Re: [tied] The phonetic value of PIE *h3 and the 'drink' root.

On Sun, 7 Jul 2002 15:13:06 +0200 (MET DST), Jens Elmegaard Rasmussen
<jer@...> wrote:

>I believe I answered this some years ago. The root 'drink' is indeed
>*peH3y- and alternates accordingly, e.g., 3sg aor. *poH3-t (Skt. á-pa:t),
>caus. *poH3y-éye-ti (pa:yáyati) with retention of -y- before a vowel and
>loss in the environment VH_C# (and VH_CC). To go with the root aorist
>there was a reduplicated present, of which píbati etc. is properly the
>refashioned subjunctive. The old injunctive would be 3sg *pi-péH3y-t >
>*pi-póH3-t, 3pl *pé-pH3y-nt > *pé-pH3-n.t (the last form with y > zero in
>CH_CC, processed before full syllabification of the sonant, as general
>for these rules). The injunctive was now structured exactly like, say,
>*dhi-dhéH1-t, 3pl *dhé-dhH1-n.t and then formed its subjunctive the same
>way, which seems to have been *dhí-dhH1-e-t(i), cf. Skt. dádhati (with
>analogical e-reduplication from dádha:ti, witness the general type
>tís.t.hati, sí:dati etc. which must come from somewhere). For this verb,
>the form would be *pí-pH3-e-t(i), whence PIE *píbeti. The easiest way out
>is to consider the change pH3 > b older than laryngeal coloration, in
>which case /-be-/ offers no problem.

I believe I have found evidence for the regular loss of *y (*i)
immediately preceding or following a stressed thematic vowel, which
would also explain **pi-ph3-é-t(i) for expected *pi-ph3y-é-t(i). I'm
thinking in particular of the eh2(a:)-stems, from thematic vowel +
*-ih2, with e.g. Nsg *-eh2 (> -a:) for expected *-o-yh2. Another form
is the o-stem NAV dual in *-eh3 (> -o:(u)) for expected *-o-yh3. With
*y before the thematic vowel we have the reduplicated (causative)
aorist, where the causative suffix *(p)éy-e/o- appears in the zero
grade as *-(p)-é- instead of expected *(p)y-é- (e.g. á-ji-jñi-p-a-t <
**(h1e)-g^i-g^nh3-py-%'-t, therefore also sis.vapas < *si-swep-é-s <
*si-sw(e)p-y-%'-s). [where *% is the thematic vowel before it split
into *e and *o]

If the loss of *y is regular before a stressed thematic vowel, then
the Greek aorist épion (*h1e-ph3i-é/ó-) must be either secondary
(Jens) or it must be derived instead from *h1e-pih3-é/ó- (Piotr).

=======================
Miguel Carrasquer Vidal
mcv@...