inflections of some one language terminate in a particular letter, the inflections of another distinct language ought likewise to terminate in that very same letter.

It seems equally contrary to every philological principle, to found an argument, with respect to the real form of words and their inflections, on the technical systems, invented by grammarians for the assistance of memory. But no person will, I think, be inclined to believe, on the simple assertion of M. Bopp, that in this respect, "les Grammariens Indiens ont fait preuve d'une in consequence blamable." On this point, however, any remarks, seem unnecessary, for all such technical systems appear to be of no utility. But, whether the Sanscrit Grammarians have employed the letters adapted for designating the terminations of words, in their inflected and proper form consistently or otherwise, it must be evident, that such artificial aids to memory cannot be received as any proof, that the inflexions, which now terminate in Visarga or h, originally terminated in s. On the contrary, the very presence of this letter in these systems is intended as a rule, in order to shew that a change in the word is necessary, before it can assume its proper form.

BUT, if the Visarga does not possess an independent existence, why has it received a place in the Sanscrit alphabet?—why does it frequently appear in all Sanscrit works, and why, when retained, does it produce an effect on Sanscrit prosody? To propose such questions to a Sanscrit scholar may appear ridiculous; but it becomes necessary in consequence of such of the Literati of Germany, as study this language, having thought proper to introduce an s or r, where other writers employ an h; and to write Ramas instead of Ramah or Rama, as the insertion or omission of Visarga is left in some measure optional by Sanscrit grammarians.† To render, however, this system uniform they ought to write Vishnus, Shivas, Lakshmis, Devis, Indras, &c. and they would thus contrive to render doubtful in Europe all the personages of the Hindu mythology, and to prevent any person, who adopted such a mode of speaking, from being intelligible to the natives of this country.‡ It might, however, be admitted, that it is most probable, that the

IN the schemes of the inflexions of nouns and verbs given by Wilkins and Yates there is not one ending in r; and therefore, what grammatical form, M. Bopp has found to end in this letter it is impossible to discover.

+ SEE Preface to Wilson's Sanscrit Dictionary, p. XLVI.

MANUS, in particular, would be totally unintelligible. This innovation, therefore, trifling as it may seem, becomes of importance in consequence of Sanscrit proper names, and words not admitting of translation, being always written in the nominative case. For it is obvious, that when the same word is made to end in a, h, or s, a person unacquainted with Sanscrit must be at a loss to determine, whether the words so written are merely one, and the same term, or distinct names. In this case, the best system of orthography, that can be adopted, is certainly the omission of the Visarga, as it is not liable to produce mistake, and conforms with sufficient accuracy to correct pronunciation.

present inhabitants of India have preserved the pronunciation of their ancestors, and that the writers who have acquired a knowledge of Sanscrit in this country are best qualified to give the Rules by which this pronunciation ought to be regulated. In which case the erroneousness of the opinion advocated by M. Bopp would, without any other consideration, be sufficiently demonstrated; for the pronunciation of the Visarga in India, is that of a strong aspirate following the preceding vowel; and in Roman characters, consequently it ought to be represented by a h.

FROM the preceding remarks it will perhaps appear that M. Bopp, in rejecting a character from the Sanscrit alphabet and in dispensing with giving the rules by which this character is affected, has acted on principles completely at variance with the real analogy and structure of the Sancrit language. The mistake, indeed, is not of much consequence, for the student who has made any proficiency in Sanscrit will immediately detect it. But it seems obvious that the study of this language can never be prosecuted with success, if innovations are so inconsiderately admitted into its grammar. For, though it be admitted that the system of Sanscrit grammarians is complicated and badly adopted for the purpose of instruction, and it may in consequence be desirable that a more simple method of teaching its grammatical Rules than any which has been hitherto published, should be submitted to the public; still these Rules in no manner affect the real form of words, but are merely intended to shew the manner in which the primitive may be inflected, and thus by establishing a few fixed principles to facilitate a knowledge of the inflexions to which a noun or verb may be subjected. Before, also they could have been invented, the language must have received its peculiar form, and the labour of the grammarian, therefore, consisted in analysing and arranging, and not in creating. His system consequently may be defective, but the elements on which he operated still remain the same; and it must hence necessarily follow that had not such a sound as the Visarga existed in the Sanscrit language, neither would a character have been invented for it, nor would grammarians have given themselves the trouble of ascertaining and specifying the permutations to which it was subject.

