Rob:
> The thing with *xnérs 'man' vs. *xnéres 'men' is that there's no
> difference in accentuation. I think this must mean that the plural *-
> es must have been added after the main changes in stress-accent.
In my theory, MIE has a nominative singular in *-sa and a plural in
*-es. Since they are both the same number of syllables in a word
either way, the accent does not alternate according to QAR. Syncope
reduces *-sa to *-s with added lengthening (Szemerenyi) while *-es
retains its original vocalism in defiance. (Otherwise it would have
merged with the thematic singular in *-&-s.)
This works perfectly in both directions, explaining how the plural
and the accent associated with it came to be in IE on the one hand,
and what IE's exact relationship is to Uralic, Altaic and EA on the
other. Since *-es can be regularly derived from ProtoSteppe *-it,
there is little left to ponder on that.
> That means that the earlier numeral forms were **trei and **kWetwor,
> respectively.
Yes, basically. I'd drop the idea that *kWetwor- has a suffix **-wor-
though.
It appears in all likelihood that *kWetwor- is an indivisible root
and I've related it to Tyrrhenian *xota from earlier *kWatWan for
what its worth. The fact that Tyrrhenian had a numeral *xota is proven
not only by Etruscan /hutH/ (sometimes said to be 'six' simply because
of a presumed numerical pattern on die) but also the fact that the city
of Tetrapolis (Greek for 'Four Cities') happens to have a Tyrrhenian
alternative name that was also recorded in Greek as Ytte:nia (presumably
from something like *xot-ena "four tribes, four peoples"). The *-ena
suffix, by the way, is a common suffix throughout Greece showing up
constantly in words and names deemed to be "Pre-Greek". It is also
coincidentally used in Etruscan to denote a person or collection of
people as with /Ras'na/ 'Etruscan' which is why I think many of these
Pre-Greek items are a mixture of not only Anatolian but Tyrrhenian
loans.
= gLeN