(1) maintain that the new Sanskrit has lost 153 much which the older Sanskrit had, and which it could only lose from the circumstance that it had died out in the intermediate period, and had now to be revived in a form which might be as intelligible as possible. To this head belong a number of roots and inflected forms which the grammarians recognize and adduce partly as current, and partly as obsolete, but of which the later Sanskrit makes next to no use. The reason of this. is that these roots, as well as these inflected forms, were either entirely lost in the vernacular dialects which existed at the time when the new Sanskrit was created, or had become so disfigured that their Sanskrit form could not have been easily discovered or understood. (2) The new Sanskrit contains in it much that the old Sanskrit could not have had. To this head belong a number of forms of roots which had become modified according to the laws of some one vernacular dialect, and which have been employed in the new Sanskrit in this modified shape, which the grammarians either hesitated to refer to its proper Sanskrit form, or did not comprehend. Every single example of this which might be adduced would, however, require detailed development and proof, which would demand too much space to be here attempted. "I will, therefore, content myself with repeating the main results of the investigations which have been here merely indicated, and in great part yet remain to be carried out. These results are: That from the period when the Sanskrit-speaking race immigrated into India down to perhaps the ninth century B.C., Sanskrit became diffused as the prevailing vernacular dialect over the whole of Hindustan, as far as the southern borders of the Mahratta country. It penetrated no further south as a vernacular tongue, but only as the language of education, and apparently at a later period. From the ninth century B.C. the Sanskrit began to die out: derivative dialects became de

153 The Sanskrit has lost a great many verbal roots, and has frequently modified the original meaning of those still in existence."-Aufrecht, Uṇādisūtras, pref. p. viii. "In the course of time some branches of literature disappeared, a number of words became antiquated, and the tradition as to their meaning was either entirely lost or corrupted. When commentators arose to explain the Uṇādisūtras," -supposed by Professor Aufrecht (p. ix.) to be considerably older than Panini,"they found the greater part of the words contained in them still employed in the literature of their age, or recorded in older dictionaries. But an unknown residuum remained, and to these, whenever tradition failed them, they were bold enough to assign quite arbitrary significations."-Ibid. pp. vi. xii.

veloped from it; and in the sixth century B.C. it had become extinct as a vernacular language. On the other hand, it maintained its ground in the schools of the Brahmans. About the third century B.C., in consequence of the regeneration of Brahmanism in Kanouj, it was brought back into public life as a sacred language, and gained a gradually increasing importance as the organ of all the higher intellectual development. About the fifth century A.D., it had become diffused in this character over the whole of India. So long as the empire of the Hindus lasted, it continued to increase in estimation; and even long after the Mahomedans had settled in India, it was almost the sole instrument for the expression of the highest intellectual efforts."

I conclude this section by quoting from an article by Mr. Beames in the Journal of the Royal Asiatic Society for 1870, vol. v., new series, pp. 149, f., the following remarks on the relation of the Prakrits to the "Old-Aryan" language:

"In assuming that the languages of the first period" (the Prakrits and Pāli) "are later than Sanskrit, I do not lose sight of the fact that historically they were contemporaneous with it. But I call them 'neo-Aryan,' because the majority of their forms exhibit a decadence from some more perfect condition. It is true that, not only in classical, but even in Vedic, Sanskrit forms are found which exhibit a perfect Prakrit type; but this does not prevent the general run of Prakrit from showing unmistakable signs of having degenerated from a purer and stronger ancient language, which we cannot call Sanskrit, because it is older still than even the language of the Vedas, and which therefore may, when necessary, be called 'Old Āryan.'

"It is a highly probable theory that the 'Old Aryan,' like all other languages, began to be modified in the mouths of the people as early as the Vedic period, and that the Brahmans, at a subsequent date, in order to prevent the further degeneration of their language, polished, elaborated, and stiffened it into the classical Sanskrit. We cannot, however, suppose that they brought any new material into the language, but simply that they reduced to rule what was till then vague and irregular, that they extended to the whole of the language euphonic laws which had been till then only of partial application, and so forth; all the while, however, only working upon already existing materials. It will, therefore, not militate against the established con

temporaneous existence of learned Sanskrit and popular Prakrit, to consider the former as in general the representative of the original Old Āryan, and, consequently, as so far older than the Prākrit; because, ex hypothesi, in Sanskrit most of what existed in Old Aryan has not only been preserved, but worked up and expanded, while in the Prakrit, on the contrary, not only has much been absolutely lost, but that which remains has been corrupted and debased. Besides, as nothing whatever of the Old Aryan has been preserved, or is likely to be discovered (although much may be, and has been, guessed at from analogy), we are driven, whether we like it or no, to look to Sanskrit for the oldest extant forms; and we do, undoubtedly, find them there, as contrasted with Prakrit and Pāli." 154

SECT. IX.-Reasons for supposing that the Sanskrit was originally a spoken language.

It appears from the passages cited from the works of Professors Lassen and Benfey, that these distinguished scholars assume that the Sanskrit (by which, no doubt, must be understood a language in some respects different from the later Sanskrit, and more akin to the Vedic dialect) was once a spoken tongue, regarding this as a fact which admits of no question: while Professor Weber is of opinon that the only Indo-Arian speech which existed at the early period to which I refer had not yet been developed into Sanskrit, but was still a vernacular tongue. 155 As, however, what seems so clear to the European scholar,-viz., that Sanskrit in its earlier form was a spoken language, may not be so plain to the Indian reader, it becomes necessary for me to adduce the most distinct evidence of the fact which I am able to discover.

