A
DICTIONARY
of the
PĀḶI LANGUAGE
Robert Caesar Childers
PREFACE
The Pāḷi language is one of the Prakrits, or Aryan vernaculars of ancient India.[1] It was spoken in the sixth century before Christ, and has therefore been a dead language for considerably over two thousand years. I see no reason to reject the Buddhist tradition that Pāḷi was the dialect of Magadha,[1] and that it was the language in which Gautama Buddha preached.[2] Originally a mere provincial idiom, the Magadhese tongue was raised by the genius of a great reformer to the dignity of a classic language,[3] and is regarded by Buddhists with the same feelings of veneration with which a Jew of the present day looks upon the language of the Pentateuch. A language is generally what its literature makes it. Had Gautama never preached, it is unlikely that the Magadhese would have been distinguished from the many other vernaculars of Hindostan, except perhaps by an inherent grace and strength which make it a sort of Tuscan among the Prakrits. The existing Pāḷi literature is of great extent and importance; it is valuable alike to the philologist, the historian, the student of folklore, and the student of comparative religion. A considerable portion of it is known to us in outline, but only the merest fraction has as yet been published textually. It may broadly be classed under three heads: first the Buddhist Scriptures, which are the oldest Buddhist writings extant; secondly the commentaries of Buddhaghosha, which date only from the fifth century a.d., but are based upon records of great antiquity; and thirdly, historical, grammatical and other works, varying in date from the second or third century to the present day.
The Buddhist Scriptures are called Tripiṭaka, "The Three Baskets or Treasuries," and are divided into Vinaya, Sūtra and Abhidharma, or Discipline, Doctrine[4] and Metaphysics. The Vinaya Piṭaka contains the laws and regulations of the Buddhist priesthood, and forms a great code of monastic discipline; it is besides rich in history and folklore, and contributes innumerable details of the life and ministry of Gautama. The Sūtra Piṭaka consists chiefly of sermons preached by Gautama, and in some instances by his apostles,[5] but it also contains other matter, as the Jātaka tales, the Niddesa attributed to the apostle Sariputra, and Theragāthā, a collection of stanzas uttered on different occasions by eminent saints. In the Abhidharma we find metaphysics pressed into the service of religion: it introduces no new dogma, but discusses the various doctrines of Buddhism from a metaphysical point of view, employing a terminology of great wealth and precision.[6] The Three Baskets form a canon of Holy Writ, and are invested by the Buddhists with all the sanctity of a canon. They are reverenced as containing the Word of Buddha, and are the ultimate appeal on all questions of belief and conduct. Owing to their great extent, estimated at eleven times that of our own Bible, they are able to treat in great detail of all the relations of life, and the doctrine they contain is consistent throughout and set forth with clearness and logical accuracy.
Upon the important question of the origin of the Buddhist Canon much has been written, and the most conflicting opinions have been expressed. The time has hardly come for dogmatising on this subject, but the tendency of all recent discoveries is to confirm the Buddhist traditions, which assign to the Canon a venerable antiquity. The Tripiṭaka bears every mark of recension, and according to the Buddhist historians this recension dates from the 3rd General Council of Buddhism, held under the emperor Asoka in the year 309 before Christ.[7] But even this is said to be a mere revival of the first recension which was made in b.c. 543, just after Gautama's death, when his words were fresh in the hearts and memories of his apostles.[8] These high pretensions have drawn down, as was inevitable, the ridicule of many Western scholars,[9] more than one of whom has held the Buddhist sacred books to be late compilations, scarcely even reflecting the teaching of Gautama. But the question has been placed on an entirely different footing since the discovery last year by General Cunningham of the Bharhut sculptures. These sculptures, which belong to the third century b.c., are illustrations in bas-relief of a great number of Buddhist scriptural subjects, and are accompanied by inscriptions in the Asoka character. Both illustrations and inscriptions are, so far as they have been identified, in perfect accord with the Buddhist Scriptures as we now have them, and in one instance a whole sentence, containing a remarkable expression, which is probably a ἃπαξ λεγόμενον, is quoted from the Vinaya Piṭaka.[10]
Next in importance to the Tripiṭaka books are the Commentaries of Buddhaghosha, the history of which is a singular one. When the great missionary Mahendra went to Ceylon in b.c. 307, he carried with him[11] not only the Tripiṭaka but the Arthakathā or Commentaries, — a whole literature, exegetical and historical, which had grown up around the Tripiṭaka during the two centuries and a half that had elapsed since Gautama Buddha's death. After accomplishing his mission of converting the island to Buddhism, he proceeded to translate these commentaries from Pāḷi into Sinhalese, and his Sinhalese version continued to exist in Ceylon for many centuries, while the Pāḷi version disappeared. In the fifth century Mahendra's Sinhalese commentaries were retranslated into Pāḷi by the famous divine Buddhaghosha, one of the most extraordinary men that Buddhism has produced, and this third version is the one we now possess, the Sinhalese original having in its turn disappeared.[12] Buddhaghosha did not confine himself to translating Mahendra, but incorporated other old Sinhalese chronicles existing in his time, and added immense contributions, chiefly exegetical, of his own. Much of the matter his commentaries contain is as old as the Tripiṭaka itself, while like the Tripiṭaka they are rich in history and folklore, and abound in narratives which shed a flood of light on the social and moral condition of ancient India.[13]
