Re: kentum/satem built-in already in PIE?

On 2006-08-03 20:32, tgpedersen wrote:

> I stand instructed.
>
> So, let me see if I got this straight:
>
> The causative suffix *-ejo-/*-eje- assumes the special
> form *-jo-/*-i:-, of which the first has an initial -j-,
> the second doesn't, which combine with a preceding -s-, here from

-k^-, to form -s^- and -s-, respectively

> which immediately refutes my suggestion.

What refutes your suggestion is rather the fact that there's no *s^ :
*s alternation before the thematic vowel *-e/o-, where *k^ and *g^(H)
are invariably reflected as *s/*z (as in *nesoN, *nesetI) irrespective
of the quality of the vowel, whereas *k/g and *kW/*gW become Slavic
*k/g before *-o- but *c^/z^ before *-e- (*pekoN, *pec^etI). And just
to clarify: in simple thematic presents, as well as *-je/o- stems, the
thematic vowel is simply retained in Slavic with its original PIE quality.

> Does this happen also after a stem in final PIE -s-?

Yes, of course, as in *gasiti 'extinguish, put out' (Lith. ge`sti,
Skt. ja:sayati): *gas^oN, *gasitI. It's only in RUKI positions that we
get invariable *s^, because its alternant *x occurred only before back
vowels.

Piotr