________________
JUNE, 1892.]
THE INSCRIPTIONS OF PIYADASI.
173
this is a mistake, and that the deductions, which have been formed on this basis, are altogether unfounded. A priori it would be extremely surprising that a single dialect should have reigned, without rival and without shades of difference, throughout the whole of north and north-eastern India, from Khálsi to Jaugada, by way of Bairât and Rapnath. Our scepticism will be found to be strengthened by several particular reasons.
According to this theory, at the time of Asoka, both at Dhauli and Jaugada, as well as at Rûpnâth and Allahabad, people must have employed a dialect which made nominatives masculine of a-bases terminate in é, and which changed r into l. This I shall, for the sake of brevity, term the Magadhi of Aboka. Now the inscription of Khandagiri, quite close to those of Dhauli and Jangada, the date of which cannot be fixed with accuracy, but is certainly not more than a century later than the monuments of Piyadasi, and which appears to emanate from a local sovereign, makes the same nominatives in ô and the locatives in é, preserves the etymological , and in a word presents none of the characteristic traits of this dialect. We are thas led to think that Asoka's dialect was not that of the country. The ancient inscriptions of che Bharhut stúpa, mid-way between Rûpnâth and Allahabad, perhaps contemporary with Piyadasi, of a surety not much later, and which are certainly expressed in a language analogous to the local idiom, present no trace of Magadhism. So also at Sanchi; yet General Cunningham has discovered there a fragment of an edict which, with a probability almost equivalent to certainty, he attributes to Piyadasi. Now, in this, fragmentary as it is, the nominatives in é, words like chilathitike leave no room for doubt. It was written in Magadhi. But all the native inscriptions found in the same looality, either contemporary with it, or belonging to a very nearly contemporary epoch, agree without exception in the use of a Prükşit free from Magadhisms. In the other localities we are not so fortunate as to be able to use parallel monuments for controlling the apparent evidence of those of Piyadasi, but these facts are sufficiently significant. Evidently, the use of the Magadhi djalect in his edicts does not prove that it was current and in vulgar use in the localities where they have been found. The conclusion readily presents itself to our minds, It was in Magadha that the head-quarters of Piyadaei's empire was situated. Magadhi must have been the language of his court, and nothing can be simpler than to suppose that he used it throughout the extent of his dominions to address his people, and moro spocially his officers, the representatives of his power,
But then, it will be said, how is it that the inscriptions of the extreme north-west and of the coast of Surashtra escape this common level ? The question appears to me to be capable of two explanations, each of which strengthens the other. No one, I think, doubts that it was in the north-west and west that a graphic system, adapted to the necessities of Indian languages, was first elaborated. At least the inscriptions of Kapur di Giri and of Girnar testify that in each case there had been already constituted a peculiar graphic system with its own traditions.
1 At the other extremity of India, in Ceylon, we find a sign which favours this theory. However great, 48 regards details, may be the exaggerations of the Sinhalese traditions with reference to the conneotion of Aboka with Tamraparpi, the testimony of Piyadasi himself would appear to indionte that he held oertain relations with that distant island. That he profited by these relations to help forward the diffusion of Buddhism, his seal and the analogy of his conduct elsewhere do not permit us to doubt. It is hence the more interesting to follow up the traces, which have, in several instances, been pointed out, of the infuence of the Magadhf dialeot on the ancient language of Ceylon. The moat ancient insoriptions which have been found in the island are without doubt of sensibly later date than Piyadasi. This interval explains the alterations which the MAgadht tradition has under. gone from the time of the earliest insoriptions. The fact itself of ita introduction, which it is diftioalt to refer to any author except Piyadasi, only stands out the more clearly from the persistence of certain traita, I do not speak merely of grammatical peculiarities: the locative in si, nominative in , &0., which have been pointed out by P. Goldschmidt (Ind. Ant., 1877, p. 318, of. Rhys Davids, Ind. Ant., 1879, pp. 188 & ff. d. Mäller. Ancient Inseript. of Ceylon, p. 8; and the recent observations of Prof. Kern in the Bijdragen tot de Tral, kunde van Vederl. India, IV. 10. p. 562). Two palieographic facts are equally characteristio. One is the adoption of the siam A before its limitation to the palatal $ (see below), and the other is the absenos of the palutal, not employed in the official writing of Piyadasi, and which we see, for example, in the inuoription of Kivindo (E. Müllor, No. 57) expressed by the compound ny, in savanyutipeta. It in, therefore, probable that Piyadasi had dirootly or indirectly transferred to Ceylon, as he had done to the provinces of his empire, the methode pooullar to hie Magadht system of orthography,