The remaining Pāḷi literature is of very varying interest. The mere titles of the books ancient and modern which it embraces would fill many pages, and it will be sufficient here to mention a few of the more noteworthy. First in importance are the two famous histories Dīpavaṁsa and Mahāvainsa,[14] the discovery of which made the name of Turnour illustrious, and which are almost our only authentic sources for the history of India previous to the Christian era. Next in order of interest should undoubtedly be named the Milinda Pañha, or Questions of Menander. Whatever be the origin of this remarkable work, there can be no doubt of its great antiquity, for it exhibits a familiarity with Greek names and places, and records a religious discussion between the Buddhist divine Nāgasena and a 'Yona' king Milinda, who can be identified with certainty with the Bactrian king Menander.[15] The latter lived towards the end of the second century b.c., and is stated by the Greek historians to have ruled over part of Hindustan. Buddhaghosha's Visuddhi Magga or Path of Holiness is next deserving of mention. It may fairly be called an encyclopaedia of Buddhist doctrine, and is a truly great work, written in terse and lucid language, and showing a marvellous grasp of the subject. The Pāḷi grammatical literature is very extensive, and centres around the famous grammar of Kachchāyana, which is unquestionably the oldest Pāḷi grammar we possess, though its exact date cannot at present be even approximately fixed.[16] There are probably as many as sixty or seventy standard grammatical works in Pāḷi, and minor ones even now from time to time issue from the native press in Ceylon and Burmah. Dictionaries in our sense of the term there are none, but in Abhidhānappadipikā we have a vocabulary of nouns of the highest authority, compiled on the model of the Sanskrit Amarakosha by a learned Sinhalese priest of the twelfth century. The Pāḷi Ṭikā's, which form quite a literature in themselves, are commentaries, chiefly exegetical, by different authors, and belonging to different periods. They are of unequal authority, but some of the more ancient ones contain an immense deal that is valuable. They comment not only upon the Tripiṭaka, but upon almost all the standard books, and the Ṭikā's on the commentaries of Buddhaghosha are often of great utility in clearing up obscure passages in those writings. The Pāḷi books on such subjects as prosody, rhetoric and medicine are mostly very modern, and formed upon Sanskrit models. Among doctrinal works may be specially mentioned Sārasaṅgaha, a modern compilation very popular in Ceylon, and Abhidhammatthasaṅgaha,[17] a masterly analysis or compendium of the Abhidharma, by a modern Burmese scholar named Anuruddha Āchārya, whose work shows that the spirit of Buddhaghosha is by no means extinct among his successors in these latter days of Buddhism.
The Pāḷi or southern version of the Buddhist Scriptures is the only genuine and original one.[18] To a great pioneer of science, Brian H. Hoqgson,[19] is due the discovery in Nepal of an extensive Buddhist literature in the Sanskrit language, which at one time was generally considered to present Buddhism in its oldest form. This view is even now not without adherents of deserved reputation, but our increasing familiarity with South Buddhism is rapidly rendering universal the belief that the North Buddhist books have no claim to originality, but are partly translations or adaptations of the Pāḷi sacred books, made several centuries after Gautama's time, and partly late outgrowths of Buddhism exhibiting that religion in an extraordinary state of corruption and travesty.
Pāḷi scholarship is a science of comparatively recent origin, and is the joint creation of two illustrious scholars, a Frenchman and a Dane. Burnouf has left us the splendid legacy of his 'Introduction à l'Histoire du Bouddhisme,' and of his 'Lotus de la Bonne Loi,' and Fausböll, still in the meridian of life, is even now crowning his great services to Pāḷi scholarship by an edition of the entire Jātaka. Among the less eminent Pāḷists the first place is due to the venerable Lassen, and the next to Spiegel, who shares with Burnouf and Lassen the gratitude felt towards a pioneer. Nor must I omit to record the name of Clough, for poor as his Pāḷi Grammar appears to us now, we must remember that it bears the date 1824, and as a grammar remained unsuperseded for more than thirty years. And to come to more recent labourers, I would venture especially to mention the services of that distinguished scholar Albrecht Weber, of Senart the first editor qf Kachchāyana, and of the younger Kuhn, the promise of whose early efforts has been amply fulfilled in his newly published treatise on Pāḷi Grammar. The brilliant erudition of Max Müller has been devoted rather to Buddhism than to Pāḷi philology, but in his 'Buddhaghosha's Parables' he has given a valuable contribution to this study, and one which I trust will not be his last.
