Oil tJic Eight Ii Verb-Class in Sanskrii. ii 



abiding by the native theory of two distinct tense-signs for 

 the sii and tlie /cz/z-verbs. 



As regards Bopp's theory, two strong objections may be 

 urged against it. In the first place, I do not think there can 

 be adduced, in the whole language, a single good instance of 

 the loss of an initial nasal, or of any initial consonant what- 

 ever of a suffix or an ending, whereas the disappearance (no 

 matter here if by direct loss or by change) of the final nasal 

 of a root or a stem before a suffix consonant is a common and 

 well-known phenomenon in Sanskrit. Witness examples such 

 as ta-td, ta-tvd, ta-ti, -ta-tya, ha-tJid, -ha-bhis, raja-bhis, jitva-sii, 

 bali-bhyas, etc., for ^tan-td, ^taii-tva, etc. Then, such a theory 

 completely ignores any influence on the root on account of the 

 accentual shift in tanonii, though such an influence is directly 

 required not only in analogy with words of the kind quoted 

 above, but also by the analogy of other j- //-verbs (cf. star' : str- 

 no-ti ; kar : kr-no-ti). 



And it is precisely in consideration of this required weak- 

 ening of the root that we are forced to explain the formation 

 of tanomi as arising from a weakening of the unaccented tan 

 to ta before the accented suffix -no. Whether we are to con- 

 sider this weakening as consisting simply in the direct loss 

 of the final nasal of the root or in its vocalization after the 

 loss of the preceding «-vowel, is immaterial to the argument, 

 and need not here be discussed. It may be said, en passant, 

 however, that the objections made by Van den Gheyn in his 

 third paper against Brugman's //-theory are nowise convinc- 

 ing. It is true that no written language has left a trace of 

 an ;/-vowel ; but written languages never perfectly represent 

 all the sounds of the spoken. No modern language has any 

 sign for either a liquid or a nasal vowel, and yet such vowels 

 are often met with (cf. English sabre, sable, fatten, button, 

 etc., pronounced like sabr, sabl, fattn, bnttn). They may 

 have existed just as well, though imperfectly represented, in 

 older dialects, the skilful Hindu phoneticians being the only 

 ones to recognize in writing any of them (rand /). Van den 

 Gheyn also objects that while the r-vowel always leaves a 



27 



