A Note on the Mahāvastu

By Dr. A. B. Keith, D.C.L., D.Litt.

[i] An indispensable preliminary to our comprehension of Buddhism is the careful analysis in detail of the great texts which mark for us at least important stages in the development of Buddhist thought. The Mahāvastu, though it is not unworthy of translation, is nevertheless one of those works which can for practical purposes be adequately and effectively represented by something more brief, in which what is valuable is less obscured by masses of unimportant detail, and we are fortunate in that Dr. Bimala Churn Law has undertaken the heavy burden of making available to us in English the substance of the Mahāvastu. Mega biblion, Mega kakon, Editor’s note: a saying of Callimachus, meaning something like ‘Big book, big evil’. was the verdict of Greece, and, if the greatness of the Buddha demanded so vast a treatise, we epigonoi, to whom Buddhism is not the end of striving and the sole abject of interest, may be excused if we prefer something more condensed and vital.

Of interest and importance of the Mahāvastu there is no doubt; it has the fascination that almost everything that is definite has yet to be discovered regarding it. Its essential character remains a matter of dispute. It claims to be a book of the Vinayapiṭaka according to the text of the Lokottaravādins, a branch of the Mahāsaṅghikas, and this [ii] claim fortunately does not contradict any evidence otherwise available. But, as we have it, the work gives hardly any of the rules of the Buddhist community, which we expect from its claim to be its content. It corresponds instead to that part of the Pāli Vinayapiṭaka which tells of the history of the coming into being of the community, and we must hold either that the Vinaya itself was still preserved by the school in some older form of speech, or that it originally followed on the Mahāvastu, as we have it, but for some reason or other has perished. Either suggestion is open to serious difficulties but it is clear that it would be absurd to suppose that the Lokottaravādins were content to regard our Mahāvastu as representing the Vinaya of their school, even if we admit that the extent of the Mahāvastu suggests that they were lacking in a sense of proportion.

The tradition which regards our Pāli texts as representing the true doctrine of the Buddha treats the Mahāsaṅghikas as the first schismatics; but this powerful school, whose strength is indicated by its name, would doubtless have repudiated any suggestion that it did not preserve the true doctrine, and the same view would have been maintained by the school of the Lokottaravādins, or Transcendentalists, of the Middle Country, the land between the Himālaya and the Vindhya. To them the Buddha was no mortal teacher, differing in nothing save superior insight from ordinary man; seers like the Buddhas stand above the world, though they [iii] accommodate their actions to its life; if they wash their feet, yet dust stains them not, if they sit in the shade, yet the rays of the sun oppress them not, if they eat, yet hunger assails them not. It is in entire harmony with this view of the nature of the Buddha that full acceptance is accorded to the doctrine of the existence of Buddhas on a wholesale scale, and that we are assured that the purity of the Buddha is such that reverence paid to him suffices to win Nirvāṇa, while abundant merit can be attained by the mere circumambulation of a Stūpa and the paying of homage by offerings of flowers. A Buddhānusmti (i. 163 ff) pays adoration to the Buddha in the same fashion as do the Brahmanical Stotras of Viṣṇu or Śiva, reminding us of the strong underlying theistic tendency of Indian thought. A feature little dwelt on in the Pāli Canon appears in the elaborate account (i. 63 ff) of the ten stages through which a Bodhisattva must pass in his advance to Buddhahood, a theme developed in the Mahāyāna texts, and, as in those texts, we are told that Bodhisattvas are not of mortal birth, but come spontaneously into existence as the result of their own merits.

