Re: [tied] h1,h2,h3 in Albanian

Abdullah Konushevci wrote:

> If you agree that *H1ep- with o-grade *H1op- yields in Alb. <hap> and
> related words <prapë> 'again', <prapa> 'back': ON after: Hett. appa: Gr.
> opithen ("Vare Arveord Etymologis Ordbok" by Harald Bjovrand & Fredrik
> Kortland, Oslo 200, pp. 20) and *H1ed- 'to eat', which thematic form
> *H1edom- yields Alb. <ha> 'I eat': OHG ezzen: Lat. edere: Sl. jesti
> 'id', then *H1ag'- 'to drive, lead' and its o-grade form seems to
> continues regular outcome *H1o > Alb. /ha/ (*H1(ed)óm, *H1op-o > Alb.
> <hap>, as well as *H1og'mo-lo > hulli 'furrow').

But, Abdullah, a PIE reconstruction must account for the forms found in
other groups apart from Albanian. In Albanian, *a and *o fell together,
but they remained distinct in several other branches. The vocalism *a
can be due to a fundamental *a (at least I think so), but such an *a
does not alternate apophonically with *o. So the /a/ found in this root
in Greek, Latin, etc. (which, unlike Albanian /a/, can't reflect PIE *o)
must go back to *e coloured by an adjacent laryngeal. In other words,
we're left with *h2ag^- as the only possibility.

Piotr