[1. Dîpavamsa VII, 40.

2. Dîpavamsa VII, 55.

3. Dr. Oldenberg, in his Introduction to the Vinaya-pitaka, p. xxxii.

4. Oldenberg, Vinaya-pitaka I, p. xvi, treats it as an extended reading of Pâtimokkha.]

p. xxviii

II. Sutta-pitaka.

[1. The Mahâparinibbâna-sutta, ed. by Childers, Journal of the Royal Asiatic Society, translated with other Suttas by Rhys Dayids (S.B.E. vol. xi). Sept Suttas Palis, par Grimblot, Paris, l876.

2. The first four are sometimes called the Four Nikâyas, the five together the five Nikâyas. They represent the Dharma, as settled at the First and Second Councils, described in the Kullavagga (Oldenberg, I, p. xi).

3. Sometimes Khuddaka-nikâya stands for the whole Vinaya and Abhidhamma- pitaka, with the fifteen divisions here given of Khuddaka-nikâya. In the commentary on the Brahmagâla-sutta it is said that the Dîghanikâya professors rehearsed the text of the Gâtaka, Mahâ and Kulla Niddesa, Patisambhidâmagga, Suttanipâta, Dhammapada, Udâna, Itivuttaka, Vimâna, and Petavatthu, Thera and Therî Gâthâ, and called it Khuddakagantha, and made it a canonical text, forming part of the Abhidhamma; while the Magghimanikâya professors assert that, with the addition of the Kariyâpitaka, Apadâna, and Buddhavamsa, the whole of this Khuddakagantha was included in the Suttapitaka. See Childers, s.v. Nikâya;. See also p. x.

4. Published by Childers, Journal of the Royal Asiatic Society, 1869.

5. Published by Fausböll, 1855.

6. Thirty translated by Sir Coomâra Swâmy; the whole by Fausböll, in Sacred Books of the East, vol. x.

7. Published by Fausböll, translated by Rhys Davids.]

p. xxix

III. Abhidhamma-pitaka.

   Taking this collection as a whole we may lay it down as self-evident that the canon, in its collected form, cannot be older than any of the events related therein.

   There are two important facts for determining the age of the Pâli canon, which, as Dr. Oldenberg^[4] has been the first to show, should take precedence of all other arguments, viz.

   1. That in the Tipitaka, as we now have it, no mention is made of the so-called Third Council, which took place at Pâtaliputta, under King Asoka, about 242 B.C.

   2. That in the Tipitaka, as we now have it, the First Council of Râgagaha (477 B.C.) and the Second Council of Vesâlî (377 B.C.) are both mentioned.

   From these two facts it may safely be concluded that the Buddhist canon, as handed down to us, was finally closed

[1. Buddhaghosa does not say whether these were recited at the First Council.

2. Partly translated by Gogerly, Journal of the Asiatic Society of Ceylon, 1852.

3. Cf. Gogerly, Journal of the Asiatic Society of Ceylon. 1848, p. 7.

4. See Oldenberg's Vinaya-pitaka, Introduction. p. xxv. The kings Agâtasatru (485-453 B.C.), Udâyin (453-437 B.C.), and Munda (437-429 B.C.) are all mentioned in the Tipitaka. See Oldenberg, Zeitschrift der D. M. G., XXXIV. pp. 752, 753.]

p. xxx after the Second and before, or possibly at, the Third Council. Nay, the fact that the description of the two Councils stands at the very end of the Kullavagga may be taken, as Dr. Oldenberg remarks, as an indication that it was one of the latest literary contributions which obtained canonical authority, while the great bulk of the canon may probably claim a date anterior to the Second Council.

   This fact, namely, that the collection of the canon, as a whole, must have preceded the Second Council rests on an argument which does great credit to the ingenuity of Dr. Oldenberg. The Second Council was convoked to consider the ten deviations^[1] from the strict discipline of the earliest times. That discipline had been laid down first in the Pâtimokkha rules, then in the commentary now included in the Vibhanga, lastly in the Mahâvagga and Kullavagga. The rules as to what was allowed or forbidden to a Bhikkhu were most minute^[2], and they were so firmly established that no one could have ventured either to take away or to add anything to them as they stood in the sacred code. In that code itself a distinction is made between the offences which were from the first visited with punishment (pârâgika and pâkittiya) and those misdemeanours and crimes which were put down as punishable at a later time (dukkata and thullakkaya). With these classes the code was considered as closed, and if any doubt arose as to the criminality of certain acts, it could be settled at once by an appeal to the Vinaya-pitaka. Now it so happens that, with one exception, the ten deviations that had to be considered at the Second Council, are not provided for in the Vinaya-pitaka; and I quite agree with Dr. Oldenberg's argument that, if they had been mentioned in the Vinaya-pitaka, the Second Council would have been objectless. A mere appeal to chapter and verse in the existing Pitaka would then have silenced all dissent. On the other side, if it had been possible to add anything to the canon, as it then existed, the ten, or nine, deviations might have been condemned

[1. Oldenberg, Introduction, p. xxix.

2. Oldenberg, loc. cit. p. xx.]

p. xxxi by a few additional paragraphs of the canon, without convoking a new Council.

   I think we may be nearly certain, therefore, that we possess the principal portion of the Vinaya-pitaka as it existed before the Council of Vesâlî.

   So far I quite agree with Dr. Oldenberg. But if he proceeds to argue^[1] that certain portions of the canon must have been finally settled before even the First Council took place, or was believed to have taken place, I do not think his arguments conclusive. He contends that in the Parinibbâna-sutta, which tells of the last days of Buddha's life, of his death, the cremation of his body, and the distribution of his relics, and of Subhadda's revolt, it would have been impossible to leave out all mention of the First Council, if that Council had then been known. It is true, no doubt, that Subhadda's disloyalty was the chief cause of the First Council, but there was no necessity to mention that Council. On the contrary, it seems to me that the unity of the Parinibbâna-sutta would have been broken if, besides telling of the last days of Buddha, it had also given a full description of the Council. The very title, the Sutta of the Great Decease, would have become inappropriate, if so important a subject as the first Sangîti had been mixed up with it. However, how little we may trust to such general arguments, is best shown by the fact that in some very early Chinese renderings of the Hînayâna text of the Mahâparinibbâna-sutta the story is actually carried on to the First Council, two (Nos. 552 and 119) mentioning the rehearsal under Kasyapa, while the third (No. 118) simply states that the Tiptaka was then collected^[2].

