> > My point was more that PIE /a/ seems to have a different function both
> > from /i,u/ and from /e,o/.
> What "different function"?
/a/ just is. Whereas /e~o~zero/ correlates with certain grammatical forms
(eg /o/ in causatives, nouns). The alternation may have been originally
phonologically conditioned, but whether it was or wasn't, it is misleading
to reconstruct PIE with the simple 5 vowel system /i u e o a/. Instead it
should be reconsutructed with
(a) vocoids which are syllabic forms of resonants: / i u m. n. r. l. H./
(b) an ablauting vowel with one of two different forms, if present: /e ~ o
~ zero/
(c) /a/.
Peter