pretty uniform. Peculiar to the Devanâgarî, on the other hand, are the vowel-like letters ri and Iri; the latter, pronounced somewhat like the Polish 1, is introduced merely for the sake of uniformity, in order that the semi-vowels, ya, ra, la, va, which are subject to conversion into vowels, may have each its corresponding element, i, ri, Iri, u. The ri, however, or vowel r, is essential to the Sanskrit, though appearing in the sister languages as a consonant, in its Guna-form, as the Indian would say; e. g. kri or kr, creare; vrit, vertere; stri, sternere; trip, zέ¬ɛom. Moreover, there is a separate sign for the lengthened sound of each vowel, and we have a and â, i and î, u and û, ri and rî, lri and Irî. A unites with i, whether either be long or short, to form the diphthong e (e as in there); a and e become ai (our long i), which is, as it were, a triphthong. In like manner comes o from a and u, and au (ou in our) from a and o. It is not for the sake of euphony only that this union takes place, when, either in composition, or to avoid the hiatus, the final vowel of a word is blended with the initial of the next following (e. g. hitopadesa for hita-upadesa, nai va for na eva); a similar change takes place, also, in the formation of derivatives, by the addition of an a to the primitive vowel; for instance, i is strengthened into e, u into 6, which first increment is called by the grammariaus Guna, coloring; again, & becomes ai; 6, au; this is termed Wriddhi, growth. By attention to these changes, we are always enabled to recognize the derivative, and refer it back to its primitive. So prauda, proud, leads us to the root, prud, to be proud; yauvana, youth, to yuvan, a young man. Applying the same process to the cognate languages, we have from cupere, first copa, then cauponari; so likewise are the Norse raudr, red, or the Lithuanian raudonas, strengthened forms, the Sanskrit rudhiras, and ¿qvogós, being simpler. Every consonant of the Devanâgarî is supposed to include and be followed by a short a, as its soul or particle (mâtrâ), unless its place is supplied by some other vowel expressed; just as, according to some authorities, in the old Roman, krus was written for karus, kra for ce ra. This a was originally signified by the perpendicular line in each consonant, which is therefore omitted when the a is silent, or doubled when it is lengthened. The exactness and consistency of this mode of vocalization, show that it could have been but once invented; its peculiarities are lost in the rounded current-hands, and are wanting even in a number of the Devanâgarî letters themselves, which may be later additions to the alphabet, being, as they are, merely modifications of other sounds. A horizontal line above the letters unites them calligraphically into words, but it is an unessential part of the character, and rarely found in ancient inscriptions. 1849.] The Consonants. 475 After the vowels and diphthongs come the consonants, arranged in five classes, and in such an order that, the classes being written above one another, the first perpendicular line will include the smooth mutes, the second their aspirates, the third the middle mutes, the fourth their aspirates, and the fifth the nasals. These latter are not in other languages, as here, denoted by separate signs, although the guttural n in angle, for instance, is obviously quite distinct from the dental in enter. A point written above the calligraphic line, termed anusvara, may stand instead of any one of the nasals. The five classes are as follows. First, the Gutturals; ka, kha, ga, gha, nga. Second, the Palatals; cha (ch in church), chha, ja (j in James), jha, na. This class is of special importance with respect to comparative investigations. Cha, even in Sanskrit, is convertible into ka; from vach, to speak, comes vâk, a word, vox; and in Latin qu often supplies its place: pancha, five, quinque; chatur, four, quatuor. In the classic languages, g stands for ja, and in all probability originally had the same pronunciation, as it still has in their dialects; râj, to rule, rajan, king, regere, rex (for regs); Italian, il rege; jânu, knee, fóru, genu; French, genou; etc. Third, the Linguals or Cerebrals, called in Sanskrit, head-tones; ta, tha, da dha, na; peculiar to this language, and uttered back in the throat. These exhibit a tendency to pass into r. Fourth, the Dentals, ta, tha, da dha, na. Fifth, the Labials, pa, pha, ba, bha, ma. Next follow the semivowels, ya, ra, la, va; then the sibilants, sa palatal, sa dental, and sha; and last of all, the aspirate, ha. The first sa varies to k, even in the Sanskrit, and yet more readily in the cognate languages; dis, to point, indicare, dɛíxvvu; das, to bite, duxw; dris, to see, dépzeσu; pasu, beast (from pas, to