Buddhist Education in Pali and Sanskrit Schools
By E. J. Thomas
The Indian Historical Quarterly, Narendra Nath Law ed.
vol 2:3, 1926.09, p.495-508
p. 495
The subject of Buddhist education is bound up
with several still unsolved problems, but it is
posibble to limit the subject by marking off some of
those questions on which scholars are still much
divided. One of these problems is the question of the
locality or localities where those schools arose that
established different forms of the writings held to
be the word of Buddha. The most accessible of the
works of these schools are the Pali Canon, and
Sanskrit works which contain Mahayana works as well
as works of Hinayana schools closely related to the
Pali tradition.
There is an article on Buddhist education in
Hastings' Encyclopaedia of Religion and Ethics, but
for the earlier period it confines itself entirely to
the reports of the Chinese pilgrims, that is to say,
it is entirely silent about the thousand years after
the death of Buddha, during which all the various
forms of the Canon had become fixed, and when the
education and instruction described by the Chinese
pilgrims had been established for centuries. Yet
there is considerable evidence both external and
internal to show what the educational methods were.
We do not need to ask how much the earliest
Buddhism borrowed from other schools. Windisch's
article on Brahmin influence on Buddhism shows how
little is really known about the actual movements in
the earliest period.(1) Windisch points out that
brahmins who entered the Order would bring their
knowledge and literary practices with them. Our
present question is what this knowledge and literary
practice was after it had become assimilated and
established in Buddhist institutions.
---------------------------------
1. In Aufsatee E. Kuhn gewidnet, Munchen, 1916
p. 496
Besides the Pali Canon a considerable body of
literature in Sanskrit of several schools is known.
Most of it has been described in the catalogues of
Rajendralal Mitra(1) and Bendall,(2) and the most
important parts have been published. Works that
survive only in Tibetan and Chinese are also now
becoming more accessible. The earlisest stage of
literary activity may be called that of
systematisation. It must be mentioned here, that
current views as to its significance are too
divergent to make it possible to say anything that
may claim to be final.
There is a view still current in the West, which
supposes that the orthodox Buddhist holds the
Scriptures to exist now in the form in which they
were uttered by Buddha, and as recited at the first
Council. The Buddhist accounts of the Councils may
not harmonise with the demands of modern historical
criticism, but they contain nothing so unhistorical
as that. Buddhaghosa knew as well as we do that the
Canon contains much that is not the direct word of
Buddha. He expressly refers to that which was recited
and that which was not recited at ther first
Council.(3) Throughout the commentaries we find notes
on passages that are said to have been added by one
of the councils. Not only have we Suttas that are
said to have been given by disciples after the death
of Buddha, but Buddhaghosa quotes a verse which says
that out of 84,000 suttas 2,000 were uttered by
bhikkhus.(4) The whole of the Niddesa is attributed
by the commentator thereon to Sariputta.(5)
On the other hand there has often been a less
excusable
---------------------
1. Sanskrit Buddhist Literature of Nepal, Calcutta,
1882.
2. Catalogue of the Buddhist Sanskrit Manuscripts
University Library, Cambridge. Cambridge, 1883.
3. Vin. Com., i, 18.
4. Ib, i, 29.
5. The commentator on Th. I, 527 quotes Nd. I, 143
(Bhagi va Bhagava) and attributes it to Sariputta
(Dhammasenapati).
p. 497
uncritical attitude among Western scholars, against
which Mrs. Rhys Davids has recently made a vigorous
protest.
"When believers in the East and. historians in
the West will come up out of the traditional
attitude, when we shall nob bear church-editing
called Buddhavacanam and thought of as
Gotama-vacanam-when we shall no more read: `The
Buddha laid down this and denied that,' but 'the
Buddhist church did so'--then we shall at last be fit
to try to pulldown super-structure and seek for the
man."(1) The fact of this editing, which is
recognised both by Buddhist commentators and moderm
critics, implies a staffs of literary activity, of
which we know nothing as to actual details. Not only
are there the variously classified compilations of
the Anguttara and Samyutta, but the Digha and
Majjhima show much elaboration also. The former is in
three vaggas, and the first vagga, although it deals
with such various subjects as the sixty-two heresies,
caste, sacrifice, brahmin ritual, and miraculous
powers, has been given an appearance of uniformity by
the insertion in each sutta of the document known as
the silas. The Majjhima is classified in much more
detail and with more reference to the subject-matter
in fifteen vaggas. The whole of the Pali Canon in
fact shows evidence of the same careful
classification.
