INTRODUCTION.

THE work of which a translation is here, for the first time, presented to the English reading public, has had a strange and interesting history. Written in Northern India, at or a little after the beginning of the Christian era, and either in Sanskrit itself or in some North Indian Prakrit, it has been entirely lost in the land of its origin, and (so far as is at present known) is not extant in any of the homes of the various sects and schools of the Buddhists, except only in Ceylon, and in those countries which have derived their Buddhism from Ceylon. It is true that General Cunningham says 1 that the name of Milinda, 'is still famous in all Buddhist countries.' But he is here drawing a very wide conclusion from an isolated fact. For in his note he refers only to Hardy, who is good evidence for Ceylon, but who does not even say that the 'Milinda' was known elsewhere.

Preserved there, and translated at a very early date into Pâli, it has become, in its southern home, a book of standard authority, is put into the hands of those who have begun to doubt the cardinal points of Buddhist doctrine, has been long a popular work in its Pâli form, has been translated into Simhalese, and occupies a unique position, second only to the Pâli Pitakas (and perhaps also to the celebrated work of Buddhaghosa, the 'Path of Purity'). From Ceylon it has been transferred, in its Pâli form, to both Burma and Siam, and in those countries also it enjoys so high a repute, that it has been commented on (if not translated). It is not merely the only work composed among the Northern Buddhists which is regarded with reverence by the orthodox Buddhists of the southern

p. xii

schools; it is the only one which has survived at all amongst them. And it is the only prose work composed in ancient India which would be considered, from the modern point of view, as a successful work of art.

The external evidence for these statements is, at present, both very slight and, for the most part, late. There appeared at Colombo in the year of Buddha 2420 (1877 A.D.) a volume of 650 pages, large 8vo.--the most considerable in point of size as yet issued from the Simhalese press--entitled MILINDA PRASNAYA. It was published at the expense of five Buddhist gentlemen whose names deserve to be here recorded. They are Karolis Pîris, Abraham Liwerâ, Luis Mendis, Nandis Mendis Amara-sekara, and Chârlis Arnolis Mendis Wijaya-ratna Amara-sekara. It is stated in the preface that the account of the celebrated discussion held between Milinda and Nâgasena, about 500 years after the death of the Buddha, was translated into the Mâgadhî language by 'teachers of old' (purwâkârin wisin);--that that Pâli version was translated into Simhalese, at the instance and under the patronage of King Kîrtti Srî Râga-simha, who came to the throne of Ceylon in the year of Buddha 2290 (1747 A. D.), by a member of the Buddhist Order named Hînati-kumburê Sumangala, a lineal successor, in the line of teacher and pupil (anusishya), of the celebrated Wœliwita Saranankara, who had been appointed Samgha-râga, or chief of the Order--that 'this priceless book, unsurpassable as a means either for learning the Buddhist doctrine, or for growth in the knowledge of it, or for the suppression of erroneous opinions,' had become corrupt by frequent copying--that, at the instigation of the well-known scholar Mohotti-watte Gunânanda, these five had had the texts corrected and restored by several learned Bhikkhus (kîpa namak lawâ), and had had indices and a glossary added, and now published the thus revised and improved edition.

The Simhalese translation, thus introduced to us, follows the Pâli throughout, except that it here and there adds, in the way of gloss, extracts from one or other of the numerous Pitaka texts referred to, and also that it starts with a prophecy,

p. xiii

put into the mouth of the Buddha when on his death-bed, that this discussion would take place about 500 years after his death, and that it inserts further, at the point indicated in my note on p. 3 of the present version, an account of how the Simhalese translator came to write his version. His own account of the matter adds to the details given above that he wrote the work at the Uposatha Ârâma of the Mahâ Wihâra near Srî-ward-hana-pura, 'a place famous for the possession of a temple containing the celebrated Tooth Relic, and a monastery which had been the residence of Wœliwita Saranankara, the Samgha-râga, and of the famous scholars and commentators Daramiti-pola Dhamma-rakkhita and Madhurasatota Dhammakkhandha.'

As Kîrtti Srî Râga-simha reigned till 1781 1, this would only prove that our Pâli work was extant in Ceylon in its present form, and there regarded as of great antiquity and high authority, towards the close of the last century. And no other mention of the work has, as yet, been discovered in any older Simhalese author. But in the present deplorable state of our ignorance of the varied and ancient literature of Ceylon, the argument ex silentio would be simply of no value. Now that the Ceylon Government have introduced into the Legislative Council a bill for the utilisation, in the interests of education, of the endowments of the Buddhist monasteries, it may be hoped that the value of the books written in those monasteries will not be forgotten, and that a sufficient yearly sum will be put aside for the editing and publication of a literature of such great historical value 2. At present we can only deplore the impossibility of tracing the history of the 'Questions of Milinda' in other works written by the scholarly natives of its southern home.

That it will be mentioned in those works there can be

p. xiv

but little doubt. For the great Indian writer, who long ago found in that beautiful and peaceful island the best scope for his industrious scholarship, is already known to have: mentioned the book no less than four times in his commentaries; and that in such a manner that we may fairly hope to find other references to it when his writings shall. have been more completely published. In his commentary on the Book of the Great Decease, VI, 3, Buddhaghosa refers to the quotation of that passage made in the conversation between Milinda and Nâgasena, translated below, at IV, 2, 1 1. And again, in his commentary on the Ambattha Sutta (D. III, 2, 12) he quotes the words of a conversation between Milinda and Nâgasena on the subject he is there discussing. The actual words he uses (they will be found at pp. 275, 276 of the edition of the Sumangala Vilâsîni, edited for the Pâli Text Society by Professor Carpenter and myself) are not the same as those of our author at the corresponding passage of Mr. Trenckner's text (pp. 168, 169; IV, 3, 11), but they are the same in substance.

