘a vendetta’ (p.lii) which led
him to ‘lash out at times at his former self and his early training. It seemed almost
to me like a type of reckoning, a settling of scores with youthful indulgences
and hesitations’ (p.li). Unfortunately for Collins, this vendetta, and the resulting
ahistorical approach to Pali texts, undermines Wisdom as a Way of Life. Prose
stories must have accompanied the Jātaka verses from the beginning; they were
not invented by Buddhaghosa or any other redactor of the commentaries.

*****

Despite these problems, Wisdom as a Way of Life is not without its merits. Collins
is right to stress the literary merit of the Jātakas, and the subtle problems these
stories address, such as the conflict of values between worldly life (especially
kingship) and renunciation, which suggests an ironic awareness of the tradition’s
sacred ideals. Even if Theravāda kings used the Jātakas for ideological purposes,
their charm and meaning resides largely in the real-world scenarios envisioned

14
On the antiquity of the Pali canon see Sujato and Brahmali (2015) and Wynne (2005, 2018).

177
 Book reviews

(if animal stories can be regarded as realistic). Collins was right to notice this,
but misguided in conceptualising it in terms of ‘dhamma 1’. It is also true that
the civilising impetus of Buddhism is often overlooked; Theravāda studies would
certainly benefit from further contributions from this perspective. Moreover, such
things as modes of piety and bodily deportment are often overlooked in studies
of Buddhist meditation. The ‘Theravāda civilisations project’ is a good idea, and
Collins has pointed towards fruitful lines of future enquiry.
But these positives must be balanced against other regrettable aspects of the
book: the many mistakes of fact and perspective, the misconceived analysis of
Part One, the lack of analysis in Part Two, and the general disconnection between
the two parts. Above all, Collins’ rejection of textual history is a serious mistake.
The synchronic study of the Pali canon, especially as essentialised into the ‘Pali
imaginaire’, is a blunt tool of analysis that can be easily misapplied. In Wisdom as
a Way of Life this approach has resulted in a misreading of the Jātakas. And this in
turn obscures what was the original project in Buddhist civilisation: the elaboration
of Buddhist universalism in the Jātakas, and its appropriation by Aśoka into a state
ideology which changed the culture of classical India and beyond.

References
Appleton, Naomi. 2019. ‘Jātaka Stories and Paccekabuddhas in Early Buddhism’.
Buddhist Path, Buddhist Teachings - Studies in Memory of L.S. Cousins: 305-
318. Equinox eBooks Publishing.
Appleton, Naomi and Shaw, Sarah. 2015. The Ten Great Birth Stories of the Buddha.
The Mahānipāta of the Jātakatthavaṇṇanā. Volume One. Chiang Mai: Silkworm
Books; Chulalongkorn University Press.
Bodhi, Bhikkhu. 2017. The Suttanipāta. An Ancient Collection of the Buddha’s
Discourses Together with Its Commentaries, Paramatthajotikā II and excerpts
from the Niddesa. Somerville, MA: Wisdom Publications.
Collins, Steven. 1998. Nirvana and Other Buddhist Felicities. Cambridge:
Cambridge University Press.
———. 2010. Nirvana: Concept, Imagery, Narrative. Cambridge: Cambridge
University Press.
Freiberger, Oliver. 2006. ‘Early Buddhism, Asceticism, and the Politics of
the Middle Way.’ In Asceticism and Its Critics. Historical Accounts and
Comparative Perspectives. Edited by Oliver Freiberger. Oxford, New York:
Oxford University Press.

178
 Book reviews

Gombrich, Richard F. 2006. Theravāda Buddhism. A social history from ancient

Benares to modern Colombo. Second edition; First Edition, 1988. London and
New York: Routledge.
v. Hinüber, Oskar. 1996. A Handbook of Pāli Literature. Berlin, New York: Walter
de Gruyter.
Jones, J. J. 1949. The Mahāvastu Volume I. London: Luzac & Company, Ltd.
Jory, Patrick. 2016. Thailand’s Theory of Monarchy: The Vessantara Jataka and the
Idea of the Perfect Man. Albany: State University of New York Press.
Norman, K. R. 1992. The Group of Discourses (Sutta-nipāta) Volume II. Revised
Translation with Introduction and Notes. Oxford: The Pali text Society.
Olivelle, Patrick. 1993. The Āśrama System. The History and Hermeneutics of a
Religious Institution. Oxford, New York: Oxford University Press.
Senart, É. 1882. Le Mahâvastu. Texte Sanscrit publié pour la première fois et
accompagné d’introductions et d’un commentaire. Tome Premier. Paris:
Imprimé par Autorisation du Garde des Sceaux à L’imprimerie Nationale.
Sujato, Bhikkhu and Brahmali, Bhikkhu. 2015. The Authenticity of the Early Buddhist
Texts. Supplement to Journal of the Oxford Centre for Buddhist Studies, vol. 5.
Wynne, Alexander. 2005. ‘The Historical Authenticity of Early Buddhist
Literature: A Critical Evaluation’. Wiener Zeitschrift für die Kunde Südasiens
XLIX: 35–70.
———. 2019. ‘Did the Buddha exist?’. Journal of the Oxford Centre for Buddhist
Studies 16: 98-148.

179
 Book reviews

Gender and the Path to Awakening: Hidden Histories of Nuns in

Modern Thai Buddhism, by Martin Seeger. Nordic Institute of Indic
Studies, Monograph Series, no.144. 342 pp. Preface.
Silkworm Press, Chiang Mai, Thailand/Nias Press, Copenhagen.
Denmark. ISBN 978-87-7694-258-8.

Reviewed by Sarah Shaw

At a time when many are speaking for, about, instead of, or on behalf of
nuns, it is pertinent and refreshing to read a book that explores the lives of six
twentieth-century nuns in detail, analysing and recounting what they had to
say about their practice and life. Seeger is in a particularly strong position to
do this. One of the few Western scholars with real knowledge of the Buddhist
background, language, culture and history of Thailand, his work, based
on years of ethnographic and bibliographic research, provides an essential
service to those of us who cannot speak Thai, but would like to find out more.
And it is always much easier to assimilate a large body of information about
practices, procedures and doctrine if we can attach it to the specificity of
detail and evocation associated with particular persons. The lives of a number
of nuns are chosen, many of whom are now considered arahants or saints:
Khunying Yai (Khunying Damrongthammasan Yai Wisetsiri, 1882–1944),
Mae Bunruean Tongbuntoem (1895–1964), Mae Chi Kaew (1901–1991), Mae
Chi Nari Karun (1876/7–1999), Mae Chi Phimpha Wongs-udom (1912-2010),
and Mae Chi Soda (1920–2009). They present admirable and awe-inspiring
examples; the histories given by Seeger offer a varied and nuanced picture
of many features peculiar not only to their personal practice, but also to the
milieu and religious background to female Thai spiritual life in general over
the last 140 years.