What this stage of Buddhist study really implied
cannot be properly answered until we know more about
the corresponding arrangements of those forms of the
Canon belonging to contemporary schools that are
extent in the Chinese. Very divergent views are at
present hold, as by Prof. Keith and Prof R. O.
Franke.(2) There can be little doubt that the system
of arrangement is earlier than the recording of the
Canon in writing, and that the chief motive was to
serve as a help to the memory. We find examples of
---------------
1 Majjhima Index, p, vi.
2 Keith, Buddhist Philosophy; Franke, Introd. to his
translation of the Digha.
p. 498
commentary already incorporated in the suttas, but
the first distinct evidence of material intended for
definite instruction tion is found in the Niddesa.
Much of the matter of this work is also found in
Abhidharma works and in the, verbal commentary of the
Vinaya, and it will be convenient to take the Niddesa
first.
As is well known this work is a commentary on the
fourth and fifth sections of the Sutta-nipata,
together with a commentary of the same nature on the
Khaggavisana sutta, which is found in the first
section, The matter of which it consists call be
divided into three types:
(1) Portions of doctrinal commentary on important
words in a style similar to the portions of
commentary occasionally found in the suttas. The
matterr and often the language is drawn from the
suttas, and in addition illustrative passages from
the suttas are frequently quoted direct, and in the
case of prose quotations regularly introduced by the
words, vuttam pi h'etam Bhagavata. Verse quotations,
which sometimes appear to be non-canonical, are more
frequently adduced without any mention of the source.
In the case.of verse 844, the Niddesa simply adopts
as its commentary a whole sutta from S. iii, 9, which
consists of a commentary on that verse.
(2) Concise definitions of individual words, such
as, sappo vuccati ahi, asnam vuccati yattha
nisidanti. The matter of this portion sometimes
corresponds with such definitions in the verbal
commentary of the Vinaya.
(3) It is in the third type that the most
characteristic feature of the Niddesa is sees. This
consists of lists of synonyms of the word commented
on. Such lists are not used to explain the meaning of
a word in a particular context. They are repeated in
the same form wherever the word occurs, and were
evidently intended to be learnt in the same may as
the more modern kosa. In the case of the verbs the
synonyms often consist of all the possible compounds
of the same verb, yutto, payutto, ayutto, samayutto,
sampayutto; vedhati, pavedhati, sampavedhati, In the
case of important
p. 499
words all the various synonyms, evidently drawn from
the scriptures, are given in long lists. The result
is that some of the synonyms are often unintelligible
apart from the context in the sutta from which they
are taken. In a long list of synonyms of tanha (Nd.
I, 8) sibbini 'sewer' occurs, and the reason for this
is seen from A. iii, 399; Sn. 1040, where it is an
epithet of tanha, and from where it has no doubt;
been taken. Among the synonyms of sada (Nd. I, 18)
occurs avici. This is evidently clue to analysing it
as a-vici "without a wave', and hence 'continuous.'
Vammiko as one of the synonyms of kaya comes from the
parable of the ant-hill in M. i, 142.
Much of this is also found in the Abhidhamma
books, but in the Niddesa it is used as general
matter applied to passages for which it was not;
immediately intended. Some of the correspondences are
as follows: chando Nd. I. 2=Dhs. 1097, Vbh. 374;
tassa Nd. I, 2=Vbb, 393; mane, piti, Nd. I, 3=: Dhs.
6, 9; tanha Nd. I, 8 =Dhs. 1059; sati Nd. T, 10 =
Pug. 25; macchariya Nd. I, 37= Dhs. 1122, Pug. 19
;panna Nd. I, 44, 77 =Pug. 25, Dhs. 16; maya Nd. I,
79=Pug. 19; gantha Nd. I, 98 =Dhs. 1135; kodha Nd. I,
215-aghata Dhs. 1154, cf. Pug 18;satheyya Nd. I,
395=Pug, 19; thiti Nd. I, 501= Dhs. 10.
