--- In
cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
> By the way, I forgot to mention that for the a:-stem gen.
> sg. and nom. pl., I could not find a way to make a
> derivation directly from *-a::s work, not for the soft
> stems. At some point, intrusive nasalization must have been
> introduced, even where we find -jê, not -jeN. Plain
> *-(j)a::s would have given -y/-ji.
Because of different (-je^ ~ -jeN) reflexes this intrusion must have
occured very late -- probably at the time after the delabialization
and subsequent denasalization *-u:N > *-y: and the merger *ei > *e.:
to *i:, so Slavic would had:
G.sg. *z^eny: ~ *zemji:
N.pl. *z^eny: ~ *zemji:
Acc.pl *z^eny: ~ *zemjiN.
Then the soft declension could have been levelled on the model of the
hard one, namely "G.sg.=N.pl.=Acc.pl". *iN was preferred to *i:
simply because introducing *i: would have created an over-homophony
in the soft declension, since there already were two more cases in *-
i: (D.sg and L.sg.).
Sergei