180
 Book reviews

Seeger has chosen only women who have died, a decision that allows him
to see their lives as a whole, including, importantly, their deaths, funerary
celebrations and subsequent reputation. For this is a distinguished group: those
whose attainments were so highly respected that in some cases stūpas and
memorials have been erected in their honour and their relics seen as objects of
devotion. After a brief survey of some parameters of the research and archives
accessible to him, Chapter 1 sets the scene: it is particularly helpful for anyone
interested in Southeast Asian culture and female participation, both lay and
monastic, over the last century or so. Seeger addresses the complexities of
the Thai attitude towards gender and spiritual practice, and provides a useful
historical account, noting the presence of an apparently strong bhikkhunī order
in ancient times, that went into decline before disappearing completely. For
the nuns’ order technically does not exist; the direct ordination line has been
lost and has traditionally been thought not to be then capable of re-instatement.
So the participants in a monastic life discussed here take eight or ten basic
precepts and live as nuns, not eating after noon, living a celibate life and so on.
Seeger explains the background economically and effectively, along with the
Pali canonical precedent for the nuns’ order. This chapter includes reference to
material which suggests the ordination could be re-established if permitted.
A number of issues are raised in this and later chapters. The first is the simple
difficulty of accessing reliable material about these women. The significance and
necessity in some Northern and Eastern Buddhist schools for autobiographies
and sacred biographies as central to communicating the lineage of a particular
guru, Rinpoche line or master is not a characteristic of Southeast Asian literary
composition. Religious autobiography is not indigenous: the genre, so embedded
in the Western literary environment since the seventeenth century, has never
historically been a natural mode though, as Seeger notes, it is there in early texts.
The interpretation of the fourth pārājika, common in South and Southeast Asia,
whereby monks – and nuns – refrain from boasting wrongly of meditational
achievement had, until the early twentieth century, led by common consensus to
a reticence on such matters; lay practitioners share this caution. It was only with
the great biographical traditions of the Northeast Thailand monastic schools
that this situation changed, and then only marginally (pp.44–45). Even Ajahn
Mun, whose arahantship was widely celebrated, did not make the claim himself
(pp.122–124). Most monks or nuns would not speak of such matters to this day
and tend to refer to meditational attainments only circuitously. Seeger explores
this reluctance to use the autobiographical mode, and then too, other attendant

181
 Book reviews

problems: not only is there a far greater reticence amongst female practitioners,
but it is sometimes difficult in such Thai literature to tell the difference between
autobiography and biography: fluidity in the use of first and third person is
frequent, with ‘mother’ (mae) or ‘teacher’ (ajan) often applied to content that
could have been written or dictated by the practitioner or students. Subsequent
editors often insert sections, or doctor and revise anecdotes that may have been
in the first place recorded by someone else, or placed after the event.
Despite these problems, Seeger has compiled a considerable archive of
personal recollections, anecdotes, hagiographies, pictures, press-cuttings,
contemporary comment and material objects such as amulets: basic facts about
the lives of the women are made succinctly but vividly accessible. We receive
a sense of the quite different backgrounds of each, as their circumstances,
education, practice and background are discussed in turn in Chapter Three.
Their routes to taking the holy life are varied and shed a fascinating light on
the obstacles, difficulties and societal pressures they had to overcome simply
to engage in the monastic life. The pulls of lay life, social pressure, opposition
from husbands and families and disapproval for their chosen path feature
frequently. Their practice interests are wide and varied. Most were highly adept
and experienced in a number of traditional meditative skills: most memorised
chants and texts not only for their teaching, but as a spiritual practice. Many
are said to have acquired considerable psychic ability and depth in samatha
meditation as well as vipassanā; some were accredited in their lifetimes with
supernatural powers (pp.75; pp.86–9). The deployment of these for healing and
inspiration to others also features (pp.248–9). Many, if not all, seem to have
been seen as approaching or reaching awakening. They are a formidable host.
These short biographies have certain recurring elements: Seeger traces
several crucial threads running through their lives, and explores them in separate
chapters. These include the notion of arahantship, and how it is validated and
assessed in modern Thailand; material objects such as relics, stūpas, amulets,
clothes and papers as manifestations and expressions of devotion; and the
prevalence of orality, memorisation and chant as an underlying mainstay of
female spiritual practice.
The themes of arahantship, relic devotion and stūpa construction are closely
related, and inevitably linked now in Thailand to the genre of autobiography. For
the twentieth-century stream of monastic biographies, from Northeast Thailand
in particular, started to broach what had previously been the domain of rumour
and speculation. The arahantship of several figures now started to be claimed

182
 Book reviews

more widely, and in print. Female subjects for this genre are rather rare, but
the notion that someone could be perceived and stated to be awakened gained
more public prominence. Stūpas were constructed for some of these women
amidst recognition that they had achieved enlightenment, though the women
involved had often requested the minimum of fuss; the relics of some were
said to have undergone crystallisations (phra that) confirming their arahant
status (eg. pp.100-101; pp.145–149). In the wake of this, material objects such
as amulets and other precious talismans emerged, validating the women as
supreme exponents of meditation and spiritual attainment: Mae Bunrean was
famously asked to consecrate amulets during her own lifetime, traditionally
a male monastic preserve (p.171). The power of such validations in Thailand
cannot be overestimated. As Seeger writes, such material objects “have often
had a much wider and deeper impact on religious practices, emotions, and
beliefs, than Pali canonical texts or sermons by contemporary figures” (p.180).
Chapter 5, on orality in Thai culture, is of particular significance for our
understanding of Buddhist culture in the region: as so often in studies of female
worlds, it opens a door onto a vibrant and hitherto hidden world of practice, ritual
and oral transmission. Women were often prime exponents of oral recitation but
such exercises were popular amongst everyone. Seeger’s discussion discloses a
way of understanding the relationship of text, ritual and personal practice that is
radically different from modern western models. Those interested in Southeast
Asian monasticism are used to hearing accounts through the lens of Anglophone
books, about or by men, designed or translated with a Western and international
readership in mind. Westerners also assume that training in Buddhist text – and to a
certain extent meditation itself – necessarily involves ‘reading’ and absorbing core
texts through solitary engagement with a book. But the detail of the lives of these
women challenge assumptions that literacy and formal education are necessary
both for the transmission of teachings and for serious practice. The importance
of ‘non-literacy’ then becomes fascinating: the meditative and cultural training of
these women often depended upon an impressively knowledgeable background
in Sutta, manual, commentary, and chant, all the more highly developed precisely
because many women at the centre of female monastic life had come late in life,
or not at all, to reading and writing themselves. In order to take ordination, some
had to learn to read before compulsory female education was instated in the 1920s;
but it is clear their mainstay remained the recited and chanted text (p.240–2). In
traditional Thailand, as Seeger explains, a text was never really a solid artefact
to read in solitude; oral teaching, learning and recitation had always underpinned

183
 Book reviews

practice. Such transmissions are equally powerful but leave less obvious traces:
more social, more suited to reflection, and working more actively on people’s
sensibilities, chants, stories, texts and recitative verses are assimilated in a way
that is quite different from our modern patterns of engagement. Texts would be
learned by heart, taught perhaps by mother to daughter, in largely undocumented
transmission (p 182–193). Practitioners of both sexes have historically applied the
mind and explored the meaning of the texts from a number of angles: Ajahn Mun
recited texts as part of his core forest practice (p.209). As Seeger shows, through
extensive quotation from monks and nuns, practitioners reflected on them, taking
them as a basis for practice and considering their meaning; the very chanting and
recollection of the text, with all its meaning and rhythms, was a primary element
in their meditation. Khunying Yai, for instance, never talked about ‘reading’ texts,
but ‘listening’ to them and ‘reciting’ them, using words like ‘recall’ (raluek) and
‘memorize’ (jam son wai) for her copious knowledge of Abhidhamma, Sutta
and vernacular training manuals, employed for her recollection of the teaching
(dhammānussati) and personal meditation (pp.185-193). It is a revelation to find
in this chapter such a living tradition of textual knowledge, amongst those who
have ‘heard much’ (bahussuto), present so recently in people who were deeply
learned, but might not have read a word.
Chapter 6, on gender, gathers these threads together, including the self-
reflexive comments, where they can be found, of the women involved in
issues pertaining to awakening and their role as nuns, alongside developments
and obstacles faced by nuns. Seeger notes the complexity of factors involved;
many issues westerners and modern scholars see as crucially important do not
assume the same prominence amongst mae chis, who perceive the real chance
in their lives as the opportunity to practise spiritual and meditative work; not
all prioritise status, wealth, public position or even education, though Seeger
carefully notes the often context-specific imbalances in these areas. Many nuns
have not invested so much importance in features such as hierarchy, a lower
social role, and domestic work: assuming a lower status and performing domestic
tasks around and about are taken as a practice, as are other obstacles (pp.164–5;
p.248). Such a position confers a power that can seem mysterious to Westerners,
but means also that those wanting to intervene on their behalf need to be mindful
of innate strengths and reserves that the order itself may already have which
are sometimes less obvious to outside commentators. At any rate the notion that
women have equal capacity for awakening, as suggested by the earliest texts, is
generally accepted, and it seems many mae chis have taken refuge in that.

