Jens:
> Since *-os contains the thematic vowel, the seeming conflict is not
> real, but the nom.pl. *-es does raise eyebrows, for unaccented short
> vowels should be lost
No. This is what I used to think until I realized that only unaccented
*a disappears, while unaccented *e remains in early Late IE as *a
(which later goes on to become alternating *e/*o). This takes care of
a lot of things that would otherwise be called "violations" if your
belief were true, such as all thematic verb stems!
> I have suggested the solution that the vowel of *-es is a post-ablaut
> epenthesis (vowel insertion),
Epenthesis without motivation. This is not a valid solution. The vowel
*e is there to disambiguate it from any singular forms, that's it.
> I have stated the conditioning of all types of IE o many times: The
> thematic vowel is stem-final and pre-voice; *H2ák^-mo:n has
> lengthening of an unaccented *-e-;
Contradiction: accusative *xék-mon-m, not **xék-men-m
> *pó:d-s has lenghtening of an *-e:- already long;
Contradictions:
accusative *pod-m, not **pe:d-m
plural *pod-es, not **pe:d-es
Look, Jens, I'm not even going to go further. Your solution isn't
explaining any of the commonmost patterns here which is why I can't
take it seriously. You need to then make up a new reason as to why
we don't find the vowels where you say we should find them. Why bother?
You say I'm the king of analogy but it looks to me that your entire
explanation here is a large bag of analogies.
My solution does away with this crap. In eLIE after Syncope, we have
*pad- "foot" with the following declension that's already much more
regular:
nom *pa:d-s *pád-es
acc *pád-m *pád-ms
gen *pad-ás *pad-ám
We still see Szemerenyi Lengthening in the nominative, which is what
we expect to find because of the clipping of the nominative ending *-sa
during Syncope. There is no vowel alternation between *o and *e yet,
since this later comes from the pretonic raising of unaccented *a -- In
closed syllables we get *e while in open syllables we find *i.
Likewise we can see that *xekmo:n- is from an earlier, much more
sensible eLIE paradigm:
nom *xékma:n-s *xékman-es
acc *xékman-m *xékman-ms
gen *xekmén-as *xekmén-am
Again, lengthening causes *a: in the nominative due to clipping,
producing later *o:, while the unaccented short *a becomes *o before
voiced *n in the nominative and accusative. In the weak cases, accent
predictably shifts to the next syllable and replaces unaccented *a with
accented *e according to normal ablaut rules at that time.
Notice how this explains things without mess AND I don't need to perform
double-lengthening voodoo either. This proto-language prestage in fact
is wonderfully normal.
= gLeN