On the Eighth Verb-Class in Sanskrit. 9 



ing of the root, especially on account of the displacement of 

 the accent, as in verbal nouns in -td -tvd, -tya {ta-td, ta-tva, 

 -ta-tya, etc.), and in nouns in -// (formed in analogy with 

 participles in -td). 



In all other cases, and they are very numerous, we find the 

 full -(?«-radical {tdn-a, tdn-as, tan-u, tdn-tu, tan-tra, etc.). It 

 seems evident that we are here dealing with principles per- 

 fectly parallel with those that have produced such root- 

 variations as in Jidji-ti : Jia-tJid ; Jidn-arn : ha-bhis ; rdjdn-am : 

 rdja-bJiis ; ds-ti : s-thd ; pitdr-ani : pitr-su : pi-tr-n, pitr-a, and so 

 on ; and it would be late in the day now to try to substitute 

 for the principle of accentual influence as causing these vari- 

 ations, the confusing principle of a variety of independent 

 roots and stems. 



b. Since there is no evidence that the so-called ' general 

 tenses ' are of a more recent formation than those of the 

 present-s3'stem, there is also no good reason to assume that 

 the -c?-roots have formed first the latter, and then with an un- 

 paralleled accretion, the others. We should expect at any rate 

 to find in the general tenses some trace of an earlier -i^-root, 

 but there is absolutely none. As for the assumption that the- 

 whole conjugation-system of the various -an-vQrhs is consist- 

 ently made up of two independent roots, which, if existing at 

 all, must have been distinguished by some shade of meaning, 

 peculiar to each, it would be hard to find anything parallel.. 

 To be sure, sporadic cases of root-mixture in the make-up of 

 verb-systems occur in all languages, but they are usually of a 

 different kind, consisting in the often traceable supplanting 

 of an earlier form by one of a different verb (as in French of 

 h'e by c'tais), and, in all events, they never take place so 

 systematically in a whole class of verbs. 



c. If the facts of the Sanskrit language itself suggest 

 clearly enough that the nasal of the /^?;/-verbs is genuine, the 

 evidence of cognate tongues also favors this view. To be 

 sure, we find even in them radical elements corresponding 

 to these roots with and without the nasal. Thus, cf. Skr. 

 \/ ksaj^i : Gr. Kreiv-oi, Kr6v-o<i ; e-KTa-rco, e-Kra-v, KTa-fievo^ ; 



25 



