On 2008-01-22 02:55, stlatos wrote:
> The oldest fem. ending that can explain this would be *puksni:x
> (analogical after *potni:x) in PIE or soon after. If s>0 / stop_$C
> (assuming for now that -st formed an onset when possible) then
> *pukni:x > *fuxi:n+ / *fuxo:n+ while *puksos > *fuxsaz, etc.
Not impossible, and even quite elegant, except that *fuxi:n- is
unattested (the <vixen> type is a derivative in *-in-jo:); but if the
loss of *s is accepted (other examples would be welcome), one can
imagine an originally nasal stem, i.e. something like
*puk^s-en-/*puk^sn- > *fuxsin-/*fuxn- > *fuxan-, fem. *fuxo:n- (just
improvising), beside the thematic type *puk^s-o- > *fuxsa-.
>> ... possibly < *puk-s-k^o-) ...
> This is much more likely to be met. ~ *puksyo+ > *puskyo+ >
> *pus^c^(y)o+ (like tus^c^ias vs tuccha- 'empty', etc.).
I can't see why it should be _much_ more likely. The suffix *-k^o- (or
exists, and the development of *sk^ and palatalised *-sk- is the same in
Skt. However, I wrote "possibly", which means that I'm willing to
consider other analyses. In either case, I like the idea that the "bushy
tail" morpheme was *peuks- rather than just *peuk-; Slavic *puxU 'down,
something fluffy' can then be analysed as *pouks-o-, a straightforward
thematic derivative.
Piotr