1887.] Dr. Mitra— On EhotibUva. 171 



however, amount to a mere begging of the question. It cannot serve 

 the purposes of a major term — a universally accepted premiss — to 

 sustain the position taken up by Mr. Growse. 



Assuming, however, for the sake of argument that the rule may 

 be, somehow or other, made to bear on ancient Sanskrit, I think it 

 necessary to enquire, what is its exact nature, and how far is it appli- 

 cable to the word under consideration ? It occurs in all the Prakrit 

 grammars that have come under my observation. As given in Cowell's 

 translation of Vararuchi, it runs thus : ' These nine consonants, k, g, 

 ch, j, t, d, p, y, v, or b, when single and non-initial, are generally elided.' 

 (Kagachajatadapayavam prdyo lopah, II, 2). 



It is always understood that when a letter is elided under this rule 

 the vowel with which it happens to be associated is left behind, and 

 when so left it is not subject to any rule regarding coalescence or sandhi. 

 This is well illustrated by the example cited by Mr. Growse. The ori- 

 ginal word being sukara, the ' bristled one ', the elision of the h leaves its 

 vocalic associate a behind, which retains its place and makes suara, and 

 Mr. Growse accepts it in that form. Had sandhi been permissible the 

 u of s'u and the a of Jca would have coalesced and produced svara. The 

 prohibition of sandhi is necessary to preserve the skeleton of the words : 

 without it there would be no limit to the process of metamorphosis. 



The qualifying term pray ah, * generally ', in the rule shows clearly 

 that the rule was not universal, and the commentator restricts it by 

 saying that " where euphony is not disturbed there should be no elision " 

 (prayograhandt yatra srutisuJchamasti tatra na bhavatiti) . No rule is 

 anywhere given to define what euphony is, nor is such a definition 

 practicable. As a matter of course it is dependent upon taste, and 

 must differ greatly in different cases. Practically, the qualifying term 

 with its commentary made the rule quite optional in its operation. 

 There is another rule which says that a simple non-initial first letter of 

 each class in Sanskrit may, at option, be replaced by the third letter of 

 its class in Prakrit. Thus h may be changed to g, ch to j t t to c?, t to d, 

 and p to b. 



Now, according to a well understood law of grammatical interpre- 

 tation two options always imply three forms : 1st, the original form ; 

 2nd, the modified form produced by the first optional rule ; and 3rd, that 

 which is the result of the second optional rule. Usage, without actual- 

 ly prohibiting the first form, is not much in favour of it. Instances, 

 however, are numerous of it, and the commentator cites several. (Cowell 

 P. P., p. 116). Generally speaking words are most frequently met with 

 in the second and the third forms : in some cases, only one form is 

 met with, 