Minor differences occur, and in some easer quite
different treatment, cf. puthujjana Nd. I, 146 and
Pug. 12. There is a triple division of puccha Nd. 339
with no reference to the fourfold division of D. iii,
229, Dhs. Mahavyut. 85.
The verbal commentary on the Vinaya is less
developed than either the Niddesa or the Abhidhamma
works. It is occupied with explaining words
concisely, in a given context without lists of
synonyms.
This shows a system for learning the vocabulary
of the Canon, and for explaining archaic forms, but
no further grammatical teaching occurs apart from the
description of certain terms as particles. Addha ti
ekamsavacanam (with seven other synonms for
ekamsavacanam; na ti patikkhepo. Even
p. 500
such a sandhi as iccayasma is not resolved into iti,
but icca is separated and explained like all such
partticles as padasandhi, padasamsaggo, padaparipuri,
akkharasamavayo, viyanjanasilitthata.
In the Niddesa we thus have direct evidence of a
general system of instruction applied to a definite
work, consisting of interpretation, doctrinal
teaching and in tile verbal expositions the
beginnings of grammar. The Abhidhamma boobs and
related works like the Patisambhidammagga give other
traces of its existence. It appears to be this system
which. is expressly referred to in the Niddesa (I,
234) and other places as the four kinds of analysis
(patisambhida): the analysis of meanings (attha), of
conditions (dhamma) , of grammatical analysis
(nirutti), and clearness of insight (patibhana).(1)
The Nirutti of the Niddesa is of the kind that we
should expect to exist when Pali was a living
language. All the grammatical analysis that was
required was a knowledge of those words in the
Scriptures that heel become obsolete, and the
explanation of unusual grammatical forms by means of
the current expression, The method was not confined
to the Pali tradition, as we find the same four
divisions called pratisamvida in the Mahavastu (iii.
321) and pratisamvit in the Mahavyutpatti (13), and
this nirutti method has reacted on the style of the
later sutras.
The practice of' learning off strings of synonyms
might be expected to influence the style of those who
passed through such a course of instruction. We
appear to find an instance of it when Buddhaghosa(2)
thus describes an earthquake: ayam mahapathavi...
kampi samkampi sampakampi sampavedhi.
---------------------
1. I They are also found ill a sutta (A, ii, 160)
which is attributed like the Niddesa itself to
Sariputta. It probably belongs to the same stratum
of scholarship. The Abhidhamma statement of
patisambhida in Vbh. ch. xi is discussed by Mrs,
Rhys Davids in the Points of Controversy, pp. 377
ff.; cf. Ps. i, 88.
2. Vin. corn. I, 30.
p. 501
Here we have the same series of compounds as we find
repeatedly in the Niddesa, and Buddhaghosa is only
using an earlier phraseology. It appears not only in
tile later commentators but also in Sanskrit and
especially Mahayana works. In several of these a
standing description of an earthquake occurs. The
synonymous verbs kamp- vidh- cal- ksubh- are given,
followed by ran and garj and each is expanded into
compounds with pra and sampra.(1) If this stood
alone, it might be taken merely as the verbosity of a
particular author, but there are other instances, and
they often correspond with series of synonyms in the
Niddesa. The Niddesa has sakkaroti garukaroti maneti
pujeti. The Avadana-sataka p. 8 exactly corresponding
has satkrto gurukrto manitah pujitah. The Mahavastu
has the same adding arcitah In Mahayana works this is
expanded, being preceded by puraskrtah and followed
by arcitah and apacayitah (Saddh-pund. 5; Karunapund.
