Re[2]: [tied] Greek labiovelars

At 4:30:33 PM on Tuesday, March 14, 2006, Sean Whalen wrote:

> --- "Brian M. Scott" <BMScott@...> wrote:

>> At 4:05:02 AM on Tuesday, March 14, 2006, Sean Whalen
>> wrote:

>>> --- "Brian M. Scott" <BMScott@...> wrote:

>> [...]

>>>> *g, *gH, and *gWH all become OIr /g/; *gW becomes OIr
>>>> /b/. Jackson says that all of this is Common Celtic, at
>>>> least initially.

>>> I looked at a few sources (most very old) but it seems
>>> like gWH remains separate from gH:

>> I presume that you mean in Common Celtic. Matasovic
>> appears to agree, and at least one of your examples looks
>> legitimate, but two definitely aren't:

>>> *gWHer-; gorim (I warm) OIr; gwre:s (heat) Welsh

>> The <w> is a vowel; this is /gure:s/.

> If you're saying that because of this gwre:s can't come
> from gw- from gW- I disagree (see below).

I was saying that in that case it wouldn't obviously be a
good candidate, but given Jim Rader's comment, that
objection is now irrelevant.

[...]

>>> *gWHen-d-; geind (wedge) Ir; guenn M. Breton

>> OIr <genn> (I presume from *<gend>) 'a wedge, a block of
>> wood'; Pokorny gives Br <genn>, MCo <genow>, Co <gedn>,
>> and W <gaing> and derives the lot from *gHe(n)d-. The MBr
>> form is clearly not representative.

> What does it matter if only one form retains w?

On the available evidence I see no reason to think that it
didn't *introduce* /w/. MacBain even mentions an OBr <gen>,
though I don't know on what evidence.

[...]

Brian