[1. Loc. cit. pp. xxvi-xxviii.

2. There are several Chinese translations of Sûtras on the subject of the Mahâparinirvâna. Three belong to the Mahâyâna school: 1. Mahâparinirvâna-sûtra, translated by Dharmaraksha, about 414-423 A.D.; afterwards revised, 424-453 (Nos. 113, 114). 2. Translation by Fa-hian and Buddhabhadra, about 415 A.D.; less complete (No. 120). 3. Translation (vaipulya) by Dharmaraksha I, i.e. Ku Fa-hu, about 261-308 A.D. (No. 116). Three belong to the Hînayâna school: 1. Mahiparinirvâna-sûtra. translated by Po-fa-tsu, about 290-306 A.D. (No. 552). 2. Translation under the Eastern Tsin dynasty, 317-42O A.D. (No. 119). 3. Translation by Fa-hian, about 415 A.D. (No. 118).]

p. xxxii

   We must be satisfied therefore, so far as I can see at present, with fixing the date, and the latest date, of a Buddhist canon at the time of the Second Council, 377 B.C. That some works were added later, we know; that many of the treatises included in the canon existed before that Council, can hardly be doubted. The second chapter of the Dhammapada, for instance, is called the Appamâda-vagga, and if the Mahâvamsa (p. 25) tells us that at the time when Asoka was converted by Nigrodha, that Buddhist priest explained to him the Appamâda-vagga, we can hardly doubt that there existed then a collection (vagga) of verses on Appamâda, such as we now possess in the Dhammapada and in the Samyutta-nikâya^[1].

   With regard to the Vinaya, I should even feel inclined to admit, with Dr. Oldenberg, that it must have existed in a more or less settled form before that time. What I doubt is whether such terms as Pitaka, basket, or Tipitaka, the three baskets, i.e. the canon, existed at that early time. They have not been met with, as yet, in any of the canonical books; and if the Dîpavamsa (IV, 32) uses the word 'Tipitaka,' when describing the First Council, this is due to its transferring new terms to older times. If Dr. Oldenberg speaks of a Dvi-pitaka^[2] as the name of the canon before the third basket, that of the Abhidhamma, was admitted, this seems to me an impossible name, because at the time when the Abhidhamma was not yet recognised as a third part of the canon, the word pitaka had probably no existence as a technical term^[3].

   We must always, I think, distinguish between the three portions of the canon, called the basket of the Suttas, the

[1. Feer, Revue Critique, 1870, No. 24, p. 377.

2. Introduction. pp. x, xii.

3. Dr. Oldenberg informs me that pitaka occurs in the Kankîsuttanta in the Magghima Nikâya (Turnour's MS., fol. the), but applied to the Veda. He also refers to the tipitakâkâryas mentioned in the Western Cave inscriptions as compared with the Pañkanekâyâka in the square Asoka character inscriptions (Cunningham, Bharhut, pl. lvi, No. 52). In the Sûtrakrid-anga of the Gainas, too, the term pidagam occurs (MS. Berol. fol. 77 a). He admils, however, that pitaka or tipitaka, as the technical name of the Buddhist canon, has not yet been met with in that canon itself, and defends Dvipitaka only as a convenient term.]

p. xxxiii basket of Vinaya, and the basket of Abhidhamma, and the three subjects of Dhamma (sutta), Vinaya, and Abhidhamma, treated in these baskets. The subjects existed and were taught long before the three baskets were definitely arranged. Dhamma had originally a much wider meaning than Sutta-pitaka. It often means the whole teaching of Buddha; and even when it refers more particularly to the Sutta-pitaka, we know that the Dhamma there taught deals largely with Vinaya and Abhidhamma doctrines. Even the fact that at the First Council, according to the description given in the Kullavagga, the Vinaya and Dhamma only were rehearsed, though proving the absence at that time of the Abhidhamma, as a separate Pitaka, by no means excludes the subject of the Abhidhamma having been taught under the head of Dhamma. In the Mahâkarunâpundarîka-sûtra the doctrine of Buddha is divided into Dharma and Vinaya; the Abhidharma is not mentioned. But the same text knows of all the twelve Dharmapravakanâni^[1], the 1. Sûtra; 2. Geya; 3. Vyâkarana; 4. Gâthâ; 5. Udâna; 6. Nidâna; 7. Avadâna; 8. Itivrittaka; 9. Gâtaka; 10. Vaipulya; 11. Adbhutadharma; 12. Upadesa; some of these being decidedly metaphysical.

   To my mind nothing shows so well the historical character both of the Kullavagga and of Buddhaghosa in the Introduction to his commentary on the Dîgha-nikâya, as that the former, in its account of the First Council, should know only of the Vinaya, as rehearsed by Upâli, and the Dhamma, as rehearsed by Ânanda, while the much later Buddhaghosa, in his account of the First Council^[2], divides the Dhamma into two parts, and states that the second part, the Abhidhamma, was rehearsed after the first part, the Dhamma. Between the time of the Kullavagga and the time of Buddhaghosa the Abhidhamma must have assumed its recognised position by the side of Vinaya and Sutta. It must be left to further researches to determine, if possible,

[1. See Academy, August 28, 1880, Division of Buddhist Scriptures.

2. Oldenberg, Introduction, p. xii; Turnour, Journal of the Asiatic Society of Bengal, vi, p. 510 seq.]

p. xxxiv the time when the name of pitaka was first used, and when Tipitaka was accepted as the title of the whole canon.

   Whenever we see such traces of growth, we feel that we are on historical ground, and in that sense Dr. Oldenberg's researches into the growth of the Vinaya, previous to the Second Council, deserve the highest credit. He shows, in opposition to other scholars, that the earliest elements of Vinaya must be looked for in the short Pâtimokkha rules, which were afterwards supplemented by explanations, by glosses and commentaries, and in that form answered for some time every practical purpose. Then followed a new generation who, not being satisfied, as it would seem, with these brief rules and comments, wished to know the occasion on which these rules had been originally promulgated. What we now call the Vibhanga, i.e. the first and second divisions of the Vinaya-pitaka, is a collection of the stories, illustrating the origin of each rule, of the rules themselves (the Pâtimokkha), and of the glosses and comments on these rules.