The above two references in Buddhaghosa to our author were pointed out by myself. Dr. Morris has pointed out two others, and in each of those also Buddhaghosa is found to quote words differing from Mr. Trenckner's text. The former of these two was mentioned in a letter to the 'Academy' of the 12th November, 1881. In the Manoratha Pûranî, his commentary on the Anguttara, on the passage marked in Dr. Morris's edition as L 5, 8, Buddhaghosa says:--

'Imasmim pan' atthe Milinda-râgâ dhamma-kathika-Nâgasenattheram pukkhi: "Bhante Nâgasena, ekasmim akkharakkhane pavattita-kitta-samkhârâ sake rûpino assa kîva mahâ-râsi bhavey-yâti?"'

And he then gives the answer:--'Vâhasatânam kho mahâ-râga vihînam addha-kûlañ ka vâhâ vîhi sattammanâni dve ka tumbâ ekakkharakkhane

p. xv

pavattitassa kittassa sankham pi na upenti kalam pi na upenti kala-bhâgam pi na upentîti.'

This passage of the Milinda, referred to by Buddhaghosa, will be found on p. 102 of Mr. Trenckner's edition, translated below at IV, 1, 19. But the question is not found there at all, and the answer, though much the same in the published text, still differs in the concluding words. Mr. Trenckner marks the passage in his text as corrupt, and it may well be that Buddhaghosa has preserved for us an older and better reading.

The other passage quoted by Dr. Morris (in the 'Academy' of the 11th January, 1881) is from the Papañka Sûdanî, Buddhaghosa's still unedited Commentary on the Magghima Nikâya. It is in the comment on the Brahmâyu Suttanta, and as it is not accessible elsewhere I give this passage also in full herb. With reference, oddly enough, to the same passage referred to above (pp. 168, 169 of the text, translated below at IV, 3, 11) Buddhaghosa, there says:--

'Vuttam etam Nâgasenattherena Milinda-rañña putthena: "Na mahârâga Bhagavâ guyham dasseti khâyam Bhagavâ dassetîti."'

In this case, as in the other quotation of the same passage, the words quoted are not quite the same as those given in the published text, and on the other hand they agree with, though they are much shorter than, the words as given in the Sumangala Vilâsinî.

It would be premature to attempt to arrive at the reason of this difference between Buddhaghosa's citations and Mr. Trenckner's edition of the text. It may be that Buddhaghosa is consciously summarising, or that he is quoting roughly from memory, or that he is himself translating or summarising from the original work, or that he is quoting from another Pâli version, or that he is quoting from another recension of the text of the existing Pâli version. We must have the full text of all his references to the 'Questions of Milinda' before us, before we try to choose between these, and possibly other, alternative explanations. What is at present certain is that when

p. xvi

[paragraph continues] Buddhaghosa wrote his great works, that is about 430 A.D., he had before him a book giving the conversations between Milinda and Nâgasena. And more than that. He introduces his comment above referred to on the Ambattha Sutta by saying, after simply quoting the words of the text he is explaining: 'What would be the use of any one else saying anything on this? For Nâgasena, the Elder, himself said as follows in reply to Milinda, the king 1'--and he then quotes Nâgasena, and adds not a word of his own. It follows that the greatest of all Buddhist writers known to us by name regarded the 'Questions of Milinda' as a work of so great authority that an opinion put by its author into the mouth of Nâgasena should be taken as decisive. And this is not only the only book, outside the Pâli Pitakas, which Buddhaghosa defers to in this way, it is the only book, except the previous commentaries, which he is known even to refer to at all. But, on the other hand, he says nothing in these passages to throw any further light on the date, or any light on the authorship, of the work to which he assigns so distinguished, even so unique, a position.

So far as to what is known about our 'Questions of Milinda' in Ceylon. The work also exists, certainly in Pâli, and probably in translations into the local dialects, in Burma and Siam. For Mr. Trenckner mentions (Introduction, p. iv) a copy in the Burmese character of the Pâli text sent to him by Dr. Rost, there is another copy in that character in the Colombo Museum 2, and Mr. J. G. Scott, of the Burmese Civil Service, has sent to England a Burmese Nissaya of the Milinda, (a kind of translation, giving the Pâli text, word for word, followed by the interpretation of those words in Burmese 3). A manuscript of the Pâli text, brought from Siam, is referred to in the Simhalese MSS. in the marginal note quoted by Mr. Trenckner at p. vi of the

p. xvii

[paragraph continues] Introduction to his edition. And there exists in the library of Trinity College, Cambridge, a complete MS., in excellent condition, in the Siamese-Pâli character 1, while there are numerous fragments in the Paris Bibliothèque Nationale of one or more MSS. of the text, in the same Kambojan character used in Siam for the writing of Pâli texts 2.

It may be noticed here that there are seven MSS. of the text written in the Ceylon character known to exist in Europe. Two of them (one a very ancient one) are in the Copenhagen University Library, two in the Bibliothèque Nationale 2, one in the Cambridge University Library 3, and two in the India Office Library 4. Three only of these seven have been used by Mr. Trenckner for his very able and accurate edition of the text, published in 1880.

