Jens:
> We do not replace *o by anything. In the stages of the history of IE
> that have *e and *o, the two are distinct phonemes. It's for the
> stages preceding the emergence of *o out of *e that we posit the
> same for both.
Yes, I understood that much.
> The rules that change *e into *o comprise: reduction of unaccented *e
> to *o; change of stem-final *e to *o before [+voice]; later,
> contraction of [e.o] (the product of lengthened /e:/) yielding PIE
> /o:/.
Hmm.
And little here is comparable to the theory I have except the *e/*o
alternation based on voiced segments which I attribute to a seperate
vowel *&. Actually there's something interesting here I can say on that.
The reason why I don't attribute *e/*o to *e instead is because this
screws up the plural in *-es. There is a minimal pair in eLIE between
*-&-s (the thematic singular > *-o-s) and the athematic plural in *-es.
This is the very reason why *-es is not reduced to **-&s as expected
in the first place! A merger would destroy thematic plurality, so these
two vowels must be kept distinct. However, I honestly can't think of an
example of unstressed **-as in eLIE, only *-&s, so perhaps one might
rather say that *& is really a surface *a. Or in your terms, that *e/*o
comes from a weakened *o instead. Of course it doesn't make sense
in your terms because your vowels are all wrong. (Damn, I'm a rude
sonuvab*ch.)
The explanation that *e: is a contracted *eo is weak. It simply suffices
to accept that *e: is a lengthened *e as we always find it. Don't the
facts as they are found in IE count for anything?
> So you don't get it? Tant pis.
C'est la vie :)
= gLeN