On Tue, 18 Jan 2005 14:31:59 +0000, Sergejus Tarasovas
<
s.tarasovas@...> wrote:
>I prefer a solution more or less along Andersen's lines: *-is, *-us >
>*-ix, *-ux, then *x analogically spreads to the *o-stems (*-as -> *-
>ax etc.; an analogical explanation is inevitable for *x in at least
>some non-RUKI environments anyway) but not to the N.sg. of the *s-
>stems, where the final *-s was supported by the oblique cases (*nebas
>retained its *s after *nebes-es). Desinential vowels then are raised
>before *-x [-h], but not before *s, and eventually the *C of -VC# is
>deleted.
I might as well throw in my solution:
*-os regularly gives -U in Slavic, -as in Lithuanian. This
explains the Slavic o-stem nom. sg. *-os > -U, dat.pl. *-mos
> -mU, verbal 1pl. *-mos > -mU. The neuter endings, which
should have given *-U (o-stems, < *-om, s-stems, < *-os),
were replaced across the board by -o (pronominal, < *-o(d)).
Lith. Dpl. -mus cannot come from *-mas (< *-mos), but it can
come from *-mo:ns (= OP -mans). In the Lithuanian 1/2pl.
endings -me:, -te:, the *-s has been replaced by length,
presumably after the dual endings *-wah2 > -va: (Sl. -vê
analogical after pronoun *weh1), *-tah2 > -ta:, Sl. -ta.
PBS had both *-mes and *-mos as the verbal 1pl. ending:
Slavic -mU besides -me [Kriv., Bulg., (Cz/Svk.)], Baltic
*-me: < *-mes besides *-mo: < *-mos (OP esmai "I am, we are"
[both from *esmo:]).
=======================
Miguel Carrasquer Vidal
mcv@...