From: Mate Kapovic
Message: 32178
Date: 2004-04-22
----- Original Message -----
From: "Sergejus Tarasovas" <S.Tarasovas@...>
To: <cybalist@yahoogroups.com>
Sent: Thursday, April 22, 2004 9:30 AM
Subject: RE: [tied] Re: Nominative Loss. A strengthened theory?
> > From: Mate Kapovic [mailto:mkapovic@...]
>
> > It does. Could it be solved this way maybe? *e (3) is
> > attested instead of *e in G. sg and n/a pl. of a-stems and a.
> > pl. of o-stems in NorthSlavic. In pre-ProtoSlavic we had G.
> > sg. *-a:s, n. pl. *-a:s and a. pl. *-a:s in a-stems and a.
> > pl. *-a:ns in o-stems. Both *-a:s and *-a:ns would yield -y
> > in later Slavic. But in later-to-be South Slavic, a. pl. of
> > a-stems becomes analogically *-a:ns instead of *-a:s and soon
> > after that G. sg. and n. pl. take the new ending being the
> > same before that as well. From this *-a:ns we have -y and -e
> > in South Slavic. But in later-to-be North Slavic we have
> > a-stem a. pl. analogically affecting o-stem a. pl. and it
> > changes from *-a:ns to *-a:s. So all four endings are now
> > *-a:s there which develops as -y (the same as *-a:ns) after
> > nonpalatal consonants but as -e (*-a:s > *-e:s > -e) after
> > palatals. The -y looks now the same because both *-a:ns and
> > *-a:s > -y but the original difference is seen in soft stems.
> >
>
> Looks attractive, but how would you explain the "double raising" *-a:s >
> *-o:s > *-u:s > *-y in phonetic terms? *s# raises, but is that enough for
> *a: to turn *u:? Or you prefer to be neo-gramatically agnostic as to the
> intermediate stages?
Yeah, that's a bit of a problem there :-) Maybe it could be solved by saying
that *a and *a: (or whatever it was phonetically (maybe [o] and [o:]) yields
*u(:) before final nasal or *-s (or both)? If *-a:s was phonetically [o:s]
there should be no problem turning it to *-u:s...
But just as an example, it seems that there was a change somewhat similar to
this one in Modern Persian. There every /a:n/ > /u:n/. I think there was no
medial stages attested and that it change pretty quickly. /a:n/ is
considered colloquial and /a:n/ is considered standard. So you have for
instance [tehru:n] for <Tehra:n>.
Mate