2). Similarly the latter sutra has tile series
harsaniya tosaniya prasadaniya avalokaniya
prahladaniya manojna. All the synonyms that we find
need not have arisen from the method that we find in
the Niddesa. Some of them were doubtless incorporated
from old texts, but the practice of compiling such
lists is certain from what we find in the Niddesa,
and the correspondences in the lists makes it
probable that there was intercourse between different
schools and common methods of teaching.(2)
Among Mahayana works there are two compendiums
which have some relation to the Niddesa. The
Dharmasamgraha is a compilation of terms, but it is
mainly doctrinal. The Mahavyutpatti was evidently
intended for grammatical instruction as well, It
gives the complete declension of vrksa (210) ,
epithets of Buddha and Boddhisattvas and their
qualities, synonyms of the teaching and names of
sections (66), epithets
-------------------------
1. Lal. v, 449; Karuna
2. It may be noticed that the term nirdesa is
frequent in Mahayana sutras.
p. 502
of Nirvana (95), terms of salutation (97), synonyms
of tusta and raudra (145-6), synonyms of sattva (207)
almost corresponding with Nd. It 12, miscellaneous
adjectives (223), a long list of all the stock words
and Phrases that occur in a sutra (241), and a list
of diseases (284), which only partially corresponds
with that in Nd. I, 17. Much of this is nirukti in
the sense of the Pali nirutti.
At present there is no general agreement as to
where the Pali language as me know it developed. It
is usually agreed that the oldest works in verse show
traces of having been composed in a different
dialect. The natural conclusion is that tile
canonical works were preserved in a monastery or
closely related group of monasteries, where a
different dialect was spoken, and. where the original
dialect of the texts was entirely effaced, except so
far as metrical facts compelled the preservation of
special forms. Doubtless this Pali language that we
know was at first a living and spoken language, but
in the course of centuries, say from the time of
Asoka to the end of the second century A. D., it
would come to be as much a learned language as
Sanskrit. The fact of the Niddesa itself seems to
show that this Pali was then a current language, but
that nirutti, grammatical analysis, was becoming
necessary for the interpretation of the texts.
Nothing profitable can be said about the earliest
date at which the Niddesa may he put. Any such theory
mould only tell us that a work of that name existed,
hut the occurrence of a geographical term in any
particular passage would only allow us to infer the
date of that passage. We can see from its different
forms and readings that it underwent changes and
received additions, and in the case of a work used
continuously for instruction this would be
inevitable, Its application of Abhidhamma material
for a general purpose seems to show that it is Inter
than the Abhidhamma books, and its reference to one
of the Alexandrias (Allasanda) founded after the
G;reek invasion, to Bengal, Burma, and Java, would
suggest that it became established and wax used ax a
text-
p. 503
book during the first two centuries B. C. It has no
reference to tha paramita, and although it gives the
37 constituents of enlightenment, it does not use the
term bodhipakkhikadhamma.
In the case of the literature of the Sanskrit
schools we can draw further information concerning
the materials and methods of education, The works are
much later than the Niddesa. They refer frequently to
writing, and the mention in the Mahavyutpatti of
Kaniska and Asvaghosa puts this work Inter than the
first century A. D., but it is probably two or three
centuries later than this, as it contains evidence of
contact with Greek astrology. The dates usually
assigned to the chief tests range from the second to
the seventh century, The four methods of analysis
with nirukti are preserved, but we may infer from the
fact that the language was Sanskrit and from the
production of' a kavya like the Buddhacarita in the
first century A. D. that grammar was a fully
developed study.
Wherever the tests of this literature originated,
we can at least assume from the accounts of the
Chinese pilgrims that down to the seventh century
Magadha was the chief district of their study.(1) Mr.
J. N. Samaddar in his interesting account of the
monasteries of Nalanda, , Vikramasila (east of
Bhagalpur) and Odandapura (Bihar) calls them
universities, and draws several remarkable parallels
with these modern institutions. The proposing of hard
questions by the keeper of the gate at Nalanda
becomes matriculation, The teaching is said to have
been both tutorial and professorial. The Master of
the Law is taken to be the Vice-Chancellor, and the
writing up of the names of famous scholars over the
gates is compared to the granting of diplomas.
