Re: [tied] Re: Bog

On Wed, 05 Sep 2001 14:23:30 -0000, "Piotr Gasiorowski"
<gpiotr@...> wrote:

>Actually, Winter's Law is quite well substantiated, and accounts
>better than any other proposal for such items as *e^d- 'eat', *se^d-
> 'sit', *nagU 'naked', *a^zU 'I', ablUko 'apple', vydra 'otter' and
>many more, traditionally explained by proposing arbitrary vrddhi not
>found in other branches. Exceptions are not many, and some are
>explicable, though there is a really troublesome residue too, the
>existence of *voda being probably the main reason why some IEists
>hesitate to accept Winter's Law. Apparently, either there is another,
>still unrecognised, complicating factor at play (like Verner's Law
>after Grimm's), or Slavic *voda is not such a straightforward reflex
>of *wod(r)- as it might seem to be. Ranko Matasovic' (a silent member
>of our list) has published an article trying to pin down the exact
>conditioning of Winter's Law. I don't agree with his conclusions but
>share his interest in that topic.

Jens Eldegaard Rasmussen in his article "Winter's Law of Balto-Slavic
lengthening" concludes that Werner lengthening was blocked by a
following sonorant (OCS <ognI> "fire", not *<agnI>), and that it
further only took place in the syllable preceding the accent, i.e. not
under the accent (this would explain *wódr > wod-a) or twice removed
from it (*maderós -> Lith. ma~daras, not *modaras). I have some
problems with the accent-restriction (BS. e:d- and s:ed- are accented
on the lengthenes vowel at least in the sg. present forms), but the
blocking by a following sonorant (usually -n(-) or -r(-)) seems real
enough, and might even account for the case of <voda> (from *wódr,
*wednós, after all).