Note on the probable age of the Dialogues


Note on this Version






(The SŃlas, 3-26. )


2. SÔMA÷÷A-PHALA SUTTANTA. INTRODUCTION (Index to the paragraphs repeated in the other Suttantas, 57-59. )








(The Arahat the true Brahaman. )




(The irony in this text, 160; Doctrine of sacrifice, 164; LokÓyata, 166. )


[\q viii/] TEXT


(The Indeterminates; Buddhist Agnosticism, 186; The Sambodhi, 190; Names in the texts, 193. )






(Method of the Dialogues, 206 TÓpasa and Bhikshu, ascetic and wandering mendicant, 208 Indian religieux in the Buddha's time, 220)












(Iddhi, 272; Buddhist Idealism, 274. )




(Ethics of Teaching)




(Union with God)





THE Dialogues of the Buddha, constituting, in the PÓli text, the DŃgha and Majjhima NikÓyas, contain a full exposition of what the early Buddhists considered the teaching of the Buddha to have been. Incidentally they contain a large number of references to the social, political, and religious condition of India at the time when they were put together. We do not know for certain what that time exactly was. But every day is adding to the number of facts on which an approximate estimate of the date may be based. And the ascertained facts are already sufficient to give us a fair working hypothesis. In the first place the numerous details and comparative tables given in the Introduction to my translation of the Milinda show without a doubt that practically the whole of the PÓli Pi˝akas were known, and regarded as final authority, at the time and place when that work was composed. The geographical details given on pp. xliii, xliv tend to show that the work was composed in the extreme North-West of India. There are two Chinese works, translations of Indian books taken to China from the North of India, which contain, in different recessions, the introduction [\q x/] and the opening chapters of the Milinda [1] For the reasons adduced (loco citato) it is evident that the work must have been composed at or about the time of the Christian era. Whether (as M. Sylvain Levy thinks) it is an enlarged work built up on the foundation of the Indian original of the Chinese books; or whether (as I am inclined to think) that original is derived from our Milinda, there is still one conclusion that must be drawnŮthe NikÓyas, nearly if not quite as we now have them in the PÓli, were known at a very early date in the North of India. Then again, the KathÓ Vatthu (according to the views prevalent, at the end of the fourth century A. D., at KÓ§cipura in South India, and at AnurÓdhapura in Ceylon; and recorded, therefore, in their commentaries, by DhammapÓla and Buddhaghosa) was composed, in the form in which we now have it, by Tissa, the son of MoggalŃ, in the middle of the third century B. C., at the court of Asoka, at PÓ˝aliputta, the modern Patna, in the North of India. It is a recognised rule of evidence in the courts of law that, if an entry be found in the books kept by a man in the ordinary course of his trade, which entry speaks against himself, then that entry is especially worthy of credence. Now at the time when they made this entry about Tissa's authorship of the KathÓ Vatthu the commentators believed, and it was an accepted tenet of those among whom they mixedŮjust as it was, mutatis mutandis, among the theologians in Europe, at the corresponding date in the history of their faithŮthat the whole of the canon was the word of the Buddha. They also held that it had been actually recited, at the Council of RÓjagaha, immediately after his decease. It is, I venture to submit, absolutely impossible, under these circumstances, that the commentators can have invented this information [\q xi/] about Tissa and the KathÓ Vatthu. They found it in the records on which their works are based. They dared not alter it. The best they could do was to try to explain. it away. And this they did by a story, evidently Legendary, attributing the first scheming out of the book to the Buddha. But they felt compelled to hand on, as they found it, the record of Tissa's authorship. And this deserves, on the ground that it is evidence against themselves, to have great weight attached to it. The text of the KathÓ Vatthu now lies before us in a scholarly edition, prepared for the PÓli Text Society by Mr. Arnold C. Taylor. It purports to be a refutation by Tissa Of 250 erroneous opinions held by Buddhists belonging to schools of thought different from his own. We have, from other sources, a considerable number of data as regards the different schools of thought among BuddhistsŮoften erroneously called `the Eighteen Sects'. [2] We are beginning to know something about the historical development of Buddhism, and to be familiar with what sort of questions are likely to have arisen. We are beginning to know something of the growth of the language, of the different PÓli styles. In all these respects the KathÓ Vatthu fits in with what we should expect as possible, and probable, in the time of Asoka, and in the North of India. Now the discussions as carried on in the KathÓ Vatthu are for the most part, and on both sides, an appeal to authority. And to what authority ? Without any exception as yet discovered, to the Pi˝akas, and as we now have them, in PÓli. Thus on p. 339 the appeal is to the passage translated below, on p. 278, ? 6 ; and it is quite evident that the quotation is from our Suttanta, and not from any other passage where [\q xii/] the same words might occur, as the very name of the Suttanta, the Kevaddha (with a difference of reading found also in our MSS.), is given. The following are other instances of quotations:

