alexandru_mg3 wrote:
> 2. *swesora: [s/w > zero; o>a] > *wesara: > [V-s-V > V-h-V] >
> *wehara: > vë-ha-rë (3 syllables: intervocalic h>zero as d,g,b>zero)
>
>>*vëarë > varë
But PIE *swéso:r had initial stress, and in none of its forms was it
shifted to the second syllable. Proto-Albanian shows many instances of
stress retraction to the initial syllable (beside rarer instances of
preserved PIE stress), but not, to my knowledge, of the reverse kind of
pattern required by your analysis.
> See also krunde 'bran' ~ variant: grundë (-> Romanian grunz)
>
> (as you have said all the variants are important ...'for sensible
> etymologies' :) )
>
> This is the `second example' (-> and there are others too showing a
> voiced /g/)
The existence of such variants doesn't strengthen you case. Being
sporadic and unconditioned (unlike the treatment of nasalised vowels,
which are regularly retained in Geg and also regularly denasalised in
Tosk), they introduce another element of arbitrariness -- enough, in my
opinion, to sink the etymology. As far as I know, Albanian uses the same
word for wasps in general, not just for hornets, so the identification
of <grenzë> etc. with *k^r.h2sen- is hardly compelling in the light of
the formal problems. So far you have ignored my examples of *r.H > ar in
words whose derivation is uncontoversial and regular (such as <bardhë>
or <parë>, where everybody agrees as to the PIE prototype), and you are
attempting to demonstrate a different development using only doubtful data.
Piotr