W dniu 2012-02-07 22:17, stlatos pisze:
> I'll say it again: opt. changes.
So you propose a houyhnhnm protoform for 'goose' to account for the
_absence_ of palatalisation in Slavic and then invoke "opt." to accunt
for the _presence_ of palatalisation in Baltic? It's way easier to
assume that the Baltic form is regular and the Slavic one is aberrant
("opt.", if you prefer). However, it seems to me that satemisation is
often cancelled in Slavic if there is already an inherited *s in the
same stem, cf. *svekUrU 'father-in-law' (Lith. s^es^uras), *kosa
'scythe' (Ved. s'as-, s'astra-, s'a:sa-, etc.). It may not be fully
regular, but then dissimilation at a distance rarely is.
Piotr