--- In
cybalist@yahoogroups.com, Mate Kapoviæ <mkapovic@...> wrote:
>
> On Pon, sijeèanj 22, 2007 10:39 pm, mcarrasquer reèe:
> > --- In cybalist@yahoogroups.com, "mandicdavid" <davidmandic@>
> > wrote:
>
> >>Wouldn't o: be raised to u: then, cf. mati < má:te:r?
> >
> > <Mati> and <mote.~> are from *máh2te~.
>
> Just a methodological point. If you assume PIE variants *-e: and
> *-e:r and you allow for to different "accents" on them, it is not
> surprising that you can get practically anything by assuming a
> reflex of one of those variants.
In the case of the n-stems, the facts clearly indicate that these
variant forms existed: cf. Skt. -a:, Latin -o: vs. Greek -ó:n, and
probably also the redistribution of these variants in Germanic
(Gothic m. -a < *-o:n, f./n. -o: < -õ, while OHG has m. -o < *-õ,
f./n. -a < *-o:n). The facts also indicate that Lith. -uõ cannot be
derived from a PIE form ending in a nasal. In the r-stems, the
overall evidence is less clear, but as far as Balto-Slavic is
concerned, all the extant forms (with the exception of vodá) can
easily be derived from the circumflex variant with loss of the
resonant, parallel to what can be seen in the n-stems:
PIE -o:r/-õ = Lith. vanduõ
PIE -e:r/-e~ = Lith. dukte.~ and Slavic dUtji with -e~ raised to -i:,
just like -õ is raised to -u: in kamy, etc.