________________
NOVEMBER, 1882.]
BOOK NOTICES.
335
Southern group or Marả thi. The Hindi is thus divided between the eastern and western groups, its two branches having more affinity with the Bangali on the one side and the Gujarati and Panjabi on the other, than they have to each other. The literary works which have reached us show that each of these groups, now broken up into numerous dialecta, formed about the twelfth or thirteenth century only one language. In examining the principal characters of these four languages we perceive that the northern group approaches the western, whilst that of the south is in closer relation with the eastern, and that consequently, at an earlier date of which the Asoka inscriptions have left us some memento, the four groups were reduced to two, which correspond to the Prakrits, Sauraseni and Magadhi,-not to the varieties of these names which grammarians speak of-(these are literary languages and more or less artificial), --but to their common dialects or Apabhramsas; and here again it is necessary to understand not the Apabhramsas of the grammarians which are themselves more or less artificial, but the true Apabhramśas which have perished, unless the Pali has preserved for us the form of one of them. As for the Maharashtri, it is a variety of the Sauraseni, that is to say, of the western group; it has nothing but the name in common with the actual Marathi, for which one often finds it pass, and its character on the contrary assigned to the eastern or Mågadhi group. Alongside of these Aryan languages, spoken by Aryans, a certain number of patois have grown up among the non-Aryan populations. These are the dialects termed Paisachi, which perished early, and of which the Paisachi of the grammarians has preserved us only certain features. These two languages, the vulgar Saurasent and vulgar Mågadhi, have both come from the west, the more eastern, the Mågadht having preceded the other, seeing that it has left traces all along the route to the valley of Kabul and even beyond. The other and later, the Sauraseni, has not advanced further towards the east and south than the actual limits of eastern Hindi and Marathi. In their course, they were only different dialects of one language, which, raised to the state of a literary language, is represented by the Sanskrit.
Such, in substance, according to Dr. Hoernle, is the history of the Aryan languages of India.' On several points, as for example on that which
relates to the respective position of the Maharashtri and the Marathi, it is quite new. The whole is charming for, at first sight, it appears simpler and better united than any that has yet appeared. Is it however free from all objection Without entering into an examination which would carry me too far, and which, to be complete, would encroach on ground where I do not feel myself competent, I ought to say that the conclusions of Dr. Hoernle appear very strong, considering the nature of his data. By the preceding summary alone one may see how many essential terms have disappeared from the series, instead of which he has only suggested approximations, which he himself suspects, and which he inserts as if confident of them. At the commencement Dr. Hoernle goes on solid ground: he has to deal with languages actually spoken; but when he goes back into the past (and that is one objection which, in passing, bears sometimes on his derivations) he deals only with literary languages, or, worse still, with languages which have served as a medium for religious movements. Because the oldest Vaishnava kirtans are neither in Bangali nor Hindi but in an idiom which partakes of both, because the western Hindi, Gujarati and Panjábi, are mingled, so to speak, in the poem of Chand, does it follow that there were then only two languages spoken from the mountains of Afghanistan to the Doab, and from the Doab to the Asam Hills? Do even the Asoka inscriptions really authorise the conclusion that a single idiom reigned at that epoch from the sources of the Jamnå to the mouths of the Ganges ? To answer these and other like questions negatively does not upset Dr. Hoernle's historic theory, but it lessens to some extent the rigour, precision and simplicity of it. Besides, do we not know how delicate the classification of dialects is, even when dealing with dialects actually existing P Dr. Hoernle finds for example that the Marathi agrees with the eastern group in four points and with the western group in eight; but the proposition is reversed if we consider the true characteristics, according to him, of the two groups. Marathi then agrees in four points with the Eastern and in only two with the Western group. This is enough to rank it among the Eastern languages, and as Maharashtri is ranked among the Western, no relation is allowed between them. Is it necessary to add that there is always something arbitrary and consequently hap-hazard in this kind
• The name according to Dr. Hoornle is to be regarded as qualificative—the language of the great kingdom.'
"The same viows indeed, supported by other arguments, have been presented by Dr. Hoernle in the preface to his beautiful edition of Chanda : The Prdkrita-Luksh. nam, or Chanda's Grammar of the Ancient (Araha) Prá.
krit, Part I. Text, with a Critical Introduction and Notes Calcutta, 1880 (Biblioth. Indica.) Of another publication by Dr. Hoornle on the same subject, but written probably with a view to a public less special,- A Sketch of the history of Prakrit Philology (Calcutta Rev., October, 1880), I know only the title.
