difference between the dialect of Gujerat and that of Cuttack. In the former we find bhavati, asti, is ; anusasti, command ; dpta, fit ; following closely upon the Sanskrit etymology : whereas in the latter we have hoti, athi, anusathi, atta, as in the modern Pali. It would be a process of inversion indeed to derive the former from the latter, while we have the instances of French, Italian and Latin before our eyes.
The dialect of Girnar, then, is intermediate between Sanskrit and Pali, or rather the pillar idiom ; for Pali, so called, agrees in some respects bet- ter with one, in some with the other, and in orthography decidedly with neither !
Thus the word idha used at Girnar for ^ iha, * here,' is correctly the Pali term as may be seen in the long quotation about the erection of a stupa in Ceylon inserted in last month's journal.
The corresponding word in the eastern dialect is curiously modified to hida, a fact I only ascertained by the collation of the two texts, and one which at once opens an important discovery to aid our studies. In several of the Dhauli inscriptions the expressions hidalokika pdralokika, —hidaloka paraloka, occur : at Girnar (13th tablet) we have also ilokikd paralokikd cha : — all these are evidently T^wfarsFT ^TCirrf^rafT^
- of this world and of the next world.' Now the opening of the pillar
inscription which so much perplexed us has the same elements hidata palate — ^ VK or T^TTKrP here and hereafter, a sense which at once renders the passage intelligible. The same may be said of hida- takaye pdlatakaye in the north compartment.
The eastern dialect is remarkable for this species of cockneyism which, as far as I know, has no parallel in any of the grammatical Pra- krits: thus the h is inserted before evam fhevamj, idam and soma other words beginning with vowels.
On the other hand (but this is also a cockneyism) the semivowel y is cut off in many words such as athd, add, ata, am which are correctly spelt at Girnar, — yathd, yadd, yata (S. yatra) and yam. In these instances the pillar language is remotest from the Sanskrit. There is a singular exception however in the feminine pronoun iyam (S. t4) which is pre- served throughout at Dhauli and on the pillars ; whereas at Girnar, ay am is made both masculine and feminine, as in modern (or rather written) Pali.
There cannot be a better test of the gradual change of language than the word prati, a prefix in Sanskrit extensively used, implying relation, direction or return. In the Pali of Girnar this is merely altered to pati j ) by omission of the r. In the language of the pillars the same prepo- sition is always written pati, j r with the cerebral *. The orthogra-