tie), pecus; satam, hundred, centum; dasa, ten, decem, dɛxa; svan, dog, xvóóv. The f is wanting among the Sanskrit elements, and the aspirated bha, performs its office; as, bhu, to be, quo, fuo; bhri, to bear, pow. This system of arrangement is of high antiquity, for we find it observed in the earliest lists of roots, and it is followed by all the vocabularies, as well as by the derived dialects. The extended use of the Devanâgarî agrees precisely in point of time with the spread of Buddhism, which during the fifth century before Christ and later, gradually made its way out of India over all the islands and through China, Japan, and northern Thibet. It has been preserved nearly pure in the written character of Cashmere, and of a northern dialect in India; is but little corrupted among the Hindostanees about Agra and Delhi, as also among the Sikhs of Punjaub and the Mahrattas. The alphabet of the Bengalee dialect along the Ganges, likewise used by the inhabitants of Assam on the Brahmaputra, is sharper and slighter. As we go southward, we find a tendency to a rounded hand more and more prevailing. Besides the systems still in use, there are many which may be regarded as dead stereotypes; as, among others, the triple character of the Pali, in which are written the sacred books of the Buddhists about Ava and Pegu. In order now to arrive at some degree of acquaintance with the internal structure and peculiarities of the Sanskrit, we must go back to the so-called roots of the language. The Indian, naturally inclined to speculate upon every subject, has ever made his own language a special study, because it was the holy tongue in which Brahma himself revealed the sacred writings. He has, in truth, made himself thoroughly master of the process of its development, as nothing better shows than that profound and remarkable step, his attempt to strip the verbs of their subjectivity, in order to arrive at the naked elements of the language; and thus, what so late as Schlozer was declared to be a useless folly, the endeavor, namely, to trace out the original beginnings of a language, he has long since effected with regard to the Sanskrit. For next after the interjection, that mere animal utterance, which seeks to express feeling only by sounds of a higher or lower key, these stems (dhâtavas) plainly constitute the first attempts to clothe with form the fleeting breath, when pantomime was found no longer sufficient. If the vowels may properly be styled the soul of language, the consonants supply its peculiar physiognomy; they are the characteristic part, and they are more or less imitations of natural sounds, in proportion as a sensual observation of Nature, or a cultivated reason, have borne the more prominent part in the formation of a language. Among the Sanskrit elements are to be found very few such imitations; instances are, tup, to strike, zúzza; pat, to fall; expressions for speaking, knowing, teaching, meditating, are comparatively very numerous; and rarest of all those for struggling and fighting; facts which bear strong testimony to the early earnest and peaceful character of the Indian; indeed, as Humboldt remarks, many and varied evidences of his propensity to abstraction and pious seclusion are traceable in his language. The roots being formed by the addition of consonants to a simple breathing, they must of course be all monosyllabic, whether containing one consonant or more; as, mâ, to measure; smi, to smile ; skand, to ascend. Simple vowel roots, Grimm disallows, and in this the acute philologist is fully borne out by the Sanskrit, which in such apparent cases shows that it has itself lost a consonant, or that the other languages have corrupted the stems; as the Greek avo, comes from the Sanskrit av and va, to blow. As to the signification of the roots, we gather from their development that they form a descriptive 1849.] Signification of the Roots. 477 part of speech, intermediate between an interjection and an adjective, and therefore the verbal stem appears generally in its purest form in the imperative, or as the final member of an adjective compound; e. g. dharmavid, skilled in law; vid being the root, meaning to understand; compare armiger, frugifer, etc. So also is the Latin termination brum to be be referred to bhri, to bear; as, candelabrum, lightbearer. The Indian grammarians regard their verbal-stems as germs, or material (prâkriti), which become verbs only when invested with intellectual predicaments, when the relation to the attribute and the existence of the subject in respect to quantity and quality are expressed; when, in short, a complete sentence is formed for instance, tupâmi, I strike, properly contains the predicate tup, the copula as (to be), and the subject mi; striking am I. They therefore define the roots by abstract ideas, as speech and