On Mon, 29 Sep 2003, Miguel Carrasquer wrote:
> >A preform *staHinos would change into *stayHnos in (i.e., before) PIE
> already.
>
> I suppose we disagree on that.
A clear and well-known example is this: There is acute in Lith.
píemeniN; the root can be assessed on the basis of Gk. pô:u 'herd' as
*poHy-; now either *poHy-men- produced an acute directly, or its
metathesis product *poyH-men- did. Neither possibility suits you.
>
> >Lith. íes^kau 'I seek' sure *is* parallel: the root is posited
> >by Pokorny as *ais-, the sk^-prs. seen in Skt. icchati is regularly
> combined with an s-aorist, so lengthened grade is expected, the
> result being *H2e:ys- which neatly explains the Baltic acute.
>
> So not parallel at all: *h2e:i dos not equal *ah2i (or *aih2).
This is getting undignified: The stimulus was this question from you
"Wouldn't a form *sta:inV- also give circumflex?"
followed by my attempt to understand and answer it:
"With lengthened-grade PIE /a:/? No, that yields acute (Lith. íes^kau)."
As I was given to understand your question this is indeed parallel, even
if it has lengthened-grade /e:/, not /a:/.
>
> Skt. ápnas- "Ertrag, Habe, Besitz" shows no indication of a laryngeal.
If it is /H1/ it should not show word-internally after a stop. I do not
venture to bring in Greek aphneós 'wealthy', because the base word
appears to be áphenos 'wealth', and an ongoing discussion seems at a
loss to get the morphology right anyway.
>
> Interesting are some forms with *o: (Skt. á:pas- "work, religious
> act", ?Osc. uupsens "fecerunt", OHG uobo "Landbauer"), which perhaps
> belong with
> *o:k^us as reflexes of *h3o-.
I have addressed the matter in two papers, still unpublished: Roots of
the structure /HeT-/ (I wouldn't know for /HeTH-/, so we can't use it
for that) do not form "o-infix derivatives" of the form *HoT-, but
instead leave the o- in its original prefixed position: *o-Hp(H)-éye-
'ueben", *o-g^H-á-H2 'leadership' (Myc. o-ka /o:ga:/, Brent Vine, basis
of other Greek ag-o:gá:); *o-H1d-éye- caus.-it. 'eat' in Arm. utem.
It's one of the things I'll have to write up very soon.
That does not save Vedic "á:pas" which, however, is "nicht zu sichern"
(Mayrhofer); I cannot make head and tail of the RV attestations,
including the variants. Pok.'s *o:k^-u- 'quick' leaves me too few
alternants to do algebra on.
Jens