I was objecting to the concept of "apophonic o" or "alternating o"
which appears nonsensical. What difference could it possibly make
that the /o/ concerned is just part of any-old alternation series as
opposed to an /o/ that is not known to appear as anything else? The
reason for lack of alternation may even be lack of knowledge about
it which makes it even more absurd.
Even so, the claim that *H3e does not yield the o-vowel that
develops into /a:/ in an open syllable in Indo-Iranian is not absurd
in itself. It would mean that *H3e developed a vowel of lesser
sonority than the other o-vowels. And if /e/ was less open than /o/
(from the many sources), then its rounded allophone in *H3e could
well have been a closer kind of [o], in which case lack of
lengthening might reflect a perfectly natural restriction on its
display of sonority.
That explanation, however, cannot be used for páti-s whose short /a/
will still have to be credited to the closed syllable of páty-e; but
then why not do the same for ávi-s/ávy-e?
Surely, the matter will have to be decided by other examples. I
still believe the best one adduced is stya:yate, but I would like
there to be more - one way or the other.
I find it a bit strange that the question of a possible PIE phonetic
difference between *H3e and *H3o has apparently not been addressed -
or have I missed it?
Jens
--- In cybalist@yahoogroups.com, "P&G" <petegray@...> wrote:
> >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