*gwh in Gmc.

[Hans-Werner Hatting]

On Tue, 23 Jan 2001 20:04:06 -0000 Douglas G Kilday wrote:

>I still think *pempwe is a better fit for Early PIE. Otherwise the Germanic
>forms require an ad-hoc assimilation of *p__kw__ to *p__p__ mirroring the
>Italo-Celtic assimilation to *kw__kw__. But if this happened, why wasn't
>*perkw- affected (Lat. quercus, OE fyrh, OHG forha)? We don't have
>*firf-trees.

>From the reflexes in IE languages I can remember offhand, Indo-Iranian,
Greek, q-Italic, and q-Celtic require a PIE */kw/. Of course, one can
operate with different stages of PIE to account for irregularities, to unite
forms which seem to belong together, or to establish links to other language
families. But in the case of "5", we do not really need a Pre-PIE */pw/. To
me, the assimilation theory looks satisfyingly convincing, and we have the
correspondent assimilation (only the other way round) in Latin and q-Celtic.
And assimilations are often singular items, not being consequently applied
to all the material in a language.

Anyway, the proposed sound change **/pw/ > */kw/ looks unusual to me. We
have a lot of /kw/ > /p/ in IE languages, but does anybody know of instances
(except assimilation) for the change proposed by Douglas Kilday?
A further argument for an old /kw/ is the nasal. By PIE phonological rules,
we would expect /m/ before /p/. But the Gmc. languages have mostly the
reflexes of /n/, which is possible before labiovelars like /kw/ (probably
being realised as (ng)), but not before true labials. Later occurences of
/m/ in Gmc. languages can be easily explained as assimilations.

DGK:
>Is there any objection to *-kwe coming from *-pwe?

If You don4t accept Gmc. /f/ out of */kw/, but retract all instances of it
to a */pw/, You cannot link the second element of "5" with the eclitic PIE
*-kwe, as it is conserved in Gmc. as Gothic _-uh_, so it should contain PIE
/kw/ even in Your model.

For my part, I think it4s simpler to assume that the reflexes of PIE /kw/
occasionally merged with those of /p/ in some stage of Proto-Gmc.
under the influence of labial consonants in the same word.

Best regards,
Hans-Werner Hatting