________________
168
THE INDIAN ANTIQUARY.
[APRIL, 1908.
We thus see that the most important points in which Marathi ngrees with the western (i..., inner) languages are the same which were common to Mâbârâshtri and Saurasên.
2. Agreement between Marathi and the eastern languages of the outer circle.
It has already been pointed on that the similar pronunciation of the palatals in Marathi and some eastern dialects cannot be urged as proving any closer relation. The same is the caso with the broad pronunciation of a short a in Konkani, which might be compared with the pronunciation of the same sound in Bengalt. A short a is pronounced in the same way in several dialects of Gujarati Bhilt, and is due to causes which we cannot as yet account for. It is foreign to Marathi proper, as are also the short e And o which are so characteristic of eastern languages and of Konkani. E and o had a short as well as a long sound in all P akrits, and the abolishing or retaining of the short variety is due to phonetical developments within the single dialect,
The d-nominative of strong masculine bases has already been dealt with, and it does not seem safe to lay any stress on it.
In the same way I do not venture to make an argument ont of the so-called pronominal suffixes which Dr Grierson has shown are need in most languages of the outer family, because I have not been able to convince myself of their existence in Mara hf. Compare Dr. Grierson in the Journal of the Asiatic Society of Bengal, Vol. LXIV., Part I., 1895, pp. 886 and ff., pp. 352 and ff.
The l-future, which Marathi shares with Rajasthani and some northern dialects, has been compared with thu l-present in Bibart by Dr. Hoerple and others. Its existence in Rajasthani, however, shows that it is not an exclusively eastern form, and I shall not, therefore, do more than mention it.
The principal points in which Marathi agrees with eastern forms of speech, and which seem to be of fundamental importance, on the other hand, are the oblique form in d and the l-sofix of the past tense. These features pervade the whole inflexional system of the language, and are accordingly of sufficient importance to prove a closer connection.
With regard to the former point, the oblique base ending in é, it must be noted that this form is much more common in the east than has been recognised by the grammarians, and forms an essential feature of those languages. Compare Bihat pahar, a guard, oblique pahará, Maratht, which is a much more conservative language than its neighbours towards the cast, uses this form in all a-bases; thus, bdp, a father; bapd-tad, of the father. The form bápd cannot be separated from båpas, which is usually a dative, but is also, dialectically, used as an ordinary oblique base. Thus bdpds-na, by the father, in the Konkan. Both forms must be derived from the old genitive; compare Maharashtri bappassa, of the father.
The L-suffix of the past tense is a secondary suffix wbich was originally added to the past participle passive. It was not originally necessary, and even at the present day it in occasion ally dispensed with, not only in the east, but also in Marashi dialects. Tbns Chitpawant máyard and mdrild, it was struck.
The 1-8ffix must be derived from & Prakrit form containing a doublo II, it being an invariable rule in Markth that every single non-initial / becomes a cerebral !, while & dental I in tbe some position goes back to a double Il in Prakrit. Compare a paper by the present writer in the Journal of the Royal Asiatic Society, 1902, pp. 417 and ff. The prototype of the 1-goflix is, therefore, the suffix illa which plays a great rôle in Maharashtri and Ardhamagadht, and probably also in Magadbi, bat is not used in Sauraseni. The oldest instance of its use in the past tense is Ardhamågadhi ánillia, brought.