   The third and fourth books, the Mahâvagga and Kullavagga, are looked upon as possibly of a slightly later date. They treat, in a similar manner as the Vibhanga, on the rules not included in that collection, and give a general picture of the outward life of the monks. While the Vibhanga deals chiefly with the original so-called pârâgika, sanghâdisesa, and pâkittiya offences, the Khandhaka, i.e. the Mahâvagga and Kullavagga, treats of the so-called dukkata and thullakkaya crimes. The arrangement is the same, story, rule, and comment succeeding each other in regular sequence. If we follow the guidance of the Vinaya-pitaka, we should be able to distinguish the following steps in the growth of Buddhism before the Second Council of Vesâlî:

p. xxxv

   The Kulavagga ascribes the settlement of the canon to the First Council, and does not even claim a revision of that canon for the Second Council. The Dîpavamsa claims a revision of the canon by the 700 Arhats for the Second Council.

CHRONOLOGY.

   In order to bring the Council of Vesâlî in connection with the chronology of the world, we must follow the Buddhist historians for another century. One hundred and eighteen years after the Council of Vesâlî they place the anointment of King Asoka, during whose reign a Third Council, under the presidency of Tissa Moggaliputta, took place at Pâtaliputta, the new capital adopted by that king, instead of Râgagaha and Vesâlî. This Council is chiefly known to us through the writings of the southern Buddhists (Dîpavamsa, Mahâvamsa, and Buddhaghosa), who belong to the school of Moggaliputta (Theravâda or Vibhaggavâda), which ruled supreme at Pâtaliputta, while Upagupta, the chief authority of the northern Buddhists, is altogether ignored in the Pâli chronicles.

   Now it is well known that Asoka was the grandson of Kandagutta, and Kandagutta the contemporary of Alexander the Great. Here we see land, and I may refer to my History of Sanskrit Literature, published in 1859, for the process by which the storm-tossed ship of Indian chronology has been landed in the harbour of real historical chronology. We are told by the monks of the Mahâvihâra in Ceylon that Asoka was crowned, according to their computation, 146 + 18 years before the accession of Dutthagâmani, 161 B.C., i.e. 325 B.C.; that between his coronation and his father's death four years had elapsed (329 B.C.); that his father Bindusâra had reigned twenty-eight years^[1] (357-329 B.C.), and Bindusâra's father, Kandagutta,

[1. Mahâvamsa, p. 21.]

p. xxxvi twenty-four years (381-357). As we know that Kandagutta, whom the Ceylonese place 381-357 B.C., was king of India after Alexander's conquest, it follows that Ceylonese chronology is wrong by more than half a century. For reasons stated in my History of Sanskrit Literature, I fix the exact fault in Ceylonese chronology as sixty-six years, assigning to Kandagutta the dates 315-291, instead of 381-357. This gives us 291-263 for Bindusâra, 259 for Asoka's abhisheka; 259 + 118 = 377 for the Council of Vesâlî, and 377 + 100 = 477 for Buddha's death, instead of 543 B.C.^[1]

   These dates are, of course, approximate only, and they depend on one or two points on which people may differ. But, with that reservation, I see no ground whatever for modifying the chronological system which I put forward more than twenty years ago. Professor Westergaard and Professor Kern, who have since suggested different dates for the death of Buddha, do not really differ from me in principle, but only in their choice of one or the other alternative, which I readily admit as possible, but not as more certain than my own. Professor Westergaard^[2], for instance, fixes Buddha's death at 368 (370), instead of 477. This seems a wide difference, but it is so in appearance only.

   Following Justinus, who says that Sandrokyptos^[3] had conquered the empire of India at the time when Seleucus laid the foundations of his own greatness, I had accepted 315^[4], half-way between the murder of Porus and the taking of Babylon by Seleucus, as the probable beginning

[1. According to Bigandet, Life of Gaudama, p. 361, the era of Buddha's death was introduced by Agâtasatru, at the conclusion of the First Council, and began in the year 46 of the older Eetzana era (p. 12). See, however, Rhys Davids, Num. Orient. vi, p. 38. In the Kâranda-vyûha, p. 96, a date is given as 300 after the Nirvâna, 'tritîye varshasate gate mama parinirvritasya.' In the Asoka-avadâna we read, mama nirvritim ârabhya satavarshagata Upagupto nâma bhikshur utpatsyati.

2. Über Buddha's Todesjahr (1860), 1862.

3. The Greek name Sandrokyptus shows that the Pâli corruption Kandagutta was not yet the recognised name of the king.

4. Mr. Rhys Davids accepts 315 B.C. as the date when, after the murder of king Nanda, Kandragupta stept into the vacant throne, though he had begun to count his reign seven or eight years before. Buddhism, p. 22O.]

p. xxxvii of Kandragupta's reign. Westergaard prefers 320 as a more likely date for Kandragupta, and therefore places the death of the last Nanda and the beginning of Asoka's royal pretensions 268. Here there is a difference between him and me of five years, which depends chiefly on the view we take as to the time when Seleucus really laid what Justinus calls the foundation of his future greatness. Secondly, Westergaard actually adopts the idea, at which I only hinted as possible, that the southern Buddhists made two Asokas out of one, and two Councils out of one. Trusting in the tradition that 118 years elapsed between Buddha's death and the Council under Asoka (at Pâtaliputra), and that the Council took place in the king's tenth year (as was the case with the imaginary Kâlâsoka's Council), he gets 268 - 10 = 258 as the date of the Council, and 368 or 370 as the date of Buddha's death^[1].

   The two points on which Westergaard differs from me, seem to me questions which should be kept before our mind in dealing with early Buddhist history, but which, for the present at least, admit of no definite solution.

   The same remark seems to me to apply to the calculations of another eminent Sanskrit scholar, Professor Kern^[2]. He lays great stress on the general untrustworthiness of Indian chronology, and I am the last to differ from him on that point. He then places the beginning of Kandragupta's reign in 322 B.C. Allowing twenty-four years to him and twenty-eight to his son Bindusâra, he places the beginning of Asoka's reign in 270. Asoka's inscriptions would fall about 258. As Asoka reigned thirty-six or thirty-seven years, his death would fall in 234 or 233 B.C. Like Westergaard, Professor Kern too eliminates Kâlâsoka, as a kind of chronological Asoka, and the Council of Vaisâlî, and therefore places Buddha's death, according to the northern tradition, 100 or 110 years before Dharmâsoka, i.e. 270 + 100 or + 110 = 370 or 380^[3]; while, according to the southern

[1. Westergaard. loc. cit. p. 128.