184
 Book reviews

Seeger does not use words like ‘ought’ or ‘should’ in his account of
their status and the issue of whether there should be fully ordained nuns, a
commendable omission. He cites the now extensive discussion on this matter;
he also carefully highlights the varied deprivations and lack of opportunities
some endure in some regions of Thailand. His caution leads to reflection:
one can understand the reticence of the sangha in Thailand to re-institute
the nuns’ order, as the existing polarity of sangha and laity has for so many
centuries been the bedrock of Thai life. But, as Seeger observes, “it seems
only a matter of time before a Theravada bhikkhunī order will take root in
Thailand” (p.31). In this light one can deeply admire nuns within Southeast
Asia, and also internationally, who exercise patience in this regard. Women
have a habit of just doing things they are not ‘supposed’ to, and doing them
so well that in the end to deny their presence is impossible. This appears to
be the case here. These determined nuns just wanted to practise the holy life
and, as their biographies show, often overcame immense obstacles to do so. It
struck me while reading that the process seems, at a completely different level
of engagement, a little like the admission of women to Oxford University in
the twentieth century: women simply worked as if undergraduates, not getting
degrees until, in the end, their achievements and presence could be denied no
longer. The nuns’ situation is different: their presence was validated a long time
ago, and lost. But at an anecdotal level, a quiet resolve that I suspect will lead
that way has always impressed me in the many nuns I have enjoyed meeting
on visits and stays at temples and monasteries in Thailand. At Wat Mahathat
in Bangkok I visited the nuns’ house, and could not have received a warmer or
more self-assuredly confident welcome: they are just getting on with it. They
have workable solutions to do what they want and, to a certain extent, are
respected and supported in that; they do not see institutionalising their order as
first priority. So you will not find nuns on high-powered committees, or with
any great status as representatives of the sangha. The nuns seem patient: my
impression is that they would appreciate respect and security and the same
meditative and educational opportunities as men, but are not in a hurry to
overturn systems; the situation is possibly changing (p.30–37). Clearly some
would rise to the challenge of a high position if the chances were there.
One of the most interesting and informative features of this book is the light
it sheds on the current debate about the re-instatement of the Buddha’s second
assembly in Southern Buddhism. Seeger’s focus on nuns who have died, and
hence undergone funerary rites, demonstrates how deeply these women are

185
 Book reviews

now revered, with the status of some as awakened beings validated amongst the
sangha and laity. The stūpas erected in their memory, and the folk mythology,
material culture and popular acceptance key to sustaining religious traditions in
Thailand have gradually absorbed them into the emotional and spiritual life of
people in general; many now take the status of some as arahants for granted. By
examining the way that the cult of relics, enlightenment stories, and accounts
of nimits (images in the mind’s eye; p.73) of these women appearing to modern
practitioners are now pervading popular culture, he shows us what will be the key
to the gradual acceptance of a bhikkhunī order – people at large are recognising
the presence of four assemblies: of monks, nuns, laymen and laywomen. The
book makes it abundantly clear that whatever the paperwork and administrative
logistics needed to accommodate the restitution of the lineage of nuns in a formal
sense, it has, for two or three hundred years been active in Thailand as a living
entity and, despite the low status of nuns, is now informally recognised as such.
However long it takes for public recognition validated by the sangha, and for the
vows that would mean so much to many such women that acknowledge their
connections in lineage to be publicly acknowledged, the nuns’ order in Southeast
Asian Buddhism already exists. As Seeger constantly reminds us, the women
concerned live in a manner entirely in accordance with early teachings.
It is to this book’s great credit that we feel this living continuity, assessed
with a non-judgemental and nuanced appreciation of the density and complexity
of the historical issues involved. And from a Buddhist Studies point of view,
Seeger’s sensitive and probing analysis of the lives of some women who have
contributed to that presence opens a door on the often neglected or simply
unknown worlds of female monastic and lay culture. He demonstrates with
admirable even-handedness that an often highly sophisticated textual and
meditative training animated the practice of women whose own preoccupations
are concerned simply with living ‘the holy life’ as wholeheartedly as they can.
One would not want future generations of nuns to live in this position; one hopes
that future generations have more recognition and support. But the fact that
these women seemed to turn disadvantages, including low status, to spiritual
attainment is a testament to their extraordinary confidence, resilience and
courage. I recommend Seeger’s book for its rich insights into popular Buddhism
in Thailand and the now deeply rooted culture of appreciation and recognition
for the distinguished mae chi line.

186
 Book reviews

Bhikkhu Bodhi, Reading the Buddha’s Discourses in Pāli.

Somerville: Wisdom Publications. 531 pages.
ISBN 978-1-61429-700-0. US $49.95

Reviewed by Niels Schoubben*

This nicely produced volume “is not intended to be … a Pāli primer” (p. viii),
but is comparable to Scharf’s (2003) edition of the Rāmopākhyāna and similar
works, where the, in this latter case, Sanskrit text is printed with a full glossary
below each verse and a literal translation, in order to help students to acquire
reading proficiency in the language. In the same vein, Bhikkhu Bodhi, who has
devoted a considerable part of his life to the translation of Pāli suttas, has done
us a great service in preparing the book under review.
After a comparatively brief, but informative introduction on Pāli (pp. 1–10),
the author gives a concise overview of Pāli grammar (pp. 11–48),1 and a separate
chapter on “common sentence patterns” (pp. 49–79), where he sketches some
syntactic peculiarities of the language which most modern readers will not be
that familiar with. The core of the book (pp. 81–501) consists of his selection
of (fragments of) Pāli suttas, all of which are chosen from the Saṃyutta Nikāya.
First, the original text is printed, based on the electronic version of the Burmese
edition, although occasionally readings from other editions, such as that of
the Pāli Text Society, are preferred (cf. p. x). Below the Pāli, a word-for-word
glossing is given, followed by a translation into more idiomatic English, but
not as idiomatic as the published translations of the author (cf. p. 8).2 Each

*Writing this review has been made possible within the framework of the European Research
Council (ERC) Starting Grant project “The Tocharian Trek” (Grant agreement ID: 758855).
1
No overview of paradigms is offered here, as Bhikkhu Bodhi’s focus is rather on the main
trends of Pāli grammar than on the details.
2
The Saṃyutta Nikāya is translated in its entirety in Bodhi (2000).

