________________
272
THE INDIAN ANTIQUARY.
(SEPTEMBER, 1892.
which we now have them, of later date than the labours of the grammarians, and consequently, than the third century.
Are we, therefore, to conclude that the dialects which the schools retouched, had never, before this epoch, been applied to literature ? Such is not my opinion. We shall see, on the contrary, that the use for which several have been specialized, the archaic form which several of them have preserved, can only be explained by the existence of certain traditions, either literary or religious. People composed stanzas in Mabârâshtri before the collection of Hala was written in its present form. Long before the Sinhalese Tripitaka was fixed in the shape in which we now read it, there existed, amongst certain sects of Buddhists, & number of formulæ, rules, and legends transmitted in a dialect in its essence closely resembling the Páli of our books. We must, nevertheless, take care not to exaggerate the accuracy or the importance of these earlier compositions. They must have remained purely oral, or, at most, had only received a written form, which was accidental and ephemeral. A sect, Buddhist, Jaina or other, which possessed, whether written, or even living in a finally established oral tradition, a definite and consecrated canon, would certainly never have consented to alter it by submitting it to a new grammatical remodelling. Moreover, this grammatical retouching must have been at first undertaken in answer to a demand, to give for the new requirements of editing and codification, the instrument which was necessary to them. The fixation and the reform of a dialect peculiar to the sect, which was used for its fundamental texts, can only be conceived as occurring at the date when they were for the first time united in a definitive collection of traditions, which had hitherto been either imperfect or dispersed. If they had been established sooner in a canonical corpus, the language of that corpus would itself have been the law. Its authority would have rendered reform both useless and impossible. This reform would, on the other hand, under the conditions in which it was produced, have been equally inexplicable, if we did not admit previous attempts at editing. Although imperfect and fragmentary, they have, in a general way, marked for each dialect the low-water mark of its phonetic development, and furnished the characteristic traits of its morphology.
It is expressly subject to this reserve that we must understand the conclusions which I have indicated. At the present moment, I am only dealing with a special class of considerations. It is unnecessary to say that there are arguments of another nature which appear to me to confirm these inductions. I here leave them aside, and only wish to point out, en passant, one interesting instance of agreement. There are reasons for believing that the stanzas of Hala represent the most ancient specimen of Prakrit literature. In the course of his learned and ingenious labours on this valuable coliection, Prof. A. Weber13 has proved that the third century is the earliest date to which it is possible to refer it.
I have now replied, so far as the documents on which I depend appear to allow, to this first question; - at what epoch did the Literary Prakrits begin to be fixed and to establish themselves? We should also like to know how and under the influence of what circumstances this blossoming forth took place.
This question has hitherto been treated as a simple problem of linguistics. Each dialect has been considered as having been, at the epoch when it received its literary form, a spoken and living idiom. Taking this principle as a foundation, a series based solely on phonetic comparisons has been converted into a chronological scale. I have protested against this confusion, and indicated why, in my opinion, we must discard a criterion which has been adopted with too ready a confidence.
The literary elaboration of the Prakrits cannot have been earlier than the second or third century. It has been in no way proved, and, indeed, it is hardly probable, that it should
13 Wober, Das Saptasatakar des Hala, p. xxiii.