Re: [tied] *gWerh3- "to devour"

From: Miguel Carrasquer
Message: 15649
Date: 2002-09-21

On Thu, 12 Sep 2002 03:38:11 +0000, "Glen Gordon" <glengordon01@...>

>Plus, I am having a hard time fitting a uvular *xW in the IE
>consonant inventory. I mean, we have *x (a uvular corresponding
>to *H2) which should be part of the uvular series (*K, *G, *GH).
>And, just as there is a plain non-uvular counterpart for each
>velar, there would be one for *x (ie: *h or *H1). The labial
>varieties of *k, *g and *gH however do not appear to be uvularized
>and so one would expect there not to be such a uvular labial as
>*xW as one would expect to be the qualities behind *H3. So I find
>myself compelled to reject *H3 and conclude that *H2 and *H3 are
>the same unrounded uvular phoneme *x.
>Am I mad?

Just wrong. The error is in equating *h1 with the plain velar, which will not
do. The fricative corresponding with the plain velar series *k, *g, *gh is
velar *x, while the uvular series *q, *G, *Gh has a corresponding fricative *X.
Just like there is no way to tell which *kW's were originally velar **kW and
which were uvular **qW, there is no way to tell which *h2's were originally
velar **x and which were uvular **X. As to *h1, it's in a different category,
neither velar or uvular, but glottal (laryngeal), and it probably represents the
merger of a separate series of two or three earlier consonants (surely **h and
**?, perhaps also **3, if we want to arrange them in a nice aspirated/
glottalized/voiced triad corresponding with pre-PIE **TH/**T'/**D and whatever
that was to give later in PIE itself).

As to *h3, it represents the labialized counterparts of *both* *h2 and *h1 (i.e.
at a deeper level of reconstruction, it can stand for **xW, **XW, **hW and
**?W). In Hittite, *h3- sometimes gives h- (like *h2-), sometimes zero (like
*h1-). It stands to reason to suppose that the first category of *h3's are in
fact *h2W's, while the second category are *h1W's. In the rest of Indo-European
(with the possible exceptions of Armenian and Albanian), all that can be seen is
the effect of *h3 on neighbouring vowels and the reflexes of a vocalized *h3,
and they are identical for *h3 = **h1W and *h3 = **h2W.

What remans puzzling is the voicing effect of *h3, as opposed to *h1 and *h2.

Miguel Carrasquer Vidal