From: stlatos
Message: 68538
Date: 2012-02-10
>Gmc. would have changed tH > t, not tH > t > T. Analogy/contamination in thanvan- = bow Av; dhánvan- S; seems likely. An expl. for this is needed independently of concerns about t-/tH- in the other root. Assuming Proto-Iranian had a word like *tengWHtlo+ = shaft/arrow > *tankHtHra+ > *tHankHtHra+ (t-tH > tH-tH), cont. between arrow and bow (and the words seen as related/der. of bow) could account for a change in initial C. Similar possibilities w slightly dif. stages are also pos.
> W dniu 2012-02-09 22:59, dgkilday57 pisze:
>
> > Also, how does Olsen account for initial unvoiced aspirates in InIr?
> > E.g. Sanskrit <phe'na-> 'foam', Ossetic <fink>; Skt. <kumbha'-> 'peak',
> > Avestan <xumba->; Av. <Tanjayenti> 'they draw', <Tanvar-> 'bow'
> > (apparently the only exx. with *tH-, Bartholomae reconstructed as
> > *tHan,gW-).
>
> I'll write more about the IIr. words later. Let me only remark now that,
> as for 'draw', even LIV has *tHengH- (a root aorist), the only entry
> there with initial *tH. The root is attested in several branches, though
> the only evidence for the initial aspirate is Iranian. *h2 is ruled out
> by Germanic cognates with uncoloured *e in the root syllable.
>