On Thu, 1 Nov 2001 12:17:02 +0100, "Piotr Gasiorowski"
<
gpiotr@...> wrote:
>This is not clear to me. Skt. has <ga:vaH>, which could of course be explained via Brugmann's Law in *gWowes, but why is there no length in the "adverbial" cases (<gava:, gave, gavi>) or in the dual? *{gWeh3-ew-es} is at least a possibility.
The dual has length (Skt. ga:va:(u) < *gWow-eh3). As to the lack of
Brugmann's Law in the weak cases, that's a good point. In an
acrostatic paradigm (Schindler's group Ia), we would expect:
N. gWou-s (> gWo:us)
A. gWou-m (> gWo:m)
obl. gWeu-
The oblique form *gWeu- does not seem to occur. In Latin or Slavic,
this is to be expected (normal development *eu > ou). In Greek or
Sanskrit, it is a problem. One can assume analogy from the strong
stem, aided by the phonological tendency to round *eu to *ou,
especially after a labiovelar. For Sanskrit, an additional
requirement is that *eu must have become *ou _before_ palatalization,
but _after_ the working of Brugmann's Law.
>On the whole, the assumption of a laryngeal explains some strange aspects of the "cow" word (especially the marked pitch accent of the nom.sg.), though for everyday purposes I am quite content with *gWo:us, etc.
>...
>From: Miguel Carrasquer Vidal
>1) There's no lengthening in the nom. pl.
>...
I simply meant that, assuming no laryngeral in the "cow" word, the
"Szemerényi" lengthening does not occur in the nom. pl.: nom.sg.
*gWo:u-s, nom.pl. *gWow-es.
There does seem to be a laryngeal in *gWou-H- "[bull]shit" (Slav.
govUno).