187
 Book reviews

section is concluded with selected grammatical explanations and at the end of

the book (pp. 503–528), a Pāli-English glossary is added as well.3 Instead of
presenting his selection of suttas in the order of their appearance in the Saṃyutta
Nikāya, the author has made the laudable decision to group passages that treat
the same basic principle of (Theravāda) Buddhism together, dividing them into
six chapters, of which the last is a very short one, with such topics as “The Four
Noble Truths” (chapter 1) or “Dependent Origination” (chapter 4).
As a consequence, the book may even be suitable for readers who have no
interest in Pāli as a language, but want to form an idea of the main tenets of Early
Buddhism from the original texts themselves rather than from an introductory
book on Buddhism, where the presentation will necessarily depend to some
extent on the personal interpretations of the author. Nevertheless, in this review,
we will rather evaluate to what extent the book fulfils its duties in helping
students to acquire the skills necessary to read Pāli texts with confidence. To
meet this aim, the author of a book such as the one under review should 1)
possess a profound familiarity with the language, i.e. the explanations given
should be correct and 2) he should be able to transmit this knowledge in such
a way that as varied a readership as possible can profit the most from engaging
with the book. In other words, the level should be high, but at the same time,
understandable to what one may want to call “interested laypeople”. It should be
kept in mind that a basic knowledge of Pāli grammar is presupposed for those
readers that really want to delve into the texts (cf. p. viii).
To come straight to the point, the author succeeds well in the aims he has
set himself. Obviously, there are always points where one may want to disagree
and I will list some of those below, but on the whole, the author should be
warmly congratulated on the work he has done. To illustrate Bhikkhu Bodhi’s
way of presenting the texts, I cite a randomly chosen example, from the
Siṃsapāvanasutta (SN 56:31; p. 112 in the book).
First, the actual Pāli text is given in a different font from the translation; the
literal translation is also printed differently from the idiomatic one.
appamattakāni, bhante, bhagavatā parittāni siṃsapāpaṇṇāni
pāṇinā gahitāni; atha kho etān’ eva bahutarāni yadidaṃ upari
siṃsapāvane ti.

3
A brief bibliography, which is unfortunately limited to Anglo-Saxon literature, can be found
on pp. 529f.

188
 Book reviews

Glossing:
“Trifling, Bhante, by the Blessed One few siṃsapā-leaves with
hand taken; but these indeed more, that is, above in the siṃsapā-
grove.”
Idiomatic translation:
“Bhante, the siṃsapā leaves that the Blessed One has taken in his
hand are few, but those above in the siṃsapā grove are indeed more
numerous.”
If I may make one comment on an otherwise sound translation, and one
that is fully in line with the others in the book, I would not have left Bhante
untranslated, as is done throughout the book. Why not simply “Sir” or
something similar instead of replicating a Pāli honorific that will be puzzling to
some readers? More generally, the necessity of glossing as well as translating
each example may be questioned, as this procedure takes a lot of space. At the
beginning of the book, I can see the usefulness of this. But once the reader
has gone through a certain amount of examples and, let us be honest, the Pāli
canon has a certain predilection for repetitions, would a translation with notes
on vocabulary and grammar not be sufficient? Even in those cases that the same
sentence is repeated just below in the text with the change of only one word
(e.g. viññāṇaṃ ‘consciousness’ instead of rūpā ‘forms’), full glossing is given
on both occasions.
What about the grammatical explanations? The author, it should be stressed,
is intimately familiar with the Pāli texts and he has done an excellent job here
as well. The comments are generally reliable and easy to follow. Two general
points should be made, however, before I list a couple of small points where I
disagree with Bhikkhu Bodhi’s explanations.
First, some infelicitous statements are made on the relationship between Pāli
and Vedic and Classical Sanskrit. The author could have chosen to present Pāli
as a language on its own, without referring to Sanskrit at all, and that would have
been completely fine. However, he occasionally does refer to Sanskrit, but not
all of his comments are fully correct and some may obscure rather than clarify
things. In the introduction (p. 1), Pāli is said “to descend” from Vedic Sanskrit,
which is not accurate because Pāli preserves linguistic archaisms (e.g. idha
‘here’) where the Vedic Sanskrit of our texts has the later form (e.g. iha ‘here’).

189
 Book reviews

Most readers will not be bothered by such details, but occasionally the clarity
of Bhikkhu Bodhi’s grammatical explanations is affected as well. For instance,
on p. 109, it is commented that in the compound sassamaṇa-brāhmaṇiyā ‘with
wandering ascetics and Brahmins’ the double -ss- “occurs through the influence
of the -śr- cluster in Skt śramaṇa”. Obviously, Sanskrit is not influencing Pāli on
this point: an older -śr- simply becomes -ss- by sound law and, when this does
not result in an over-heavy syllable this geminate is preserved and otherwise
simplified to a single consonant.
Second, Bhikkhu Bodhi makes the case system of Pāli more complicated than
it actually is, by promoting the dative to a position it no longer has in Middle Indo-
Aryan languages, where, the dative, apart from relic forms (on which, for Pāli,
see e.g. Oberlies 2019 I: 207; Spencer 2020: 121f.), merges with the genitive. As
a consequence, I would not call a form like tassa < Skt tasya a dative, as Bhikkhu
Bodhi does on several occasions (e.g. p. 18; 63; 107; 153 etc.). This is simply
a genitive used as an indirect object, as is possible in Sanskrit as well and it is
not clear to me why the author, who otherwise follows the standard grammars
quite faithfully, has opted for this idiosyncratic deviation from them. More such
examples can be found in the book: on p. 166, e.g., bhagavato is said to be “a
genitive with the function of an ablative”, but genitive and ablative simply have the
same form in the vant-declension, so that one should call such a form an “ablative”.
A few minor comments, including mere typo’s, are listed below:4
• p. xiii: Rhy > Rhys
• p. 91: tasmāt is not only Vedic Sanskrit and the “probably” may
be deleted.
• p. 91 l. 4 of the Pāli text: abhisambhujjhissati >
abhisambhujjhissanti
• p. 94; 109; 110: the two options for the translation of ariyasacca
are discussed: “Noble Truth” or “Truth of the Noble One”. The
author simply notes that some passages support interpretation
one and others interpretation two. See on this also Norman (1990
= 1993: 171–174), who argues that both meanings are intended
at the same time. On p. 251, SN 35: 228 [187] is cited, where one

4
I will not list here all those cases where I disagree with the use of the term “dative” or with
the way Sanskrit etymologies are presented.

190
 Book reviews

reads ariyassa vinaye ‘in the discipline of the Noble One’. This
could also be cited as additional support for a translation “Truth
of the Noble One”.
• p. 95 fn. 73: Māgadhi > Māgadhī
• p. 105: pativijjhati > paṭivijjhati
• p. 115: Even though paññāya is correctly translated, it is
explained wrongly as an absolutive, whereas it is here a dat.f.sg.
of the noun paññā- ‘understanding’.
• p. 158: Because, as is accurately discussed on p. 160, sadevakā
… sadevamanussāya are ablatives, they should not be translated
as if they are locatives. The punctuation of the Pāli could also be
improved here.
• p. 180 (et passim): Tradition is followed and diṭṭhe’va dhamme
is translated as “in this present life”, but Gombrich’ (2006²: 116
fn. 14) “when he has seen the truth” seems more likely to me.
• p. 192: The English word “monk” has intruded in the Pāli text
instead of bhikkhu.
• p. 213: yoniso ‘thoroughly’ is confusingly called an “ablative”,
but -so is an adverbial suffix < Skt -śaḥ (On p. 302, the correct
identification of yoniso as an adverb is given and on p. 342,
sabbaso ‘entirely’ is rightly explained as well).
• p. 373 with fn. 186: For the occasional use of nominative phrases
to introduce places in Pāli, cf. von Hinüber (2006: 198–200 with
further ref.). The explanation cited from the Sumaṅgalavilāsinī
seems rather far-fetched.
• p. 398ff: ‘to enter the rains’ is too literal a translation for vassaṃ
gacchati. For readers who do not know the expression, the meaning
only becomes clear at p. 405, where it is explained that this idiom
refers to the three-month retreat of monks during the rainy season.
• p. 405: upagacchāmi is of course first person and not third person
singular.
• p. 444: Sākata > Sāketa

