Re: Grimm's Law is about to expire (Collinge 1985, p. 267, Thundy 1

From: Richard Wordingham
Message: 47846
Date: 2007-03-14

--- In, "mkelkar2003" <swatimkelkar@...> wrote:

> The reason I (Thundy) am skeptical of many proto-Indo-European roots
> based on Grimm's law is that most Indo-Europeanist apply Grimm's law
> universally and claim that there is no exception to it. For the Old
> English habban, dictionaries give *kap as the root and ignore the
> Latin habere as cognate, because Grimm's Law does not allow for any
> exception! Then we have the classic cases of Sanskrit pibati, of the
> Greek pachus versus the Sanskrit bahu, and the Latin bibit versus the
> Praenstine pipafo (Collinge 259-65). By postulating laryngeals we can
> get around these problems, as scholars have done.

We don't need any laryngeals to explain the consonants of Greek
_pa:khus_ / _pe:khus_, Sanskrit _ba:hu_ and English _bough_. It looks
like a straightforward case of Grassmann's Law to me.

What might be helpful is to collect together the cases of inconsistent
voicing between IE families. However, please leave out the examples
that can be explained by Grassman's law - they will swamp the data and
serve no useful purpose.

> even though Grimm's law is about to expire
> (Collinge 267) with the radical revision of the proto-Indo-European
> obstruent system by the glottalicists Gamkrelidze, Ivanov, Hopper and
> Bromhard.

The other Nostratic camp (e.g. Shevoroshkin) is currently
reconstructing the PIE, Altaic and Nostratic stop system as having the
phonations [th] (PIE *t), [t] (PIE *d) and [d] (PIE *dH). The old
(Indocentric?) reconstruction has a lot of life left in it.

As Peter Gray said, it is important to separate the two aspects of
Grimm's law - the correspondence and the reconstruction. I think
everyone realises that reconstructing the precise phonetic reality is
, which is why I like to think of the 'traditional' reconstruction as
an orthography.

> Since every language comprises dialects, consistent voicing
> of consonants is not necessarily an infallible criterion for
> distinguishing dialects or language families—in American English, for
> instance, the voicing of /s/ in words like greasy is at best a
> variable dialectical or ideolectal feature.

This is not a particularly good example. The variation in _greasy_
simply results from the tension between regular sound change (yielding
/z/) and analogy (restoring /s/ from _grease_). Compare _lousy_,
whose semantic distancing from _louse_ helps preserve the /z/ in the
adjective. He'd've done better to pick on Romance examples like the
English cognates _gaol_ and _cage_ - a threefold development from
Latin _cavea_! (The first form is a diminutive.)