This is what is inferred from Hiuen Tsiang, but
it is
----------------------
1. The vihara of Vikramasila is mentioned in the
colophon of one Ms, (Mitra, p, 229), and according
to Mr, Samaddar Nalanda occurs (Glories of
Magadha, p, 104 ff.).
p. 504
I-tsing who describes the actual studies.(1) From
Prof. Takakusu's account it appears that grammar was
based entirely on works of brahminical schools, the
Sutras of Panini, the Kasika of Jayaditya, the
Mahabhasya of Patanjali, and three works by
Bhartrhari. It is not clear from this whether the
Sutras were those of Panini in their original form,
but Panini was certainly known to the Buddhists.(2)
He is mentioned in the Lankavatara-sutra, and
Taranatha in his history tells a wonderful story of
his acquirement of grammatical science. The chief
form however in which the Paninean grammar was
studied by these schools appears to have been
Candragomin's Candravyakarana, which is put at the
beginning of the seventh century. This is the only
grammar which is mentioned in Bendall's list along
with commentaries on it, chiefly that of Anandadatta,
and in the Tanjur the matical works as given by Csoma
are either Candragomin's work or others still later.
The Niddesa also shows the beginnings of
lexicography, and its continuation appears in the
Dharmasamgraha and Mahavyutpatti. Its full
development is seen in the Amarakosa of Amarasimha,
who was a Buddhist himself: It is not mentioned by
I-tsing, and Winternitz puts it between the sixth and
eight centuries, There are several copies of it in
Bendall's list, and it is also in the Tanjur.
Apart from philosophy, which formed part of' the
doctrinal
--------------------
1. Ch. 34. ed. Takakusu.
2. Dr. B. C. Law has pointed out in Buddhaghosa a
passage reminscent of Panini, V, 2, 93. It may be
asked whether this comes directly from Panini or
from some adaptation, but it certainly corresponds
much more closely with Parini than with the
corresponding sutra and vrtti of Candragomin, IV.
2, 97. The Pall grammar of Kaccayana is later than
Buddhaghosa and belongs to the literature of
Ceylon. Later works, says Geiger, follow the
models of Sanskrit grammar and lexicography
slavishly, and apply their system mechanically to
Pall. Geiger, Pali Lit. Pend Sprache; Franke,
Gesch. und. Krit, der einheimt. Pali-grammatik.
p. 505
teaching, two important secular subjects are medicine
and astronomy. That medicine must have been studied
early we know from the Vinaya, as the sixth chapter
of the Mahavagga is devoted to medicines and surgery,
I-tsing mentions cibitsavidya, but there is nothing
in the surviving literature to indicate that if ever
became an independent study. He does not mention
jyotisa among the vidyas, and it is clear that as
astrology was an integral part of astronomy and the
chief motive of its study, the latter science could
not be expected to flourish so long as Buddhism
forbade interpretation of the stars (e, g. Sn. 927
and Nd. 1, 381).1 It came in when the practice of
astrology revived. The only astronomical work
mentioned in Mitra's list is a tika on the Jain work
Suryaprajnapti. Among the Buddhist fragments from
Central Asia, edited by Hoernle is an astrological
work which shows that it is based on Greek astrology,
and that Buddhism had come to adopt astrological
practices. There is also evidence ef Greek. influeace
in the list of the nine planets in Mahavyut. 164. The
first seven of them beginning with Aditya are in the
order of the days of the week, and this order, which
depends on an elaborate assignment of a planet to
each of the 24 hours of the day, came from Greece.(2)
It is certain that the monasteries of Magadha
mere the chief places where this Sanskrit literature
was studied, and probably also the region of its
origin. It represents the product of several schools
and shows certain relations with Pali works. But the
views of scholars concerning the district where Pali,
as we know it, originated are so divergent that
--------------------
1. The knowledge of astronomy among the Buddhists
has been treated in the writer's article Sun,
Moon, and Stars (Buddhist) in Hastings' Ency. of
Rel. and Ethics.
2 The Ptolemaic order of the planets is Saturn,
Jupiter, Mars, Sun, Venus, Mercury, Moon, The lord
of the first hour of Sunday is the Sun, of the
second hour Venus, and so on. This makes the Moon
the lord of the first hour of Monday, and so on
throughout the week.
p. 506
it is impossible to do more than draw attention to a
problem still in need of solution. It is the question
not of the original language of Buddha, and his first
disciples, but of the Pali of Ceylon. The Pali of the
time of Buddhaghosa was no living language, except in
so far as it may have been learnt and used within
each monastery. The commentaries of that time were
translations and adaptations in Pali of there already
existing in Singhalese. The traces of an earlier
dialect surviving in the Canon may be survivals of
the dialect in which it existed when it was taken to
Ceylon. But it is the Pali as used by Buddhaghosa
which the Singhalese tradition calls Magadhi.(1) The
view that Pali really was the language of Magadha is
generally rejected, and various attempts to fix the
district in India where P ii developed have been made
on the assumption that it must have been somewhere
else than Magadha.
