On 2006-08-12 21:12, Andrew Jarrette wrote:
> I personally like the idea presented in the paper written by the
> Finnish academe that you directed me to, that Slavic *-o in neuters
> comes from a variant ending IE *-o instead of *-om that Slavic
> developed, most likely to parallel the endingless i-stem and u-stem
> neuters, at a time when neuter gender was still an important
> distinction in the ancestor of Slavic and thus an ending different
> from the masc. acc. sg. might have been desirable.
> Andrew
Somehow the lack of this important distinction was no problem for the
Italic-speakers, the Greeks, the early Celts or the Indo-Iranians, all
of whom spoke languages with a fully preserved three-gender system. The
*-om of thematic neuters _is_ anomalous, given the consistently
endingless form of athematic neuters and the pronominal *-o-d. However,
whatever its true prehistory, it isn't likely to have developed directly
from an endingless noun stem, since the thematic vowel, when exposed in
final position would have had the e-timbre, as in the vocative. There is
no sufficient comparative ground for the reconstruction of *-o as a
variant of *-om already in PIE. I doubt if the ancestor of Slavic could
have generalised the pattern of neuter stems in *-i and *-u, since both
types had been completely eliminated by proto-Slavic times and their
members were either shifted to the masculine gender (*medHu [n.] -->
*medU [m.] 'honey, mead') or converted into productive declensional
types (*mori --> *mori-o- > *morje 'sea').
See also
http://groups.yahoo.com/group/cybalist/message/31827
and the discussion that followed.
Piotr