________________
APRIL, 1903.)
MAHARASHTRI AND MARATHI.
188
The Prakrit grammarians make one important exception from the rule. At between vowels becomes d in Saurasêni and Magadhi, but is dropped in other dialects. It will be remembered that this change of t to d and the corresponding one of th to dh in Saurasêni and Magadhi is the most striking feature in which those dialects differ from Mâhârâshtri and Ardhamagadhi. Compare Sanskrit jándti, Saurasêni jánddi; Magadhi yanadi; Maharashtri and Ardhamagadbi jándi, he knows ; Sanskrit gata, Sanrasêni and Magadhi gada, Maharashtrt and Ardhamågadhi gua or gaya, gone, etc. In reality, however, this is not a difference of dialect but of time, the Boft consonant being the intermediary step between the hard one and the dropping of the whole sound. To take a parallel from a distant language, every d between vowels is dropped in Norwegian. That language has been largely influenced by Danish, in which form of epeech every t between vowels is softened to a d, while t in the same position remains in Norwegian. In such words, however, which have been borrowed in the Danish form, a d between vowels is dropped even when it represents an old t. Thus Danish lade, from late, to let, Norwegian la.
The change of t to a which is so characteristic of Sacrasêni and Migadhf is exactly analogous to the change of k to g in Ardhamagadhi and Jaina Maharashtrt ;' thus, aadga, aśoks. The grammarian Vararuchi (II. 7) also allows the change of t to d in Maharashtri in some words. It thus follows that the different treatment of a between vowels cannot be taken as the starting point of a classification of the old Prakrits.
The common theory among English scholars seems to be that the Prakrits can be divided into two groups, one western comprising Maharashtri and Saurasóni, and the other, tho eastern, comprising Magadht alone, Ardhamagadhf is then considered as a link between these two groups. Compare Dr. Hoernle, l. o., and Dr. Grierson, Seven Grammars of the Dialects and Sub-dialeots of the Bihári Language. Part I., Calcutta, 1883, p. 5, and in other later publications.
The principal points in which both groups differ are as follows:
1. All o-sounds become in the west and á in the east; thos, Saurasên kdolmah, Magadht késére, in the hairs.
2. Every r is changed to an l in the east; thus, Saurasêni rdá, Magadhi idd, a king. .
8. Every initial; is changed to a y in the east, while the opposite change from y toj is the role in the west. Thus, Saurasêại jdņé, I know; jadhd, as; Magadht ydņ8, I know ; yadhd, as. The manuscripts usually write j also in Magadhl.
4. The nominative singular of masonline a bases ends in 8 in the west and in & in the east. Thus, Saurasêni purisd, Magadht pulis, a man.
Ardhamagadht agrees with the western languages in the three first points, while in the fourth it usually has the eastern form. The nominative ending in d, however, is also used in the oldest text.
It will be seen that the features on which this classification is based are mostly of a superficial kind. The different pronunciation of various sounds cannot properly be taken as the • starting point for a philological classification. We should then, for instance, be justified in
using the different pronunciation of an original D.and the treatment of the old s in the future in Gujarati and Western Hindi in order to prove that both languages belong to quite different groups of Indo-Aryan Vernaculars. With regard to the treatment of 8-sounds in the east we know that it differed in different localities. In Dhakkt, which shares some of the characteristio features of Magadht and would certainly have to be classed as an eastern Prakrit, wo have two 8-sounds, a dentals representing the dental and cerebral 8-sounds in Sanskrit, and #palatal corresponding to Sanskrit é. Thus, pulied, a man; dasa, ten.
Jains Maharashtrt is the dialeat of non-canonical literature of the Artimbar Jains. For our prosent purpones it may be considered m identioal with ordinary McAhArlahtrt
