Jens:
> No, the root of the old collective is accented, cf. Gk. húdo:r,
> levelled from *wéd-/*ud-´.
No way. The first syllable in the Greek word is reduced. In Hittite,
it clearly isn't. This suggests that Greek had reduced the vowel. Why?
Obviously because it was unaccented. This automatically implies
*wedó:r with final accentuation. Afterall, now we have a pattern
whereby *wedó:r > Hittite /witar/ and *wednós > /witenas/ show pretonic
*e becoming /i/. On the other hand, accented *e is reflected undeniably
in examples like *esti > /eszi/ where *é becomes /e/, never /i/. In your
account, neither Hittite /watar/ nor /witar/ would properly reflect
**wédo:r which demands */wetar/.
So back to sanity, the Greek accent must have been secondarily placed
on the first syllable, perhaps by influence of agent nouns ending in
unaccented *-to:r in the nominative.
>> Latin shows genitive /pedis/ for a root clearly
>> otherwise in *o, as in /podium/.
>
> Hey, podium is a Greek loanword.
So? IE *o underlies the form nonetheless. It changes nothing.
>> So how anyone can deny the *ó/*e pattern sensibly is beyond me.
>
> Maybe it is. Still, I have explained it in the preceding posting.
> There is no immediate alternation *ó/e governed by the accent. The
> weak form of the -ó- of the perfect, of the intensive and of the
> reduplicated aorist is zero, not /e/.
Again, so? There were many patterns of vowel alternations in IE
going on all at the same time. In the paradigm of *es-, we see
the reduction of unaccented *e in the plural, hence *?s-énti and
yet another athematic paradigm such as that of *ed- shows *ed-énti
without reduction. On a third hand, the noun *?dont- 'tooth' shows
the original *e/ZERO pattern again!
The IE ablaut system was not intuitive and one had to learn the
proper ablaut pattern for any given paradigm or any given conversion
rule to sift through this madness. So the fact that *o doesn't
alternate in the above paradigms has no bearing whatsoever on whether
*o alternated with *e in *o-grade duratives or the paradigms of
mobile-accented nouns with *o-vocalism, where it must have surely done
this.
Considering the complexity of the ablaut system that IE had, there
would certainly be a motivation to abolish *ó/*e alternation but it
persisted in case pairs like *po:dm/*pedos and *wodr/*wednos
nonetheless. You can't skirt around this. I've already mentioned above
why it's paramount that we reconstruct *wednós as the only possible
source of Hittite /witenas/. So this shows conclusively to me among
other things that the *ó/*e pattern did indeed exist.
> Greek pédon and péde: have root accent, so does Hitt. pe:dan 'place'
> but surely one of them had final accent, since we have Skt. padá-m,
Humurously if /pe:dan/ means "place", it negates your expressed faith
that *pedóm meant "footprint" in a previous post!
We reconstruct *pedóm in the end to account for /padám/. I would like
to know the exact context of /pe:dan/ considering other forms /pieti/
and /pidi/ (mentioned at
http://www.wordgumbo.com/ie/cmp/hitt.htm).
I'll have to look this up.
We also have *yugóm, also proven to have had final accent and it's
of the same construction as *pedóm. So this all shows that Greek has
screwed around with the original accent and can't be relied on without
getting a confirmation of initial accent from other languages. We don't
see this confirmation in *wedó:r since Hittite reflects the final accent.
We should expect that Greek likewise switched the accent placement
in /hudo:r/.
= gLeN