motion, which they regard as original categories; and moreover express the definition in a case which has the least direct connection with an object, namely, the locative; e. g. î, to go, is explained by kânti-gati-vyâpti-kshepa-prajanakhadaneshu, in loving, going, obtaining, throwing, begetting, biting, a single compound, whose last member bears the case-ending. There are about five hundred roots of motion; next in number are those of sound; but not until they become organized and inflected verbs, do they come within the province of general usage, which limits more narrowly the nature of the motion or sound; e. g. ling, to move, inflected or united with a prefix, means to embrace, umschillingen; ag, to set in motion, agere, applies particularly to fire; hence agnis, fire, ignis: valg, to move, is as a verb modified to express wry or crooked motion; compare the Latin valgus, the Greek q2yo, flecto, the English to walk; the latter, however, has lost the accessory idea. So tus, to sound, in common use, becomes to cough, tussire; vrih, to sound, properly refers to the cry of the elephant; and thus all these lighter shades of meaning lie dormant, as it were, in the roots, and are waked into life by being brought forth into the actual language. It is easy to see how much this analysis of the verbs must facilitate our researches into the inmost structure of the language, and a similar treatment of the classic tongues were a thing much to be desired, in order that by closely comparing the roots of kindred dialects, we might be enabled to trace out their historical development. The Latin and Lithuanian seem more to resemble the Sanskrit in keeping fast hold of their verbal stems than does the Greek, because the noble tongue of the Hellenes, earlier left to itself, concealed its Asiatic physiognomy under the veil of Grecian individuality; but even here, with the aid of the Sanskrit, we find it possible to trace and separate a great number of pure roots. Almost every native Indian grammar contains, in the form of an appendix, a catalogue of roots (dhâtupâta), and of these, two in particular are highly esteemed for completeness; that of Kasinatha, which has been edited by Wilkins; and that of Vopadeva (Kavikalpadruma), which Carey embodied in his Grammar. Rosen's learned and accurate work, with examples from the ancient writings, is made up from them both. The number of verbal-stems is 2352; to these are to be added the so-called Sautra-roots, which are assumed by the grammatical rules (sûtrâni), in order to furnish a derivation for the few nouns which cannot be referred back to the others. This number might be reduced one half, if roots of similar signification, but differently conjugated, were to be reduced to the same original. The common usage of the language seems to content itself with a much less number; there are more than 1800 for which Rosen fails to find examples, and therefore in the earlier productions of Sanskrit literature, only five hundred appear to be commonly used; these, however, with the aid of prefixes, suffice to express an inexhaustible variety of meanings. These prefixes, eighteen in number, are all of them related to the prepositions of the allied languages, as for instance pra (pro), pari (nɛgí), uрa (vzó), etc.; they are mostly inseparable in the Sanskrit, and are only elevated to the rank of distinct prepositions in the kindred dialects, as the case-endings go out of use; so the Slavonian needs no with or by to express the instrument, for he has his instrumental case. Examples of the manner in which the prefixes modify the verbs are pat, to fall; utpat, to fly up, nέrɛoda; âpat, to fall headlong; lambh, to obtain, upalambh, to understand, to take upon one's self, as it were, vñolaμßáve; svas, to breathe, âsvas, to console, to give free breath; nirsvas, to sigh, expire, visvas, to be quiet, Rosen's investigations show that no root receives four prefixes; instances of the addition of two and three are, however, frequent. etc. If we now take a general view of the verbal-stems of the Sanskrit, we shall conclude that they were reduced at a time when the language was still fresh and vigorous, but after it had begun in some measure to extend and modify itself with respect to these its fundamental elements. It is, perhaps, still possible to follow these efforts of the language toward a more copious development, for we can detect among the roots certain classes of modifying changes which could hardly have been suddenly effected. The first attempt was to acquire a new supply of roots, by rejecting a final consonant, and lengthening the radical vowel; this change is evidently of very ancient date, for the cognate languages often divide the double roots among them, and now find their point of union in the Sanskrit; examples are, gam |