____________________

That is all the external evidence at present available. What can be inferred from the book itself is about as follows. It consists of the discussion of a number of points of Buddhist doctrine treated in the form of conversations between King Milinda and Nâgasena the Elder (Thera). It must be plain to every reader of the following pages that these are not real conversations. What we have before us is really an historical romance, though the didactic aim overshadows the story. Men of straw, often very skilfully put together, are set up for the purpose, not so much of knocking them down again, as of elucidating some points of ethical or psychological belief while doing so. The king himself plays a very subordinate part. The questions raised, or dilemmas stated, are put into his mouth. But the solutions, to give opportunity for which the questions or dilemmas are invented, are the really important part of the work, and these are put into the mouth of Nâgasena. The dialogues are introduced by a carefully constructed

p. xviii

preliminary story, in which the reader's interest in them is aroused by anticipation. And the ability of this part of the work is very great. For in spite of the facts that all the praise lavished therein upon both Milinda and Nâgasena is in reality only praise of the book itself, and that the reader knows this very well, yet he will find it almost impossible to escape from the influence of the eloquent words in which importance and dignity are lent to the occasion of their meeting; and of the charm and skill with which the whole fiction is maintained.

The question then arises whether the personages were any more real than the conversations. Milinda is supposed to be the Menander, who appears in the list of the Greek kings of Baktria, since he is described in the book as being a king of the Yonakas reigning at Sâgala (the Euthydemia of the Greeks), and there is no other name in the list which comes so near to Milinda. This identification of the two names is certainly correct. For whether it was our author who deliberately made the change in adapting the Greek name to the Indian dialect in which he wrote, or whether the change is due to a natural phonetic decay, the same causes will have been of influence. Indra or Inda is a not uncommon termination of Indian names, and meaning king is so appropriate to a king, that a foreign king's name ending in -ander would almost inevitably come to end in -inda. Then the sequence of the liquids of m-n-n would tend in an Indian dialect to be altered in some way by dissimilation, and Mr. Trenckner adduces seven instances in Pâli of l taking the place of n, or n of l, in similar circumstances 1.

There remains only the change of the first E in Menander to I. Now in the Indian part of the inscription, on undoubted coins of Menander, the oldest authorities read Minanda as the king's name 2, and though that interpretation has now, on the authority of better specimens, been given up, there is no doubt that Milinda runs more easily

p. xix

from the tongue than Melinda, and Mil may well have seemed as appropriate a commencement for a Milakkha's name as -inda is for the ending of a king's name. So Men-ander became Mil-inda.

It may be added here that other Greek names are mentioned by our author--Devamantiya at I, 42, and the same officer, together with Anantakâya, Mankura, and Sabbadinna, at II, 3. There is a similar effort in these other Pâli forms of Greek words to make them give some approach to a meaning in the Indian dialect: but in each case the new forms remain as really unintelligible to an Indian as Mil-inda would be. Thus Deva-mantiya, which may be formed on Demetrios, looks, at first sight, Indian enough. But if it meant anything, it could only mean 'counsellor of the gods.' And so also both Ananta and Kâya are Indian words. But the compound Ananta-kâya would mean 'having an infinite body,' which is absurd as the name of a courtier. It may possibly be made up to represent Antiochos. What Mankura and Sabbadinna (called simply Dinna at p. 87) may be supposed to be intended for it is difficult to say 1. But the identification of Milinda with Menander is as certain as that of Kandagutta with Sandrokottos.

____________________

Very little is told us, in the Greek or Roman writers, about any of the Greek kings of Baktria. It is a significant fact that it is precisely of Menander-Milinda that they tell us most, though this most is unfortunately not much.

Strabo, in his Geography 2, mentions Menander as one of the two Baktrian kings who were instrumental in spreading the Greek dominion furthest to the East into India. He crossed the Hypanis (that is the Sutlej) and penetrated as far as the Isamos (probably the Jumna).

Then in the title of the lost forty-first book of Justin's work, Menander and Apollodotus are mentioned as 'Indian kings.'

Finally, Plutarch 3 tells us an anecdote of Menander.

p. xx

[paragraph continues] He was, he says, as a ruler noted for justice, and enjoyed such popularity with his subjects, that upon his death, which took place in camp, diverse cities contended for the possession of his ashes. The dispute was only adjusted by the representatives of the cities agreeing that the relics should be divided amongst them, and that they should severally erect monuments (μνημεῖα, no doubt dâgabas or sthûpas) to his memory.

This last statement is very curious as being precisely analogous to the statement in the 'Book of the Great Decease 1,' as to what occurred after the death of the Buddha himself. But it would be very hazardous to draw any conclusion from this coincidence.

The only remaining ancient evidence about Menander-Milinda (apart from what is said by our author himself), is that of coins. And, as is usually the case, the evidence of the coins will be found to confirm, but to add very little to, what is otherwise known.

As many as twenty-two 2 different coins have been discovered, some of them in very considerable numbers, bearing the name, and eight of them the effigy, of Menander. They have been found over a very wide extent of country, as far west as Kâbul, as far east as Mathurâ, and one of them as far north as Kashmir. Curiously enough we find a confirmation of this wide currency of Menander-Milinda's coins in the work of the anonymous author of the 'Periplus Maris Erythræi.' He says 3 that Menander's coins, together with those of Apollodotos, were current, many years after his death, at Barygaza, the modern Baroach, on the coast of Gujarat.

