>Is there for instance a special reason that Sanskrit
> /i/ and /u/ cannot have the same status as *i and *u in PIE? Any thoughts?
Sanskrit is sometimes offered as "proof" that a real language can be
analysed as having only one vowel. To a large extent this works, but
remember that there are exceptions. There are sequences that contrast
(e.g.) -u: a- with -v a-. There are also other places where the vowels
/i/ and /u/ never ever appear as consonants.
But I don't think the exceptions disprove the basic points:
(a) perhaps there isn't anything to be afraid of in an analysis which allows
PIE only one vowel at some underlying level, provided it accepts more than
one at a phonetic level.
(b) Sanskrit. like PIE, gives one vowel a very different role from any
other.
Peter