Miguel:
>Yes, it must be *-oh3, or at least a labialized laryngeal. The NA form
>varies in Skt. between -a: and -a:u. The same /u/ appears in the genitive
>*-ous (< *-&3-s) and locative *-ou (*-&3-u < *-xW-i), as well as in the
>Greek o-stem oblique form -oiin < -oiun (*-oy-&3-m).
You're going to have to distinguish between your own theories and
reality. I assume you're talking about paradigms like that of *okto:u
"eight" and *dwo:u "two" where there is no clear evidence in favour
of *h3. These roots are not normally reconstructed this way because
most other people favour Occam's Razor as a valuable principle of logic.
It is more economical to reconstruct a long vowel *o: here.
= gLeN
_________________________________________________________________
The new MSN 8: smart spam protection and 2 months FREE*
http://join.msn.com/?page=features/junkmail