The portrait on the coins is very characteristic, with a long face and an intelligent expression, and is sometimes that of a young man, and at other times that of a very old man. It may be inferred therefore that his reign

p. xxi

was as long as his power was extensive. All the coins have a legend in Greek letters on one side, and a corresponding legend in Ariano-pâli letters on the other side. On twenty-one out of the twenty-two, the inscriptions, according to the latest interpretations from a comparison of the best examples, are respectively,

BASILEÔS SÔTÊROS MENANDROU

and

MAHARAGASA TRADATASA MENANDRASA 1.

Wilson read 2 the last word Minadasa. But when he wrote, in 1840, the alphabet was neither so well known as it is now, nor had such good examples come to hand. So that though the Mi- is plain enough on several coins, it is almost certainly a mere mistake for Me, from which it only differs by the centre vowel stroke being slightly prolonged.

Fifteen of the coins have a figure of Pallas either on one side or the other. A 'victory,' a horse jumping, a dolphin, a head (perhaps of a god), a two-humped camel, an elephant goad, a boar, a wheel, and a palm branch are each found on one side or the other of one of the coins; and an elephant, an owl, and a bull's head each occur twice. These are all the emblems or figures on the coins. None of them are distinctively Buddhist, though the wheel might be claimed as the Buddhist wheel, and the palm branch and the elephant would be quite in place on Buddhist coins. It may be said, therefore, that the bulk of the coins are clearly pagan, and not Buddhist; and that though two or three are doubtful, even they are probably not Buddhist,

One coin, however, a very rare one, differs, as to its inscription, from all the rest that have the legend. It has on one side

BASILEÔS DIKATOU MENANDROU,

and on the other,

MAHARAGASA DHARMIKASA 3 MENANDRASA.

p. xxii

Is any reference intended here to the Buddhist Dharma as distinct from the ordinary righteousness of kings? I think not. The coin is one of those with the figure of Pallas on the side which bears the Greek legend, and five others of the Baktrian Greek kings use a similar legend on their coins. These are Agathocles, Heliokles, Archebios, Strato, and Zoilos. There is also another coin in the series with a legend into which the word Dharma enters, but which has not yet been deciphered with certainty--that bearing in the Greek legend the name of Sy-Hermaios, and supposed to have been struck by Kadphises I. If there is anything Buddhist in this coin of Menander's, then the others also must be Buddhist. But it is much simpler to take the word dharmikasa in the sense of the word used in the corresponding Greek legend, and to translate it simply 'the Righteous,' or, better still, 'the Just.' Only when we call to mind how frequent in the Pâli texts is the description of the ideal king (whether Buddhist or not) as dhammiko dhamma-râga, we cannot refuse to see the connection between this phrase and the legend of the coins, and to note how at least six of the Greek kings, one of whom is Menander, are sufficiently desirous to meet the views of their Buddhist subjects to fix upon 'Righteousness' or 'Justice' as the characteristic by which they wish to be known. The use of this epithet is very probably the foundation of the tradition preserved by Plutarch, that Menander was, as a ruler, noted for justice; and it is certainly evidence of the Buddhist influences by which he was surrounded. But it is no evidence at all that he actually became a Buddhist.

To sum up.--Menander-Milinda was one of those Greek kings who carried on in Baktria the Greek dominion founded by Alexander the Great. He was certainly one of the most important, probably the most important, of those kings. He carried the Greek arms further into India than any of his predecessors had done, and everything confirms the view given by our author at I, 9 of his justice and his power, of his ability and his wealth. He must have reigned for a considerable time in the latter

p. xxiii

part of the second century B.C., probably from about 140 to about 115, or even 110 B.C. 1 His fame extended, as did that of no other Baktrian king, to the West, and he is the only Baktrian Greek king who has been remembered in India. Our author makes him say, incidentally 2, that he was born at Kalasi in Alasanda (= Alexandria), a name given to an island presumably in the Indus. And, as was referred to above, Plutarch has preserved the tradition that he died in camp, in a campaign against the Indians in the valley of the Ganges.

[It is interesting to point out, in this connection, that the town (gâma) of Kalasi has not been found mentioned elsewhere. Now among the very numerous coins of the Baktrian kings there is one, and only one, giving in the legend, not the name of a king, but the name of a city, the city of Karisi. As this coin was struck about 180 B.C. by Eukratides, who was probably the first of these kings to obtain a settlement on the banks of the Indus, it is possible that the two names, one in the Pâli form (or more probably in the form of the dialect used by our author), the other in the local form, are identical; and that the coin was struck in commemoration of the fact of the Greeks having reached the Indus. If that be so, then that they gave the name Alasanda (Alexandria) to the island on which the town was built, and not to the town itself, seems to show that the town was not founded by them, but was already an important place when they took it.]

____________________

Beyond this all is conjecture. When our author says that Milinda, was converted to Buddhism 3, he may be either relating an actual tradition, or he may be inventing for his own purposes. There is nothing inherently impossible, or even improbable, in the story. We know that all the Baktrians, kings and people alike, eventually became

p. xxiv

[paragraph continues] Buddhist. But the passage occurs in a part of the book which is open to much doubt. We have to place against it the negative evidence that none of Menander's coins show any decisive signs of his conversion. And the passage in question goes much further. It says that he afterwards gave up the kingdom to his son, and having entered the Buddhist Order, attained to Arahatship. The Simhalese MSS. add a marginal note to the effect that the whole of this passage with its context was derived from a MS. brought from Siam. Mr. Trenckner is therefore of opinion 1 that it belongs to a spurious supplement. That may be so, in spite of the fact that it is quite in our author's style, and forms an appropriate close to the book. But it is incredible that an author of the literary skill so evident throughout the work should have closed his book deliberately in the middle of a paragraph, without any closing words to round it off. The Siamese MS. may after all have preserved the reading of older and better MSS. than those in Ceylon, and the last leaf of the book may have been lost there. There must have been some conclusion, if not in the manner of the paragraph under discussion, then in some other words which we may not be able to trace. But even if our author actually wrote that Menander did become a Bhikkhu and an Arahat, that is very poor evidence of the fact, unless he not only intended what he states to be taken quite literally, but also wrote soon after the events he thus deliberately records.

