On 05-01-21 17:50, Miguel Carrasquer wrote:
> Have you compared notes with Jens' theory about these
> *VHi/*VH/*ViH/*Hi/*&-forms? I seem to recall a discussion a
> long time ago, but I can't remember if anything was
> resolved.
Oh, yes. I was on summer vacation at the time, and unable to cope with
the e-mail traffic on my occasional visits to Cybalist, so there wasn't
much of a discussion on my side. My "smoothing theory" left Jens
unconvinced (and I myself wouldn't any longer defend all its elements),
although I think we both, as well as other people, agree at least about
one thing: the last word about the "long diphthong roots" has not been
said yet.
I still think that, given the normal structure of PIE verb roots, with
final *RC being normal and *CR anomalous, we should expect *...RH rather
than *...HR as well; I also still believe my theory is phonetically
plausible, and that the fact that it works with no recourse to
metathesis in cases like *ph3-j-tó- > *pih3-tó- is a merit (for me,
*pih3- is just the nil grade of *peih3-). For Jens, *póh3-tlom and
*póh3-mn. are regular forms resulting from loss of root-final *j before
two consonants. For me, they are equally regular forms resulting from
the smoothing of *oi before a syllable-final laryngeal (other than *h1,
for this particular diphthong). In *pi-pih3-eti I assume root
simplification (in reduplication as well as compounding) through loss of
an unstressed high vowel (after a liquid, the final laryngeal would have
been lost instead, as in *pi-plh1-e-ti > *pi-pl-e-ti). This might also
explain Gmc. *kwikwa- as a reflex of reduplicated *gWi/e-gWih3-o-
(details and parallel examples on demand).
Jens has a different explanation from mine (and, I must admit, a
convincing one) for *-ih2-, gen. *-j-ah2-s (the <devi:> type), but I
haven't heard his opinion on some other of my examples, e.g. *trej-es
vs. *tri-h2 (< my *trei-h2), or the apparent nil grade in *píH-wr.,
*píH-wo:n 'fat'.
Piotr