191
 Book reviews

One final point. Even though the author justifies his choice (p. 5; 8), it is still
a pity that only passages from the Saṃyutta Nikāya are included in the book.
At least parts of some other Nikāyas could have been incorporated to present
the student with a more representative sample of Pāli literature. In fact, I think I
would not be alone in welcoming a second volume that would not only include
samples from the other Nikāyas, but also from the Vinaya, the Abhidhamma and
other Pāli texts, such as the Jātakas, the Dhammapada, etc. Such a book would
be another significant tool for more advanced students of Pāli.
Saṃkhittena, an occasional point of criticism aside, there should be no doubt
that this book will form a major help for the happy few who want to read the
Buddha’s teachings in Pāli.

Bibliography.
Bodhi, Bhikkhu (2000). The Connected Discourses of the Buddha: A New Translation
of the Saṃyutta Nikāya. Boston: Wisdom Publications.
Gombrich, Richard Francis (2006²). How Buddhism Began: The Conditioned
Genesis of the Early Teachings. 2nd Edition. London: Routledge.
Hinüber, Oskar von (2006). “Hoary Past and Hazy Memory. On the History of Early
Buddhist Texts.” Journal of the International Association of Buddhist Studies,
29 (2), 193–210.
Norman, Kenneth Roy (1990). “Why Are the Four Noble Truths Called ‘Noble’?”
In Yakupitiyage Karunadasa (ed.), Ānanda: Essays in Honour of Ananda W.P.
Guruge. Colombo, 11–13 (= Collected Papers IV, 1993, no. 90, 171–174).
Oberlies, Thomas (2019). Pāli Grammar: The Language of the Canonical Texts of
Theravāda Buddhism. Revised and Enlarged Edition in 2 Volumes. Bristol: The
Pali Text Society.
Scharf, Peter (2003). Rāmopākhyāna: The Story of Rāma in the Mahābhārata: An
Independent-Study Reader in Sanskrit. London: RoutledgeCurzon.
Spencer, Matthew (2020). “Review Article of Pāli Grammar: The Language of the
Canonical Texts of Theravāda Buddhism (Volume I), by Thomas Oberlies.”
Buddhist Studies Review, 37 (1), 117–126.

192
 Book reviews

Genèse d’un apocryphe bouddhique: le Sūtra de la pure délivrance

by Costantino Moretti.
Paris: IHE Chinoise, December 2016.
ISBN-10: 2857570767
ISBN-13: 978-2857570769

Reviewed by Frederick Shih-Chung Chen

Chinese Buddhist apocrypha have long been depreciated by orthodox scholarly

monks and Buddhist cataloguers for their inauthentic nature as forgeries of
Indian Buddhist scriptures. In the late 1980s and early 1990s, however, leading
scholars such as Robert Buswell called for a reassessment of their value, arguing
that they actually provide significant material relating to less-explored facets of
the development of Buddhism outside traditional philosophical and doctrine-
centred study. Noting that the Chinese apocryphal scriptures are made up of
diverse traditions, Buswell in particular reminded us that, “Exclusive focus
on national traditions all too easily conceals the manifold points of symbiosis
between those traditions, which can be of immense value in detailing both the
evolution of the national varieties of Buddhism and the indigenous texts that
reflect that evolution.”1
Coincidentally, a collection of medieval manuscripts of Buddhist scriptures
was uncovered in 1990 at the Nanatsu-dera 七寺, Nagoya, Japan. Among them
is the previously missing second juan of the three-juan version of the Jingdu
sanmei jing 淨度三昧經 (‘Samādhi-Sūtra of Liberation through Purification’).
This sūtra, composed from miscellaneous origins, has long been regarded by
scholars as one of the most important Chinese indigenous apocryphal scriptures
epitomizing the Sinification and popularization of Buddhism during the period

1
Buswell, Robert E. ed. Chinese Buddhist Apocrypha. Honolulu: University of Hawaii Press,
1990, p. 22.

193
 Book reviews

of the Northern-Southern dynasties. Classified as apocryphal during the

medieval period, the sūtra was excluded from the Chinese Tripiṭaka, and the
full text was subsequently lost, although modern scholars had tried to recover it
from quotations in secondary texts and the fragmentary manuscripts excavated
at Dunhuang. The discovery of the Nanatsu-dera manuscripts has now made it
possible to reconstruct a complete version of all three juans.
Costantino Moretti’s Genèse d’un apocryphe bouddhique: le Sūtra de la
pure délivrance (‘The Genesis of a Buddhist Apocrypha’) is a monograph based
on his PhD thesis, which seeks to update our understanding of the making of
the Jingdu sanmei jing. It is one of several attempts in Western languages to
reassess this particular sūtra at the PhD level since the discovery of the Nanatsu-
dera manuscripts.
The reconstructed Jingdu sanmei jing was first studied by the research team
of The Manuscripts of Nanatsu-dera in Japan.2 Harumi Hirano Ziegler’s PhD
thesis at UCLA in 2001 (supervised by R. Buswell and published by UMI), “The
Sinification of Buddhism as Found in an Early Chinese Indigenous Sūtra: A Study
and Translation of the Fo-shuo Ching-tu San-mei Ching,” is the pioneering work
in a Western language on the recovered three-juan version. It surveys the origin
and composition of the sūtra, and aspects of Sinification, and it includes a full
English translation of all three juans. Ziegler challenges a view held by some
previous scholars that the sūtra was probably composed by Tanyao 曇曜 (fl. 453
– 499 CE), a leading figure in the resurrection of Buddhism after its persecution by
Emperor Taiwu 太武帝 of the North Wei 北魏 during 446 – 452 CE. Tanyao had
first been mentioned as one of several translators of different versions of the sūtra
in the Lidai sanbao ji 歷代三寶記, a somewhat unreliable Buddhist catalogue by
Fei Changfang 費長房 (fl. 562 – 598 CE). The assumption was generally based
on comparison of the Jingdu sanmei jing with another significant apocryphal
scripture, the Tiwei jing 提謂經 (‘The Sūtra of Trapuṣa and Bhallika’), composed
by Tanjing 曇靖 during 453 – 464 CE in the Northern Wei; both scriptures display
a similar mixture of Buddhist and indigenous Chinese beliefs and they share some
content, such as the “Account of the Days of the Eight Kings 八王日”. From a
close reading of the earliest existing bibliographical source, the Chu sanzangji

2
Ochiai Toshinori 落合俊典 ed. The manuscripts of Nanatsu-dera. Kyoto: Italian School
of East Asia Studies, 1991. Makita Tairyō 牧田諦亮 ed. in chief & Ochiai Toshinori 落合俊典
managing ed. Nanatsudera koitsu kyōten kenkyū sōsho chūgoku senjutsu kyōten (sono2) 七寺古
逸經典研究叢書 中國撰述經典(其之二). Tokyo: Daitō Shuppansha, 1996.