Now the opinion has been expressed above that we have to deal with a book of didactic ethics and religious controversy cast into the form of historical romance. If this is correct no one would be more astonished than the author himself at the inconsistency of modern critics if they took his historical statements au grand serieux, while they made light of his ethical arguments. It is true that he would scarcely have been guilty of anything that seemed grossly improbable, at the time when he wrote, to the readers whom he addressed. But if, as is most probable, he wrote in North-Western

p. xxv

[paragraph continues] India when the memory of the actual facts of Menander's reign was fading away--that is, some generations after his death--he may well have converted him to Buddhism, as the most fitting close to the discussion he records, without intending at all to convey thereby any real historical event.

This brings us to the next point of our argument.

____________________

We have seen that the work must have been written some considerable time before Buddhaghosa, and after the death of Menander. Can its date be determined with greater accuracy than this? The story of Nâgasena introduces to us his father Sonuttara, his teachers Rohana, Assagutta of the Vattaniya hermitage, and Dhamma-rakkhita of the Asoka Ârâma near Pâtaliputta, and there is also mention of a teacher named Âyupâla dwelling at the Sankheyya hermitage near Sâgala. None of these persons and none of these places are read of elsewhere in any Buddhist text, whether Sanskrit or Pâli. For the Asvagupta referred to in passing at p. 351 of the Divyâvadâna has nothing in common (except the name) with our Assagutta, the Rohana of Anguttara, III, 66, is quite distinct from our Rohana, and there is not the slightest reason for supposing Nâgasena to be another form of the name Nâgârguna, found in both the Chinese and Tibetan Buddhist literatures 1, and in the Jain lists 2. The famous Buddhist scholar so called was the reputed founder of the Mahâyâna school of Buddhism. Our Nâgasena represents throughout the older teaching. If there is any connection at all between the two names, Nâgasena must have been invented as a contrast to Nâgârguna, and not with the least idea of identifying two men whose doctrines are so radically opposed. Even were there any reason to believe this to be the case, it would not help us much, for the date

p. xxvi

of Nâgârguna is quite as much open to dispute as that of the author of the 'Questions of Milinda 1.'

I ought to mention here that an opinion of a Nâgasena is, according to Burnouf 2, discussed at length in the Abhidharma Kosa Vyâkhyâ; and that Schiefner 3 quotes from a Tibetan work, the Bu-ston, the statement that a schism took place under a Thera Nâgasena 137 years after the Buddha's death. It would be very interesting if the former were our Nâgasena. And if Schiefner's restoration of the name found in his Tibetan authority be correct, and the authority itself be trustworthy, it is possibly the fading memory of that Nâgasena which induced our author to adopt the name as that of the principal interlocutor in his 'Questions of Milinda.'

Finally, Professor Kern, of Leiden--who believes that Buddha is the sun, and most of his principal disciples stars--believes also not only that our Nâgasena is an historical person, but also that there never was a Buddhist cleric of that name; and that Nâgasena is simply Patañgali, the author of the Yoga philosophy, under another name. If this is not a joke, it is a strange piece of credulity.

The only reason alleged in support of it is that Patañgali has the epithets of Nâgesa and of Phanin. That he was a Hindu who believed in the soul-theory of the current animistic creed, while all the opinions put into Nâgasena's mouth are those of a thorough-going Buddhist and non-individualist, is to count as nothing against this chance similarity, not of names, but of the name on one side with an epithet on the other. To identify John Stuart Mill with Dean Milman would be sober sense compared with this proposal.

p. xxvii

[paragraph continues] But it is deliberately put forward to support an accusation against the Buddhists of having falsely appropriated to themselves every famous man in India 1. Any mud, it would seem, is good enough to pelt the Buddhists with. Yet who is it, after all, who really makes the 'appropriation,' the Buddhists or Professor Kern himself?

____________________

It would seem, therefore, that most of our author's person and place names are probably inventions of his own 2.

But it is quite different with the books quoted by our author. In several passages he has evidently in his mind certain Pâli texts which deal with similar matters. So far as yet ascertained the texts thus silently referred to, either in. the present volume or in the subsequent untranslated portion of the book, are as follows:

p. xxviii

In several other passages he refers to a Pâli book, or a chapter in a Pâli book, by name. This is much more valuable for our purposes than the silent, and sometimes doubtful, references in the last list. So far as is yet ascertained, these references are as follows:

p. xxix

p. xxx

p. xxxi

Lastly, our author quotes a large number of passages from the Pitaka texts, which he introduces (without naming any book) by the formulas: 'It was said by the Blessed One;' or, 'It is said by you' (you in the plural, you members of the Order); or, 'It was said by so and so' (naming some particular member of the Order). A great many of these quotations have already been traced, either by Mr. Trenckner or myself. Occasionally words thus attributed, by our author, to the Buddha, are, in the Pitakas, attributed to some one else. Such passages are distinguished in the following list by an asterisk added to the letter B, which marks those of them attributed by our author to the Buddha. The women quoted are distinguished by the title 'Sister.'

p. xxxii

p. xxxiii

p. xxxiv

p. xxxv

p. xxxvi

Now the Pâli Pitakas consist of the following twenty-nine books:

This shows the total extent of the three Pitakas to be about 10,000 pages 8vo. as printed, or to be printed, by the Pâli Text Society 1. If our English Bible, in the older authorised version, were to be printed in the same manner and type and on the same size of page, it would occupy about 5,000 pages. So that the Buddhist Bible without its repetitions (some of which are very frequent, and others very long), would only occupy about double the space of the English Bible. This would not have been a literature too large to be familiarly known to our author. What is the conclusion which can fairly be drawn, from a comparison of the last list with those preceding it, as to his knowledge of those books now held, by living Buddhists, to be canonical?

