Brian:
> Nonsense. A reconstruction offers much weaker evidence than
> a directly observable language, but it is not wholly without
> evidentiary value.
You're speaking nonsense. It has no value in determining language
universals because the theory itself is _based_ on these language
universals which are supported by attestable data in modern languages
as well as written ones in six-thousand years of historical records.
Proto-Semitic isn't the source for this evidence, it is the result.
Pure and simple. We simply don't know whether there are stems with
lexical vowels in Proto-Semitic (and all we need is one hard example)
because it is a theory in progress. We can't possibly know what we
will find in the future until we find it! With a living language,
it's all there before us.
= gLeN