Miguel:
> Yet another solution that yields circumflex -รต, but
> we need acute -o:.
No. The shortening of /o(w)e~/ to *o: wouldn't cause
circumflex. The cases of circumflex involve an
intermediary *h1, as in the example of the thematic
genitive plural, *-o-hom. Here, it is the combination
of thematic vowel *-e/o- and vowel-initial suffix *-om
which causes the intrusive *h1. In the case of the
1ps thematic pronom. ending, *-o-mi causes no such
pattern for obvious reasons.
So the loss of *m is apparently irregular and the
two vowels instead collide without *h1 having a
reason to appear.
> Sorry, but I completely miss that. The nominative
> *-s (from *-z) has nothing to do with the genitive
> *-os/*-es/*-s, [...]
Early IE speakers misanalysed genitival derivatives
in *-os as *-o-s (them.vowel plus _nominative_ *-s).
I've repeated this so many times, you can't possibly
have missed it. Drink some coffee :P
> The -m of the thematic neuter is simply the
> accusative ending.
Then it shouldn't appear in the vocative... oh but
it does! Ergo, your theory is wrong and should be
forgotten.
So back to my theory.
As I said, inanimate stems in genitive *-om were
missegmented as thematic stems in *-o-m.
This misanalysis is the only thing that can credibly
cause the creation of the pseudosuffix *-m in *all*
the cases we *actually* see them in -- The ones that
coincidentally were once _unmarked_ in the pre-IE
inanimate paradigm as suggested by all the other
attested inanimate stems.
= gLeN
____________________________________________________
Start your day with Yahoo! - make it your home page
http://www.yahoo.com/r/hs