2. Jaartelling der Zuidelijke Buddhisten, 1873.

3. See Professor Kern's remark in Indian Antiquary, 1874, p. 79.]

p. xxxviii tradition, that 118 years elapsed between Asoka's accession and Buddha's death, the Ceylonese monks would seem originally to have retained 270 + 118^[1] = 388 B.C. as Buddha's Nirvâna, a date which, as Professor Kern holds, happens to coincide with the date assigned to the death of Mahâvira, the founder of the Gaina religion.

   Here we see again that the moot point is the beginning of Kandragupta's reign in accordance with the information supplied by Greek historians. Professor Kern places it in 322, Westergaard in 320, I myself in 315. That difference once granted, Dr. Kern's reasoning is the same as my own. According to the traditions which we follow, Buddha's death took place 100, 110, 118, or 228 years before Asoka. Hence Professor Westergaard arrives at 368 or 370 B.C., Professor Kern at 370 (380) or 388 B.C., I myself at 477 B.C. Every one of these dates is liable to certain objections, and if I prefer my own date, 477 B. C., it is simply because it seems to me liable to neither more nor less reservations than those of Professor Westergaard and Professor Kern, and because, so long as we always remember the grounds of our differences, namely, the beginning of Kandragupta's reign, and the additional century, every one of these dates furnishes a good hypothesis to work on, until we can arrive at greater certainty in the ancient chronology of India. To my mind all dates beyond Kandragupta are as yet purely tentative, resting far more on a chronological theory than on actual tradition; and though I do not doubt the historical chatacter of the Council of Vaisâlî, I look upon the date assigned to it, on the authority of the Dîpavamsa and Mahâvamsa, as, for the present, hypothetical only.

[1. When Professor Kern states that the Mahâvamsa (p. 22) places the Third Council 218 years after Buddha's death, this is not so. Asoka's abhisheka takes place in that year. The prophecy that a calamity would befall their religion, 118 years after the Second Council (Mahâvamsa, p. 28), does not refer to the Council, but to Kandâsoka's accession, 477 - 218 = 259 B.C.]

p. xxxix

[1. Westergaard, 320-296; Kern, 322-298.]

p. xl

   Though the preceding table, embodying in the main the results at which I arrived in my History of Ancient Sanskrit Literature, still represents what I hold to be true or most probable with respect to Indian chronology, previous to the beginning of our era, yet I suppose I may be expected to say here a few words on the two latest attempts to fix the date of Buddha's death; the one by Mr. Rhys Davids in the Numismata Orientalia, Part VI, 1877, the other by Dr. Bühler in the Indian Antiquary, 1877 and 1878^[1]. Mr. Rhys Davids, to whom we owe so much for the elucidation of the history of Buddha's religion, accepts Westergaard's date for the beginning of Kandragupta's reign, 320 B.C., instead of 322 (Kern), 315 (myself); and as he assigns (p. 41) to Bindusâra 25 years instead of 28 (Mahâvamsa, p. 21), he arrives at 268 as the year of Asoka's coronation^[2]. He admits that the argument derived from the mention of the five foreign kings in one of Asoka's inscriptions, dated the twelfth year of his reign, is too precarious to enable us to fix the date of Asoka's reign more definitely, and though, in a general way, that inscription confirms the date assigned by nearly all scholars to Asoka in the middle of the third century B.C., yet there is nothing in it that Asoka might not have written in 247 quite as well as in 258-261. What chiefly distinguishes Mr. Rhys Davids' chronology from that of his predecessors is the shortness of the period between Asoka's coronation and Buddha's death. On the strength of an examination of the list of kings and the list of the so-called patriarchs, he reduces the traditional 218 years to 140 or 150, and thus arrives at 412 B.C. as the probable beginning of the Buddhist era.

   In this, however, I cannot follow him, but have to follow Dr. Bühler. As soon as I saw Dr. Bühler's first essay on the Three New Edicts of Asoka, I naturally felt delighted at the unexpected confirmation which he furnished of the date which I had assigned to Buddha's death, 477 B.C. And though I am quite aware of the

[1. Three New Edicts of Asoka, Bombay, 1877; Second Notice, Bombay, 1878.

2. Mr. Rhys Davids on p. 50 assigns the 35 years of Bindusâra rightly to the Purânas, the 38 years to the Ceylon Chronicles.]

p. xli danger of unexpected confirmations of one's own views, yet, after carefully weighing the objections raised by Mr. Rhys Davids and Professor Pischel against Dr. Bühler's arguments, I cannot think that they have shaken Dr. Bühler's position. I fully admit the difficulties in the phraseology of these inscriptions: but I ask, Who could have written these inscriptions, if not Asoka? And how, if written by Asoka, can the date which they contain mean anything but 256 years after Buddha's Nirvâna? These points, however, have been argued in so masterly a manner by Dr. Bühler in his 'Second Notice,' that I should be afraid of weakening his case by adding anything of my own, and must refer my readers to his 'Second Notice.' Allowing that latitude which, owing to the doubtful readings of MSS., and the constant neglect of odd months, we must allow in the interpretation of Buddhist chronology, Asoka is the only king we know of who could have spoken of a thirty-fourth year since the beginning of his reign and since his conversion to Buddhism. And if he calls that year, say the very last of his reign (212 B.C.), 256 after the departure of the Master, we have a right to say that as early as Asoka's time, Buddha was believed to have died about 477 B.C. Whether the inscriptions have been accurately copied and rightly read is, however, a more serious question, and the doubts raised by Dr. Oldenberg (Mahâvagga, p. xxxviii) make a new collation of the originals absolutely indispensable, before we can definitely accept Dr. Bühler's interpretation.

   I cannot share Dr. Bühler's opinion^[1] as to the entire worthlessness of the Gaina chronology in confirming the date of Buddha's death. If the Svetâmbara Gainas place the death of Mahâvîra 470 before Vikramâditya, i.e. 56 B.C. + 470 = 526 B.C.,and the Digambaras 605, i.e. 78 A.D. deducted from 605 = 527 B.C., this so far confirms Dr. Bühler's and Dr. Jacobi's brilliant discovery that Mahâvîra was the same as Nigantha Nâtaputta, who died at Pâvâ during Buddha's lifetime^[2]. Most likely 527 is too early a date, while another

[1. Three Edicts. p. 21; Second Notice. pp. 9, 10.

