Re: [tied] Re: Nominative: A hybrid view

> [...]I'm refering to the number of distinct (ie: contrastive) vocalic
> phonemes exist in a language. There is no language I know of
> where there is one distinct vowel, not even Sanskrit. People
> had tried with Kabardian too but this failed.
>
> Sanskrit is not monovocalic in this sense because there ARE contrasts
> and we can see quite clearly that the language has five short vowels
> "a", "e", "i", "o" and "u" with long counterparts.
>
> A single minimal pair like "Shivah/Shivaah" suffices to prove that your
> statement is false and that Sanskrit is _not_ monovocalic.

Of course /a/ and /aa/ contrast, but so do /t/ and /tt/; we do not posit a
unit /t:/ for that reason. So, as far as that goes, it really is
monovocalic.

I am not saying there were no niceties where /i/ and /y/ do in fact appear
in a non-predictable way, for there are, but they are marginal. And PIE
has the same little holes in its skeleton, even under the most extremist
one-vowel theory. But mostly we do not *know* PIE well enough to make
statements about these things, so for practically ALL legitimate purposes
PIE *does* become a one-vowel language if one removes /a/ and makes /o/ a
predictable phenotype of /e/ (which, however, one cannot do and I do not
advocate).

Jens