KathÓThe Vatthu. NikÓyas. page. 344 =A. II , 50. 345 =S. I , 33. 345 =A. 1 I, 54. 347 =Kh. P. VI 1, 6, 7. 348 =A. III, 43. 351 =Kh. P. VIII, 9. 369 =M. 1, 85, 92, &C. 404 =M. 1, 4. 413 =S. IV, 362. 426 =D. I, 70. 440 =S. I, 3 3. 457 =D. (M. P. S. 23). 457 =A. II, 172. 459 =M. I, 94. 2. 481=D. I, 83, 84. 483 =D. I, 84. 484 =A. II, 126, 494 =S.1, 206 J. IV, 496. 505 =M. I, 490. 506 =M. I, 485 = S. IV, 393 (nearly). 513=A. I, 197, 522 =M. I, 389. 525 =Dhp. 164. 528 =M. I, 447. 549=S. N. 227 = Kh VI, 6. 554=S. 1, 233. 554=Vim. V. XXXIV, 5-2 7. 565 =D. I, 156. 588, 9 =P. P. pp. 71, 72. 591 =M. I, 169. 597, 8 =A. I, 141, 2. 602=Dh. C. P. Sutta, ?? 9-23.

There are many more quotations from the older Pi˝aka books in the KathÓ Vatthu, about three or four times as many as are contained in this list. But this is enough to show that, at the time when the KathÓ Vatthu was composed, all the Five NikÓyas were extant; and were considered to be final authorities in any question that was being discussed. They must themselves, therefore, be considerably older. Thirdly, Hofrath Bühler and Dr. Hultsch have called attention [3] to the fact that in inscriptions of the third century B. C. we find, as descriptions of donors to the dÓgabas, the expressions dhammakathika, pe˝akŃ, suttantika, suttantakinŃ, and pa§ca-nekÓyika. [\q xiii/] The Dhamma, the Pi˝akas, the Suttantas, and the five NikÓyas must have existed for some time before the brethren and sisters could be described as preachers of the Dhamma, as reciters of the Pi˝aka, and as guardians of the Suttantas or of the NikÓyas (which were not yet written, and were only kept alive in the memory of living men and women). Simple as they seem, the exact force of these technical designations is not, as yet, determined. Dr. K. Neumann thinks that Pe˝akŃ does not mean `knowing the Pi˝akas,' but `knowing the Pi˝aka,' that is, the NikÓyasŮa single Pi˝aka, in the sense of the Dhamma, having been known before the expression `the Pi˝akas' came into use. [4] As he points out, the title of the old work Pe˝akopadesa, which is an exposition, not of the three Pi˝akas, but only of the NikÓyas, supports his view. So again the Dialogues are the only parts or passages of the canonical books called, in our MSS., suttantas. Was then a suttantika one who knew precisely the Dialogues by heart ? This was no doubt the earliest use of the term. But it should be recollected that the KathÓ vatthu, of about the same date, uses the word suttanta also for passages from other parts of the scriptures. However this may be, the terms are conclusive proof of the existence, some considerable time before the date of the inscriptions, of a Buddhist literature called either a Pi˝aka or the Pi˝akas, containing suttantas, and divided into Five NikÓyas. Fourthly, on Asoka's Bhabra Edict he recommends to the communities of the brethren and sisters of the Order, and to the lay disciples of either sex, frequently to hear and to meditate upon seven selected passages. These are as follows:

1. Vinaya-samukkaŘsa.
2. Ariya-vÓsÓni from the DŃgha (SaŘgŃti Suttanta).
3. AnÓgata-bhayÓni from the A´guttara III, 105- 108.
4. Muni-gÓthÓ from the Sutta NipÓta 206-220.
5. [\q xiv/] Moneyya Sutta from the Iti Vuttaka 67=A. I, 272.
6. Upatissa-pasina.
7. RÓhulovÓda = RÓhulovÓda Suttanta (M. I, 414-420).