2. See Jacobi, Kalpa-sûtra of Bhadrabâhu, and Oldenberg, Zeitschrift der D.M.G., XXXIV, p. 749.]

p. xlii tradition fixing Mahâvira's death 155 years before Kandragupta^[1], 470 B.C., is too late. Yet they both show that the distance between Asoka (259-222 B.C.), the grandson of Kandragupta (315-291 B.C.), and the contemporaries of Buddha was by the Gainas also believed to be one of two rather than one century.

   When I saw that the date of Buddha's death, 477 B.C., which in my History of Ancient Sanskrit Literature (1859) I had myself tried to support by such arguments as were then accessible, had received so powerful a support by the discovery of the inscriptions of Sahasrâm, Rûpnâth, and Bairât, due to General Cunningham, who had himself always been an advocate of the date 477 B.C., and through their careful decipherment by Dr. Bühler, I lost no time in testing that date once more by the Dîpavamsa, that Ceylonese chronicle having lately become accessible through Dr. Oldenberg's edition and translation^[2]. And here I am able to say that, before having read Dr. Bühler's Second Notice, I arrived, though by a somewhat different way, at nearly the same conclusions as those so well worked out by Dr. Bühler in his restoration of the Episcopal Succession (therâvali) of the Buddhists, and therefore feel convinced that, making all such allowances as the case requires, we know now as much of early Buddhist chronology as could be known at the time of Asoka's Council, 242 B.C.

   Taking the date of Buddha's death 477 B.C. for granted, I found that Upâli, who rehearsed the Vinaya at the First Council, 477 B.C., had been in orders sixty years in the twenty-fourth year of Agâtasatru, i.e. 461 B.C., which was the sixteenth year A.B. He must therefore^[3] have been born in 541 B.C., and he died 447 B.C., i.e. thirty years A.D., at the age of 94. This is said to have been the sixth year of Udâyi, and so it is, 453 - 6 = 447 B.C.

   In the year 461 B.C. Dâsaka received orders from Upâli, who was then 80 years of age; and when Dâsaka had been

[1. Oldenberg, loc. cit. p. 750.

2. The Dîpavamsa, an ancient Buddhist historical record. London, 1879.

3. Assuming twenty to be the minimum age at which a man could be ordained.]

p. xliii in orders forty-five years (Dîpavamsa IV, 41), he ordained Saunaka. This would give us 461 - 45 = 416 B.C., while the tenth year of Nâgadâsa, 429 - 10, would give us 419 A.D. Later on the Dîpavamsa (V, 78) allows an interval of forty years between the ordinations of Dâsaka and Saunaka, which would bring the date of Saunaka's ordination to 421 B.C., instead of 419 or 416 B.C. Here there is a fault which must be noted. Dâsaka died 461 - 64 = 397 A.D., which is called the eighth year of Sisunâga, and so it is 405 - 8 = 397 A.D.

   When Saunaka had been in orders forty years, i.e. 416 - 40 = 376, Kâlâsoka is said to have reigned a little over ten years, i.e. 387 - 11 = 376 A.D., and in that year Saunaka ordained Siggava. He died 416 - 66 = 350 A.D., which is called the sixth year of the Ten, while in reality it is the ninth, 359 - 6 = 353 A.D. If, however, we take 419 as the year of Saunaka's ordination, his death would fall 419 - 66 = 353 B.C.

   Siggava, when he had been in orders sixty-four years, ordained Tishya Maudgalîputra. This date 376 - 64 = 312 B.C. is called more than two years after Kandragupta's accession, and so it very nearly is, 315 - 2 = 313.

   Siggava died when he had been in orders seventy-six years, i.e. 376 - 76 = 300 A.D. This year is called the fourteenth year of Kandragupta, which it very nearly is, 315 - 14 = 301.

   When Tishya had been in orders sixty^[1] years, he ordained Mahendra, 312 - 60 = 252 B.C. This is called six years after Asoka's coronation, 259 - 6 = 253, and so it very nearly is. He died 312 - 80 = 232 B.C., which is called the twenty-sixth year of Asoka, and so it very nearly is.

[1. I take 60 (80), as given in Dîpavamsa V, 95, 107, instead of 66 (86), given in Dîpavamsa V. 94.]

p. xliv

BUDDHIST PATRIARCHS.

{not all of the formatting could be reproduced in this table; all cells with three number are joined with a brace on the right in the original}

   If we test the dates of this table by the length of time assigned to each patriarchate, we find that Upâli ruled thirty years, from Buddha's death, 477 to 447; Dâsaka fifty years. To Saunaka forty-four years are assigned, instead of forty-seven, owing to a fault pointed out before; and to Siggava fifty-two years, or fifty-five^[1] instead of fifty. Tishya's patriarchate is said to have lasted sixty-eight years, which agrees with previous statements.

   Lastly, the years of the death of the six patriarchs, as fixed according to the reigns of the kings of Magadha, agree extremely well.

   Upâli died in the sixth year of Udâyi, i.e. 453 - 6 = 447 B.C.

   Dâsaka died in the eighth year of Sisunâga, i.e. 405 - 8 = 397 B.C.

   Saunaka died in the sixth year of the Ten, i.e. 359 - 6 = 353 B.C., showing again the difference of three years.

[1. The combined patriarchates of Saunaka and Siggava are given as 99 by the Dîpavamsa.]

p. xlv

   Siggava died in the fourteenth year of Kandragupta, i.e. 315 - 14 = 301 B.C.

   Tishya died in the twenty-sixth or twenty-seventh year of Asoka, i.e. 259 - 27 = 233 B.C.

   This general and more than general agreement between dates taken from the history of the kings and the history of the patriarchs leaves on my mind a decided impression of a tradition which, though not strictly historical, in our sense of the word, represents at all events the result of such enquiries as could be made into the past ages of Buddhism at the time of Asoka. There are difficulties in that tradition which would certainly have been avoided, if the whole chronology had been simply made up: but there is no doubt a certain method too perceptible throughout, which warns us that we must not mistake a smooth chronology for solid history.

THE TITLE OF DHAMMAPADA.

   The title of Dhammapada has been interpreted in various ways. It is an ambiguous word, and has been accepted as such by the Buddhists themselves. Dhamma has many meanings. Under one aspect it means religion, particularly the religion taught by Buddha, the law which every Buddhist should accept and observe. Under another aspect dhamma is virtue, or the realisation of the law.

   Pada also has many meanings. In the Abhidhânapadîpikâ it is explained by place, protection, Nirvâna, cause, word, thing, portion, foot, footstep.