194
 Book reviews

ji 出三藏記集 (‘Collection of notes on the translated Tripitika’), compiled by

Sengyou 僧祐 (445 – 518 CE), Ziegler established that, while several versions
of the Jingdu sanmei jing were available in South China prior to 515 CE, three
sūtras translated by Kiṅkara 吉迦夜 and Tanyao in 472 CE in the Northern Wei
had not at that date been transmitted there, due to the division of territory by the
northern and southern polities. The later attribution to Tanyao in the Lidai sanbao
ji therefore seems unlikely. Given that the three-juan version of the Jingdu sanmei
jing contains a wide range of Buddhist ideas and references, Ziegler held that it
could only have been composed in South China, where scholarly sources were
more easily available than in the Northern Wei after the persecution of Buddhism.
Ziegler’s approach to dating the sūtra rests on the assumption that this
reconstructed three-juan version is the original text. However, the three-juan version
was not first mentioned in the Chu sanzangji ji, but attributed to Baoyun 寶雲 (376-
449 CE) in the Zhongjing mulu 眾經目錄 by Fajing 法經 in 594 CE, and versions
of the Jingdu sanmei jing with different numbers of juan had already been recorded
earlier, some supposedly having been subject to later modification. The Chu
sanzangji ji includes a note asserting that the “Record of the Origin of the Abstinence
Days of the Eight Kings” 八王日齋緣記 comes from the Jingdu sanmei jing. The
same account is also in the Tiwei jing. This suggests to me that the “Record of the
Days of the Eight Kings” was once in the earliest stratum of the sūtra. In my DPhil
thesis (University of Oxford 2010),3 I inspect the origin of the Jingdu sanmei jing by
examining this particular feature. Textual analysis of the “Record of the Days of the
Eight Kings” shows that its content was probably formulated and developed through
the integration of the Buddhist text of the Four Great Kings 四天王經 with the eight
seasonal days stipulated by the Laozi zhong jing 老子中經 (‘Central Scripture of
Laozi’). The Laozi zhong jing is an early Daoist scripture on self-cultivation (dated
by Kristofer Schipper not later than the fourth-century Baopuzi 抱朴子),4 which
proclaims that the human lifespan can be prolonged by the visualization of deities on
the associated part or organs of the body on associated festive days, particularly the
eight seasonal days. Both texts were based on parallel religious ideas about periodic
abstinence days and associated metaphors of the inspection and recording of human
deeds by otherworld bureaucratic deities.

3
Chen, Frederick Shih-Chung. The Transformation of Concepts of Bureaucratization of the
Other World in Early Medieval China: From Buddhist Perspectives. DPhil diss., University of
Oxford, 2010, pp. 92 – 176.
4
Schipper, Kristofer. “The Inner World of the Lao-Tzu Chung–Ching.” in Huang, Chun-Chieh
and Zürcher, Erik ed. Time and Space in Chinese Culture. Leiden: Brill, 1995, pp. 118 – 119.

195
 Book reviews

Analysis also reveals that the Eight King Messengers 八王使者 in the Jingdu
sanmei jing are actually a Buddhist adoption of the Eight Trigram Deities, the
invocatory deities recommended by the Laozi zhong jing for life-prolonging
visualizations on the eight seasonal days. Furthermore, this account is probably
the textual source of the identities of the enigmatic Eight Trigram Deities
depicted on the bottom of the Northern Liang votive stūpa 北涼石塔 below
the line of the Foshuo shi’er yinyuan jing 佛說十二因緣經 (‘Sūtra Spoken by
the Buddha on Twelve Co-dependent Originations’) and also of the images of
the seven past Buddhas and the Buddha of the future — Maitreya. The function
of stūpas, as Peter Harvey has concluded, is a visualization of “representing
the Dhamma (teaching, path and realizations) and the enlightened personality
embodying the culmination of Dhamma-practice.”5 As the doctrines of Twelve
Co-dependent Originations and the Eight Buddhas are both meditative objects
for enlightenment according to Buddhist practice,6 the Eight Trigram Deities
below them on the Northern Liang stūpa therefore epitomize a Buddhist
adoption of visualization objects from Daoism. The inclusion of the meditative
term samādhi 三昧 in the title of the apocryphal scripture might imply such
religious practices in Northern Liang. The early core stratum of the Jingdu
sanmei jing was very likely composed during this period by monks from the
Northern Liang, such as Baoyun who later moved to the south.7
I was not able to access Moretti’s thesis before my article on the Eight Kings
in the Jingdu sanmei jing appeared in Asia Major,8 as it was still under revision.
Therefore I was extremely excited to learn about this publication. It is a well-
researched book that presents in encyclopedic detail an extensive range of
primary and secondary sources relating not only to the Jingdu sanmei jing, but
also to issues concerning other early medieval and medieval indigenous Chinese
Buddhist scriptures. Moretti explores the origin of the sūtra by considering

5
Harvey, Peter. “The Symbolism of the Early Stūpa.” in The Journal of the International
Association of Buddhist Studies, vol. 7, no. 2, 1984, pp. 67 – 93.
6
Zuochan sanmei jing 坐禪三昧經 by Kumārajīva 鳩摩羅什. T. 15, no. 614, 282c – 284; Fo
shuo guanfo sanmei hai jing 佛說觀佛三昧海經. Buddhabhadra 佛陀跋陀羅 (359 – 429 CE) T.
15 no. 643, ch.10, p. 693.
7
Leading features in the resurrection of Buddhism in the Northern Wei, such as Tanyao, were
also originally from the Northern Liang.
8
Chen, Frederick Shih-Chung. “Who Are the Eight Kings in the Samādhi-Sūtra of Liberation
through Purification? Otherworld Bureaucrats in India and China.” Asia Major, 3rd ser., 26, no.
1 (2013): 55–78.

196
 Book reviews

four aspects: its place in bibliographic catalogues, its content and philological
borrowings, the narration of hells, and specific elements such as the pantheon,
practice and worship. The book also includes a French translation of the first
juan of the Jingdu sanmei jing.
The first chapter surveys bibliographic catalogues relating to the Jingdu
sanmei jing. Of these, Moretti considers that the Lidai sanbao ji by Fei
Changfang, the first to mention the one-juan version translated by Tanyao, gives
the most information about its origin. His argument, based on the views of such
scholars as Tsukamoto Zenryū 塚本善隆 and Makita Tairyō 牧田諦亮, posits a
strong link between the Jingdu sanmei jing and the Tiwei jing. Their similarities
of style, content and doctrine, not to mention philological and linguistic features,
suggest that the two texts could have been conceived in the same environment.
Moretti is confident that Tanyao, if not the true “editor” of the Jingdu sanmei
jing, was at least the person responsible for its “making”.
The second chapter elucidates the content and philological borrowings in
the Jingdu sanmei jing in three sections. It starts with a detailed illustration of
how the sūtra of three juans was reconstructed from the surviving manuscripts
in Dunhuang and the Nanatsu-dera and summarizes the content. Secondly, it
traces the content and context of quotations from the Jingdu sanmi jing that
survived in secondary sources (encyclopedic works, religious commentaries
and treatises). The survey is summed up in a meticulous illustrative table of the
contents of the Jingdu samei jing and corresponding quotations from associated
manuscripts and secondary sources (pp. 110 – 115). The third section analyzes
the linguistic and stylistic borrowings manifest in the sūtra. The overall
examination shows heterogeneity in the choice of translation styles and forms,
ranging from the very complex phonetic transcriptions that characterize some
translators to the Sinicized forms that mark the style of others. For example, in
certain cases, instead of using the translation style of Kumārajīva (344 – 413
or 350 – 409 CE), which was closer in time to the formation of the sūtra, more
archaic forms by Dharmarakṣa 竺法護 (230? – 316 CE) were chosen. These
inconsistencies, from Moretti’s viewpoint, confirm the apocryphal character
and heterogeneous features of the sūtra. With regard to the usage of certain
terms that are not specifically Buddhist, Moretti points out that these typically
Chinese religious expressions, which some translators chose to use and others
tried to avoid, were generally familiar to lay people. Their inclusion suggests
to him that this apocryphal sūtra was aimed at an audience mainly composed
of lay people.

