--- In
cybalist@yahoogroups.com, glen gordon <glengordon01@...> wrote:
> Piotr:
> > Actually, if *h3 was distinctively voiced (note its
> > voicing effect in the Hoffmann compounds and in
> > *pi-bh3-e-), one would expect the thematic vowel to
> > have become *o prior to, and independently of, the
> > laryngeal coloration.
>
> I'm convinced now that the original pronunciation of
> *h3 was still /hW/. I can see how /hW/ can be then
> later voiced to /H/ (IPA hooked-h) in much the same
> way as English "wh" has been voiced in dialects like
> my own.
Okay, crazy idea time. Supposing the glottalic theory is right, could
*H3 actually be *p'? *p' doesn't appear word initially (except maybe
twice). Now I know it's typologically going to be rare, but that rare?
It could disappear initially with rounding of the following vowel. PIE
*p does that in Celtic (except presumably with no rounding) so this
wouldn't be surprising. In Armenian *p > h so we'd almost have a
precedent for *p' becoming a laryngeal. That would seem to agree with
the Anatolian and Armenian (if indeed H3 is preserved in Armenian)
evidence. For the most part *p' only survives after non-syllabic
resonants. As an ejective we wouldn't expect it in suffixes, and it
seems that we don't. Also, by being an ejective, it would eventually
become voiced.