But taken all in all, it is not so much for novel philosophic views that the Mahāvastu is attractive as for the rich variety of its literary content. There is in it no trace of arrangement by a master hand; confusion, practically inextricable, reigns, and the main theme, the legend of the Buddha, is constantly broken by the interpolation of a [iv] Jātaka or Avadāna, or even a dogmatic Sūtra. Repetitions abound; the tale of the Buddha’s birth is repeated no less than four times, and the same episode or Jātaka may appear first in prose, then in verse. Thanks to this wealth of material we can find abundant parallels to important passages already known to us in the Pāli Canon. Thus we have an old version of the famous legend of Siddhārtha’s departure from his home, parallel with the account in the Majjhimanikāya (26 and 36); there are versions of the Pabbajjā, Padhāna and Khaggavisāṇa Suttas of the Suttanipāta, of the Khuddakapāṭha, of the Sahassavagga of the Dhammapada, of the great sermon at Benares, of the Mahāgovinda Sutta of the Dīghanikāya, of the Dīghanakha Sutta of the Majjhimanikāya, etc. We have old ballads regarding the birth of the Buddha, which carry us back to the early days of the faith; we have also parallels for the late Vimānavatthu and Buddhavaṁsa. The Jātakas and similar tales, which make up a good half of the text, present us with the familiar picture of the Bodhisattva as now a king exercising universal sovereignty, now a merchant’s son, a Brāhman, a Nāga prince, or the king of beasts; the stories constantly present close analogy with those in the Pāli Jātaka book, differing often in significant detail, or presenting important variation of substance; thus the Kusa Jātaka is told once (iii. lff) in a metrical version showing close connection with the Gāthās of the Pāli Jātaka (531) and once (ii. 420 ff) with marked variation. [v]

The value of these semi-parallels is very great; thus the tale of Nalinī and Ekaśṅga as it appears in the Mahāvastu (ii. 143 ff) is no more than a pious and edifying legend of no special interest, yet it preserves some traces of antiquity which have disappeared from the Pāli legend of Isisiṅga, in the Nalinikā Jātaka. Other tales remind us of Brahmanical literary tradition and breathe the spirit of the Purāṇas. The childless Brahmadatta (i 272 ff) has recourse to the seers, and three birds are born for him, who speak with human voices and deliver sayings of wisdom; the seer Rakṣita attains as a hermit such magic powers that he can touch the sun and the moon with his hands (i. 283 ff); the royal genealogy of Śākyamuni begins with the normal account of the creation (i. 338 ff). There are parallelisms between the account of the hells at the beginning of the Mahāvastu and that of the Mārkaṇḍeya Purāṇa, which testify to the inter-relation of the Buddhist and Brahmanical schools of thought. From the point of view of literary form we have a certain parallelism between the Jātakas of the Mahāvastu and the Pāli Canon; some of the Mahāvastu legends are in prose, some in prose with verses intermingled, some in prose, followed by verses covering practically the same ground.

Such a work, it is plain, cannot be the product of one period, and it is as necessary as it is difficult to seek to resolve it into its constituent parts. The language offers some aid in the process; the whole [vi] is written in what may for lack of a better term be styled “mixed Sanskrit”, but this form of language is not the same throughout; in part it approximates more closely to Sanskrit proper, and it is doubtless plausible to hold that, as a rule, the more correct the diction the later the text in its present form, as opposed to the matter. This view is based on the belief that “mixed Sanskrit” represents an effort on the part of those, who used in their intercourse and teaching a Prākrit of some sort, to adopt Sanskrit as the language of their sacred texts, their success, however, being at first strictly limited. The alternative view is that which regards “mixed Sanskrit” as representing a genuine development in some form of proto-Sanskrit, and it is a striking proof of our imperfect knowledge of the early development of language in India that we are unable definitely to disprove this view. But, even if we adopt it, it would seem probable that the passages written in a language more closely approximating to Sanskrit are later than those which show more freedom, and the test of language, therefore, is probably valid. Unfortunately however, it carries us but a little way. Nor can we derive much aid from references of a historical character, for these are few and far between. The halo of the Buddha carries the passage in which it is mentioned down to the Christian era, for it indicates influence by the sculpture of Gandhāra. It is much more doubtful whether we can hold that we are carried down to the fourth century A.D. for [vii] those passages in which the Huns occur, Chinese speech and writing are alluded to, and the Yogācāras are mentioned; the Huns are known to the Mahābhārata and we are probably mistaken if we assume that their name penetrated to India only in the fourth or fifth century A.D.; the date of the term Cīna is still uncertain and problematic, and the Yogācāras (i. 120) are in all likelihood not the idealist school which later bears that style. We are on firmer footing when we find the term Horāpāṭhaka applied to an astrologer (iii. 178), for we are reasonably assured that the term Horā in this use was not known to India before the third century A.D.