Of these passages Nos. 1 and 6 have not yet been satisfactorily identified. The others may be regarded certain, for the reasons I have set out elsewhere [5] No. 2 also occurs in the tenth book of the A´guttara. It is clear that in Asoka's time there was acknowledged to be an authoritative literature, probably a collection of books, containing what was then believed to be the words of the Buddha: and that it comprised passages already known by the titles given in his Edict. Five out of the seven having been found in the published portions of what we now call the Pi˝akas, and in the portion of them called the Five NikÓyas, raises the presumption that when the now unpublished portions are printed the other two will also, probably, be identified. We have no evidence that any other Buddhist literature was in existence at that date. What is perhaps still more important is the point to which M. Senart [6] has called attention, and supported by numerous details: the very clear analogy between the general tone and the principal points of the moral teaching, on the one hand of the Asoka edicts as a whole, and on the other of the Dhammapada, an anthology of edifying verses taken, in great part, from the Five NikÓyas, The particular verses selected by M. Senart, as being especially characteristic of Asoka's ideas, include extracts from each of the Five. Fifthly, the four great NikÓyas contain a number of stock passages, which are constantly recurring, and in which some ethical state is set out or described. Many of these are also found in the prose passages [\q xv/] of the various books collected together in the Fifth, the Khuddaka NikÓya. A number of them are found in each of the thirteen Suttantas translated in this volume. There is great probability that such passages already existed, as ethical sayings or teachings, not only before the NikÓyas were put together, but even before the Suttantas were put together. There are also entire episodes, containing not only ethical teaching, but names of persons and places and accounts, of events, which are found, in identical terms, at two or more places. These should be distinguished from the last. But they are also probably older than our existing texts. Most of the parallel passages, found in both PÓli and Sanskrit Buddhist texts, come under one or other of these two divisions. Sixthly, the SaŘyutta NikÓya (III, 13) quotes one Suttanta in the Dialogues by name; and both the SaŘyutta and the A´guttara NikÓyas quote, by name and chapter, certain poems now found only in a particular chapter of the Sutta NipÓta. This Suttanta, and these poems, must therefore be older, and older in their present arrangement, than the final settlement of, the text of these two NikÓyas. Seventhly, several of the Dialogues purport to relate conversations that took place between people, contemporaries with the Buddha, but after the Buddha's death. One Sutta in the A´guttara is based on the death of the wife of Mu§ˇa, king of Magadha, who began to reign about forty years after the death of the Buddha. There is no reason at all to suspect an interpolation. It follows that, not only the Sutta itself, but the date of the compilation of the A´guttara, must be subsequent to that event. There is a story in Peta Vatthu IV, 3, 1 about a King Pingalaka. DhammapÓla, in his commentary, informs us that this king, of whom nothing is otherwise known, lived two hundred years after the Buddha. It follows that this poem, and also the Peta Vatthu in which it is found, and also the VimÓna Vatthu, with which the Peta Vatthu really forms one whole work, [\q xvi/] are later than the date of Pingalaka. And there is no reason to believe that the commentator's date, although it is evidently only a round number, is very far wrong. These books are evidently, from their contents, the very latest compositions in all the Five NikÓyas. There is also included among the Thera GÓthÓ, another book in the Fifth NikÓya, verses said, by DhammapÓla the commentator [7], to have been composed by a thera of the time of King BindusÓra, the father of Asoka, and to have been added to the collection at the time of Asoka's Council. Eighthly, several Sanskrit Buddhist texts have now been made accessible to scholars. We know the real titles, given in the MSS. themselves, of nearly 200 more. [8] And the catalogues in which the names occur give us a considerable amount of detailed information as to their contents. No one of them is a translation, or even a recession, of any one of the twenty-seven canonical books. They are independent works; and seem to bear. to the canonical books a relation similar, in many respects, to that borne by the works of the Christian Fathers to the Bible. But though they do not reproduce any complete texts, they contain numerous verses, some whole poems, numerous sentences in prose, and some complete episodes, found in the PÓli books. And about half a dozen instances have been already found in which such passages are stated, or inferred, to be from older texts, and are quoted as authorities. Most fortunately we may hope, owing to the enlightened liberality of the Academy of St. Petersburg, and the zeal and scholarship of Professor d'Oldenbourg and his co-workers, to have a considerable number of Buddhist Sanskrit Texts in the near future. And this is just what, in the present state of our knowledge of the history of Buddhist writings, is so great a desideratum.