If we compare Pāḷi with classical Sanskrit, we find that about two-fifths of the vocabulary consist of words identical in form with their Sanskrit equivalents, as nāga, Buddha, nidāna. Nearly all the remaining words present a more or less late or corrupted form. The change is in some instances slight, as when Sūtra becomes sutta or Prajāpati becomes Pajāpati; but there are extreme cases in which the change is so great that the identity is not at first sight apparent.[20] Words of the above two classes nearly exhaust the Pāḷi vocabulary; but there remains a small though important residuum of forms distinctly older than classical Sanskrit, and found only in the oldest known Sanskrit, that of the Vedas.[21] Nay, I do not feel sure that Pāḷi does not retain a few precious relics older than the most ancient Sanskrit, and only to be explained through the allied Indo-Germanic languages.[22]
It results from all this that Pāḷi cannot be derived from Sanskrit; both, though most intimately connected, being independent corruptions of the lost Aryan speech which is their common parent; but that Pāḷi is on the whole in a decidedly later stage than Sanskrit, and, to adopt a metaphor popularised by Max Müller, stands to it in the relation of a younger sister. If the proud boast that the Magadhese is the one primeval language fades in the light of comparative philology, Buddhists may console themselves with the thought that the teaching of Gautama confers upon it a greater lustre than it can derive from any fancied antiquity.[23]
The parallel between Italian in its relation to Latin and Pāḷi in its relation to Sanskrit, is striking enough to deserve special notice. In the thirteenth century the literary language of Italy, the language of culture and science, was Latin, which however had long died out as the spoken tongue of cultivated society, and was probably reserved for the drama, and for occasions of state and ceremony. The spoken language of Italy was to be found in a number of provincial dialects, each with its own characteristics, the Piedmontese harsh, the Neapolitan nasal, the Tuscan soft and flowing. These dialects had long been rising in importance as Latin declined, the birth-time of a new literary language was imminent. Then came Dante, and choosing for his immortal Commedia the finest and most cultivated of the vernaculars, raised it at once to the position of dignity which it still retains. Read Sanskrit for Latin, Magadhese for Tuscan, Gautama for Dante, and the Three Baskets for the Divina Commedia, and the parallel is complete. There is strong evidence that in Gautama's time Magadha was one of the most important centres of Hindu civilization, and it is far from improbable that its language was the most esteemed of the Prakrits, just as the Tuscan was the most esteemed of the Italian vernaculars. Like Italian, Pāḷi is at once flowing and sonorous: it is a characteristic of both languages that nearly every word ends in a vowel,[24] and that all harsh conjunctions are softened down by assimilation, elision or crasis, while on the other hand both lend themselves easily to the expression of sublime and vigorous thought.[25]
We have seen that historically Pāḷi was a vernacular or language of the people, and this is fully confirmed by internal evidence. A close examination of its grammar and vocabulary reveals all the distinctive peculiarities of a vernacular. At every turn we meet with words like atraja for Sanskrit Atmaja, vimamsā for mīmāṁsā,[26] nisadā for dṛishad, jalābu for jarāyu, pārupana for prāvarana, makasa for maṣaka, aggini for agni, piñja for piccha, bhamu for bhrū, suṁsumāra for sisumāra, — vocables racy of the soil, and dear to the comparative philologist. Again, the artificial regularity of Sanskrit sandhi finds no plaoe in the free and easy prose of Magadha, and though sandhi is certainly used in Pāḷi it is hardly more used than in Italian or English. Another well-known feature of a vernacular is the frequency of double forms, like dvādasa and bārasa 'twelve,' rasmi and raṁsi 'ray,' pappoti and pāpuṇāti 'to obtain.' Not uncommonly these divergencies are utilized to differentiate meaning, as in the case of aññatra and aññattha, the former meaning 'except,' and the latter 'elsewhere,' while their Sanskrit original anyatra has both meanings.[27] Words in common use sometimes even appear under three or more forms, as when agni becomes aggi, aggini, gini, or svāna becomes sāna, soṇa, sūna, svana and suvāna.[28] But by far the most striking evidence of the vernacular character of Pāḷi is its wealth of idiom and colloquial expression. Sanskrit is essentially a formal and scientific language: poetry and the drama, science, philosophy and exegesis, take up almost the whole of its literature, leaving but a small space for the light narrative and conversational writing which alone can make us acquainted with the inner life of an ancient people. But with Pāḷi the case is entirely different. Here a very large proportion of the literature consists of stories of Gautama's ministry among the people, of narratives and dialogues of the most varied description, of sermons addressed to all classes of men, and abounding in homely yet forcible illustrations drawn from the incidents of everyday life. Whole strata of Hindu life and character are opened up and explored which are hardly more than touched by Sanskrit literature, and the colloquial idiom of ancient Hindustan is for the first time revealed to us.[29]
The change which Pāḷi has undergone relatively to Sanskrit, though considerable, is almost wholly confined to the vocabulary. And here the parallel between Pāḷi and Italian stops short, for the latter, owing chiefly to foreign influence, has passed into an entirely new grammatical stage; and even looking only at its vocabulary, it is decidedly in a more advanced stage of phonetic decay than Pāḷi.[30] The losses which Pāḷi has undergone are by no means inconsiderable. Its alphabet is deficient in the vowels ṛi, ṛī, ḷi and ḷī, the diphthongs ai and au, and the consonants ṣ sh and visarga. The dual is lost in both declension and conjugation,[31] and two of the tenses (the Periphrastic Future and the Benedictive) are wanting. Some of the verbal roots are unrepresented in Pāḷi, of others only traces remain, and a host of verbal forms have disappeared. A large number of nouns are also lost, and such agencies as assimilation, vowel-shortening and the elimination of one out of two or more conjunct consonants has brought about a real impoverishment of the vocabulary.[32] But all that Pāḷi loses in one direction, it regains, and more than regains, in another. The dual and the two tenses are easily spared. If some roots are little used, others have sprung into unexpected importance. If many nouns are lost, their place is supplied by a greater number of new ones,[33] while false analogy has brought into existence new verbal forms that may almost be reckoned by thousands,[34] and latitude of phonetic change makes up for all the losses caused by assimilation and other causes.[35] The softening or breaking up of groups of consonants, the dropping of final consonants, the absence of rigid rules of sandhi, the absence of sounds like ṛi, ṣ and au, — all this gives to Pāḷi a softness and flexibility for which we may gladly exchange the stately but harsh regularity of Sanskrit.
