> Some time ago, I sent you an email with a couple of questions about the
> claims on the Nostratica web site. You said you would reply later, when
> you had the time. I haven't heard anything from you yet and I would like
> to make sure I haven't lost your reply somehow. There's absolutely no
> rush - please take your time if you're busy.
Oops, sorry, I haven't replied yet. I'll try to make up for it now. As
you probably suspect, I believe almost all long-range research produces
illusory results, and Nostratic is no exception. It isn't because I
dogmatically believe that all evidence for genetic relationship
evaporates after, say, ten milennia. Some models of lexical replacement
make it possible to estimate expected lifetimes of words depending on
their meaning (which in turn determines their frequency of use), and the
predicted lifetimes are of the order of tens of thousands of years for
the top-frequency words -- which unfortunately happen to be mostly
pronouns and other grammatical items, plus possibly just a few content
words like, for example, family terms and one or two of the lower
numerals. What we can get when we apply the comparative method to _very_
distantly related languages is a small set of correspondences which may
be suggestive (like the "mi/ti" phenomenon in the pronoun system) byt
insufficient to demonstate relatedness "beyond reasonable doubt" (which
normally requires a consistent reconstruction of a sizeable part of the
protolanguage system). I agree with Larry Trask that even some of the
widely recognised "families" (like Altaic and Afroasiatic) are rather
nebulous constructs, based on vague etymological equations and
questionable correspondences.
And yet ambitious Nostraticists like Dolgopolsky confidently reconstruct
Nostratic words for things like "sinciput", "leopard", "arrow", "pot"
etc., of course "reaching down" ad libitum and making extensive use of
phonological wildcards like *K(')U{r/l}c^ (where *K is "any velar"
etc.). This makes no sense to me. The _only_ way to get anywhere would
be to start with _really_ basic words and proceed _very_ cautiously to
rule out phantom correspondences. But if you do that, forget Nostratic
poems in the manner of Illich-Svitych and all that romantic nonsense.
Piotr