Miguel:
> That's impossible, because there is nothing there to
> lengthen the */o/.
That's a biased and unproven assertion. Yes, it is
possible because...
IE *-o: < *-owi < unaccented *-omi
This occured just after Vowel Shift, so it is a
change that is quite late. Dialects were probably
already beginning to break away from the core by this
point, in fact. Perhaps, the occurences of *-o-mi
in post-IE dialects is a lingering archaicism from a
more 'innovative-resistant' dialect that got
swallowed up by the IE expansion? Food for thought.
> We only have */o:/ out of oH or -oRs/-oRh2
Nope. We have not only the 1ps thematic *-o: as
proof of the loss of *m but also locative *yugoi,
again with the loss of *m seen in nominative *yugom.
The difference between the two merely lies in
accent placement. In both, however, *m disappears
between *o and *i, as if first becoming *w. So
we can formulate a thorough post-Vowel-Shift rule:
*-ómi > *-ówi > *-ói
*-omi > *-owi > *-o:
It's reminiscent of the Hittite /m/~/w/ alternation.
>> 2) *-oh2
>
> That's impossible, because the thematic vowel is
> */e/ before */h2/.
Exactly. We expect **-eh2 (identical with the
feminine ending) because all laryngeals remain
unvoiced during Final Voicing.
> The only possibility is *-oh3.
That is not a possibility at all because there are
no securely reconstructed suffixes with a word-final
labialized phoneme anywhere.
= gLeN
__________________________________
Yahoo! Mail for Mobile
Take Yahoo! Mail with you! Check email on your mobile phone.
http://mobile.yahoo.com/learn/mail