On Tue, 23 Jan 2001 20:08:20 +0100, "Piotr Gasiorowski"
<
gpiotr@...> wrote:
>The fact is that if one relaxes the standard constraints on regularity even a little (by ignoring vowels, allowing "doublets" and other n-tuplets, permitting ad hoc consonantal permutations, etc.) the likelihood of finding spurious cognates increases dramatically, since a *vast* number of new sound combinations can be treated as "equivalent".
But that's not what I was talking about. Both Illich-Svitych and
Bomhard (don't know about Møller, as I haven't really studied his
proposals) stick rigourously to their regularities, most of the time,
and obtain results which, not always, but in a number of cases, seem
rather convincing, or at least suggestive of a real genetic connection
between the "Nostratic" languages. The "doublets", which, even if we
don't "allow" them, are still there [look under the carpet!], explain
the curious fact that sometimes both proposals overlap semantically,
but not, strictly, phonetically. What the significance, if any, is of
this, I couldn't say. The only thing that seems obvious is that if
Nostratic exists, the exact sound correspondences (as yet unknown)
will hardly be as simple as, say, Bomhard's PIE ~ PAA *t ~ *t, *d ~
*t', *dh ~ *d. That just may be correct as a Grimmian first
approximation, but there's a whole lot of Verner still required.
=======================
Miguel Carrasquer Vidal
mcv@...