________________
228
THE INDIAN ANTIQUARY
[ SEPTEMBER, 1918
thus giving ay, av as the causative principle of the broad sound." Translated into practice, this means that O. W. Rajasthani karaï to become modern Marwari-Gujarati karè, had to pass through the stages: karaya >karay, the entire process being as follows:
(1) karai (2) karaya >(3) karay>(4) kare.
There is no room for the karai of the manuscripts here, but this is no stumbling-block for Mr. Divatia as he has already disposed of the inconvenient form karai by denying its existence. Anyhow, one would like to ask, in what does karay differ from karai? For it is clear that it must differ in something, otherwise the third phase would represent no progress in respect to the first. My opponent's reply to this question can hardly be expected to be any other but this: that the last letter of karaï is a distinct i separated by hiatus from the preceding a, whereas the last letter of karay is an indistinct i attached to and forming one syllable with the preceding a. Well, if it is so, is this not tantamount to admitting that the second syllable of karay is a diphthong? And if it is a diphthong, is not ai its proper expression?
I think I can guess whence Mr. Divatia's idea of the intermediate phase aya, ava has sprung from. He has seen that in modern Gujarati the ai, au of tat-samas (e. g. daiva, gaurava) is pronounced differently from the è, è of tadbhavas, while on the other hand aya, (ava) of tatsamas and semi-tatsamas (e. g. samaya, paya, nayana, kavari) is pronounced very much like è, ò, and has concluded that aya, ava are akin to è, ò, and ai, au remote from it. If this was Mr. Divatia's line of thought, he has made here a double mistake: firstly in assuming that tadbhava ai, au were necessarily pronounced in exactly the same way as tatsama ai, au, 10 and secondly in imagining that aya, ava are correctly written in all cases when they are pronounced è, ò. Forms like samaya, paya, nayana, etc., as are commonly met with in O. W. Rajasthani and modern Marwari-Gujarati, are really incorrect spellings for samaï, paï, naïna or samai, pai, naina, respectively. In my article mentioned above I had suggested that in the case of all these tatsamas or semi-tatsamas the transition of aya to è must have taken place through the intermediate step aï, but I had been unable to adduce any instances of this passing of aya into aï then, as up to the time of writing that article I had met with none in the manuscripts I had examined. Since then I have found many instances of aï<aya in the Râu Jaita Si rau Chanda by Vithû Sûjò, a Diùgala poem, whereof a copy dated Samvat 1629 is preserved in the Darbar Library in the Fort of Bikaner, 11 and in a few other manuscripts.
Continuing, Mr. Divatia quotes some etymologies which in his opinion prove that O. W. Rajasthani aï, aii when accented on the a, give aya, ava and hence è, ò in MarwariGujarati, and when accented on the i, u, give é, ó. Unfortunately, a large proportion of these etymologies are incorrect, and some instead of proving what they are intended to prove, prove exactly the contrary. To point out only a few inaccuracies:
aneru is not from anaïrau, but from annaerü (Ap.),
the resultant of maigala is not mégala, but mègala,
verè does not come from vaïraï but from *viaraï (Ap.)>*virai, verai (O. W. Raj.), Rathora is not from Rathaûra, but from Rathauṛa.
10 Probably they were pronounced in a way similar to the ai, au of Hindi.
11 See Descr. Cat. of Bard. and Histl. MSS., Sect. ii, pt. i, No. 15. The instances include: pai (<paya), has (< haya), maigala (<mayagala), haivara (<hayavara), vijas (<vijaya), and other similar forms. Forms like these are also met with in the adespotic Jaita Si rau Chanda, whereof a copy dated Samvat 1672 is likewise found in the Darbar Library at Bikaner.
