--- In
cybalist@yahoogroups.com, "Rob" <magwich78@...> wrote:
> --- In cybalist@yahoogroups.com, "elmeras2000" <jer@...> wrote:
>
> > I don't know, but my guess would be prehistoric sandhi between
stem
> > and flexive. The unmarked vowel is certainly /e/ in the stage we
> > reconstruct. We find /o/ caused by accent weakening and by
sonority
> > increase. The common denominator would seem to be tonal
lowering.
> And
> > [o] does have a lower tone than [e], so that's my guess at a
> > causation.
>
> The UTexas IE site mentioned some kind of tonic accent scheme as a
> possible origin for Ablaut, which I think is likely. As you
> said, /o/ has a lower tone than /e/. If the rule in (pre-)IE
words
> was that each word could have only one high-tone syllable, then
any
> vowels preserved with low tone would become /o/. What remains is
a
> way to determine where low-tone vowels were preserved and not
> apocopated.
I have offered that too. I gave some of it in the preceding posting.
It should be mentioned that Miguel and Glen also offer explanations
for these things, very different from mine.
Without engaging in polemics I can list my rules in crude wording:
Only short unaccented vowels are lost. That loss occurred after the
change of unaccented e to o, but before the lengthening caused by
the nominative marker (and the collective marker). I assume that
some stems have underlying long //e:// which is shortened when not
originally accented; that caused ablaut e:/e. After ablaut the
accent was retracted to any remaining pretonic full vowel; that
caused "acrostatic" é:/é. I assume (naively perhaps) that
lengthening of already-long /e:/ yielded a vowel with two syllabic
peaks, one of which was unaccented and so became [o], whereupon
contraction led to /o:/; that gave acrostatic ó:/é. A long vowel was
shortened in the nominative and collective if the stem ended in two
consonants: *nókWt-s, *-ont-s, *-ont-&2; these had long vowels at
the time of the loss of unstressed short vowels.
The thematic vowel is a different story: -o-/-e- according to +/-
voice in the following segment, independent of accent. May have
contained a special phonetic feature when unaccented short vowels
were lost.
Also different is ó/zero as in some reduplicated structures: *kWe-
kWór-e, *kWe-kWr.-mé. Grave error on the Texas site: Greek pépoitha
does not reflect e > o after accent, for the accent was *on* the /o/
in PIE: Greek does not show the IE accent in finite verbs.
Further different is the o-infix story of the causative and the type
*bhoró-s. Not another word about it here.
Much of this is subjected to the vicissitudes of the "contrastive
accent": *bhoró-s 'carrying, carrier' => *bhóro-s 'act of carrying'.
Jens