On 2008-02-05 09:26, fournet.arnaud wrote:
> About d and s,
> an example is : *r_w "to flow"
> s-rew
> d-rew
> rava
> Are there not river-names with the root under its different forms ?
> I forgot to mention this example before.
I don't think it's a valid analysis. If I were to segment *dreu-
morphologically, I'd prefer *dr-eu- because of the synonymous *dr-ah2-
and *dr-em-. I actually plead guilty of conjecturing, in an article just
published (Indogermanische Forschungen 112), that *sreu-, *dreu- and a
few similar roots may contain a pre-PIE extension indicating forward
movement. From the purely IE point of view, however, *sreu- is
indivisible, and in particular has no s mobile. River names like Rava
can be explained otherwise and hydronymic evidence is too weak on its
own to prove the existence of *reu- as a byform of *sreu-.
Piotr