greatly corrupted, not only by the admixture of many Persian and Hindustáni terms and phrases, but even by a foreign idiom and construction. To omit altogether the laws of versification, in the grammar of a language whose best and most ancient authors wrote only in metre, is an unpardonable neglect, of which Mr. Halhed alone, in a manner, has not been guilty; though it must be owned, he is neither altogether correct, nor full in any degree equal to the requirement of the case. But it must be recollected that he tried an altogether untrodden path, without the aid of a single landmark left by any preceding traveller. He is rather to be extolled for having done so much, than greatly blamed for not having done more. The ground of real astonishment is, that with so intelligent and diligent a precursor, those who followed should have been generally both so incorrect and so deficient, as to leave it true to the full, that the only clear and safe guide as yet in Bengáli grammar is Halhed's volume, now long out of print and rarely indeed to be met with; surprising is it that it should never have been reprinted, and that while still unequalled it should have passed almost into oblivion; most surprising that subsequent authors should not have profited by his labours, of which one might even suppose most or all of them to have been in total ignorance!

Still, as must obviously be the case in the instance of a first Grammar in any language, Mr. Halhed's is doubtless imperfect, and sometimes incorrect: it must be owned too that numerous typographical, or press errors, greatly deform this otherwise elegant volume. These were however to a great degree unavoidable, printed as the work was at an out-station, by the agency of ignorant and careless natives alone, and without the superintendence of a professional printer. As instances of error, we notice the supposed "neuter names of animals, &c. to which may be added at pleasure different terminations of (sexual) specification," in p. 48; but which, however ingenious the notion, have yet no real existence :-" the third gerund (so called) in &ta,” in p. 115, which is really the genitive of the verbal noun in 1, under government ;-the passive signification given to the part: f, in p. 121, which is either erroneously rendered passively, as the pandits contend, or is incorrectly written for দেওয়ান The most remarkable oversight of all, is the misapprehending of the verbal terminations expressive of respect or inferiority respectively, for a singular and plural! No doubt the striking peculiarity of the Bengálí verb in this respect, occasioned the misconception; it was supposed that respect was marked by the use of a plural verb in connection with a singular pronoun. Take it however all in all, we venture to assure the inquiring student that he will obtain a far more extensive and more accurate acquaintance with the forms and genius of the Bengáli language, and a more expeditious and tasteful introduction to its purest literature, from this grammar than from all together that have succeeded it. One important exception to this remark must be made however, namely that Mr. Halhed has altogether passed over the very important rules for permutation, which must be sought in Wilkins's Sanskrit, or in Haughton's or Carey's Bengálí Grammar.

A full and complete grammar of the Bengáli dialect is, still a desideratum. We have nothing as yet, for instance, in any of the published grammars, upon the accentuation of words; next to nothing on the government and dependencies of cases; nearly as little on the idiomatic uses of the tenses; not a syllable upon the regular formation of lower from higher terms; very little upon the laws of metre and various species of versification, and other matters of the greatest moment. We know however of one individual, who has a grammar in a considerable state of forwardness, which it is believed will enter fully into most or all of these. Meanwhile there can be no hesitation in recommending Halhed is longe optimus among the guides to whom a learner should commit himself. The original price of his work has not been ascer tained; its present cost in Calcutta, is various; we have known copies, now and then at long intervals, sold for 4 rupees or thereabouts : it would however be a cheap purchase at double the highest price of any grammar noticed in this index. Parbury's London Catalogue gives its price there, £2. 2.

No. 2. A GRAMMAR of the Bengalee Language, by W. Carey, Teacher of the Sungskrit, Bengalee, and Mahratta Languages in the College of Fort William. Serampore, printed at the Mission Press. 1st Ed. 1801, 2nd Ed. 1805, 3rd Ed. 1815, 4th Ed. 1818.

The first edition of this work we have not met with. This is doubt. less of the less consequence as in the preface to the Second it is stated, that" since the 1st edition was published, the writer has had an opportunity of obtaining a more accurate knowledge of this language. The result of his application to it he has endeavoured to give in the following pages, which, on account of the variations from the former edition, may be esteemed a new work." The same precisely is stated of the 3rd and 4th editions in regard to the 2nd. Between these however, we do not observe, on a close comparison, any such differences as should entitle the latest even to be considered " a new work." Substantially the work is one; no new views are expressed, no additional light is thrown upon the main principles of the Bengáli accidence or construction. Still, as being fuller on some points and somewhat variously divided and arranged, the 4th edition is of course preferable to its predecessors, and may be taken to exhibit the result of the author's maturest judgment and most extended acquaintance with the language.