The answer to this question will be of some importance for another reason beyond the help it will afford towards settling the date of the: original 'Questions of Milinda.' As is well known, Asoka, in the only one of his edicts, addressed specially to the members of the Buddhist Order of mendicants, selects seven portions of the Buddhist Scriptures, which he mentions by name, and expresses his desire that not only the brethren and sisters of the Order, but also the laity, should constantly learn by heart and reflect upon those seven. Now not one of the seven titles which occur in the edict is identical with any of the twenty-nine in the last list. Whereupon certain Indianists have rejoiced at being able to score a point, as they think, against these

p. xxxviii

unbrahmanical Buddhists, and have jumped to the conclusion that the Buddhist canon must be late and spurious; and that the Buddhism of Asoka's time must have been very different from the Buddhism of the Pâli Pitakas. That would be much the same as if a Japanese scholar, at a time when he knew little or nothing of Christianity, except the names of the books in the Bible, were to have found an open letter of Constantine's in which he urges both the clergy and laity to look upon the Word of God as their only authority, and to constantly repeat and earnestly meditate upon the Psalm of the Shepherd, the words of Lemuel, the Prophecy of the Servant of the Lord, the Sermon on the Mount, the Exaltation of Charity, the Question of Nicodemus, and the story of the Prodigal Son--and that our Oriental critic should jump to the conclusion that the canonical books of the Christians could not have been known in the time of Constantine, and that the Christianity of Constantine was really quite different from, and much more simple than the Christianity of the Bible. As a matter of fact the existence of such a letter would prove very little, either way, as to the date of the books in the Bible as we now have them. If our Japanese scholar were to discover afterwards a Christian work, even much later than the time of Constantine, in which the canonical books of the Christians were both quoted and referred to, he would have much surer ground for a sounder historical criticism. And he would possibly come to see that the seven portions selected for special honour and commendation were not intended as an exhaustive list even of remarkable passages, much less for an exhaustive list of canonical books, but that the number seven was merely chosen in deference to the sacred character attaching to that number in the sacred literature.

Such a book is our Milinda. It is, as we have seen, later than the canonical books of the Pâli Pitakas, and on the other hand, not only older than the great commentaries, but the only book, outside the canon, regarded in them as an authority which may be implicitly followed. And I venture to think that the most simple working hypothesis

p. xxxix

by which to explain the numerous and varied references and quotations it makes, as shown in the preceding lists, from the Pitakas as a whole, and from the various books contained in them, is that the Pâli Pitakas were known, in their entirety, and very nearly, if not quite, as we now have them, to our author. For out of the twenty-nine books of the Pitakas, we find in the lists of works referred to by him the three Pitakas as a whole, the Vinaya Pitaka as a whole, and all of its component books except the Parivâra (which was composed in Ceylon), the Sutta Pitaka and each of the four great Nikâyas, the Abhidhamma Pitaka and each of its seven component books, and the Khuddaka Nikâya as a whole and several of its separate books. And when we further recollect the very large number of quotations appearing in my lists as not yet traced in the Pitakas, we see the necessity of being very chary in drawing any argument ex silentio with respect to those books not occurring in the lists.

To sum up.--It may be said generally that while the Sutta Vibhanga and the Khandhakas, the four great Nikâyas, and the Abhidhamma were certainly known to our author, he very likely had no knowledge of the Parivâra; and it remains to be seen how far his knowledge of the Khuddaka Nikâya, which he happens to mention once 1 as a whole by name, did actually extend. At present it is only clear that he knew the Khuddaka Pâtha, the Dhammapada collection of sacred verses, the Sutta Nipâta, the Thera and Theri-gâthâ, the Gâtakas, and the Kariyâ Pitaka. I hope to return to this question in the Introduction to my second volume, only pointing out here that the doubtful books (those concerning which our author is apparently silent) would occupy about two thousand pages octavo, out of the ten thousand of which the three Pitakas would, if printed, consist: and that those two thousand pages belong, for the most part, precisely to that part of the Pitakas which have not yet been edited, so that there they may very likely, after all, be quoted in one or other

p. xl

of the numerous quotations entered as 'not traced' in my lists 1.

____________________

Such being the extent, so far as can at present be shown, of our author's knowledge of the three Pitakas, the question arises as to the degree and accuracy of his knowledge. In the great majority of cases his quotations or references entirely agree with the readings shown by our texts. But there are a few exceptions. And as these are both interesting and instructive, it will be advisable to point them out in detail.

The reference to the Avîki Hell as being outside the earth, if not at variance with, is at least an addition to the teaching of the Pitakas as to cosmogony 2. But there is some reason to believe that the passage may be an interpolation, and the difference itself is not only doubtful but also of no particular importance.