[\q xvii/] It is possible to construct, in accordance with these facts, a working hypothesis as to the history of the literature. It is also possible to object that the evidence drawn from the Milinda may be disregarded on the ground that there is nothing to show that that work, excepting only the elaborate and stately introduction and a few of the opening chapters, is not an impudent forgery, and a late one, concocted by some Buddhist in Ceylon. So the evidence drawn from the KathÓ Vatthu may be disregarded on the ground that there is nothing to show that that work is not an impudent forgery, and a late one, concocted by some Buddhist in Ceylon. The evidence drawn from the inscriptions may be put aside on the ground that they do not explicitly state that the Suttantas and NikÓyas to which they refer, and the passages they mention, are the same as those we now have. And the fact that the commentators point out, as peculiar, that certain passages are nearly as late, and one whole book quite as late, as Asoka, is no proof that the rest are older. It may even be maintained that the PÓli Pi˝akas are not therefore Indian books at all: that they are all Ceylon forgeries, and should be rightly called `the Southern Recension' or `the SiŘhalese Canon.' Each of these propositions, taken by itself, has the appearance of careful scruple. And a healthy and reasonable scepticism is a valuable aid to historical criticism. But can that be said of a scepticism that involves belief in things far more incredible than those it rejects? In one breath we are reminded of the scholastic dulness, the sectarian narrowness, the literary incapacity, even the senile imbecility of the Ceylon Buddhists. In the next we are asked to accept propositions implying that they were capable of forging extensive documents so well, with such historical accuracy, with so delicate a discrimination between ideas current among themselves and those held centuries before, with so great a literary skill in expressing the ancient views, that not only did [\q xviii/] they deceive their contemporaries and opponents but European scholars have not been able to point out a single discrepancy in their work [9] . It is not unreasonable to hesitate in adopting a scepticism which involves belief in so unique, and therefore so incredible, a performance. The hesitation will seem the more reasonable if we consider that to accept this literature for what it purports to beŮthat is, as North Indian [10] and for the most part pre-AsokanŮnot only involves no such absurdity, but is really just what one would apriori expect, just what the history of similar literatures elsewhere would lead one to suppose likely. The Buddha, like other Indian teachers of his time, taught by conversation. A highly educated man (according to the education current at the time), speaking constantly to men of similar education, he followed the literary habit of his time by embodying his doctrines in set phrases, sňtras, on which he enlarged on different occasions in different ways. In the absence of booksŮfor though writing was widely known, the lack of writing materials made any lengthy written books impossible [11] such sňtras were the recognised form of preserving and communicating opinion. These particular ones were not in Sanskrit, but in the ordinary conversational idiom of the day, that is to say, in a sort of PÓli.

