________________
NOVEMBER, 1882.]
ASIATIC SOCIETIES.
329
by sevane, 'serve, honor'; and there are groups of identical finals with almost every consonant in the alphabet as initial. Under this head are considered at some length the causes which may be conjectured to have led to the fabrication of such groupe. 4. The discrepancy between the number of the two classes represented in cognate languages :-Fick finds evidence for regarding about 450 of the authenticated radicals as belong. ing to the Indo-European period; of the others, only 80, and many of these on very unsatisfactory grounds.
While the general conclusion from the facts and argumente presented is that the vast majority of the unauthenticated roots are pure figments of the grammarians, the probability still remains that a certain percentage of them are real, and either stowed away in some unexplored part of the literature or never recorded there.
The paper embraces an alphabetical list of the authenticated roots, stating under each whether it occurs in the Rig Veda alone, in the later literature alone, or in both, also whether it is combined with prepositions, and whether deriva. tives are made from it. To this list is added an index of the same roots arranged alphabetically according to their finals.
The second article is "On the Accentuation of the Vocative Case in the Rig and Atharva Vedas" by Dr. W. Haskell.
The third paper is also by Dr. Edgren, and is “On the Relation in the Rig Veda between the palatal and labial vowels (i, t, u, ú) and their corresponding semi-vowels (y, v)."
Dr. Edgren points out the difference between the Vedic dialect and the classical Sanskrit in regard to the treatment and occurrence before dissimilar vowels of i, u or y, v: the semi-vowels being alone found (by conversion or otherwise) in the classical language, but the two vowels being of very frequent occurrence, as proved by metri. cal evidence, in the Veda. A careful examination of the whole field shows beyond doubt that, whatever share arbitrary usage and corruption of the texts may have in the varied occurrence of vowels or semi-vowels, it is in the main of organic nature, and gives additional support to the theory that the semi-vowels in question are only later developments of the more primitive vowels i and u, and that we meet in the Rig Veda with a transitional state. Dr. Edgren tries to demonstrate by an exhaustive statistical account of all cases in the Rig Veda in which i, (, u, 1 or y, v occur before vowels, that the more primitive sounds have been retained as a rule, or prevailingly, wherever they occurred at the end of a word or stem, and thus helped to preserve the individuality of the
word; and, on the other hand, that the semivowels are found to prevail in all combinations the original independence and significance of which were dimmed and forgotten (as in derivative and especially inflectional suffixes, and in radical elements). The whole subject is considered under three heads : 1. The treatment of final i, t, u, ú, of words or themes before dissimilar vowels; 2. The occurrence of i or y, w or v in formative elements; and 3. Their occurrence in the radical part of the word.
1. In the collocation of words in sentences, i and u are retained almost without exception. In 1294 verses chosen from all the Mandalas, i and u occur together 391 times, y and v only 6 times (in praty, dno, addhu). An examination of a number of other passages confirmed the fact that only a few such less independent words as prepositions have begun to show a tendency to convert into a semi-vowel the fina i or u before a dissimilar vowel. In compounds the case is nearly the same. Final i and u occur altogether in 553 instances, but their corresponding semi-vowels only 52 times; and it is especially the prepositions ati, abhi which convert their vowels. Two words (gavyati, ritvij) occur not less than 39 times of the 52, but at least the former of them (gdvyúti) is of doubtful formation. In noun-stems ending in i, i, u or 4, the i() is retained in 392 instances, but consonantized in 240 instances; and the u (4) is retained 285 times, but consonantized 241 times (chiefly, or 110 times, in the two forms madhvas, vásvas). If each stem alone be considered, the difference in the occurrence of vowel or semi. vowel is much more marked, the vowel (i oru) being found then about twice as often as the semi-vowel. In both cases, the final long vowel is preserved more tenaciously than the short : the t-stems, indeed, never consonantizing & before a vowel-ending; and further, thematici (1) is found to occur mostly after a long, and y after a short syllable. In verb-roots the final iand u-vowels are generally combined with the following vowel through the medium of guna. strengthening or the insertion of a semi-vowel, less frequently by conversion of the final. The vowel i is retained in 51 instances, the vowel u never.
2. Of the formative elements, the derivative suffixes are taken up first; and of them the suffix -ia (-ya) is by far the most frequent. The form -ia occurs 2033 times, and -ya 1628 times. There are 47 words which are found in different passages with both form, -ia and -ya, but as a rule even these show very prevailingly one of the forms (in two-thirds of the cases it is -ia), and the exceptionally used termination is in one half