154 With reference to a question already discussed, see pp. 31, ff., I add the following sentences from Mr. Beames's article, p. 150:—“With regard to the languages of the second period, it must be explained that I do not intend to touch on the obscure question of how far non-Aryan elements enter into their composition. Much there is which is still doubtful, but this is admitted on all hands, that a very large proportion of their constituent parts is of Aryan origin."

155 Indische Literaturgeschichte, p. 1. His words, as translated, are these:"In its earliest period the Indo-âryan speech had not yet become Sanskrit, i.e. the language of cultivated men, but remained still a vernacular tongue, whilst in its second period, the people spoke not Sanskrit but Prakritic dialects, which had been developed out of the ancient Indo-âryan vernacular contemporaneously with the Sanskrit."

First:-Even though we assume, as we must do, that there were, from the earliest times, other forms of spoken language current in India besides the Sanskrit; yet these would be the dialects of the Dasyus, or non-Ārian tribes; while the upper classes of the population of the Ārian race, the same order of persons who in after times spoke Prākṛit, must have been in the habit of speaking Sanskrit (by which must be understood the then current form or forms of the Old

Ārian speech) a few ages previously; for, in fact, no other Arian language then existed in India which they could have used. If languages with such a complicated structure as the Pali and the Prakrits were employed in common conversation, there is no difficulty in supposing that Sanskrit too, which was not much more complex, should have been spoken by ordinary persons. We must not, of course (as Professor Benfey has well remarked above, p. 140, f.), imagine that all the refined rules for the permutation of letters which were used in later Sanskrit composition were then employed in daily discourse, though some few of them might have been; for the use of these rules is by no means essential to the intelligible or grammatical employment of the language; and at the time to which I refer, they had not been developed or systematized. Many, too, of the more complicated inflections of Sanskrit verbs would be then little used in conversation; as, in fact, they are now comparatively little used in most literary compositions. 156

156 The remark in this last sentence probably rests on a misapprehension of the character of the language vernacularly employed by men in the earlier stages of society. But I leave it as it stood, in order to make the following remarks on it by Professor Bentey, in the review above referred to, more easily intelligible. He writes, p. 135: "Here, nevertheless, I should like to see much otherwise understood. Thus it is said in p. 154-Many, too, of the more complicated inflections of Sanskrit verbs would be then [at the time when the Sanskrit existed as a spoken language] little used in conversation;' which, as appears to me, leads to an erroneous understanding. It is precisely the deficiency of so many forms in the regenerated Sanskrit, as, for example, the want of a conjunctive generally, of the moods for the different tenses, the unfrequent employment of the aorists as compared with the Vedic Sanskrit, the disuse of so many double forms, as e.g. the substitution of the single form of ais for ais and ebhis, as the ending of instrumental cases of nouns in a, the limitation of the strong case-forms, which in the Vedas are used very irregularly, the regulation of the reduplication and many other differences of this description between the Vedic, or ancient, and the regenerated Sanskrit,-it is just these points which determine us to explain the latter (the modern Sanskrit) principally through the predominance of the vernacular dialects: those persons who wrote the regenerated Sanskrit were too much accustomed to these vernaculars to do more in general than to turn the speech

It is true that we cannot point out the exact forms of all the Sanskrit words in use at the latest period at which it was so employed as a spoken tongue; especially as the language of conversation always differs to some extent from the language of formal composition or of books, and the vernacular Sanskrit was no doubt undergoing a perpetual alteration till it merged into Prākṛit.

Second:-The case which I have supposed here of Sanskrit having been once a spoken language, and having at length ceased to be employed in ordinary discourse, while the provincial dialects which sprang out of it, and gradually diverged more and more from it and from each other, have taken its place as the popular vehicles of conversation,-is by no means a singular occurrence, unprecedented in the history of language; on the contrary, the manner in which the Italian, French, and Spanish languages (to which Burnouf and Lassen refer in a passage cited above, p. 69) have been formed out of Latin, presents a very close parallel to the mode in which the various medieval Indian Prakrit bhāshās (which in their turn have given birth to the modern Bengali, Hindi, Mahrattī, etc.) grew out of Sanskrit. During the existence of the Roman empire, Latin, as is quite well known, was the spoken language of Italy, and other western portions of Europe. It is now in nearly all those countries a dead language, and is only known to the learned who study the works of the Latin philosophers, historians, and poets; just as it is only the Pandits of India and other scholars who can understand the Sanskrit Sastras. But while Latin has itself ceased to be a spoken language for eight hundred or a thousand years, various vernacular dialects have (as I have said) sprung out of it, such as Italian and the other modern tongues already specified; the Latin words which compose the greater part of their Vocabulary being variously modified, and the ancient Latin inflections being either corrupted, or dropped, and replaced by particles and auxiliary verbs. Of these derivative dialects, the Italian, which is with which they were familiar into Sanskrit according to the reflex rules (Reflexgesetze) which were known to them. It was only a constantly deeper study of the old remains of the genuine vernacular Sanskrit and the compositions which were more closely connected with it that brought back many of its at first neglected peculiarities into the regenerated Sanskrit, a point which can be proved by the express testimony of Panini himself in reference to the participles of the reduplicated perfect. (Compare Pān. iii. 2. 108; and my complete Sanskrit Grammar, p. 413, note 13, and sorter Grammar, § 361, 369.)"