Re: [tied] *gWerh3- "to devour" and this odd feeling of deja-vu

Once again, I recall this is something that we discussed at some depth a long tome ago, in the early months of Cybalist. I actually raised it myself, trying to argue for a two-laryngeal system. A lot of such stuff is buried somewhere in the archives and will be more and more difficult to recover until somebody finds a way to conduct a full search.

More recently (Message #13972), Jens Rasmussen discussed the evidence in favour of *h3 being distinctively voiced; he mentioned Hamp's analysis of Celtic *abon- < *h2ap-h3on- plus a group of examples discovered by Birgit Olsen and involving the so-called Hoffmann suffix *-HVn, which seems to voice preceding stops in Latin (as in vora:x, -a:cis --> vora:go:, -inis). I have my doubts -- not concerning the effect of the laryngeal, which appears to be real enough, but its identity, since I prefer the reconstruction of the Hoffmann suffix as *-h1en-, chiefly on the basis of its reflexes in Balto-Slavic. I don't mind discussing it all again.

Piotr

----- Original Message -----
From: Glen Gordon

To: cybalist@yahoogroups.com

Sent: Saturday, September 14, 2002 6:24 AM

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

Peter: >Yes, but also delightful. I am intrigued by your idea that H2 and H3 were >one phoneme, and unrounded. I am sure you have a way of explaining the -o- >colouring which led to the positing of H3 in the first place. How does it >go? Well, if I understand, what one would call "undeniable" instances of *H3 are few and far between. No doubt, there appear to be many examples of *H3 but are they really *H3, or are they in reality *H2 with o-grade? We know that while Anatolian preserves laryngeals, even this branch doesn't preserve the distinction of *H2 and *H3... so does it really exist or is it a throwback from the pre-Hittite days of yore when laryngeals were first being hypothesized without evidence to back it up? For example, we have traditionally written *deH3- "to give" but we could just as well write *dox- (*doH2-). When zero-graded, it would yield *dx- which should normally end up as **da- in many IE languages that lose laryngeals, but analogical pressures that would associate this zero-grade with the full-grade form *do:- would probably resist such a normal development since if gone unchecked it would create unnecessary ablaut irregularities. Thus we end up with an affected *do- that makes it seem as though we are dealing with the instance of an *H3 when it may not be the case. Are there any suffixes with *H3? Are there any particles with *H3? Can we be sure that *H3 in these instances cannot be interpreted as *H2 with o-grade?