Oldenberg sought it in South India, probably in
the kingdoms of the Andhras or Kalingas.(2) According
to Prof. R. O. Franke its original home was in a
district somewhere in the middle to the west of the
vindhya mountains. "Accordingly it is not impossible,
though naturally a pure supposition, that the city of
Ujjen, which evidently had become a centre of culture
comparatively early, also formed the centre of the
dialect-area of literary Pali."(3) This was also the
view of Westergaard and E. Kuhn, which Oldenberg
expressly reject- ed. Sir George Grierson holds that
"we have a strong reason
------------------
1. Buddhaghosa was told to go and translate the
Atthakatha ainto Magadhanam nirutti, Mhvs. p, 251
(Tumour), quoted by Dr, R. C. Law, The life and
work of Buddhaghosa, p. 75.
2. Vinaya, Introd., p.I.
3. Pali und Sanskrit, p. 138. By literary Pall Dr,
Franke merely means the Pali as generally
understood, The reason is that he uses the general
term Pali to include the spoken Aryan languages of
the whole of sub-Himalayan India and Ceylon; ib, p
vi. There is nothing to be said against the
terminology except that it has not won general
acceptance, and that scholars still call these
languages Prakrit.
p. 507
for concluding that literary Pali is the literary
form of the Magadhi language, the then koine of
India, as it was spoken and as it was used as a
medium of literary instruction in the Taksasila
University.(1) The conclusion of Rhys Davids was that
"Buddhism born in Nepal, received the garb in which
we now know it in Avanti, in the far West of India,"
and he held that this was nearer to the other view
"so often put forward as convenient that Buddhism
arose in Magadha and that its original tongue was
Magadhi."(2) These are the results of thirty years of
research.
Geiger has taken the unpopular course of holding
that the tradition of the Chronicles and commentaries
is the true one, and that what they call Magadhi is
Magadhi.(3) Oldenberg's statement that "it is certain
that the Pali language is not the Magadhi language",
merely means that it is not the language of the
Asokan inscriptions. There is not slightest reason
why the texts of the Canon should have been adapted
to the spoken language of the time of Asoka. It is
far more likely that the dialect of the texts had
already begun to form a sacred language,and we know
that there wax a rule in the Vinaya saying that the
monks were to learn the word of Buddha in its own
grammar or dialect, anujanami bhikkhave skayaa
niruttiya buddhavacanam pariyapunitum, and
Buddhaghosa understands this as meaning in the
Magadhi language. It is true that this sentence has
been understood against both grammar and tradition in
a quite opposite sense, but this does not now need
discussion.
The latest attempt to solve the question has been
made by Dr. M. Walleser,(4) who also decides for
Magadha, but it cannot be said that within the space
of twenty-four pages he
--------------------
1. Commemorative Essays presented to Sir. R. G.
Bhandarkar, p. 123.
2. Cambridge History of India, I. 187.
3.Pall Litteratur and Sprache, p. 3.
4 Sprache undJ Heimat des Pali-Kanons. Heidelberg,
1924.
p. 508
has done justice to the arguments of his
predecessors. He further prejudices his own cars by
assertiug that Pali means not the body of sacred
texts but the language in which they were composed.
However, his evidence for the phrase paliIhasa rests
merely on Childers, and ignores such decisive
passages as that of the Mahavamsa referred to above,
and thus translated by Dr. B. C. Law: "The Pali (text
of the Tripitaka) only (palimattam) has been brought
over here. The Ceylon commentary is current among the
people of Ceylon. Please go there and study it, and
then translate it into Magadhi (magadhanam niruttiya
parivattehi)."(1) But Dr. Walleser has certainly made
the claims of Magadha more probable, and it may be
hoped that deeper investigation of the geographical
question will lead to the establishing of further
links in the history of Buddhist scholarship.
-------------------------
1. Life and Work of Buddhaghosa. P.74.