--- In
cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
> >[JER:]
> >> >The word ve^dró is a vrddhi formation *we:d-r-ó- 'associated
with
> >> >water'.
> >>
> >> I doubt it. The length looks like Winter's law.
> >
> >Length looks like length, sure. A cluster /dr/ is one of the
> >barriers to the operation of Winter's lengthening.
>
> Another fine example is *udráh2 > vy"dra (Latv. ûdr(i)s,
> Ved. udrá-).
You make it sound as if this is just another one out of a million. I
don't think I know of any examples of apparent lengthening before
media + sonant other than vêdró and vy"dra. For vêdró independent
length is a very obvious possibility. For vy"dra I only notice that
the feminine (or generic) form of an animal name is combined with
vrddhi in Lith. várna (1) and Germanic Huhn. I would therefore
assume that the most likely explanation of its length is that the
vowel was long already before any working of Winter's law. If Lith.
ú:dras (3) and Latv. ûdrs have taken over the long vowel from ú:dra
(1), then vy"dra does not need to have ever changed the position of
its accent. The BSl. forms would be masc. *u:drá-s, fem. *ú:dra:,
levelled from *udrá-s, *ú:dra:; if one distinguishes the genders it
is precisely masc. udrá-, fem. húdra: that is quotable from Sanskrit
and Greek (whose masc. húdros is then due to levelling).
The failure of vêdró to cause accent retraction is harder to account
for if one accept a theory saying that clusters of this kind should.
One of your own recent postings (Febr.25, on the Nostratic-L list)
contained the very interesting observation that Slavic iterative
verbs with lengthened root vowels show acute on high vowels and
circumflex on non-high vowels. My guess, which you did not like, was
that they were not of equal length; that was only an immediate
reaction to a piece of information that was new to me, but it could
perhaps come in handy here. Suppose the accent retraction caused by
(non-strident) clusters did not work in vêdró because the vowel was
*too* long, perhaps disyllabic in some sense (I simply do not know
what a syllable is), then the weight of the segment that mattered
for the question of retraction could have been too light, but heavy
enough in other cases.
> >> I wasn't just referring to peró of course. It was shorthand
> >> for (from Zaliznjak's Old Russian list): licé, loz^é, mytó,
> >> peró, plec^é, pljuc^é, bIrvInó, veretenó, volokUnó, govInó,
> >> gumInó, kopIjé, okUnó, pisImó, polotInó, pIs^enó, res^etó,
> >> sedIló, sukUnó, sIrdIcé, tenetó, tolokUnó, jajIcé.
> >
> >Those that had final accent would keep it there
>
> So you agree with me that there were three AP's in
> Proto-Balto-Slavic.
At least. Longer words offer room for more variation. Was that ever
in doubt?
Jens