The description of the contents of the Puggala Paññatti given in I, 26, does not really agree with the text. The book, in its first section, sets out six different sorts of discrimination or distinction. One paragraph only is devoted to each of the first five discriminations, and the author or authors then proceed, in the rest of the book, to deal with the details of the last of the six. Our author gives the six as the divisions of the book itself.

But I think it is clear that so far as the description is inaccurate, the error is due, not to any difference between the text as he had it and that which we now possess, but simply to our author laying too great a stress upon the opening paragraphs of the book.

In the reference to the Buddha's first sermon, the Foundation of the Kingdom of Righteousness (in 1, 38), our author says that 'eighteen kotis of Brahma gods, and an innumerable company of other gods, attained to comprehension

p. xli

of the truth.' There is no statement of the kind in the Pitaka account of this event (see my translation in 'Buddhist Suttas,' pp. 146-155). But it is not inconsistent with the Pâli, and is doubtless added from some edifying commentary.

There is a difference of reading between the lines put into Sâriputta's mouth, at II, 2, 4, and those ascribed to Sâriputta in the Thera Gâthâ (1002, 1003). If the Milinda reading is not found in some hitherto unpublished passage, we have here a real case of divergence.

Perhaps the most important apparent variation between our author and the Pitaka texts is the statement put by him, in IV, 4, 9, into the mouth of the Buddha, that a deliberate lie is one of the offences called Pârâgika, that is, involving exclusion from the Order. Now in the old Canon Law there are only four Pârâgika offences--breach of chastity, theft, murder, and a false claim to extraordinary spiritual powers (see my translation in vol. i, pp. 1-5 of the 'Vinaya Texts'); and falsehood is placed quite distinctly under another category, that of the Pâkittiyas, offences requiring repentance (see p. 32 of the same translation). If our author was a member of the Order, as he almost certainly was, it' would seem almost incredible that he should make an error in a matter of such common knowledge, and of such vital importance, as the number and nature of the Pârâgikas. And indeed, in the immediate context, he refers to the Pâkittiya rule, though not in the exact words used in the text of the Pâtimokkha. I think that he must have known very well what he was talking about. And that a passage, not yet traced, will be found in the unpublished parts of the Pitakas, in which the Buddha is made to say that falsehood is a Pârâgika--just as a Christian might maintain that falsehood is forbidden in the Ten Commandments, and yet be perfectly aware of the exact phraseology of the Ten Words.

In IV, 4, 26, our author identifies the learned pig in the Takkha-sûkara Gâtaka with the Bodisat. He differs here from the Gâtaka Commentary, in which the Bodisat is identified with the tree-god, who acts as a kind of Greek chorus in the story. And the summaries in IV, 4, 28 of

p. xlii

[paragraph continues] Ruru Gâtaka, and in IV, 4, 30 of the Sabba-dâtha Gâtaka, do not exactly agree with Professor Fausböll's text 1. But the commentary is not the text; and it is well known that there are numerous such light variations in the different expansions of the verses, which latter alone form the actual text.

In IV, 4, 44 we find our author giving a version of a well-known incident in the Buddhist Gospel story different from the oldest version of it in the Pitaka texts. This is another instance of an expansion of the original adopted from some unknown commentator, and does not argue an ignorance of the text as we have it.

I have noticed in the untranslated portion of our author, four or five cases of readings apparently different from the Pitaka texts he refers to. These I hope to deal with in my next volume. But I may notice here that two stanzas, given on p. 414 of the text, and said on p. 413 to be 'in the Sutta Nipâta,' are not found in Professor Fausböll's edition of that work; and we have there, in all probability, another case of real divergence. But the reading in the Milinda may possibly be found to be incorrect.

The general result of this comparison, when we remember the very large number of passages quoted, will be held, I trust, to confirm the conclusion reached above, that our author knew the Pitakas practically as we now have them, that is as they have been handed down in Ceylon.

Outside the Pitakas there are unfortunately no references to actual books. But there are several references to countries and persons which are of importance, in as much as they show a knowledge in our author of places or occurrences not mentioned in the sacred books. It will be most convenient to arrange these passages first in an alphabetical list, and then to make a few remarks on the conclusions the list suggests. They are as follows:

It will be noticed that the only names of persons, besides those occurring in the story itself, are, in one passage, Asoka and Bindumatî the courtesan, and in another Kandragupta and Bhaddasâla who fought against him. Of places, besides those in the story, we have a considerable number of names referring to the Panjâb, and adjacent countries; and besides these the names only of a few places or countries on

p. xliv

the sea coast. The island Alasanda in the Indus, and the town of Kalasi situated in that island, have been discussed above. The country of the Sakas and Yavanas, Gandhâra, Kashmir, Bharukakkha, Surat, and Madhura, explain themselves. Nikumba and Vilâta were probably in the same neighbourhood, but these names have not been met with elsewhere, and I can suggest no identification of them. The places on the sea coast, to which a merchant ship could sail, mentioned on p. 359, are mostly well known. Kolapattana must, I think, be some place on the Koromandel coast, and Suvanna-bhûmi be meant for the seaboard of Burma and Siam. The author mentions no places in the interior south of the Ganges.

At four places he gives lists of famous rivers. In three out of the four he simply repeats the list of five--Gangâ, Yamunâ, Akiravatî, Sarabhû, and Mahî--so often enumerated together in the Pitakas 1. In the fourth passage (p. 114) he adds five others--the Sindhu, the Sarassatî, the Vetravatî, the Vîtamsâ, and the Kandabhâgâ. Of these the first two are well known. Professor Eduard Müller suggests 2 that the Vîtamsâ is the same as the Vitastâ (the Hydaspes of the Greeks and the modern Bihat). The Vetravatî is one of the principal affluents of the Jumna; and the Kandabhâgâ rises in the North-West Himâlayas, and is not unfrequently referred to as the Asiknî of the Vedas, the Akesines of the Greek geographers, the modern Kînâb 3.