To the above brief sketch I have only to add that, with the exception of a very few imported Dravidian nouns like chātīī and chumbaṭa, there is no foreign element in Pāḷi.[36] It is on the whole in the same inflectional stage as Sanskrit, and everything in its vocabulary, grammar and syntax can be explained from the sister tongue.[37] But at the same time it exhibits a remarkable elasticity, a power of enriching itself by throwing out new forms; we may perhaps even detect in it adumbrations of a tendency to pass into a later phonetic stage. What Pāḷi would have become had it run on unchecked in its course of decay and regeneration may be seen from the modern Sinhalese, which springs from an idiom closely allied to Pāḷi, and has long passed into the analytical stage.[38] To a great extent Sinhalese may for practical purposes be viewed as a lineal descendant of Pāḷi, and it has worked out a whole legion of grammatical forms the germs of which may often be detected in Pāḷi, and which make it a rich, though as yet almost unexplored, philological mine.[39]
It now only remains for me to express my thanks to the friends who have lent me their help and encouragement in my studies, and first of all to Dr. Rost, to whom I have dedicated this work, and but for whom I should never have written a line. I am proud to be able to call myself the pupil and friend of that eminent Pāḷist Mr. V. Fausböli. Towards another Dane, Mr. V. Trenckner, a ripe and graceful Pāḷi scholar, I shall ever entertain feelings of gratitude and respect: from the perusal of no single work do I remember to have derived greater advantage at an early period of my studies than from his masterly edition of the first chapter of Milinda Pañha, the manuscript of which (still I regret to say unpublished) was in my hands for several months. I owe a debt of gratitude to my friend Mr. N. Trübner for his enterprise in undertaking the publication of my Dictionary at a time when its success was, to say the least, uncertain; and to my friend Mr. Stephen Austin for the ready zeal with which he has all along seconded my efforts to carry the work quickly and satisfactorily through the press. From three Sinhalese Buddhists I have received valuable contributions in the shape of letters replying to questions on points of scholarship and interpretation. They are, first the priest Dhammārāma of Yātrāmullē, whose premature death in January, 1872, deprived the Buddhist Church of one of its brightest ornaments; next the priest Subhūti of Vaskaḍuvē, well known to European Pāḷists as the able editor of Abhidhānappadīpikā; and lastly the Mudliar L. Corneille Vijēsiṁha, a scholar of much learning and originality. During the progress of this work I have received from almost all communities in Ceylon proofs of sympathy and appreciation, but from none more than the Buddhist clergy, a generous and enlightened body of men, towards whom I am under many and deep obligations.
R. C. CHILDERS.
[1] The true or geographical name of the Pāḷi language is Māgadhī, 'Magadhese language,' or Magadhāsā 'language of the Magadha people.' The word pāḷi in Sanskrit means 'line, row, series,' and by the South Buddhists is extended to mean the series of books which form the text of the Buddhist Scriptures. Thence it comes to mean the text of the scriptures as opposed to the commentaries, and at last any text, or even portion of a text, of either scriptures or commentaries. Pāḷibhāsā therefore means 'language of the texts,' which of course is equivalent to saying 'Māgadhi language.' The term Pāḷi in the sense of sacred text is ancient enough, but the expression Pāḷibhāsā is of modern introduction, and Māgadhi is the only name used in the old South Buddhist texts for the sacred language of Buddhism. The English use of the word Pāḷi is derived from the Sinhalese, who use it exactly as we do.
[2] This tradition is generally dismissed in a very summary manner, on the ground that Pāḷi does not possess the phonetic characteristics of Vararuchi's Māgadhī. It is curious to see those who are to ready to discredit one tradition accept without examination another tradition resting on evidence not a tithe as good. For that Vararuchi's Māgadhī was really a Magadha vernacular is after all only a tradition like the Buddhist one. Considering the great interval that separates Gautama and Vararuchi, the discrepancy may be explained in a way that will suggest itself to those who are familiar with the migrations of languages and the names of languages in historical times. Moreover the Magadha territory may have varied greatly in extent at different periods, and have included several dialects. One of the much-despised Buddhist traditions is that Ceylon was colonised from a district of Magadha called Lāla, which is evidently meant to be an outlying district, or at least not that in which Gautama preached. If then Pāḷi and Sinhalese are both dialects of Magadha, we should expect them to resemble each other closely, while at the same time presenting dialectic differences. That this is actually the case I have shown in my 'Note on the Sinhalese Language,' in the Journal of the Royal Asiatic Society for 1874. So great are the straits to which those who deny the Magadhese origin of Pāḷi are driven that Kern is compelled to declare Pāḷi a literary manufacture. His argument that the Asoka edicts are not Pāḷi, and that therefore Pāḷi cannot be Māgadhī, rests on the assumption that the edicts are Māgadhī.
[3] A parallel will be fouud in the elevation of the dialect of Western Arabia through the influence of the Kuran. Muhammad did for Arabic what Gautama did for Magadhese. See also p. xiii for the influence of Dante's Divina Commedia upon the Italian language.
[4] This is a free rendering, but most of the Sūtras expositions of doctrinc (see art. Suttaṁ).
[5] E.g. the Sangīti Sūtra was preached by Sāriputra.
[6] See the specimen under Viññāṇāṃ. The Buddhist philosophy is of great interest, and has anticipated an immense deal of modern speculation. Curiously enough Buddhism, like the Kantian philosophy, has four great Problems; they are the First Cause (Karma), the Supernatural, the Origin of Matter, and the attributes of a Buddha (Man. B. 9). These four subjects Gautama declared to be unthinkable (achinteyya), and he forbad his priests to dwell upon them, lest they should lose their reason.
