Torsten:
> Does your belief in PIE /a/ include its being
> 'native' PIE? Surely, if pre-PIE /a/ -> {/e/,
> /o/, zero}, PIE would have had a hole where /a/
> was supposed to be (except for the environment of
> 'plain' /k/, /h2/, etc).
It _would_ have had a hole there if it weren't for
an exception to Vowel Shift where *a normally
becomes *o just before the final stage of IE.
In some cases, the change is resisted. No sound
change rule is absolute afterall. Hence *a exists
and we don't need to think that it's because of
those big bad foreign people :P
= gLeN
__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com