________________
DECEMBER, 1917]
THE WIDE SOUND OF E AND O
303
these three instances is preceded by for and not by t; none the less they instance reversion and prati-samprasara na together. The fact is that, as in the case of other changes, 12 this change of 3 back to 13 is found side by side with a different process undergone by the same double vowel. Thus, while it gave it on the one hand, it also gave on the other; 794 gave 774 (the parent of fra) on the one hand, while it gave a (the parent of Ta ) on the other, and so forth. How or why this double operation came into play will be explained further on below. (3) To come to Dr. Tessitori's third and last ground. It is this
(a) and sit, derived from 1 and 3, are found in all the earliest manuscripts
of both Gujarati and Maravadi; (6) When 43-373 began to be written as t-sht, it was because they were
pronounced as diphthongs, and only afterwards they were reduced to
long wide vowels Ti.e., f and ); (c) If अइ-अउ had really passed into अब-भव, manuscripts would have written
them as 372-37 instead of writing them, as they do, as it, especially
as they show a tendency to write - for 3-3. I should like to make my position clear before taking up each of these three sub-heads. But it will be convenient to touch one point under (a) just now, viz., the state of things in earliest Gujarati Manuscripts. So far as I have been able to ascertain, -ir, as evolutes of 37-38, are not seen in Gujarati Manuscripts of any period. Dr. Tessitori puts the rise of Gujarati as a separate offshoot of O. W. Rajasthani somewhere about the beginning of the seventeenth century of the Christian era 1! (i.e. from V. S. 1656 downward). Manu
11 It may be contended that after all this from a direct, and a Sanskrit tatmama. But a careful consideration of the probabilities based on the place of these words in the language as words of such frequent currency as can only be acquired by tadbhava formations, will go against such contention. Besides, gas is only a potential step.
In some cases the co-existence of apparently different stages of formation can be accounted for; e. 9..
अहोतरम बुद्धडी रावण तण कपालि।
एकूबुद्धि न सांपडी लंका भंजण कालि ॥ (Munja Rdsd, quoted from in Sastri Vra jalal's Gujarati Bhashano Itihasa, p. 44.) Here the 45 in T is due to a final termination, while the in s and is is the result of the absence of that termination. Similarly TT er in Bhuwanadipa bhashantara (Gujarat-Saia-Patra, March, 1910, p. 112.)
This would be good in the case of nouns and adjectives. In the case of verbs the and cannot De so explained, and inust be regarded as forins different in nature.
13 I must note that anti-samprasdran (or YTTETETT) does not mean that the which undergo that process have in all cases been derived by samprasierant. They may have been evolved differently as well ; e.g., Af -
-.(72754-27 -40; afe -
- (09)
(5) -- TUT( Tar *T; - H -- ( )-AT; गौरी - Tae - TT (Tu)-fra;
-- T* - 9 (79 ) ; et cetera. All that is meant is a process which is the opposite of samprasarana. (In fact, where is affected by samprasarana, there is no prati-sampreadrans, the 73 remains in hiatus or contracts into a narrow , or forms the diphthong ti e.ge or 3*, qiqrat, (from TT).
* V de Dr. Tessitori's" Notes", p. 5 of the Introduction.