[7] In arguing against the historical reality of the three Councils Kern observes that the name Kālāṣoka, 'Chrnological Aṣoka,' is in itself suspicious. But the Pāḷi spellig with the Vedic ḷ (see Dict.) proves that Kāla in this name means 'black' or 'dark-featured,' and so the argument falls to the ground. In answer to another argument of Kern's I may point out that it is not more wonderful that two Aṣokas should have held councils than that two Constantines should have held councils.
[8] Mahavansa states that the sacred books were handed down orally till the first century B.C., when they were committed to writing. This statement has been frequently pointed to as vitiating all the claims of the Tripiṭaka to real antiquity. In 1870 I wrote to the Sinhalese priest Subhūti to ask his views on this point, and received from him a letter dated Nov. 25th of that year, from which I translate the following reply: "There is no Sinhalese or Pāḷi book which tells us anything different from what is said in Mahavansa respecting the time at which they wrote the Tripiṭaka in books. But my own opinion is that though all the doctrines of Buddhism and of the Tripiṭaka began to be written and used in books in the time of king Vaṭṭagāmanī, it is not that the doctrines were not at all written before that time. It is said that in king Vaṭṭagāmanī's time all the sacred books were systematically written down, but it is said nowhere whatever that the doctrines had not been written at all before that period. We are told that at the first Council the pupils of the different Theras undertook to preserve their respective Nikāyas by committing them to memory. They may have done so for the most part, but such as were unequal to the task must have written down the words in books, and used them both privately and even publicly. It is probable that in process of time every one found it was difficult to get on in this way, and then in Vaṭṭagāmanī's time they all assembled in consultation and publicly entered all the doctrines in books." Now the same difficulty exists with regard to the transmission of the Veda, and I will place side by side with the above extract a passage from Whitney's Oriental and Linguistic Studies, putting forward an hypothesis identical with Subhūti's: "Thus while oral tradition continued to be the eioteric practice, writing might still be resorted to esoterically; collections might be made and arranged, treatises composed, texts compared and studied, by the initiated, while the results were communicated to the school by oral teaching, and memorised by the neophytes" (p. 87). See also Böthlingk's short essay in vol. iii. of the Melanges Asiatiques, where the same view is advanced.
[9] Vassilyeff says, "There can, it would seem, be no doubt that Çakyamuni actually existed, but what his actions were, and wherein his teaching consisted, these are questions in dealing with which we cannot rely upon the assertions of Buddhists" (Buddhismus, Germ. ed. p. 10). Elsewhere (p. 9) he says, "The Buddha appears less as a person than as a term or dogma." It would be unfair to press this unfortunate expression too far, but to those who are familiar with the Pāḷi sacred books nothing is more striking than the intense personality of Gautama, as the way in which he impresses his individuality on every detail of his system. A masterly defence of the antiquity of the Buddhist canon will be found in Max Müller's Preface to Buddhaghosha's Parables, pp. x-xxiv.
[10] I allude to the bas-relief representing the purchase of the site of Jetavana and the presentation of the monastery to Buddha. The inscription is, JETAVANA ANĀDHAPEḌIKO KOṬISANTHATENA KETĀ, and a pavilion forming part of the bas-relief is inscribed GANDHAKUṬI. The Tripiṭaka account of the purchase and presentation of Jetavana is in the Chūla Vagga of the Vinaya. It is unfortunately not at present accessible to me, but a summary of it by Buddhaghosha will be found at p. 92 of Fausböll's newly published first volume of the Jātaka, whence I extract the following passages, — Tasmiṁ samaye ANĀTHAPIṆḌIKO gahapati .. JETAVANAM KOṬISANTHĀRENA aṭṭhārasahiraññakoṭīhi KIṆITVĀ navakammaṁ paṭṭhapesi, so majjhe Dasabalassa GANDHAKUṬIṀ kāreei .. Buddhapamukhassa saṅghassa dammīti ADĀSI. For details see my letters in the Academy of Nov. 28, Dec. 6, and Dec. 12, 1874, and of May 1,1875.
[11] Buddhaghosha uses the expression ābhata, which seems to imply that they were brought in writing, and this is confirmed by a Ṭīīkā quoted by Vijesinha, which uses the word ānetvā in the same connection (see Vijesinha's article in the Journal of the Royal Asiatic Society, vol. v. N.S., p. 289, which contains the best information we have on the commentaries),
[12] The disappearance of the Sinhalese Commentaries appears to me to be easily accounted for. The Malabar princes who invaded Ceylon in the twelfth century, and more than one of whom temporarily obtained sovereignty over the island, were the most determined foes of Buddhism, and are stated to have systematically effected the destruction of all the sacred books they could lay hands on. Among these would be the Tripiṭaka books, the Pāḷi Commentaries of Buddhaghosha, and the Sinhalese Commentaries of Mahendra. The two first existed in Burmah, and were replaced, as history tells us, from that country, while the Sinhalese Commentaries existed only in Ceylon, and once destroyed could not be replaced. All we can hope is that here and there a manuscript may have escaped the destructiye fury of the conquerors, and may yet come to light, like a new Codex Sinaiticus, in some secluded monastery.
[13] See art. Aṭṭhakathā.
[14] They belong to the fourth and fifth centuries of the Christian era.
[15] See art. Yono.