   Hence dhammapada may mean 'footstep of religion,' and thus the title was first rendered by Gogerly, only that he used the plural instead of the singular, and called it 'The Footsteps of Religion,' while Spence Hardy still more freely called it 'The Paths of Religion.' It may be quite true, as pointed out by Childers, that pada by itself never means path. But it means footstep, and the footstep towards a thing is much the same as what we call the path to a thing. Thus we read, verse 21, 'appamâdo amatapadam,' earnestness is the step, i.e. the path that leads to immortality. p. xlvi Again, 'pamâdo makkuno padam' can hardly mean anything but that thoughtlessness is the path of death, is the path that leads to death. The commentator, too, rightly explains it here by amatasya adhigamupâya, the means of obtaining immortality, i.e. Nirvâna, or simply by upâyo, and even by maggo, the way. If we compare verses 92 and 93 of our text, and verses 254 and 255, we see that pada is used synonymously with gati, going. In the same manner dhammapada would mean the footstep or the footpath of virtue, i.e. the path that leads to virtue, and supply a very appropriate title for a collection of moral precepts. In verses 44 and 45 'path of virtue' seems to be the most appropriate meaning for dhammapada^[1], and it is hardly possible to assign any other meaning to it in the following verse (Kundasutta, v. 6):

'He who lives restrained and attentive in the way that has been well pointed out, in the path of the law, cultivating blameless words, such a Bhikkhu they call a Maggagîvi (living in the way).'

   I therefore think that 'Path of Virtue,' or 'Footstep of the Law,' was the idea most prominent in the mind of those who originally framed the title of this collection of verses. It seems to me that Buddhaghosa also took the same view, for the verse which D'Alwis^[2] quotes from the introduction of Buddhaghosa's commentary,--

Sampatta-saddhammapado satthâ dhammapadam subham Desesi,

and which he translates, 'The Teacher who had reached the very depths (lit. bottom) of Saddhamma, preached this holy Dhammapada,'--lends itself far better to another translation, viz. 'The Teacher who had gained a firm

[1. Cf. Dhammapada, v. 285, nibbânam sugatena desitam.

2. Buddhist Nirvâna, p. 62.]

p. xlvii footing in the Good Law, showed (preached) the holy Path of the Law.'

   Gogerly, again, who may generally be taken as a faithful representative of the tradition of the Buddhists still preserved in Ceylon, translates the title by the 'Footsteps of Religion,' so that there can be little doubt that the priests of that island accept Dhammapada in the sense of 'Vestiges of Religion,' or, from a different point of view, 'The Path of Virtue.'

   M. L. Feer^[1] takes a slightly different view, and assigning to pada the meaning of foot or base, he translates Dhammapada by Loi fondamentale, or Base de la Religion.

   But it cannot be denied that the title of Dhammapada was very soon understood in a different sense also, namely, as 'Sentences of Religion.' Pada means certainly a foot of a verse, a verse, or a line, and dhammapadam actually occurs in the sense of a 'religious sentence.' Thus we read in verse 102, 'Though a man recite a hundred Gâthâs made up of senseless words, one dhammapadam, i.e. one single word or line of the law, is better, which if a man hears, he becomes quiet.' But here we see at once the difficulty of translating the title of 'dhammapadam' by 'religious sentences.' Dhammapadam means one law verse, or wise saw, not many. Professor Fausböll, who in his excellent edition of the Dhammapada translated that title by 'a collection of verses on religion,' appeals to such passages as verses 44 and 102 in support of his interpretation. But in verse 42 dhammapadam sudesitam, even if it does not mean the path of the law, could never mean 'versus legis bene enarratos,' but only versum legis bene enarratum, as Dr. Fausböll himself renders ekam dhammapadam, in verse 102, by unus legis versus. Buddhaghosa, too, when he speaks of many law verses uses the plural, for instance^[2], 'Be it known that the Gâthâ consists of the Dhammapadâni, Theragâthâ, Therîgâthâ, and those unmixed (detached) Gâthâ not comprehended in any of the above-named Suttânta.'

[1. Revue Critique, 1870, p. 378.

2. D'Alwis, Pâli Grammar, p. 61.]

p. xlviii

   The only way in which Dhammapada could be defended in the sense of 'Collection of Verses of the Law,' would be if we took it for an aggregate compound. But such aggregate compounds, in Sanskrit at least, are possible with numerals only; for instance, tribhuvanam, the three worlds; katuryugam, the four ages^[1]. It might therefore be possible in Pâli, too, to form such compounds as dasapadam, a collection of ten padas, a work consisting of ten padas, a decamerone, but it would in no wise follow that we could in that language attempt such a compound as Dhammapadam, in order to express a collection of law verses^[2]. Mr. Beal^[3] informs us that the Chinese seem to have taken Dhammapada in the sense of 'stanzas of law,' 'law texts,' or 'scripture texts.'

   It should be remembered, also, that the idea of representing life, and particularly the life of the faithful, as a path of duty or virtue leading to deliverance, (in Sanskrit dharmapatha,) is very familiar to Buddhists. The four great truths of their religion^[4] consist in the recognition of the following principles: 1. that there is suffering; 2. that there is a cause of that suffering; 3. that such cause can be removed; 4. that there is a way of deliverance, viz. the doctrine of Buddha. This way is the ashtânga-mârga, the eightfold way^[5], taught by Buddha, and leading to Nirvâna^[6]. The faithful advances on that road, padât padam,

[1. See M. M.'s Sanskrit Grammar, § 519.

2. Mr. D'Alwis' arguments (Buddhist Nirvâna, pp. 63-67) in support of this view, viz. the dhammapada may be a collective term, do not seem to me to strengthen my own conjecture.

3. Dhammapada from Chinese, p. 4.

4. Spence Hardy, Manual, p. 496.

5. Burnouf, Lotus, p. 520, 'Ajoutons, pour terminer ce que nous trouvons à dire sur le mot magga, quelque commentaire qu'on en donne d'ailleurs, que suivant une définition rapportée par Turnour, le magga renferme une sous-division que l'on nomme patipadâ, en sanscrit pratipad. Le magga, dit Tumour, est la voie qui conduit au Nibbâna, la patipadâ, littéralement "la marche pas à pas, ou le degré," est la vie de rectitude qu'on doit suivre, quand on marche dans la voie du magga.'

6. See Spence Hardy, Manual, p. 496. Should not katurvidha-dharmapada, mentioned on p. 497, be translated by 'the fourfold path of the Law?' It can hardly be the fourfold word of the Law.]

p. xlix step by step, and it is therefore called patipadâ, lit. the step by step.