ART. V.—Remarks on Professor Dugald Stewart's Theory of the Origin of the Sanscrit Language.

THE third volume of Stewart's Philosophy of the Human Mind opens with observations on Language, as an intellectual faculty, in the course of which, some speculations are advanced, respecting the origin of the Sanscrit, or Classical Language of the Hindus. This language, the Professor, adverting to the remarkable similarity, that subsists between it and Greek, supposes to have

H

been fabricated from the Greek by the Bramins, within two generations after Alexander's invasion of India, and to have been gradually brought to its utmost perfection upon its Greek model, by the first century anterior to the Christian Era.

As Professor Stewart's estimate of the affinity between Sanscrit and Greek depends upon information, derived from other sources, and not upon his own knowledge of Sanscrit, it is necessary, before investigating the accuracy of his conclusions, to enquire into the soundness of his premises, and to determine the weight of that authority, by which he has been guided. In doing this, it is impossible not to be struck with the very vague estimate of the affinity between Greek and Sanscrit, which the authorities he cites afford, and their utter incompetency to be regarded as conclusive evidence on the subject.

MR. WILKINS is one authority, upon which the affinity rests, and in some respects, might be held sufficient; but, as Mr. Stewart admits, "he has not made any mention of the coincidence between Greek and Latin in any of his publications;" and his testimony appears in the form of some general, and occasional conversation, with Lord Monboddo, which was repeated by the latter in the fourth volume of his Ancient Metaphysics, and which amounts to little more, than that the use of the Alpha privativa was as common in Sanscrit, as in Greek.

THE next authority quoted is Mr. Halhed, who states in the introduction to his Bengali grammar, that " every Sanscrit verb has a form, equivalent to the middle voice of the Greek, and that all the Greek verbs in mi are formed exactly upon the same principle with the Sanscrit conjugations, even in the minutest particulars." The latter assertion is by no means correct, although the resemblance is very striking. Mr. Halhed is known as a good Bengali scholar, but his acquirement of Sanscrit was evidently only general, and he is not an authority in any question regarding its structure.

NEXT comes the Rev. Dr. Brown, who in a letter to a friend remarks "Sanscrit answers to Greek, as face to face in a glass." That this practice of quoting the unguarded expressions of persons in private letters is most mischievous, is here abundantly proved when such a man as Stewart, is misled by it, to conclude the writer, to be an authority of weight: he lays great stress on this testimony of Dr. Brown's, which, as that divine knew nothing of Sanscrit himself, is vastly inferior to the evidence of Wilkins and Halhed, and is in fact of no value whatever.

WE next have the suspicion of Gibbon, that "some, perhaps much of the Indian Science, was derived from the Greeks of Bactriana,"-to which hint, however, the Professor very properly

paid little attention, untill he found the same opinion stated with considerable confidence by the very learned Meiners in his Historia de Vero Deo. We should see no reason to prefer Meiners to Gibbon, except that the former supports his opinion by reference to Bayer. On looking into Bayer, however, Mr. Stewart was much disappointed by these proofs; and after all, therefore, the opinion of Meiners is a mere conjectural notion, which it is marvellous a writer on the Philosophy of the human mind could mistake for authority.

