--- In
cybalist@yahoogroups.com, "Mate Kapovic" <mkapovic@...>
wrote:
> This is totally unnecessary. The word can be derived from PIE
regularly with
> no problem at all.
The reconstruction "*k^orHkeh2" has the following weaknesses:
1. It does not account for the Sl. forms with sv- which match Alb.
sorrë in a most intriguing way.
2. The -i- of Skt. s'a:rika: is not from schwa, cf. the Iranian
borrowing in Armenian sarik. Also, Brugmann's law appears not to
produce length before -C&-, cf. IIr. *bharamïnas (Av. bar&mno:) vs.
Gk. pherómenos.
3. The comparanda do not fit the form.
Nor of course does a putative Balto-Slavic *c'(v)a:rka: match Alb.
sorrë exactly, but it begins and ends right, and the middle may eb
dispensable. For a prestage of the Albanian word would certainly
have a sequence "rounded palatal + /a:/ (from older /e:/) + /r/",
which is exactly what the BSl. word has. The final /-ka:/ would be
natural in hypocoristics, which often take the shape "shortened word-
form + suffixal -ko-/-ka:" (I find seka for 'sister' in Skok's SCr.
etym. dict.). So it would be a perfect form of a word
meaning 'little blackbird', supposing there was a word *c´va:r(s)na:
in the relevant prestage of Albanian. Now, that would result
regularly from IE *kWe:rsnaH2 (or *kWersnaH2 in case rs is reduced
with compensatory lengthening, which I am beginning to like).
Jens