[16] Professor Eggeling has shown that some of Kachchāyana'a rules are found almost verbatim in the Sanskrit grammar Kātantra, and his forthcoming edition of that work will probably throw mnch light upon the age of Kachchāyana.
[17] See art. Viññāṇaṃ, p. 577 (a), note.
[18] In an interesting preface to his Pāḷi Grammar, Minayeff says (Fr. ed., p. xlii) that "the early Buddhist literature, orally handed down, must have become modified according to the language of each country." He supports this view by quoting from the Vinaya Piṭaka a saying of Buddha that "the word of Buddha is to be understood by every one in his own dialect." Unfortunately the words thus translated have an exactly opposite meaning. The passage is a very important one, and Dr. Minayeff deserves the credit of having first brought it to light, see his Prātimoksha Sūtra, p. xlii. The following is the correct translation of the whole passage: "Two brothers (came to Gautama and said), 'Lord, at the present time there are monks who have taken orders from various tribes and castes and families, these distort the word of Buddha from its own proper dialect; suppose, Lord, that we render the word of Buddha into Sanskrit' (chhandaso āropema)." Here the comment says, "Cchhandaso āropema means, Let us adopt the practice of recitation in the Sanskrit language (Sakkatabhāsā), like the Vedas." Buddha replies, "Priests, the word of Buddha is not to be turned into Sanskrit, let him who so turns it be guilty of an offence: I command you, priests, to learn the word of Buddha in its own dialect." The comment adds, "Here its own dialect (sakā nirutti) means the Maqadha vernacular as spoken by Buddha" (see art. Nirutti). In proof of his statement that "the word of Buddha long remained oral and was transmitted from mouth to mouth to different countries not in one particular dialect, but in several dialects simultaneously," Minayeff plaoes side by side a number of stanzas from Mahāvastu, a North Buddhist Sanskrit text, and corresponding stanzas from the Pāḷi Canon. An examination however of the two texts makes it quite clear that the Mahāvastu stanzas are merely clumsy translations of the Pāḷi ones, made at a very late period by men who in some instances did not understand the expressions they were translating. Thus unable to make anything of the purely Pāḷi word vanatha, the translator turns vanathaṃ na kayirā, "let him not be lustful," into satatam na gacche (!); again he adopts ujjugatesu unaltered, and the exigencies of metre force him to admit such a monstrosity as çilavantasya (p. xxx). But tbe question was practically long ago set at rest when Burnouf in his 'Lotus de la Bonne Loi' printed a number of parallel passages from North and South Buddhist texts (p. 860); with regard to which I have only to repeat what I have said under my art. Paṭisambhidā, "No one can doubt that one set are translations of the other, and I have difficulty in understanding how any one can believe the Pāḷi to be a translation of the Sanskrit." See also the articles Opapātiko, Reuaggasā, Sakkāyo, Ubbillāpito, Phāsu, Uposatho, Pātimokkhaṁ, Iddhipādo, Upādiseso. At B. Lot. 307 we find the Pāḷi vedhita 'shaken,' which is really from vyath, adopted by the North Buddhist translators unaltered, under the idea of its being from vyadh.
[19] Mr. Hodgson, who has lived to see a new edition of his Essays after a lapse of upwards of forty years since their first appearance, may fairly be called the discoverer of Buddhist literature. His 'Notices of the Languages, Literature, and Religion of Nepal and Tibet' appeared in 1828, while Gogerly's essays began to appear in 1837, and Csoma Körös i's Analysis of the Dulva was printed in the Asiatic Researches for 1836.
[20] E.g. heṭṭhā = adhasthāt, pārupati = prāvarati, alla = ārdra, talisa = chatvāriṁṣat.
[21] The following are some of the Vedic forms in Pāḷi. Infinitive in -tave, as netave, kātave, hetave (from bhū, appearing at Jāt. p. 4, line 1, under the form hetuye). Ger. in -tvāna, as katvāna, sutvāna. The form imassa. the Vedic imasya, as the gen. and dat. from ayaṁ. Gonaṁ, gen. pl.from go, is Vedic, and so is tiṇṇaṁ (trīṇām), gen. pl. from tayo. Vidū is doubtless the Vedic vidus. In Pāḷi div, "the sky" (see Divo), is masc. as in the Veda. Forms like yamāmase, kasāmase, retain the Vedic s, which in classical Sanskrit is softened to h (see Dham. p. 110). The imperf. akā from karoti is the Vedic akat. Pāḷi has the Vedic ḷ. Kuhaṁ is the Vedic kuha with added anuswāra (as in cirassuaṁ, kudāchanaṁ, etc.). It is usual to say that Pāḷi has preserved the Vedic instr. in -ebhis, but this is not really the case, as in all the conjugations we find in Pāḷi that the instr. plural is assimilated to the abl. plural, and buddhehi both instr. and abl. is really the Sanskrit abl. plur. buddhebhyas.
[22] In the oldest Sanskrit we find the secodary (assimilated) form guu, ut in Pāḷi we have garu, to account for which we must go to Greek and Latin, where we find βαρύς and gravis respectively (traces of the original a are fond even in Sanskrit in the derivatives garīyas, agaru, etc.). Again, I cannot help thinking that in the Pāḷi opt. assa, "let him be," we have a true archaic form, corresponding to the Greek εῐη for ὲσίη, and retaining the initial vowel of the root which is lost in the Sanskrit syāt. Again, how is the remarkable form sabba-dhi "everywhere," to be explained? and is not the Pāḷi and Prakrit idha, which we find in the Zend, an older form than the Sanskrit iha?
