On 2009-05-28 19:34, Carl Edlund Anderson wrote:
> I have recently been wondering about Proto-Italic *mari (Latin mare)
> alongside PIE *mori, and happened across this old thread while trying to
> figure out the apparent /o/ > /a/ change. However, the thread seems to
> raise more questions than it answers! :)
One possible explanation (quite widely accepted, I think) is that Lat.
/a/ reflects a schwa secundum generalised from *m&rei-, the putative
weak allomorph of *mori. I don't find it very convincing, as the
vocalism of *mori would suggest an acrostatic paradigm, *móri/*méri-.
Jens Rasmussen posits a special acrostatic pattern with *ó/*á (his other
examples are *stom-(n-) 'mouth' and *loku- 'lake'). Some blame the
strange behaviour of *mori- on its substratal origin (but can it be
substratal if it's found in Ossetic?). Italic is the only branch which
shows an unexpected development.
Piotr