>The "thematic vowel" (/o/ and /e/, not connected with the accent,
stem-final)
>The causative (-o- always unaccented)
>The perfect (-o- accented, unaccented zero)
>These three are totally different. There is therefore no sense in
>combining their o-variants into a common concept of "ablauting o".
Yes, you are right to point to significant differences between them.
But:
(a) they all occur in verb stems which (usually) also show -e- grade
forms;
(b) they all appear (usually) in Skt as long /a:/.
The fact that that there are differences between them does not invalidate
the claim (b) - though other things might!
If (b) is right, and long Skt /a:/ does not regularly appear from other
PIE sources of *o, that seems to me to be significant information about the
pre-history of the -o- grade in verbs, and the prehistory of Skt, even if
what it means it not yet clear.
Peter