[23] The authorship of the well-known stanza asserting Pāḷi to be the original language is still unknown. Turnour (Mah. xxvii) says it comes from Payoga Siddhi, a grammar of the fourteenth century; but this is a mistake, for on examining a MS. of that work I find that the stanza is merely referred to, the first pāda only being quoted. It may possibly be in Moggallāna Vyākaraṇa, a twelfth-century work, but I am inclined to think it is yet older. I venture to quote it here:
sā Māgadhī mūlabhāsā narā yāy' ādikappikā
Brahmāno c'assutālāpā sambuddhā cāpi bhāsare.
Which means, "The Magadhese is the original language, in which men of former ages, and Brahma angels, and those who have never heard speech, and supreme Buddhas speak" (assutālāpā = assuta-ālāpā, yāya is instr.). Even Buddhaghosha (reminding one of Herodotus' story) says that a child brought up without hearing the human voice would instinctively speak Māgadhī (Alw. I. cvii).
[24] The principal exception in Pāḷl is that a small proportion of words in every page end in annswāra, which however is not a full consonant like k or d, and is called by Kuhn a 'nasal vowel.'
[25] Nothing can be grander in diction than the well-known passage of Tasso, "Chiama gli abitator delle eterne ombre," etc.; and compare with it the splendid lines, yadā have pātubhavanti dhammā, etc., or paññāpāsādam druyka, etc, (Dh. p. 6). It is strange that no one should have pointed out the remarkable similarity of the latter passage to Lucretins' "suave mari magno." I render it thus, "Climbing the terraced heights of wisdom, the wise man looks down upon the fools, serene he looks upon the toiling crowd, as one that stands upon a mountain looks down upon them that stand upon the plain."
[26] Kern gives these two words as proofs of his theory that Pāḷi is an artificial language; "It is obvious," he says, "that they are clumsy fabrications" (dat zulke woorden gefabriceerd zijn, en wel op zeer onhandige wijze, springt in't oog. — Jaartelling der Znidelijke Buddhisten, p. 15). The fact is that like the othen I have mentioned they are extremely interesting provincial or rustic forms, vulgarisms if you will, which could easily be paralleled from almost any language oriental or western. AtrjA has passed through a form ātnaja, and vīmaṁsā is a case of consonant dissimilation, like takkola, kipila, nalāta, nisaḍā, vitaehchhikā, phṣāsulikā, tikichehhati, and many others. As to appābādha, it is clear (as Burnouf has shown) that the reading apāb- of the inscription is one of the instances in which a single consonant is made to do duty for a double one: alpābādha 'well,' is as good an adjective as alpajña 'ignorant.'
[27] Other examples of differentiation are, assa 'to him,' and imassa 'to this man;' chaṇa 'festival,' and khaṇa 'moment' (both = kshaṇa); ānā 'command,' and aññā 'knowledge' (both = ājña); attha 'thing,' and aṭṭa 'lawsuit' (both = artha); Saṅkhata 'composed,' and sakkata 'Sanskrit;' vattati 'to be,' and vaṭṭati 'to behove;' pavatteti 'to set going,' and pavaṭṭeti 'to roll;' amuka 'this,' and asuka 'a certain;' pabhavati 'to arise,' and pahoti 'to suffice,' etc.
[28] Again, chaluddasa, chuddasa, choddasa = chaturdaṣan; suṇisā, suṇhā, husa = snushā; bhavissati, hessati, hehiti, (anu)bhossati, (pa)hossati = bhavishyati.
[29] Even a cursory inspection of this dictionary will reveal innumerable words, meanings and expressions unknown in Sanskrit. Among new words are: saṅgaṇikā 'association,' paṇṇākāra 'a present,' kittaka 'how much,' Sampavaṅko 'a friend,' vemajjha 'middle,' vevachana 'synonym,' nikkujjita 'overturned,' aññadatthu 'certainly,' kathikā 'talk,' sākachchhā 'conversation,' sahavyatā 'company,' anudisā 'intermediate direction,' santaka 'belonging,' vītisāreti 'to remind,' sappāya 'beneficial,' sārāṇiya 'that should be called to mind,' pāramī 'perfection,' sambahula 'many,' odakantika 'a deep pit,' vebhassa 'bullying,' upaḍḍha 'half,' samaṅgī 'possessed of,' ekaṁsa 'certainty,' and innumerable others. Sometimes it is a new combination of a preposition with a root, as paṭisāmeti 'to put away,' nipajjati 'to lie d»wn,' nibbedheti 'to pierce,' nijigiṁsati 'to covet,' uppaṇḍti 'to ridicule,' vyantikaroti 'to abolish,' paggharati 'to trickle,' pachchupaṭṭhita 'imminent,' uyyuta 'busy,' opunāti 'to winnow.' Sometimes a root or noun is combined with a different preposition, to convey the same meaning, as adhi-ppāya = abhi-prāya, ni-ssaya and ni-ssitm = ā-ṣraya and ā-srita, paṭi-pāṭī = pari-pāṭī, rājābhirāja = rājādhirāja, sachchhikaroti (sākshīkri) = sākshātkri, nibbuddha (niryuddha) = niyuddha, nir-abbuda = nyarbuda (ni-arbuda), ni-gaṇṭha = nir-grantha. Sometimes we have new derivatives of well-known roots, as vachi from vach, ragā