Re: [tied] Re: ka and k^a

From: glen gordon
Message: 40873
Date: 2005-09-29

Rob:
> Personally, I'd say that such an inventory is less
> likely, phonologically speaking.

Yes, logically, we should only side with the *more*
common unless there is adequate basis to reconstruct
something more unique. That then means that if we
need a three-vowel system, it should be *a, *i and *u
until evidence shows something "odder" like
centralization, say.


= gLeN




__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com