>*H3, or in reality *H2 with o-grade?
I take your point, but these arguments seem apposite:
(a) ablaut patterns
If we assume H2 with o-grade, then we have to assume o-grade in some verb
forms where we expect e-grade. I admit that we don't actually know a whole
bunch about the e/o ablaut, but it would be nice to be consistent, so that
the pattern in:
dhi-dh-e-H1-mi
and si-st-e-H2-mi
were repeated in di-d-e-H3-mi, rather than having an ad hoc di-do-H2-mi.
So also Bhleh3 "talk"
Likewise, why should we suggest an o-grade in these roots which have an
initial H3e- or H2o-:
h3ed smell
h3eid swell
h3ed hate
h3elh1 destroy (hardly a "stative" or "inner quality")
h3emh swear
etc - about 8 verbs. (There might be more of an argument for the nouns,
such h3es branch.)
(b) Forms where -o- cannot be analogy or o-grade (are there any?)
Will h3reg' "rule" do? The e-grade is in the following
syllable.
Likewise h3reiH "get moving"
and a couple of other such verbs, pattern HCeC-
Visitors are about to call, and my brain has run out. More later, perhaps!
Peter