The list is meagre enough. An ethical treatise is scarcely the place to look for much geographical or historical matter. But unless our author deliberately concealed his knowledge, and made all the remarks he put into the mouth of Nâgasena correspond with what that teacher might fairly be expected to have known, the whole list points to the definite conclusion that the writer of the 'Questions of Milinda' resided in the far North-West of

p. xlv

India, or in the Panjâb itself. And this is confirmed by the great improbability of any memory of Menander having survived elsewhere, and more especially in Ceylon, where we should naturally look for our author's residence if he did not live in the region thus suggested.

____________________

As my space is here limited, I postpone to the next volume the discussion as to how far the knowledge displayed by our author, the conditions of society with which he shows himself acquainted, and the religious beliefs he gives utterance to, afford evidence of his date. I will only say here that on all these points his work shows clear signs of being later than the Pitaka texts. And in the present state of our knowledge, or rather of our ignorance, of Pâli, there is very little to be drawn from the language used by our author. In the first place we do not know for certain whether we have the original before us, or a translation from the Sanskrit or from some Northern dialect. And if, as is probably the case, we have a translation, it would be very difficult to say whether any peculiarity we may find in it is really due to the translator, or to the original author. No doubt a translator, finding in his original a word not existing in Pâli, but formed according to rules of derivation obtaining in Pâli, would coin the corresponding Pâli form. And in doing so he might very likely be led into mistake, if his original were Prakrit, by misunderstanding the derivation of the Prakrit word before him. Childers in comparing Buddhist Sanskrit with Pâli, has pointed out several cases where such mistakes have occurred, and has supposed that in every case the Sanskrit translator misunderstood a Pâli word before him 1. As I have suggested elsewhere it is, to say the least, quite as likely that the Sanskrit Buddhist texts are often founded on older works, not in Pâli, but in some other Prakrit 2. And it may be possible hereafter to form some opinion as to what that dialect was which the Sanskrit writers must have had before

p. xlvi

them, to lead them into the particular blunders they have made. In the same way an argument may be drawn from the words found exclusively in Milinda as to the dialect which he spoke, and in which he probably wrote. A list of the words our author uses, and not found in the Pitakas, can only be tentative, as we have not as yet the whole of the Pitaka texts in print. But it will be useful, even now, to give the following imperfect list of such as I have noted in my copy of Childers' 'Dictionary.'

p. xlviii

This list might be considerably extended if words were included which differ from those used in the Pitakas only by the addition of well-known suffixes or prefixes--such, for instance, as viparivattati, at p. 117, only found as yet elsewhere in the Tela Katâha Gâthâ, verse 37. But such words are really only a further utilisation of the existing resources of the language, and would afford little or no ground for argument as to the time and place at which our author wrote. I have thought it best, therefore, to omit them, at least at present.

If we turn from isolated words to the evidence of style it will be acknowledged by every reader that the Milinda has a marked style of its own, different alike from the formal exactness of most of the Pitaka texts, and from the later manner of any other Pâli or Sanskrit-Buddhist authors as yet published. It is no doubt the charm of its style which has been one of the principal reasons for the great popularity of the book. Even a reader who takes no interest in the points that are raised, or in the method in which the questions are discussed, will be able, I trust, to see, even through the dark veil of a lame and wooden translation, what the merits of the original must be. And to a devout Buddhist, in whose eyes the book he was reading offered a correct solution of the most serious difficulties in religion, of the deepest problems of life,--to whose whole intellectual training and sympathies the way in which the puzzles are put, and solved, so exactly appealed,--to such a reader both the easy grace of the opening dialogue, as of a ship sailing in calm waters, and the real eloquence of occasional passages, more especially of the perorations by which the solutions are sometimes closed, must have been a continual feast. I venture to think that the 'Questions of Milinda' is undoubtedly the master-piece of Indian prose; and indeed is the best book of its class, from a literary point of view, that had then been produced in any country. Limits of space prevent the discussion of this last proposition, however interesting: and it would be, no doubt, difficult to prove that anything from India was better than the corresponding thing produced by our noble selves, or by those

p. xlix

whose Karma we inherit. But in ancient Indian literature there are only two or three works which can at all compare with it. It ought not to seem odd that these also are Buddhist and Pâli; that is, that they come from the same school. And while the Dîgha Nikâya may be held to excel it in stately dignity, the Visuddhi Magga in sustained power, and the Gâtaka book in varied humour, the palm will probably be eventually given to the 'Questions of Milinda' as a work of art.

I am aware that this conclusion is entirely at variance with the often repeated depreciation of Buddhist literature. But the fact is that this depreciation rests upon ignorance, and is supported by prejudice. As a critical judgment it will not survive the publication and translation of those great Buddhist works which it overlooks or ignores. Some Sanskrit scholars, familiar with the Brahmin estimate of matters Indian, and filled with a very rational and proper admiration for the many fine qualities which the old Brahmins possessed, may find it hard to recognise the merits of sectarian works written in dialects which violate their most cherished laws of speech. But the historical student of the evolution of thought, and of the rise of literature in India, will more and more look upon the question as a whole, and will estimate at its right value all Indian work, irrespective of dialect or creed.

T. W. RHYS DAVIDS.

TEMPLE,

August, 1889.