   If we make allowance for these ambiguities, inherent in the name of Dhammapada, we may well understand how the Buddhists themselves play with the word pada (see v. 45). Thus we read in Mr. Beal's translation of a Chinese version of the Prâtimoksha^[1]:

TRANSLATION.

   In translating the verses of the Dhammapada, I have followed the edition of the Pâli text, published in 1855 by Dr. Fausböll, and I have derived great advantage from his Latin translation, his notes, and his copious extracts from Buddhaghosa's commentary. I have also consulted translations, either of the whole of the Dhammapada, or of portions of it, by Burnouf, Gogerly^[2], Upham, Weber, and others. Though it will be seen that in many places my translation differs from those of my predecessors, I can only claim for myself the name of a very humble gleaner in this field of Pâli literature. The greatest credit is due to Dr. Fausböll, whose editio princeps of the Dhammapada will mark for ever an important epoch in the history of Pâli scholarship; and though later critics have been able to point out some mistakes, both in his text and in his translation, the value of their labours is not to be compared with that of the work accomplished single-handed by that eminent Danish scholar.

   In revising my translation, first published in 1870^[3], for

[1. Catena, p. 207.

2. Several of the chapters have been translated by Mr. Gogerly, and have appeared in The Friend, vol. iv, 1840. (Spence Hardy, Eastern Monachism, p. 169.)

3. Buddhaghosha's Parables, translated from Burmese by Captain T. Rogers, R. E. With an Introduction, containing Buddha's Dhammapada, translated from Pâli by F. Max Müller. London, 1870.]

p. l the Sacred Books of the East, I have been able to avail myself of 'Notes on Dhammapada,' published by Childers in the Journal of the Royal Asiatic Society (May, 1871), and of valuable hints as to the meaning of certain words and verses scattered about in the Pâli Dictionary of that much regretted scholar, 1875. I have carefully weighed the remarks of Mr. James D'Alwis in his 'Buddhist Nirvâna, a review of Max Müller's Dhammapada' (Colombo, 1871), and accepted some of his suggestions. Some very successful renderings of a number of verses by Mr. Rhys Davids in his ('Buddhism,' and a French translation, too, of the Dhammapada, published by Fernand Hû^[1], have been consulted with advantage.

   It was hoped for a time that much assistance for a more accurate understanding of this work might be derived from a Chinese translation of the Dhammapada^[2], of which Mr. S. Beal published an English translation in 1878. But this hope has not been entirely fulfilled. It was, no doubt, a discovery of great interest, when Mr. Beal announced that the text of the Dhammapada was not restricted to the southern Buddhists only, but that similar collections existed in the north, and had been translated into Chinese. It was equally important when Schiefner proved the existence of the same work in the sacred canon of the Tibetans. But as yet neither a Chinese nor a Tibetan translation of the Pâli Dhammapada has been rendered accessible to us by translations of these translations into English or German, and what we have received instead, cannot make up for what we had hoped for.

   The state of the case is this. There are, as Mr. Beal informs us, four principal copies of what may be called Dhammapada in Chinese, the first dating from the Wu dynasty, about the beginning of the third century A.D. This translation, called Fa-kheu-king, is the work of a

[1. Le Dhammapada avec introduction et notes par Fernand Hû, suivi du Sutra en 42 articles, traduit du Tibetain, par Léon Feer. Paris, 1878.

2. Texts from the Buddhist Canon, commonly known as Dhammapada, translated from the Chinese by Samuel Beal. London, 1878.]

p. li Shaman Wei-ki-lan and others. Its title means 'the Sûtra of Law verses,' kheu being explained by gâthâ, a verse, a word which we shall meet with again in the Tibetan title, Gâthâsangraha. In the preface the Chinese translator states that the Shamans in after ages copied from the canonical scriptures various gâthâs, some of four lines and some of six, and attached to each set of verses a title, according to the subject therein explained. This work of extracting and collecting is ascribed to Tsun-ke-Fa-kieou, i.e. Ârya-Dharmatrâta, the author of the Samyuktâbhidharma-sâstra and other works, and the uncle of Vasumitra. If this Vasumitra was the patriarch who took a prominent part in the Council under Kanishka, Dharmatrâta's collection would belong to the first century B.C.; but this is, as yet, very doubtful.

   In the preface to the Fa-kheu-king we are told that the original, which consisted of 500 verses, was brought from India by Wai-ki-lan in 223 A.D., and that it was translated into Chinese with the help of another Indian called Tsiang-sin. After the translation was finished, thirteen sections were added, making up the whole to 752 verses, 14,580 words, and 39 chapters^[1].

   If the Chinese translation is compared with the Pâli text, it appears that the two agree from the 9th to the 35th chapter (with the exception of the 33rd), so far as their subjects are concerned, though the Chinese has in these chapters 79 verses more than the Pâli. But the Chinese translation has eight additional chapters in the beginning (viz. On Intemperance, Inciting to Wisdom; The Srâvaka, Simple Faith, Observance of Duty, Reflection, Loving-kindness, Conversation), and four at the end (viz. Nirvâna, Birth and Death, Profit of Religion, and Good Fortune), and one between the 24th and 25th chapter of the Pâli text (viz. Advantageous Seivice), all of which are absent in our Pâli texts. This, the most ancient

[1. Beal, Dhammapada, p. 30. The real number of verses, however, is 760. In the Pâli text, too, there are five verses more than stated in the Index; see M.M., Buddhaghosha's Parables, p. ix, note; Beal, loc. cit. p. 11. note.]

p. lii Chinese translation of Dharmatrâta's work, has not been rendered into English by Mr. Beal, but he assures us that it is a faithful reproduction of the original. The book which he has chosen for translation is the Fa-kheu-pi-ü, i.e. parables connected with the Dhammapada, and translated into Chinese by two Shamans of the western Tsin dynasty (A.D. 265-313). These parables are meant to illustrate the teaching of the verses, like the parables of Buddhaghosa, but they are not the same parables, nor do they illustrate all the verses.

   A third Chinese version is called Kuh-yan-king, i.e. the Sûtra of the Dawn (avadâna?), consisting of seven volumes. Its author was Dharmatrâta, its translator Ku-fo-nien (Buddhasmriti), about 410 A.D. The MS. of the work is said to have been broght from India by a Shaman Sanghabhadanga of Kipin (Cabul), about 345 A.D. It is a much more extensive work in 33 chapters, the last being, as in the Pâli text, on the Brâhmana.

