Jens:
> That does not matter here, for the statement I protested against was
> one comparing Sanskrit with PIE. And if an abstract analysis in
> terms of a system with a single vowel phoneme is possible (or
> closely so) for one of these, the same goes for the other.
That wasn't what I was denying so no wonder Richard didn't understand
what you were objecting to. Nor did I. I think I understand now what
you evidently misread in my post.
I was objecting to claiming any language has a one-vowel system outside
of these clever little analysis games. On the surface, sure, but so
what? Yes, Sanskrit _and_ IE can be analysed to some extent as having
what appears to be one underlying vowel. However, phonetically and in all
honesty, we have [i], [u], [o], [e] and [a] with long counterparts at
the very least. So... we have some ten vowels despite all the parlor
tricks.
Similarly, in IE, we have [e], [o], [a], [i] and [u] also with long
counterparts despite your fun little games. So since we know that
despite the success or failure of a monovocalic 'abstract analysis' of
Sanskrit, the preceding stage (namely IE) is known to have had at least
ten distinct vowels too.
Therefore, I object to anyone who tries to reason that if IE can
be analysed as underlyingly monovocalic (which may or may not be
justified in itself) then that implies that pre-IE must have had only
one vowel. No. This is absurd. It's illogical assumption. It does
not imply such a thing. I see you doing just that, Jens, and it irks
my sense of reasoning because there is a gap that is not being filled.
= gLeN