________________
183
ently. E.g. v- in vuccai (Skt ucyate) is from vac-. Alternatively viva may be a frozen sandhi-form. It is difficult, though, to separate viva from the form via of the other dialects. Admittedly, the origin of the medial -v- presents a problem, for which I am unable to offer a satisfactory explanation. One might suppose contamination with iva, found in AMg. texts besides viva. I am in any case hesitant to connect it with Gāndhārt Dhammapada Pkt viva, about which much is still uncertain; see Brough (1962: 110-2).
The conditioning factor for the variant piva beside viva is strikingly similar to that for pi and vi. It has even been suggested by, for instance, Weber (Ed., p. 1), Goldschmidt (Setubandha, index, s.v. via) and Jacobi (Kalpasūtra, p. 100) that it is a compound particle made up of pi/vi plus (i)va. This derivation is doubtful. Admittedly, compound enclitic particles are not entirely unknown in Pkt, which has (c)cia, from caiva, but the latter combination is already known in Skt, whereas api plus iva is not. : This derivation has earlier been dismissed by Pischel (5 336) who, instead, suggests that the equation pi : vi may in fact have been responsible for the origin of piva from viva. A logical extension of Pischel's reasoning would be that piva is in reality but a hypercorrect form or a scholarly invention. For, an original initial v- remains vafter Anus vāra, and in this connection it does not matter whether it is a proclitic consonant, a glide or a medial -v- transposed through metathesis. Note that in the case of pi/vi the situation is reversed, the P- being the original and the v- its phonologically conditioned variant.
The fact that it is well established from the oldest Amg. texts onwards would seem to speak against this view of piva. This may, however, also be taken as an indication that these texts were subjected to rigorous redactional interventions. That this is indeed the case I shall try to demonstrate elsewhere.
It remains to explain the occurrence of piva in M. Pkt, in which, at least in the Sattasat, viva does not occur. It should be noted that at the beginning of the present manuscript tradition of the Sattasat stands a MS, or a line of MSS, written in Jaina Nāgars, which most likely will have been written by someone well acquainted with AMg. and/or JM. It is not unlikely that at this stage piva was entered into the text for via oc