The author of this grammar having held, for many years, the influential office of a Professor in the College of Fort William, his pupils would as a matter of course not only be required to make use of it, but find it necessary to do so in order to derive the full benefit from his lectures in the language: he being on the spot also, there was no chance of its being allowed to go out of print. Hence it is easy to perceive that it would obtain an extensive circulation, the more so as the only work that could have come into competition with it. Halhed's grammar, had been so long out of print, none having either an interest in its reimpression or influence to give it currency, under the circumstances stated; otherwise, we confess, we should find extreme diffi

culty in accounting for so ill-digested, jejune, and unphilosophical a compilation thrusting a work, every way so superior to it, out of use. Our task obliges us to strict truth and perfect impartiality; and we feel that we should violate the double obligation were we to suppress the expression of our unhesitating judgment in this matter: we will now simply state what we deem the merits and demerits of this work, and leave the reader to form his own opinion upon the whole.

1. Dr. Carey has supplied the greatest of his predecessor's omissions, by inserting the rules of Sandhi, i. e. for the permutation of letters, agreeably to the principles regulating the composition of words in the parent Sanskrit. He has also correctly discarded the notion of a plural termination in verbs. He has given many of the native grammatical terms, by which communication between the learner and his pandit is much facilitated. These are the chief recommendations of his work. But on the other hand, his nomenclature of the tenses is most confused and unphilosophical. A grammatical term should be a definition, conveying some clear and distinct notion to the mind of the learner. Why the term aorist should have been borrowed from the Greek, (where however it has an application in coincidence with its etymological meaning in some sort,) to the Bengáli whose tenses are as definite as in any language under the sun. it is not easy to imagine. Again, a " present definite" supposes a present indefinite: if farofa therefore be the former, as it is, then af must be the latter, yet is in this grammar termed "the first aorist!" Again, fast, which is called the 2nd aorist, has two distinct uses; as a past frequentative or term of past habit, like the latin amabam, or as a conditional like the French ferais: in neither application is it an aorist. Again, if afrofa be a present definite, fast should, in consistency, have been a past definite, but yet is named a Pluperfect; which term even excludes the specification of time altogether, and is therefore a misnomer as denoting a tense, i. e. time: the full compound term, preter-pluperfect at least, should have been given. Then fata, which is made an aorist when indicative, becomes solely a præter, in the subjunctive, though in fact no such mood as a subjunctive exists in the Bengáli. The whole scheme of the tenses is, as to their nomenclature, a tissue of inconsistencies, and involves the use of the times of the verb in the greatest possible uncertainty and confusion: and this with Halhed's clear and philosophical arrangement and phraseology too as a guide!

Dr. Carey takes no notice of the real meaning and use of fast as a plural affix; a simple apprehension of which solves at once some of the most contested matters in Bengáli composition, especially the use of a genitive or a samás (†), contended for respectively by different schools, one writing always মনুষ্যেরদিগের or the contracted মনুষ্যেরদের, the other as invariably for a Both are in fact alike grammatically correct, and only wrong in the exclusion of a discretionary employment of the one form or the other. This too Halhed had perceived, though he had not drawn it out into its consequences, nor indeed has any other grammarian hitherto.

Dr. Carey has omitted the entire subject of Prosody, including as well the laws of versification and varieties of metre, as the poetic licences,

&c. although without these it is utterly impossible to read, with intelligence and pleasure, 9-10ths of the whole literature of the language! Carey's syntax is mostly common-place, and gives very little insight into the idiomatic construction of the Bengali dialect, in which consist its chief peculiarities and its difficulty to Europeans, and without a correct perception and facility in the employment of which no learner will ever be able to express himself correctly and efficiently either in conversation or written composition. The whole doctrince, for instance, of the relative, including its position and idiomatic usage, is unnoticed! though that too involves one of the most sharply contested proprie. ties of the language, the employment of the conjunctive (.

Dr. Carey's notions of grammatical government are quite original, if we may judge by his 2nd rule for the use of gerunds, (another Latin term borrowed with equal absence of necessity or propriety,) where he tells us that "the verb in this form is governed by the accusative case." Here, besides the extraordinary assertion of an accusative case exercising a government over a verb, there is the singular omission of all notice of the word governing that case itself!