   A fourth translation dates from the Sung dynasty (800 or 900 A.D.), and in it, too, the authorship of the text is ascribed to Ârya-Dharmatrâta.

   A Tibetan translation of a Dhammapada was discovered by Schiefner in the 28th volume of the Sûtras, in the collection called Udânavarga. It contains 33 chapters, and more than 1000 verses, of which about one-fourth only can be traced in the Pâli text. The same collection is found also in the Tangur, vol. 71 of the Sûtras, foll. 1-53, followed by a commentary, the Udânavarga-vivarana by the Âkârya Pragñâvarman. Unfortunately Schiefner's intention of publishing a translation of it (Mélanges Asiatiques, tom. viii. p. 560) has been frustrated by his death. All that he gives us in his last paper is the Tibetan text with translation of another shorter collection, the Gâthâsangraha by Vasubandhu, equally published in the 72nd volume of the Sûtras in the Tangur, and accompanied by a commentary.

p. liii

SPELLING OF BUDDHIST TERMS.

   I had on a former occasion^[1] pleaded so strongly in favour of retaining, as much as possible, the original Sanskrit forms of Pâli Buddhist terms, that I feel bound to confess openly that I hold this opinion no longer, or, at all events, that I see it is hopeless to expect that Pâli scholars will accept my proposal. My arguments were these: 'Most of the technical terms employed by Buddhist writers come from Sanskrit; and in the eyes of the philologist the various forms which they have assumed in Pâli, in Burmese, in Tibetan, in Chinese, in Mongolian, are only so many corruptions of the same original forms. Everything, therefore, would seem to be in favour of retaining the Sanskrit forms throughout, and of writing, for instance, Nirvâna instead of the Pâli Nibbâna, the Burmese Niban or Nepbhân, the Siamese Niruphan, the Chinese Nipan. The only hope, in fact, that writers on Buddhism will ever arrive at a uniform and generally intelligible phraseology seems to lie in their agreeing to use throughout the Sanskrit terms in their original form, instead of the various local disguises and disfigurements which they present in Ceylon, Burmah, Siam, Tibet, China, and Mongolia.'

   I fully admitted that many Buddhist words have assumed such a strongly marked local or national character in the different countries and in the different languages in which the religion of Buddha has found a new home, that to translate them back into Sanskrit might seem as affected, nay, prove in certain cases as misleading, as if, in speaking of priests and kings, we were to speak of presbyters and cynings. The rule by which I meant mainly to be guided was to use the Sanskrit forms as much as possible; in fact, everywhere except where it seemed affected to do so. I therefore wrote Buddhaghosha instead of the Pâli Buddhaghosa, because the name of that famous theologian, 'the Voice of Buddha,' seemed to lose its significance if turned

[1. Introduction to Buddhaghosha's Parables, 1870. p. l.]

p. liv into Buddhaghosa. But I was well aware what may be said on the other side. The name of Buddhaghosa, 'Voice of Buddha,' was given him after he had been converted from Brahmanism to Buddhism, and it was given to him by people to whom the Pâli word ghosa conveyed the same meaning as ghosha does to us. On the other hand, I retained the Pâli Dhammapada instead of Dharmapada, simply because, as the title of a Pâli book, it has become so familiar that to speak of it as Dharmapada seemed like speaking of another work. We are accustomed to speak of Samanas instead of Sramanas, for even in the days of Alexander's conquest, the Sanskrit word Sramana had assumed the prakritized or vulgar form which we find in Pâli, and which alone could have been rendered by the later Greek writers (first by Alexander Polyhistor, 80-60 B.C.) by {Greek: samanaioi}^[1]. As a Buddhist term, the Pâli form Samana has so entirely supplanted that of Sramana that, even in the Dhammapada (v. 388), we find an etymology of Samana as derived from sam, 'to be quiet,' and not from sram, 'to toil.' But if we speak of Samanas, we ought also to speak of Bâhmanas instead of Brâhmanas, for this word had been replaced by bâhmana at so early a time, that in the Dhammapada it is derived from a root vah, 'to remove, to separate, to cleanse^[2].'

   I still believe that it would be best if writers on Buddhist literature and religion were to adopt Sanskrit throughout as the lingua franca. For an accurate understanding of the original meaning of most of the technical terms of Buddhism a knowledge of their Sanskrit form is indispensable; and nothing is lost, while much would be gained, if, even in the treating of southern Buddhism, we were to

[1. See Lassen, Indische Alterthumskunde, vol. ii. p. 700, note. That Lassen is right in taking the {Greek: Sarmanai}, mentioned by Megasthenes, for Brahmanic, not for Buddhist ascetics, might be proved also by their dress. Dresses made of the bark of trees are not strictly Buddhistic.

2. See Dhammapada, v. 388; Bastian. Völker des östlichen Asien, vol iii. p. 412: 'Ein buddhistischer Mönch erklärte mir, dass die Brahmanen ihren Namen führten, als Leute, die ihre Sünden abgespült hätten.' See also Lalita-Vistara, p. 551, line 1; p. 553, line 7.]

p. lv speak of the town of Srâvastî instead of Sâvatthi in Pâli, Sevet in Sinhalese; of Tripitaka, 'the three baskets,' instead of Tipitaka in Pâli, Tunpitaka in Sinhalese; of Arthakathâ, 'commentary,' instead of Atthakathâ in Pâli, Atuwâva in Sinhalese; and therefore also of Dharmapada, 'the path of virtue,' instead of Dhammapada.

   But inclinations are stronger than arguments. Pâli scholars prefer their Pâli terms, and I cannot blame them for it. Mr. D'Alwis (Buddhist Nirvâna, p. 68) says: 'It will be seen how very difficult it is to follow the rule rigidly. We are, therefore, inclined to believe that in translating Pâli works, at least, much inconvenience may not be felt by the retention of the forms of the language in which the Buddhist doctrines were originally delivered.' For the sake of uniformity, therefore, I have given up my former plan. I use the Pâli forms when I quote from Pâli, but I still prefer the Sanskrit forms, not only when I quote from Sanskrit Buddhist books, but also when I have to speak of Buddhism in general. I speak of Nirvâna, dharma, and bhikshu, rather than of Nibbâna, dhamma, and bhikkhu, when discussing the meaning of these words without special reference to southern Buddhism; but when treating of the literature and religion of the Theravâda school I must so far yield to the arguments of Pâli scholars as to admit that it is but fair to use their language when speaking of their opinions.