On the Eighth Verb-Class in Sanskrit. 7 



It is of some importance here to distinguish between that 

 stage of the Indo-European language when the radicals which 

 are now deduced by comparative analysis as those underlying 

 its formal and inflectional development were already evolved, 

 and a yet earlier stage when these radicals had not assumed 

 the form they then had. There is, indeed, every reason to 

 believe that such an evolution of roots from earlier germs, or 

 perhaps, in some cases, from earlier polysyllabic and com- 

 pounded entymons, took place long before the language 

 passed into its inflectional stage, and in some instances we 

 may even yet discover the probable or evident traces of such 

 a development. Thus, it seems probable enough that coup- 

 lets or groups of roots like i : i7z 'go,' ci : cit 'observe,' -mar 

 'grind' : marri 'crush' : mard 'grind' : inarch 'hurt,' etc., 

 are cognates of the same origin. If their original germ is 

 actually represented by any one of the forms preserved to 

 us, or if it is entirely lost, cannot be decided. There are 

 some faint indications, indeed, that the shortest form may, 

 ordinarily, be the most original, but they are after all 

 uncertain.^ 



Now it so happens that for every one of the roots in -an 

 enumerated above has been suggested also, on more or less 

 convincing evidence, a co-ordinate radical lacking the nasal 

 and terminating in -a or, usually, in -a. Thus, cf. ksan 

 'hurt' : *ksa (in ksapay- 'destroy,' t?ivi-ksa 'much destroy- 

 ing,' etc.), and, perhaps, ksd 'burn' ; — tan 'stretch' : td (in 

 the pass, tdydte, and in tayate ' stretches ') ; — inan ' think ' : 

 (.'') vid ' measure ' ; — vaji ' win, like ' : vd ' desire ' (in the parti- 

 ciple z'dtd and the desid. vivdsati), va (only in vasiniahi) ; — san 

 'procure ' : sd 'procure ' (in sdtd, sis-dsati, etc., and in aqva-sd 

 'horse-acquiring,' etc.), sa (only in sa-sa-vahs). It is evident 

 that the assumption of (7-roots finds a very meagre and doubt- 

 ful support in these comparisons (about which later). But a 

 host of forms, participles or derivatives, with a radical in -a, 



"^ Cf.Yi. Edgren, "On the Verbal Roots of the Sanskrit Language and the San- 

 skrit Grammarians," Journal of Amet-ican Oriental Society, XL, p. 5, etc. 



2Z 



