On 04-10-05 12:36, Sean Whalen wrote:
> I'm not sure I'm convinced, but if so might it
> merely be a case of "double-locative" from -i-en? A
> few common phrases could be lexicalized and perhaps
> these would be generalized from to form some other
> words. The example might even be a triple-locative
> from *g^Heim-en-i-en (at each stage of addition the
> whole phrase is lexicalized, repetition).
But the first *-en- is in fact a stem-forming suffix, not a case ending,
let alone a postposition: nom. *g^Héim-o:n, gen. *g^Him-én-s, etc. (pace
Hirt's analyses, for example). It seems to be connected with *-nt- stems
(cf. Hitt. gimant-) rather than any locative formations, and doesn't
show any characteristically locative semantics. Gk. kHio:n (< *g^Hjo:m)
means 'fallen snow, ice-cold water' etc., kHeima (< *g^Heim-n.(t)) means
'winter weather, frost', and kHeimo:n (< *g^Heim-o:n) means chiefly
'winter' as a season. Locatives of *-n. and *-o:n words in *-en (WITHOUT
a case ending) or in *-en-i are commonplace, but I've never seen an
example of *-en-i-en .
Piotr