On Thu, 25 Aug 2005 11:58:28 +0000, elmeras2000
<
jer@...> wrote:
>--- In cybalist@yahoogroups.com, "pielewe" <wrvermeer@...> wrote:
>
>> Actually, the situation is more complex in that several other
>> possibilities are attested in the manuscripts alongside the textbook
>> ending -y, most spectacularly the use of a glagolitic letter that
>was
>> apparently developed for the express purpose of writing this very
>> ending and nothing else.
>
>Omygawd! That is of course highly relevant information, of which I was
>not aware. It does change the picture, and the unanimous -a of ORuss.,
>Czech and OPol. indeed looks significant after all. If the final vowel
>of nesy was a vowel different from other cases of Cyrillic y, there is
>of course very good reason to equate it directly with ORuss. etc. -a.
>That may indicate a more open pronunciation, i.e. something like an
>unrounded o (ë). It must somehow reflect a difference in the outcome
>of *-onts and *-o:ns (or *-ons?). In hindsight, it is highly
>reminiscent of the Lithuanian difference between veda~Ns and vilkùs.
I think vilkùs definitively proves that the PIE prototype
was *-o:ns (with acute ó:) > -uó(N)s > -ùs. The regular
outcome of *-ons would have been *-ons > *-ans > *-a:(N)s >
*-às.
In the participle (*-onts), the nasal was protected by the
/t/, which also kept the vowel from lengthening: *-an~ts >
*-aN~ts > -aN~s (with circumflex nasal diphthong).
Something similar may indeed have happened in Slavic. In
the acc.pl. I believe that (non-front) vowels were raised
twice before final -N and -h, which explains the difference
between acc.sg. *-ah2m > *-a:N > *-o:N > *-aN > -oN and
acc.pl. *-ah2ns > *-a:Nh > *-o:Nh > *-u:Nh > *-u:N > *-u: >
-y.
If the /t/ of the ptc. was lost early on, we should have:
*-onts > *-ants > *-ans and further with raising *-uNh, with
lengthening before -Rh -u:Nh > *-u:N > *-u: > -y (and in the
soft stems *-jans > *-juNh > *-ju:Nh (> *-jo:Nh) > *-je:Nh >
*-jê/*-jeN).
On the other hand, if the /t/ remained until _after_ the
raisings (but was lost before the lengthening: both nesy and
nesa have long endings after all), we would have: *-ants >
*-aNts > *-aNs > *-a:Nh, with a long nasal /a:N/ (or /o:N/,
see below), corresponding to Kortlandt's /aN/, which might
conceivably have given /a/ (the other long nasal vowels
develop as: *-i:Nh > -i, *-u:Nh > -y, *-(j)e:Nh >
-(j)ê/-(j)eN, so loss of nasalization is more or less the
rule, unlike what we see in the short nasal vowels: *-iN >
-I, *-uN > -U, but *-eN = -eN, *-aN = -oN).
The soft stems would then have: *-jants > *-jaNts > *-jaNs >
*-ja:Nh, which would then have to have been affected by
j-umlaut to *-je:N(h) > -jeN. We have a problem here,
because neither *-jaN nor *-ja: are affected by j-umlaut
(e.g. acc.sg. zemjaN, nom.sg. zemja:). However, just like
we have spontaneous lowering in the case of *-ju:Nh (>
*-jo:Nh) > *je:Nh (Gsg., NApl. zemjeN/zemjê), we could have
had spontaneous raising in the case of *-(j)a:Nh >
*-(j)o:Nh, and after -j- merger with the result of *-ju:Nh
into *-je:Nh.
Alternation in the m./n. N(A)sg. between *-ants (> -a) and
*-ans (> -y) is the kind of variation which is not
surprising to find within dialects of a single language, so
that seems totally plausible.
=======================
Miguel Carrasquer Vidal
mcv@...