We can't. There's no historical continuity
between the two. Even if the PIE ablaut triad *e/*o/Ø (plus *a derived from
*h2-coloured *e) are reflexes of a single underlying _pre_-PIE vowel (let's
call it *A) -- and this is indeed what internal reconstruction within PIE
suggests -- Vedic (as well as Proto-Iranian) /a/ does not derive directly from
it, but from the merger of _more recent_ (but still PIE) *e, *o and *a. I
won't bore the list with the proof: the releveant facts have been presented more
than once, and anyone interested in them may search the archive. Additionally,
there are morphemes containing *o which does not undergo qualitative ablaut
(and, if *h3-colouring is ruled out, may derive from something
different than *A). Finally, Vedic /a/ is also a reflex of PIE syllabic nasals
(as in <s'atam> < *k^mtom).
There are about a dozen possible PIE
reconstructions of the initial <a> in <as'va-> (*h2a-, *h3o-, *n-,
*h1e-, *h1m- ...), which shows that the sound represents the convergent
development of many different sounds or (here, mostly) sequences of sounds. We
canot reconstruct PIE forms on the basis of Vedic or even Vedic plus Greek. More
witnesses are usually needed to disambiguate the reconstruction.
Piotr
----- Original Message -----
Sent: Thursday, March 21, 2002 9:06 AM
Subject: [tied] Re: Sanskrit and e, a, o
--- In cybalist@......,
"tgpedersen" <tgpedersen@......>
wrote:
> > According to the good old "Sanskrit consonants
and Greek
vowels" rule, PIE had a vowel (how come this "Ablaut vowel"
doesn't
have a name?) which was realised as e/o/nothing.
How about
calling it Vedic a.