On 2006-04-20 18:58, mandicdavid wrote:
> How do all these sounds fit into PIE phonotactics?
They are basically obstruents, but they pattern with sonorants in that
they can be syllabic when flanked by other obstruents. In some branches
(e.g. in Greek) they were vocalised also in other interconsonantal
environments.
> Another thing, I can't work out how these fricatives (back, and maybe
> even low) could yield an 'i' in indo-iranian, corresponding to 'a' in
> other language groups, when vocalised.
The result of the vocalisation in most IE branches was a kind of central
schwa *[&], which could easily merge with *a. In Indo-Iranian, the
vocalisation of the laryngeals yielded *[(h)&], with the consonantal
element *[h] still present in the reflex of PIE *h2, and the
accompanying vowel different from PIIr. *a. Let's sidestep Iranian,
where the development was slightly more complicated, and focus on Indic:
*[&] changed into high central *[I] and eventually yielded /i/ plus
aspiration of the preceding stop if the *[h] survived sufficiently long.
This is why in words like <duhitar-> the initial syllable may count as
heavy in the Vedic metre: it _was_ closed and therefore heavy in
Proto-Indic *dug[hI]tar-.
Piotr