197
 Book reviews

The third chapter investigates the enumeration of hells in the Jingdu sanmei
jing, particularly the thirty hells, drawing comparisons with previous Chinese
Buddhist texts. First indicated by Saitō Takanobu 齊藤隆信, the thirty hells
were formulated through a combination of parallel narrations mostly from three
earlier Buddhist scriptures, namely the Tiecheng nili jing 鐵城泥犁經 (T.1, no.
42), the Nili jing 泥犁經 (T. 1, no. 86), and the chapter on the visualization
of hells as meditative objects for liberation in the Xiuxing daodi jing 修行道
地經 (the Yogācārabhūmi, the ‘Ground of Meditation Practitioners’ T. 15, no.
606).9 Moretti explores in minute detail the parallels between each of the thirty
hells enumerated in the Jingdu sanmei jing and those in the previous scriptures,
summarizing his findings in a clear diagram. He elucidates the description of
each hell and the religious moral and practice that lay behind it.
The fourth chapter surveys featured elements of the Jingdu sanmei jing,
including the pantheon, practice and worship, via a threefold examination of the
five precepts, the Days of the Eight Kings and the concept of self-salvation. First,
it traces the association between the Buddhist five precepts and the five officials
of Chinese indigenous deities and the twenty-five guardian deities who protect
keepers of the precepts. While the five officials were probably an expansion
of the three celestial officials in early Chinese religions to match the religious
symbolism of the number five, the twenty-five guardian deities of the five
precepts were first mentioned in earlier Buddhist scripture. Such associations
with the five precepts are further extended to other symbolic instances of the
number five in Chinese religion, such as the five viscera, five elements, and
so on, presented in the Tiwei jing and other similar Buddho-Daoist scriptures.
Secondly, Moretti examines content relating to the Abstinence Days of the Eight
Kings in this three-juan version of the Jingdu sanmei jing, considering such
issues as the observance of precepts, consequent reward and punishment in
terms of increased or decreased lifespan, and the bureaucratic deities involved
in inspecting and recording human actions. Although all these examples show
clearly that the Days of the Eight Kings, derived from the eight seasonal days in
Daoism, were considered particularly important by the Daoist tradition, Moretti
maintains that the complexity of the interplay between Buddhism, Daoism and

9
Saitō Takanobu 齊藤隆信. “Jōdo sanmaikyō no kenkyū: Anrakushū to Kannen hōmon no
baai 『浄度三昧経』の研究―『安楽集』と『観念法門』の場合.” Bukkyō Daigaku Sōgō
Kenkyūjo kiyō 佛教大学総合研究所紀要, no. 3, 1996, pp. 218 – 219. The Xiuxing daodi jing is
a Śrāvakayāna 聲聞乘 treatise for meditation practitioners (yogācāra) on the practice of calm and
insight 寂觀 (śamatha-vipaśyanā) for attaining nirvāṇa.

198
 Book reviews

Chinese popular religious beliefs makes it difficult or even dangerous to trace

the roots of “Daoist contamination” on the basis of insufficient information: “We
should avoid claiming that elements traditionally considered typical of Chinese
thought are mixed up in the Jingdu jing with Indian ideas, like the idea of the
inspection of deeds by deities, which are associated in this text with certain days
of the year, the eight seasonal days, where different ceremonies, assimilable in
part into popular religion and into Daoism, but also into Buddhism, took place”
(p. 328). The chapter ends with a brief discussion of the issue of self-salvation.
In his conclusion, Moretti reiterates that the Jingdu sanmei jing should be
considered a treatise (in the form of a sūtra) for lay disciples or else a textbook
by which the Buddhist clergy could teach and convert lay people. He assumes
that the concept of samādhi is not approached in an explicitly doctrinal way in
the sūtra, so that the use of the term in the title is merely emblematic, intended to
add Indian colour and an authentically “exotic” stamp to the Chinese apocrypha.
In his view, the highlighting of such practices as the observance of precepts
during abstinence days and the making of offerings at non-Buddhist as well as
Buddhist festivals associates the text with the development of organized Buddhist
communities and lay associations such as yiyi 邑義 and yihui 邑會 during the
same period. These heterogeneous doctrines and practices, expounded in the
Jingdu sanmei jing, laid the foundation for the further development of popular
religions in medieval China.
The three-juan version of the Jingdu sanmei jing is made up of diverse and
complex religious texts derived from heterogeneous traditions. With painstaking
effort, Moretti pieces them together and demonstrates a clear and detailed map
of the structure of the sūtra and related references. This provides a very rich
and useful guide to the study of the Jingdu sanmei jing and Chinese apocryphal
scriptures. Such hard work deserves tremendous credit.
Given that the title refers to the genesis of this Buddhist apocryphal scripture,
it is clear that Moretti’s conclusions on this central issue are quite at odds with
those of the other two most recent works on the Jingdu sanmei jing. Makita Tairyō
cautiously assumed that it is not possible to identify its translator solely from
bibliographic catalogues and suggested that the attribution to Tanyao is probably
due to his having been the leading figure in the restoration of Buddhism when
the Tiwei jing was translated by Tanjing, also in the Northern Wei.10 Moretti, by

10
Makita Teiryō 牧田諦亮. Gikyō kenkyū 疑經研究. Kyoto: Kyōto Daigaku Jinbun Kagaku
Kenkyūjo, 1976, p. 249.

199
 Book reviews

contrast, is more confident that Tanyao was at least responsible for the “making”
of the Jingdu sanmei jing. In this, he apparently disagrees with Ziegler’s point
that, while scriptures co-translated by Tanyao and Kiṅkara were not available in
South China due to the dynastic division, the Jingdu sanmei jing had already been
circulating in the South. Moretti is also unusual in placing so much weight on
the Lidai sanbao ji by Fei Changfang as the key bibliographic catalogue. Most
scholars regard the Lidai sanbao ji as less trustworthy, particularly because of the
many new ascriptions for canonical texts seemed to be added arbitrarily by Fei.
Michael Radich has recently voiced serious concern at its careless use.11 Moreover,
it should be noticed that, in the Lidai Sanbao ji, Fei actually made a note on the
Jingdu sanmei jing attributed to Tanyao, saying that this Northern version of one
juan was the second translation. Although roughly abbreviated, it is essentially the
same as the two-juan version translated by Baoyun (which was based on an Indic
manuscript brought by Faxian 法顯, d. 418 – 423 CE).12 See the catalogue by
Daozu. 淨度三昧經一卷 (第二出。與寶雲譯二卷者同。廣略異耳。見道祖
錄).13 Therefore Fei Changfang’s comment does not support but in fact undermines
Moretti’s idea that Tanyao was the most likely editor, if not translator, of the three-
juan sūtra which comprises such a broad range of contents and doctrines. Moretti’s
claim is based not on any substantial newfound sources but on wishful thinking.
On the issue of the specific messenger and other secondary deities mentioned
in the Account of the Days of the Eight Kings in the Jingdu sanmei jing, Moretti
is insightful in comparing the parallel narrative sentences about “The Lord of
the Grand One 太一君”, who is the also Lord of human beings, residing in the
human navel, along with the Grand General of the Pillar of Heaven 柱天大將軍,
Specially Promoted War King or Lord King 特進兵王 (特進君王) and the eight
messengers, the Eight Trigram Deities (pp. 290 – 291) mentioned in the Tiwei
jing and the Shichan boluomi cidifamen 釋禪波羅蜜次第法門 (Understanding
Dhyāna Pāramitā: A Method in Stages), written by the founder of the Tiantai
School 天台宗, Zhiyi 智顗 (538 – 597 CE). On the other hand, he fails to note
that this parallel narration is actually an abbreviated quotation from the thirteenth
chapter (the thirteenth Immortal 第十三神仙) of the Laozi zhong jing about the