A far more fruitful source of discrimination between the elements of the work is presented by stylistic considerations, attention to which was first definitely drawn by Oldenberg (Nachrichten von der Königl. Gesellschaft der Wissenschaften zu Göttingen, 1912, pp. l23ff.), in continuation of E. Windisch’s important treatise, Die Komposition des Mahāvastu. He distinguishes between two distinct styles (A and B): the former is marked formally by the predominance of the nominal style, in which the copula is often omitted, and participles in –ta take the place of finite verbs. Such verbal forms as occur are usually in the present, and are often common word such as bhavanti or gacchanti. Sentences are frequently combined by the particle dāni often following some form of derivative of the pronominal base ta, as in so dāni, tatra dāni. The [viii] B style, on the other hand, uses freely finite verbal forms, including many past tenses; the nominal style is rare in narrative, more frequent when a speaker or thinker lays down some statement of fact. Dāni as a particle of connection is rare, while, on the other hand, atha khalu frequently serves this end, and an answer to a question or speech is frequently prefaced by evam ukte. Frequent also is the taking up of a narrative tense by a gerund or participle corresponding. Nor is the distinction between the styles formal only; the A style is far superior to the B style; its movement is free and unfettered, the structure of the sentence changes freely, it expresses much in little space, and in its long compounds it can compress a vast amount of matter, though with the usual risk of ambiguity. The B style, on the other band, is stiff and monotonous; it occupies much space in telling little; it abounds in repetitions, and it lacks the power of subordination; the essential is described with no more emphasis than the idle detail of the mode in which the actors are dressed, meet one another, and sit down beside one another. There can be no doubt as to the true parallels of the two styles; A reminds us of the Tantrākhyāyika, B of the Pāli Canon, and stylistic grounds alone suggest the earlier date of passages in the B style. Confirmation may be drawn from the nature of the subject-matter of the passages in which either style is used. The B style is often found in passages for which we [ix] have Pāli parallels, the A style in passages which suggest additions and workings up of simpler accounts, narratives, or descriptions; it is in an A passage that Horāpāṭhaka occurs. From this we may deduce the conclusion that the Mahāvastu represents the working over of material which is preserved to us often in a simpler and more primitive form in the Pāli Canon, though neither the Mahāvastu nor that Canon can be said to represent faithfully the oldest Buddhist Canon in Māgadhī or Old Ardhamāgadhī.

Oldenberg admits that there has been much admixture of style through working over, and he distinguishes among the B passages a number (B⁰¹) which have no Pāli parallels but represent later work on the B model, such as the first Avalokita Sūtra (ii. 257-293). In his view Sūtra and Vinaya materials of this kind, in part parallel with the Pāli texts, in part later developments in the old style, were worked up into a whole with additions by an author using style A. The suggestion is attractive and ingenious, but the facts are extremely complex, and we may seriously doubt whether the redaction of the main body of the work apart from smaller interpolations such as the second Avalokita Sūtra (ii. 293-397) which admits itself to be a supplement, was the work of one hand. But for decision on these and on many other points the time is not ripe; least of all can we say how far the doctrines of the Mahāvastu as it stands represent the views of the Lokottaravādins pure and simple, how far they are [x] Mahāyāna ideas adopted by that school or simply interpolated in their great text. Sufficient has been said to show how important and how interesting the study of the Mahāvastu is from the point of view of philosophy and literature alike.