In the formation of the verb, we are told that "the passive voice is made in two ways." Subsequently, under the syntax of the verb, we are informed "there are three ways of expressing the passive voice." Two of these, the 2nd and 3rd, are identical with those given in the former place. The remaining, or 1st mode, is quite unique and expressed with singular näiveté :-"in the first, the form of the verb is the same as in the active voice, but the agent is in the instrumental case and the subject in the nominative : er. বাঘে or বাঘেতে মানুষ খাইয়াছে, a man was devoured by a tiger!! One should have supposed that the Doctor, who must have read many thousand Bengáli shloks in his time, could hardly have cast his eye down half a page of the Rámáyana, for instance, without knowing-to say nothing of the absurdity of turning an active into a passive voice for the nonce-that the enclitics and 43, in poetry especially, no more distinguish the instrumental than they do the nominative case. In truth they are in this instance nothing but a nominative, and being the object not the subject, the simple rendering is the tiger has eaten the man!! A glance at one of Halhed's examples would have prevented the putting forth of so absurd a rule.

But let this suffice. We judged it necessary fully to substantiate our judgment, or we should not have written at so much length. The student has now the question of comparative merit fully before him, and will decide for himself. We are confident that to use Carey's grammar only, would leave the learner but very imperfectly instructed at the best, and in many cases would lead him into undoubted and serious error upon matters of the first importance. It is cheap however and always to be procured. The fourth edition is sold by Mr. P. S. D'Rozario, Church Mission Press, at 2 Co.'s Rs.

No. 3. RUDIMENTS OF BENGALI GRAMMAR, by Sir Graves Chamney Haughton, M. A. and Professor of Sanskrit and Bengáli in the Hon. E. I. Company's College, London, 1821. 4to. pp. 168. Sold by Par bury and Co., Leadenhall-street.

This is an elegantly printed volume, on excellent paper, and in a good clear type, possessing every external recommendation.

The author states it to have been composed "under the immediate pressure of a want, which admitted of no delay," Dr. Carey's Grammar being no longer to be procured in England at the time; and also that "the sheets were generally written over night and printed the next day." It were certainly not to be expected that a grammatical work compiled under such circumstances of extreme haste, should be free from many and serious faults and deficiencies. Yet, as the author must, in the exercise of his professional duties at the E. India College, be presumed to have been familiar with the details, and long previously to have settled in his own mind the principles of Bengáli Grammar, the defects, it is to be concluded, should not be either so many or serious as might be apprehended on the first announcement of the rapidity with which the sheets were worked off. Besides, as Sir G. C. Haughton acknowledges, he had both Halhed and Carey before him yet certainly a more meagre grammar never was put forth under so imposing an aspect. It may be excellently adapted to aid the H. C. embryo writers to get up a smattering of Bengáli so as to pass muster at an examination; but whoever should carry his study of the language no further than this volume would serve as a guide, would remain lamentably deficient indeed in all that constitutes sound scholarship and practical facility in the application of his grammatical furniture. E. g. the author tells us that" the subject of Prosody has been altogether omitted, as a matter more of curiosity than of real utility in a language like the Bengáli,"-like the Bengali! more than fourfifths of the entire literature of which is poetry, and poetry, too, admitting of a great variety of licence, and of transpositions of which the prose writings exhibit almost nothing!!

This author revives the plural terminations in the verb, though opposed to the whole character of the language.

He has, in another form, exhibited the same anomalous view of the enclitic nominative in 4 or 3, although he had himself only a page before (Par. 59) expressly asserted it (after Halled), e. g. statt effe লইয়া এক দুর্গ পর্বতে রাখিয়াছে ‘a fairy has carried off your daughter and placed her on an inaccessible mountain,' he renders-' your daughter has been carried away by a fairy, &c.' His explanation is curious"In the above instance the word can only be considered as the nom. of the verb here employed passively, as by the fairy, is in the instrumental case and is therefore the agent of the passive verb." It would not assuredly be easy to get a greater mass of anomalous and contradictory assertion into a single sentence. Objectives turned into nominatives, active verbs into passives, agents into instruments, and all for what? to support an unmeaning hypothesis that the affix ( corresponds to the Persian, and is employed to mark the word to which it is subjoined as the main object in the speaker's mind!

Still we think this Grammar has borrowed so much from Halhed's as to be on the whole greatly preferable to Carey's, the single instance only of the plural forms of verbs (instead of the honorific) excepted.