[\q xix/] When the Buddha died these sayings were collected together by his disciples into the Four Great NikÓyas. They cannot have reached their final form till about fifty years afterwards. Other sayings and verses, most of them ascribed not to the Buddha himself, but to the disciples, were put into a supplementary NikÓya. We know of slight additions made to this NikÓya as late as the time of Asoka. And the developed doctrine found in certain short books in itŮnotably in the BuddhavaŘsa and CariyÓ Pi˝aka, and in the PetaŮ and VimÓna-VatthusŮshow that these are later than the four old NikÓyas. For a generation or two the books as originally put together were handed down by memory. And they were doubtless accompanied from the first, as they, were being taught, by a running commentary. About 100 years after the Buddha's death there was a schism in the community. Each of the two schools kept an arrangement of the canonŮstill in PÓli (or possibly some allied dialect). Sanskrit was not used for any Buddhist works till long afterwards, and never used at all, so far as we know, for the canonical books. Each of these two schools broke up, in the following centuries, into others; and several of them had their different arrangements of the canonical books, differing also no doubt in minor details. Even as late as the first century after the Christian Era, at the Council of Kanishka, these books, among many others then extant, remained the only authorities [12] . But they all, except only our present PÓli NikÓyas, have been lost in India. Of the stock passages of ethical statement, and of early episodes, used in the composition of them, and of the Suttas now extant, numerous fragments have been preserved in the HinayÓna Sanskrit texts. And some of the Suttas, and of' the separate books, as used in other schools, are represented in Chinese translations of the fourth and [\q xx/] fifth centuries A. D. A careful and detailed comparison of these remains with the PÓli NikÓyas, after the method adopted in Windisch's `MÓra und Buddha,' cannot fail to throw much light on the history, and on the method of composition, of the canonical books, which in style and method, in language and contents and tone, bear all the marks of so considerable an antiquity. Hofrath Dr. Bühler, in the last work he published, expressed the opinion that these books, as we have them in the PÓli, are good evidence, certainty for the fifth, probably for the sixth, century B. C. Subject to what has been said above, that will probably become, more and more, the accepted opinion. And it is this which gives to all they tell us, either directly or by implication, of the social, political, and religious life of India, so great a value [13] .

It is necessary, in spite of the limitations of our space, to add a few words on the method followed in this version. We talk of PÓli books. They are not books in the modern sense. They are memorial sentences intended to be learnt by heart ; and the whole style, and method of arrangement, is entirely subordinated to this primary necessity. The leading ideas in any one of our Suttantas, for instance, are expressed in short phrases not intended to convey to a European reader the argument underlying them. These are often repeated with slight variations. But neither the repetitions nor the variationsŮintroduced, and necessarily introduced, as aids to memoryŮhelp the modern reader very much. That of course was not their object. For the object they were intended [\q xxi/] to serve they are singularly well chosen, and aptly introduced. Other expedients were adopted with a similar aim. Ideas were formulated, not in logically co-ordinated sentences, but in numbered groups; and lists were drawn up such as those found in the tract called the SŃlas, and in the passer on the rejected forms of asceticism, both translated below. These groups and lists, again, must have been accompanied from the first by a running verbal commentary, given, in his own words, by the teacher to his pupils. Without such a comment they are often quite unintelligible, and always difficult. The inclusion of such memoria technica makes the Four NikÓyas strikingly different from modern treatises on ethics or psychology. As they stand they were never intended to be read. And a version in English, repeating all the repetitions, rendering each item in the lists and groups as they stand, by a single English word, without commentary, would quite fail to convey the meaning, often intrinsically interesting, always historically valuable, of these curious old documents. It is no doubt partly the result of the burden of such memoria technica, but partly also owing to the methods of exposition then current in North India, that the leading theses of each Suttanta are not worked out in the way in which we should expect to find similar theses worked out now in Europe. A proposition or two or three, are put forward, restated with slight additions or variations, and placed as it were in contrast with the contrary proposition (often at first put forward by the interlocutor). There the matter is usually left. There is no elaborate logical argument. The choice is offered to the hearer; and, of course, he usually accepts the proposition as maintained by the Buddha. The statement of this is often so curt, enigmatic, and even Ůowing not seldom simply to our ignorance, as yet, of the exact force of the technical terms usedŮso [\q xxii/] ambiguous, that a knowledge of the state of opinion on the particular point, in North India, at the time, or a comparison of other NikÓya passages on the subject, is necessary to remove the uncertainty. It would seem therefore most desirable that a scholar attempting to render these Suttantas into a European languageŮevolved in the process of expressing a very different, and often contradictory, set of conceptionsŮshould give the reasons of the faith that is in him. He should state why he holds such and such an expression to be the least inappropriate rendering: and quote parallel passages from other NikÓya texts in support of his reasons. He should explain the real significance of the thesis put forward by a statement of what, in his opinion, was the point of view from which it was put forward, the stage of opinion into which it fits, the current views it supports or controverts. In regard to technical terms, for which there can be no exact equivalent, he should give the PÓli. And in regard to the mnemonic lists and groups, each word in which is usually a crux, he should give cross-references, and wherever he ventures to differ from the Buddhist explanations, as handed down in the schools, should state the fact, and give his reasons. It is only by such discussions that we can hope to make. progress in the interpretation of the history of Buddhist and Indian thought. Bare versions are of no use to scholars, and even to the general reader they can only convey loose, inadequate, and inaccurate ideas. These considerations will, I trust, meet with the approval of my fellow workers. Each scholar would of course, in considering the limitations of his space, make a different choice as to the points he regarded most pressing to dwell upon in his commentary, as to the points he would leave to explain themselves. It may, I am afraid, be considered that my choice in these respects has not been happy, and especially that too many words or phrases have been left without comment, where reasons were necessary. But I have, [\q xxiii/] endeavoured, in the notes and introductions, to emphasise those points on which further elucidation is desirable; and to raise some of the most important of those historical questions which will have to be settled before these Suttantas can finally be considered as having been rightly understood.

