On Sat, 7 Apr 2001 12:47:59 +0200, "Guillaume JACQUES" <xiang@...>
wrote:

>Les trois laryngales correspondent a k en ouralique dans la couche la plus profonde (la seule qui soit susceptible de ne pas etre empruntee). Ca pourrait aller avec l'idee de Miguel que les laryngales remontent a *k dans certaines conditions (j'avoue d'ailleurs ne pas avoir completement bien compris ton systeme, Miguel).

Alors:

One of the "typological" abnormalities found in reconstructed PIE
phonotactics is the near absence in word-final position of the
unmarked stops *-p, *-t, *-k. The only clear example is verbal
3rd.sg. *-(e-)t, but that one may easily be analogical after present
tense *-ti. That a final -/k/ can easily lead to -/x/ needs no
further comment, and the corpus of IE cases where *k alternates with
*H (mainly *h2), as given by Martinet and others, can at least in part
be re-interpreted as evidence for a development **-k > *-h2 (which is
not to say that there are no cases of "hardening" of laryngeals,
**h1/2/3 > *k: we certainly have to take that possibility into account
as well). On the basis of Hitt. ins.sg. -et vs. *-eh1 elsewhere (and
other cases, like *met- ~ *meh1- "to measure"), I concluded that *-t
could have developed into *-h1 (though not until the "post-Anatolian"
stage). For *-p (/p/ of course being the most marked voiceless stop)
I have no evidence one way or another. *If* there were cases of final
**-p at all, and *if* they too developed into laryngeals (but why
should they?), then the only possibility would be the labialized
laryngeal *-h3. In the course of my investigations on the internal
reconstruction of pre-PIE morphology, I stumbled on further
alternations that might be explained through the transformation of
final voiceless stops, in particular the plural variants in -es(w)
(Armenian -k`!) and -i- (from *-a:tu ~ *-ati, cf. also the numeral
*trey- < *tiláti > *TalaT-), and the duals in *-eh3 ~ *-h1 (from
*-a:ku ~ *-aki). To list my hypothesized soundlaws concerning final
voiceless stops in tabular form, we have:

*-pi > *-p^ > ?
*-pa > *-p > ? (*h3?)
*-pu > *-pw > *-kw > *-h3

*-ti > *-t^ > *-y
*-ta > *-t > Anat. *-t, non-Anat. *-h1
*-tu > *-tw > *-sw > *-s

*-ki > *-k^ > *-h1
*-ka > *-k > *-h2
*-ku > *-kw > *-h3

I have only considered my theory for the case of grammatical endings.
In so far as "roots" could function as independent words, a root like
*dheh1- might derive from **dheki or **dheta, besides **dhehV or
**dhe?V, with "original" laryngeals (assuming **h and **? both gave
*h1).

=======================
Miguel Carrasquer Vidal
mcv@...