Indian Antiquary Vol 47

Author(s): Richard Carnac Temple, Devadatta Ramkrishna Bhandarkar
Publisher: Swati Publications

Page 243

SEPTEMBER. 1918 1
E AND O IN MARWARI AND GUJARATI
227
that languages " go backwards and forwards”, but the generally accepted principle is very clifferent from this. 8
My third and last argument was that when the Marwari and Gujarati scribes found that the spelling aï, aü no longer corresponded to the actual pronunciation, they did not alter it into aya, ava, but into ai, au. Evidently, by the time when ai, au were introduced into use about the sixteenth century A.D.--the two elements in the vocal compounds aï, aü had been blended together into diphthongs and were then pronounced as diphthongs. This is, perhaps, the strongest and most decisive of all my arguments in that it proves that during the period of transition from O. W. Rajasthani to modern Marwari-Gujarati, if not earlier, the tendency of the language was to fuse the two elements in the groups aï, ai into one, not to divaricate them further by amplifying them into aya, ava. But Mr. Divatia easily disposes of this argument by refusing to believe that early Gujarati manuscripts contain the spelling ni, au. Even if this was the case, it would suffice to know that the spelling is found in Marwari manuscripts, but that it is found in Gujarati manuscripts as well is a matter that can be easily ascertained by Mr. Divatia himself if he only cares to complete his researches, which, as he states, are "limited in extent in this respect."
In conclusion, none of Mr. Divatia's replies to the arguments given by me against his theory, does really hit the point, much less can these replies demolish my criticism. However, Mr. Divatia has satisfied himself if no others, and thinking that he has cleared his path of all obstacles, proceeds on. I shall not follow him into all his details, but will confine myself to examining the two or three main points in his discussion and conclusions. He begins by suggesting that if aya, ava (as developments of aï, aü) were not actually written, except in a few cases, "they were potential developments as precedent conditions requisite for the production of the wide sound (è, ô) which comes on the final a being lost through want of accent
• The other examples with which Mr. Divatis tries to show that a va of the Apabhramsa after becoming u in O. W. Rajasthani can revert to va in Gujarati, are : desdura > desavara, deula > devala, and deura > devara. Here the reversion of the samprasarana is only apparent. In several old Marwari manuscripts (e.g., MS. No. 15 of Descr. Cat. of Bard. and Histl. MSS., Sect. ii, pt. i, Samvat 1615-34). I have found tho spelling vu for u coming after a long vowel. Thus: ravu for rau, no rârula for rdula, vituli for vdu!, Sekhavrata for Sekhi uta, etc. Evidently, we have here insertion of va-śruti between 1 and the preceding long vowel, and it is this va-sruti that has given rise to the modern va. Thus 0. W. Rajasthani deula first becomes devula, through insertion of va-sruti, and then, by dropping the u, devala. There is no question of reversion of samprasdrana hero.
Not only is the spelling ai, au found in early Gujarati manuscripts, but it is very often found side by side with the old spelling ai, ai, a circumstance that shows better than anything else that the former spelling is the immediate successor of the latter and that there are no intermediate steps like aya, ava between them. Here is an illustration of the above-mentioned case, taken from the first page of a manuscript in my possession, written, apparently, towards the middle, if not the end, of the seventeenth century A.! . and containing a Gujarati bálavabodha to a “Jambucaritra," a Jain work : s a ET 3977 सांभले छंद तिवारह एक देवता महडिक भगवंत नई बांदई..... बांदी प्रश्न करई माहरी देवलोका विडां केतलो
ret . .. ., etc. It will be noticed that in the above extract, ai, ai are used side by side; whereas au is constantly represented by o. This is not a mere graphic peculiarity of the manuscript, but it is a general fact that while early Gujarati manuscripts as a rule always represent d by ai, they very seldom represent d by au, but either use the old form aü or the newer form o. Marwari manuscripts are more consistent in this respect and use both ai and au. The Gujarati manuscript cited above is only one of many I could cite in which ai is used side by side with ai. Indeed, the practice of writing ai is so prevalent in early Gujarati manuscripts that I am very much puzzled to explain how & Gujarati roholar can assert that he has never come across any instance thereof!