T. W. Rhys Davids
`NÓlandÓ,' April, 1899.


Buddhist Canonical Books


A´guttara NikÓya



Buddha VaŘsa



DŃgha NikÓya









Khuddaka PÓ˝ha



Majjhima NikÓya


M. P. S.

MahÓparinibbÓna Sutta



Samyutta NikÓya


S. N.

Sutta NipÓta






VimÓna Vatthu


V. (or Vin.)





Abh. Pad.

AbhidÓna PadŃpikÓ



Attha SalinŃ


Ath. V.

Atharva Veda



BrihadÓra§yaka Upanishad


Dhp. Cy. 

Dhammapada commentary





Ep. Ind. 

Epigraphia Indica


J. P. T. S.

Journal of the PÓli Text Society


J. R. A. S.

Journal of the Royal Asiatic Society


ChÓnd. Up.

ChÓndogya Upanishad


M. B. V.

MahÓ Bodhi VaŘsa



Milinda Pa§ha. Par. DŃp. Paramattha DŃpanŃ


S. B. E

.Sacred Books of the East



Sumangala VilÓsinŃ


Sat. Br. 




TaittirŃya Upanishad



Zeitschrift der deutschen morgenlandischen Gesellschaft.


[1] See the authors quoted in the Introduction to vol. II of my translation. Professor Takakusu, in an article in the J. R. A. S. for 1896, has added important details.

[2] They are not `sects' at all, in the modern European sense of the word. Some of the more important of these data are collected in two articles by the present writer in the `Journal of the Royal Asiatic Society' for 1891 and 1892.

[3] `Epigr. Ind.,' II, 93, and Z. D. M. G.,' xi, p. 58.

[4] `Reden des Gotamo,' pp. x, xi

[5] `Journal of the PÓli Text Society,' 1896; journal of the Royal Asiatic Society,' 1898, p. 639. Compare, `Milinda' (S. B. E., vol. xxxv), pp xxxvii foll.

[6] `Inscriptions de Piyadasi,' II, 314-322.

[7] Quoted by Prof. Oldenberg at p. 46 of his edition.

[8] Miss C. Hughes is preparing a complete alphabetical list of all these works for the journal of the Royal Asiatic Society,' 1899.

[9] As is well-known, the single instance of such a discrepancy, which Prof. Minayeff made so much of, is a mare's nest. The blunder is on the part of the European professor, not of the Ceylon pa§ˇits. No critical scholar will accept the proposition that because the commentary on the KathÓ Vatthu mentions the VetulyakÓ, therefore the KathÓ Vatthu itself must be later than the rise of that school.

[10] North Indian that is, from the modern European point of view. In the books themselves the reference is to the Middle Country (Majjhima Desa). To them the country to the south of theVindhyas simply did not come into the calculation. How suggestive this is as to the real place of origin of these documents!

 [11] Very probably memoranda were used. But the earliest records of any extent were the Asoka Edicts, and they had to be written on stone.

[12] On the often repeated error that a Sanskrit canon was established at Kanishka's Council, see my `Milinda,' vol. ii, pp. xv, xvi.

[13] No reference has been made, in these slight and imperfect remarks, to the history of the Vinaya. There is nothing to add, on that point, to the able and lucid exposition of Prof Oldenberg in the Introduction to his edition of the text.