HAD the character of Stewart's authorities been higher than it really is, there would still be reason to object to their being recognised. They are, to say the least, obsolete, dating from 20 to 40 years back, in which interval Sanscrit has been much more studied, and the principles of its construction more fully and more profoundly investigated. The two Schlegels, G. Humbold, Bopp, Frank, and Klaproth have directed equal learning and industry to the subject, and no person can competently meddle with it, especially if ignorant of the Language, without referring to the detailed comparison, which they have instituted. Their writings, it is true, are in the German Language, of which Stewart regrets his ignorance, but notices of some of their labours have appeared in the Journal Asiatique, the Journal des Savans, and in the English Reviews, and from one of the latter the Professor might have derived more profit, than he seems to have done, with respect to the affinities detailed by Bopp, and reviewed by Hamilton in the Edinburgh Review. There is also an English article on Sanscrit affinities by Bopp in the Annals of Oriental Literature, with which the Professor is wholly unacquainted, nor does he take any notice of two able articles in the Quarterly Review on the affiliation of Language, nor of the article "Language" in the Supplement to the Encyclopædia Britannica in all which, the connexion of Sanscrit with other tongues makes a conspicuous figure. It is difficult to understand, why he has preferred going back to Halhed and Meiners, when so much more modern and better materials were at hand, unless we suppose the opinions, now published, were formed some twenty years ago, and have never since been revised; and it is melancholy to observe a man like Dugald Stewart inert amidst the progress of improvement, and making a full stop, whilst all around him were in rapid advance.

As the correctness of his theory depends upon the nature, and duration of the intercourse between the Greeks and Hindus, the Professor next adverts to the historical proofs of that intercourse, in which he either identifies Bactria with India, in a very singular manner, or he asserts what no history authorises, that India was subject to Alexander and his successors. Thus, he says; "It is not to be doubted, that Alexander followed the policy of blending the Greeks with the natives, in his Indian, as well as his Persian dominions;" and he adds, that, "the measures he had taken for the

security of his conquests had been so well concerted, that India quietly submitted to Pytho, the son of Agenor, and afterwards to Seleucus." This last sentence is from Robertson, to whose precision it does no credit; for the Historian had previously shewn, that Alexander's invasion of India was transient, and his conquests limited, that he had explored, rather than subdued the country, and had explored only a small part of it, and that his operations did not extend beyond the modern province of Lahore, and the countries on the banks of the Indus, from Multan to the sea. Even in these countries, however, Alexander attempted no permanent settlement. He left Taxila and Porus in possession of their hereditary dominions, nominally subject to him, it is true, and perhaps so also to his successors. The subjection must, however, have been very indefinite and unreal, as soon as a powerful Greek force was withdrawn, and could not have led to such international intercourse, as to have influenced the whole body of the Hindus. As far as any change followed the presence of the garrisons that Alexander is said to have left in two or three towns, consisting of a handful of worn out veteran solders, it is more likely, that they were imperceptibly absorbed into the mass of the Hindu population, than that they converted the latter into Greeks.

As far as written testimony is to be received, the Greeks with Alexander appear to have been learners, not teachers. The philosophy, self-sufficiency and ascetic practices of the Gymnosophists are well described by Arrian and Strabo, and two of the Indian sages Calanus and Zarmanochages accompanied the army on its return to Persia and Greece. The author of the Dabistan also, as quoted by Sir Wm. Jones, asserts that Callisthenes sent to his uncle Aristotle, a technical system of logic, which is supposed to have laid the foundation of the Aristotelian method. We do not attach much weight to this testimony of the Dabistan, nor is it likely, that the retired and uncommunicative abstractions of Indian Sanyasis exercised much influence upon Greek philosophy; but as far as the evidence goes, it is decidedly in favour of the communication of Indian knowledge to Greece. Stewart thinks it more probable, that Callisthenes communicated the Syllogism to the Hindus, than that he received it from them. We do not see why; and the supposition is in the teeth of the tradition, which if it be received at all, is to be received as it stands, not cut and carved so as to suit any preconceived opinions.

THAT the brief period of Alexander's invasion, and the nominal subjection of one or two Indian princes to his successors, in Syria, could have produced any marked, and permanent impression upon the people of India, is to conceive effects wholly disproportional to their causes, and cannot be admitted in the absence of any thing like positive proof. That the connexion between India and Bactria was more likely to leave traces of its existence, is possible in proportion to its longer duration, but the circumstan