11
Radich, Michael. “Fei Changfang’s Treatment of Sengyou’s Anonymous Texts.” Journal of
the American Oriental Society 139.4 (2019): 819 – 841.
12
淨度三昧經二卷 (法顯齎。梵本來。見竺道祖雜錄). Lidai sanbao ji 歷代三寶紀. T. 49
no.2034: 89c18
13
淨度三昧經一卷 (第二出。與寶雲譯二卷者同。廣略異耳。見道祖錄). Lidai sanbao ji
T. 49 no.2034: 85a 24.

200
 Book reviews

Eight Trigram Deity who reports the record of human beings to the Grand One on
“the eight seasonal days.” My own article had already pointed out that, not only
this passage about the Lord of the Grand One and his Eight Trigram Messengers
in the Tiwei jing (col. 107 – 109 of Dunhuang Manuscript P. 3732), but also the
whole paragraph addressing the correspondence between deities and human organs
(col. 105 – 115) are abbreviated quotations from the 13th, 17th, 18th, 19th, etc.,
chapters of the Laozi zhong jing.14 The quotation in the Shichan boluomi cidifamen
also comes from a passage which consists of similar abbreviated sentences from
the Laozi zhong jing. As this quotation by Zhiyi includes the following sentence
“Together, they are the (so-called) Nine Ministers 合為九卿” in the Laozi zhong
jing, it appears that Zhiyi’s comment derives from his knowledge of more direct
sources of the Laozi zhong jing, rather than from the Tiwei jing. Also, the Jingdu
sanmei jing and the Tiwei jing are probably the two earliest Buddhist scriptures
to mention both the Grand General of the Pillar of Heaven and the Specially
Promoted War King or Lord King along with the Grand One.
These parallels suggest a close link between the content related to the eight
seasonal days in the Laozi zhong jing and the Account of the Days of the Eight
Kings in both the Jingdu sanmei jing and the Tiwei jing. Nevertheless, apart
from one brief reference to the two deities on shoulders 肩背神二人 included
in Yao Changshou’s 姚長壽 article (p.283), there is no further mention of the
Laozi zhong jing in this book.15 Moretti does not even include the Laozi zhong
jing in his bibliography, despite listing quite a number of textual sources and
information related to the eight seasonal days, but mostly from the later period.
While he advises us against attempting to trace the origins of the account in
the Jingdu sanmei jing, due to the complexity of the textual sources on the
eight seasonal days in his list and lack of sufficient information, it seems to me
incomprehensible that this book should totally omit such an early and closely
related primary source on the eight seasonal days as the Laozi zhong jing. His
summary directly disagrees with the argument and approach of my article in Asia
Major which highlights the importance of the Laozi zhong jing and the Buddhist
text of the Four Great Kings in the formation of the Account of the Days of the
Eight Kings (although my article is not mentioned in this context, but is merely
noted as “a hypothesis” in a brief footnote in another part of this book (p. 248)).

14
Chen (2013), p.66. Makita (1976), pp. 186 – 187.
15
Yao Changshou 姚長壽. “Jingdu sanmei jing yu rentianjiao 淨度三昧經與人天教.” in the
Zhonghua foxue xuebao 中華佛學學報, no. 12, Oct. 1999, p. 89.

201
 Book reviews

Furthermore, Ziegler’s thesis included a specific analysis of a range of

Buddhist scriptures with the title of samādhi-sūtra.16 Had Moretti paid attention
to it, he would probably have gained a broader understanding of scriptures with
samādhi-sūtra in the title and thought twice before jumping to the common and
convenient conclusion that the term merely functions as an emblem of “exotic”
origin, simply based on preoccupation with national traditions.

Bibliography
Primary sources
Fo shuo guanfo sanmei hai jing 佛說觀佛三昧海經. Buddhabhadra 佛陀跋陀羅.
T 15, no. 643.
Lidai sanbao ji 歷代三寶紀. Fei Changfang 費長房. T. 49, no. 2034.
Zuochan sanmei jing 坐禪三昧經 by Kumārajīva 鳩摩羅什. T. 15, no. 614.

Secondary sources
Buswell, Robert E. ed. Chinese Buddhist Apocrypha. Honolulu: University of
Hawaii Press, 1990.
Chen, Frederick Shih-Chung. The Transformation of Concepts of Bureaucratization
of the Other World in Early Medieval China: From Buddhist Perspectives.
DPhil diss., University of Oxford, 2010.
Chen, Frederick Shih-Chung. “Who Are the Eight Kings in the Samādhi-Sūtra of
Liberation through Purification? Otherworld Bureaucrats in India and China.”
Asia Major, 3rd ser., 26, no. 1 (2013): 55 – 78.
Harvey, Peter. “The Symbolism of the Early Stūpa.” in The Journal of the
International Association of Buddhist Studies, vol. 7, no. 2, 1984, pp. 67 – 93.
Makita Teiryō 牧田諦亮. Gikyō kenkyū 疑經研究. Kyoto: Kyōto Daigaku Jinbun
Kagaku Kenkyūjo, 1976.
Makita Tairyō 牧田諦亮 ed. in chief & Ochiai Toshinori 落合俊典 managing ed.
Nanatsudera koitsu kyōten kenkyū sōsho chūgoku senjutsu kyōten (sono2) 七寺
古逸經典研究叢書 中國撰述經典(其之二). Tokyo: Daitō Shuppansha, 1996.
Ochiai Toshinori 落合俊典 ed. The manuscripts of Nanatsu-dera. Kyoto: Italian
School of East Asia Studies, 1991.

16
Ziegler (2001), pp. 205 – 237.

202
 Book reviews

Radich, Michael. “Fei Changfang’s Treatment of Sengyou’s Anonymous Texts.”

Journal of the American Oriental Society 139.4 (2019): 819 – 841.
Saitō Takanobu 齊藤隆信. “Jōdo sanmaikyō no kenkyū: Anrakushū to Kannen
hōmon no baai 『浄度三昧経』の研究―『安楽集』と『観念法門』の場
合.” Bukkyō Daigaku Sōgō Kenkyūjo kiyō 佛教大学総合研究所紀要, no. 3,
1996, pp. 208 – 229.
Schipper, Kristofer. “The Inner World of the Lao-Tzu Chung – Ching.” In Huang,
Chun-Chieh and Zürcher, Erik ed. Time and Space in Chinese Culture. Leiden:
Brill, 1995, pp.114 – 131.
Yao Changshou 姚長壽. “Jingdu sanmei jing yu rentianjiao 淨度三昧經與人天教.”
in the Zhonghua foxue xuebao 中華佛學學報, no. 12, Oct. 1999, pp. 79 – 95.
Ziegler, Harumi Hirano. The Sinification of Buddhism as Found in an Early Chinese
Indigenous Sūtra: A Study and Translation of the Fo-shuo Ching-tu Sab-
mei Ching (Samādhi-Sūtra of Liberation through Purification). PhD. Thesis.
University of California, Los Angeles, 2001.

203