It is my intention,' says he, (Preface, page iii.) in the present work to attempt a comparative description of the organism of the languages enumerated in the title-page, to investigate their physical and mechanical laws, and the origin of those forms which express grammatical relations. The mystery of the roots only (or the reason why particular primary notions were expressed by such or such combinations of sounds) we leave untouched; we do not inquire why, for instance, the root i signifies to go and not to stand; or why sta or stha means to stand and not to go. But, excepting this, we shall try to follow language as it were in its growth, and in the process of its developement, so conducting our inquiry that those who are averse to explanations of things considered by them inexplicable, will perhaps find less here to offend them than they would expect in a book written with the views just stated.'

6

The part of the work now before the public comprises three divisions. The first of these (pp. 1-104,) is headed Schrift und Laut-system, system of writing and of sounds.' The author first discusses the sounds expressed by the Sanskrit alphabet (§ 1-29); next those of the Zend (§ 30—65), and Gothic (66-92); after which he reverts to the Sanskrit, in order to point out such phonetic laws observable in the transmutation of its letters, as could not be noticed in the analysis of its alphabet. To the consideration of the alphabets of the other languages here inquired into, no separate section is given, but they are elucidated by occasional references.

The second division of the work (pp. 105-132) contains general remarks on the common character of the roots in the languages compared; and the third (pp. 133-288) is on declension, or on the crude stems, and the formation of the numbers and cases of nouns.

Every part of this work is so replete with novel and interesting matter, that it would be fruitless to attempt to introduce into our notice even a succinct outline of all the new information which it contains, or which it will lead an attentive reader to discover. We must be satisfied with a few extracts, which we shall intersperse with some remarks of our own.

The order in which the Hindu grammarians have arranged the very copious Sanskrit alphabet is remarkably systematic, and renders it an excellent basis for the comparison of the notation of sounds in other languages. We have first the simple vowels, five in number, each of which may be short or long:

a, â; i, î; u, ú; ri, rí; lri, lrí.*

* The vowel lri is only found in the verb kirip, and its derivatives. We remember no instance of the long iri in any Sanskrit word : Bhattoji, in commenting on Pânini, I, 1, 9, says that it is not used. Siddhântakuumudî, fol. 1. b.

By contraction with a short a placed before them, the four last of the simple vowels form the compound sounds,

é; ó; ar; al,

which being again contracted with another short a placed before them, form the diphthongs

ai, au, ár, ál.

The former change is called the guna, or corroboration,' and the latter the vriddhi or increment' of the respective simple vowels. The distinction of both is important in Sanskrit etymology. The guna is applied generally in forming primary, and the vriddhi in secondary derivatives from verbal roots. A few examples will render further comments unnecessary.

Root.

Vid, to know.

Form with guna, (primary Form with vriddhi (se-
derivative.)
condary derivative.)

vêda, knowledge, the vaidika, relating to the

Klrip, to prepare, to kalpa, regulation, pre- kâlpika, regulative, pre

cept.

ceptive.

regulate. Of the above vowels and diphthongs, the following are represented by distinct letters in the Sanskrit alphabet,

a, â; i, î; u, ú; rì, rî; lri, lrí; é, ai; ó, au.

The corresponding vowels of the Zend language, which are distinguished in writing, if arranged on the same plan, will stand thus:

a, ă, e, ê, ú ; i, í; u, Ö, ú; ê ; Ô, âu.

The diphthongs ói, ái and áu are in Zend (as in Italian) expressed by a juxtaposition of the component vowel-characters. The vowels of the Greek language that are represented by separate characters, if put in the same order, will stand then,

α, &, o; 1; u; n; w.

The consonants in the Sanskrit alphabet are divided into three groups: mutes, semivowels, and sibilants. The mutes are subdivided into five orders, according to the part of the organs of speech by means of which they are uttered, and each of these orders contains two hard (surd) and two soft (sonant) consonants, each being either unaspirated or aspirated, besides a character expressive of the modification of the nasal sound adapted to it. The separate characters are arranged as follows:

The sibilants are, two s, the one designated the palatal, and the other the lingual s, and sh. To these is added the h, the last letter of the Sanskrit alphabet.

