Re: [tied] The disappearance of *-s -- The saga continues

On Tue, 13 Apr 2004 02:03:19 +0200, Mate Kapovic
<mkapovic@...> wrote:

>
>----- Original Message -----
>From: "Miguel Carrasquer" <mcv@...>

>> But -a:s and -a:n _are_ the same thing, basically,
>
>How? I don't understand...
>
> >cf.
>> i-stem acc.pl. masc. -i:n vs. fem. -i:s, u-stem -u:n/-u:s
>> and r-stem -r:n/-r:s (from *-ins, *-uns, *-rns).
>
>Yeah, but couldn't this be analogical to o- and eh2-stems? Whence else -i:n
>and -i:s etc. from the same ending *-ins?

One could think of grammaticalization of different sandhi
variants.

>> >, and in Gothic they would also give the same thing and
>> >we have -o:s and -ans (directly attesting PIE short *-o-!).
>>
>> Could be Osthoff shortening.
>
>In Gothic? Why not in *-eh2ns > *-a:ns > *-o:ns then? Since *a: and *o:
>merge in Germanic shouldn't *-eh2ns > *-a:ns and *-o:ns produce the same
>ending?

Not necessarily. The two forms may have developed
differently. Jens has already mentioned the possibility
that *-m- was lost after *h2. My suggestion of Osthoff's
law would also have applied differently (*-o:ms > *-oms vs.
*-eh2ms > *-eh2ms, resulting in Gmc. *-anz vs. *-o:z).

Another possibility is that the acc.pl., like the gen.pl.,
showed hesitation between short *o and long *o: from the
very beginning: the peculiar sequence *-oym(s) was
contracted to *-om(s) in some dialects, to *-o:m(s) in
others; or even to *-om vs. *-o:ms (apparently in Slavic) or
*-o:m vs. *-oms (apparently in Germanic).

=======================
Miguel Carrasquer Vidal
mcv@...