At 5:36:29 PM on Saturday, February 16, 2008, fournet.arnaud
wrote:
>> One of Miguel's posts includes such
>> reconstructions as
>> *ú:d-an > *wódr.
>> with stressed long vowels and unstressed short ones.
> *wodr is from
> *H2ut?-
> with -o- intrusion
This is completely irrelevant to the issue that I was
addressing, namely, your inaccurate claim that Miguel's
reconstructions have only stressed short and unstressed long
vowels.
Brian
==============
Ok
You are right on this point.
Now
I consider it farfetched to
work with non-laryngeal *long* vowels
in order to *erase* o.
I think PIE had *a/e *i *o *u as vowels
with no long/short contrasts.
I think the system (*a *a:) *i *i: *e *e:
*u *u: is uselessly complicated
and I think it's not even supported by PIE itself
not to speak about macro-comparative data
Arnaud
==================