The consonants represented by distinct letters in the Zend alphabet, if arranged on the same plan, will stand thus:

The simple consonants of the Greek alphabet, if put in the same order, will appear as follows:

It is a remark which cannot fail to strike every one who studies Sanskrit, that very few of the words of that language which contain any one of the letters above designated linguals possess etymological equivalents in the Greek, Latin, or other cognate languages. We are told that the present. Hindus pronounce these by applying the point of the tongue to the root of the teeth, or the upper part of the mouth. In some of the vernacular idioms derived from the Sanskrit, they often sound like ; the name of the giant Hidimba, well known to the readers of Bopp's episodes from the Mahâbhârata, is thus pronounced, Hirimba-and the name of the month Ashadha be

*The Greek y before x, x, y, which is generally considered as representing the sound of, in fact represents a modified sound of », analogous to the Sanskrit guttural and (perhaps) palatal n. See Buttmann's 'Larger Greek Grammar.'

A

comes Ashárha. That the lingual consonants were pronounced with this leaning towards r many centuries ago, appears from the circumstance that in the Tarikh-al-Hokamá, an Arabic work on literary history, often quoted by Casiri, the name of the Hindu astronomer Aryabhatta is reported under the form Arjabahar. (Casiri, Bibliotheca Arabico - Hispana, Vol. I. p. 426.) In the obsolete dialect of the vedas, l is frequently found instead of the lingual d, and lh seems to take the place of dh.

The transition of the common or dental n into the n of the lingual order, if preceded by r or sh, is familiar to every Sanskrit student. But we think it has not yet been observed, that in a few instances the dental mutes d and dh are also liable to be changed into the lingual d and dh, owing to the influence of a preceding r or sh. The numeral adverb shóḍhâ, sixfold, is derived from shash, six (here changed into shổ, as if it had been shas, after the common rule of euphony), and the same suffix dha, which we have in ekadhá, tridha, &c. ; the dh of this word, shodha, has become lingual through the influence of the initial sh of the primitive word. In the same manner and for the same reason, the d of dasan, 'ten,' in the compound numeral shodasan, sixteen' (whence shola in the present Bengali), and that of dat (for danta), a tooth,' in the compound adjective shôḍat, six-toothed' (like chaturdat, ubhayatódat, &c.), is made lingual. We can adduce but one instance of this change being effected through the letter r; it occurs in the word puródása (written purólása in the prose part of the Rig-veda), 'an introductory ceremony,' which is derived from puras, 'before,' and the verb das, 'to worship, to perform a sacrifice.' In Pali, thirteen' is terasa: here the first of the two component parts, the element signifying 'three,' te (in Sanskrit tri), has lost its original r; but its former existence is manifest from the influence which it has had upon the subjoined numeral for ten,' dasa, in changing its initial d, first into the lingual d, and thence into r.

6

The remarks with which our author illustrates his analysis. of the different alphabets, and the manner in which he points out the mutual correspondence and transition of sounds in the languages compared, will be read and studied with advantage by every philologist. It is the first time that etymological investigations have been reduced to certain principles derived from observations on an extended scale, and forming the basis on which a lofty fabric may with safety be raised.

In one of the earliest paragraphs of the work we meet with an observation on vowels which we extract

With regard to the greater or less weight of different vowels, I have become convinced, from a few but safe observations which I

shall point out in the theory of inflexion, that in Sanskrit a and â are weightier than I and i respectively. This remark is of importance, as well for every special, as for the general and comparative grammar. In Latin, also, i is accounted lighter than a, and usually takes the place of the latter, when a word containing a primitive a is burdened at its beginning by some prefix or by reduplication. The Latin radical a is subject to a twofold modification when the root is increased by syllables or words placed before it; it is changed into i in open syllables,* and into e if the vowel is confined by a following consonant: hence tubicen, abjectus, but tubicinis, abjicis; and inermis, imberbis, not inirmis, imbirbis; and again, inimicus, insipidus, not inemicus, insepidus.'

We add a few more examples, to draw attention to the extensive applicability of this remark. They require no comment.

How does it happen that pario deviates from the analogy, forming its preter-perfect tense peperi instead of pepiri? We are at a loss how to account for this anomaly; but have no doubt that others, better conversant than we profess ourselves to be with the archæology of Latin grammar, will soon discover the reason.

In etymological discovery much depends on accident, by which an entire series of combinations will often seem to present itself ready made to the mind. It is therefore by no means surprising that all the transitions or correspondences of sounds in the different languages here examined, of which good examples might be produced, are not found recorded in Mr. Bopp's work. We perceive that in speaking of the guttural letters of the Sanskrit (§ 13), our author says nothing about the sounds etymologically corresponding in Latin or Greek to the Sanskrit compound consonant ksh. The following ex

* By an open syllable is here meant, one terminating in a vowel. The context shows that Mr. Bopp divides the syllables, in the examples which he adduces, on the following plan: tu-bi- cen, ab- jec- tus, &c.