On 2007-05-31 22:20, Sean Whalen wrote:
> Again, there's no ev. anything like this happened in
> PIE. For example, ts>ss in most IE, but not Sanskrit.
> Should I then assume PIE t+s > ts+s and s>0/t_s in
> Skt?
What happened in Indic is clear: the affricate *{ts} (or any other
dental affricate), when followed by an obstruent, lost its fricative
component, becoming a stop. In most other IE languages it lost its stop
component, becoming a fricative. The whole affricate is preserved in
Hittite. The Indic development is demonstrated by the behaviour of the
sigmatic aorist, as in áruddHa = á + rudH+s+ta, with the same kind of
simplification. The development of *t+s was therefore into *{ts}s >
Indic /ts/ (but /ss/ in most branches).
Piotr