Re: [tied] 'Dog' revisited

25-11-03 03:08, Glen Gordon wrote:

> However, you said that "the compound analysis of the 'dog' word explains
> in a simple and natural way its apparently aberrant -- actually archaic --
> forms
> and accounts for the observed range of variants (*k^won- fares worse in
> these respects)." How do you mean by "observed range of variants"? Are
> you referring merely to the declension (which can be explained otherwise)?
> Are you referring to variants of the root itself whereby *kwon- isn't the
> only
> reconstructed phonetics of this root? I didn't quite follow the jist of that
> last
> paragraph. If you could elaborate, that would be dandy.

I mean the stress on *u as in Greek <kúo:n>, which is not normal in true
Lindeman forms (we have *dijé:us, not *díje:us), but matches the regular
pattern of compounds, and especially of *h2jú-ho:n (Skt. yúvan-). True,
we have Gk. dúo: as well, but that's another case where the disyllabic
form (*du[w]-o:) may be older than the monosyllabic one (*dw-o:). While
*k^wó:n is explicable as a contraction of *k^úwo:n, the latter form is
difficult to understand if we start with *k^won- (*k^wó:n/*k^unós). And
if the Avestan forms with <su:n-> are inherited, they are also
problematic from the point of view of the "null hypothesis".

Piotr