from rañj, ravā from ru, āhā from ūh, parittā from paritrā, virūlhi from viruh; or of well-known nouns, adjectives, etc., as veramaṇīī, vāritta, pāramī, orima, heṭṭhima, pahonaka, padipetyya, aññathatta, ottappa, vanatha, daratha, sabbadhi, dāsavya. Among idioms and familiar expressions unknown in Sanskrit are: nimitaṁ gaṇhāti 'to fall in love,' saññaṁ na karoti 'to make no sign,' obhāsaṁ karoti 'to drop a hint,' kheḷo chalalati 'my mouth waters' (Trenckner), mukhaṁ olokkti 'to be a respecter of persons,' ujukaṁ oloketi 'to look a person straight in the face,' khiraṁ muccati 'the milk cnrdles,' niṭṭhitaṁ bhattaṁ 'dinner's ready,' svātanāya nimanteti 'to invite a person to dinner for next day,' kin te aphāsukaṁ 'what's the matter with you?' kīdisaṁ bhadde 'how are you, madam?' sarīri adhimuchchati 'to possess a man' (of an evil spirit). Sometimes the same word has a different meaning in Pāḷi: thus kāṇksh in S. means 'to desire,' in Pāḷi to 'doubt,' itaretara in S. means 'mutual,' in Pāḷi 'any whatever;' psāta in S. means 'eaten,' in Pāḷi (chhāta) 'hungry;' the caus. fr. adhivas in S. means 'to cause to inhabit,' in Pāḷi 'to consent;' sāmagrī in S. means 'goods,' in Pāḷi 'concord;' kākapeya in S. means 'shallow,' in Pāḷi 'brimfull;' nikri in S. means 'to illtreat,' in Pāḷi 'to deceive;' pradhānam in S. means 'chief thing,' in Pāḷi 'effort' (comp, also paṇidhānaṁ); avamriṣ in S. means 'to touch,' in Pāḷi 'to revile;' niyāma in S. means 'restraint,' in Pāḷi 'manner.' Sometimes a new meaning is added to the Sanskrit ones, as when paṇita means 'savoury,' as well as 'exalted;' or when parigaṇhāti means 'to explore,' as well as 'to embrace;' or when vikati means 'sort,' as well as 'change;' or when obhāsa means 'hint,' as well as 'lustre;' or when āvuṇāti means 'to string' as well as 'to cover.'
[30] Grammarians have amused themselves by constructing long sentences to read either as Latin or Italian, and a specimen of this sort of exercise on the part of a Pāḷi grammarian will be found at Alw. I. c.
[31] Ubho 'both' is I think the only unquestionable relic in Pāḷi of the Sansk. dual; pitaro 'parents,' is a plural.
[32] Here are a few of the many examples of two or more different Sanskrit words assnming the same form in Pāḷi. Dosa = dvesha and dosha, eṭṭha = ushṭra and oshṭha, ahosi aor. from hu and from bhu, diṭṭha = dvishṭa and dṛshṭa, rukkha = vṛoksha and rūksha, aṭṭa = aṭṭa, artha and ārta, jhāyati = kshāyati and dhyāyati, achehha = achchha and ṛiksha, vassati = varshati and vāṣyate, ratana = ratna and ratni, muddikā = mudrikā and mṛidhvikā, kvii = kavi and kapi, jeyyo = jyāyas and jeya, bhusa = busa and bhṛiṣa, aññāta = ājṇāta and ajñāta, patta = pattra, prāpta and pātra, sattha = ṣāstra, ṣastra and sārtha, appamatta = alpamātra and apramatta, khipati 'to sneeze' from kshīv, and khipati 'to throw' from kship.
[33] See examples at p. xv, note 1.
[34] Sometimes the older or regular form only is in use, as gacchati, dissati, dassati, bhavati (or hoti). Sometimes the regular form is lost and its place supplied by an irregular one due to false analogy, as pachissati compared with pakshyati. But in innumerable cases regular and irregular forms co-exist, to the great enrichment of the language, as dakkhati and passissati, dajjā and dadeyya. How much poetry gains from double verbal forms may be seen from the use of kayirā and kare at Dh. v. 42, jahe and jahtyya at v. 221, jine and jeyya at v. 103.
[35] See examples at p. xiv.
[36] These two words I have only met with in late texts.
[37] I must of course except an insignificant number of forms like those mentioned at p. xiii, note 1. I have been obliged to leave a considerable number of words unidentified in my dictionary, but as our knowledge increases the list will steadily diminish; and if some words should finally remain unidentified (which is extremely probable) we must remember the vernacular character of Pāḷi, which would explain its possessing many undoubted Aryan words which have not crept into Sanskrit literature. Thus the Pāḷi name for white ant, upachikā, which is almost certainly a derivative of upachi, does not occur in Sanskrit, because, I suppose, the white ant does not happen to be mentioned in Sanskrit literature. The same argument applies to words like karavika, kachavara, nālipaṭṭa, oḍḍeti, niyura, kakkārī, kakaṇṭaka, kaṭṭhissa, pulava, jaiogi, kusi, kukkuha, kukutthaka, and many others.
[38] See p. i, note 2.
[39] See my 'Note on the Sinhalese Language' in Journ. Roy. As. Soc. 1874. I shall not go into the subject further here, as I hope shortly to resume my 'Notes.' A careful study of Sinhalese affords a complete answer to the arguments of those who hold Pāḷi to be a 'fabricated' language.
