Re: [tied] Grimm and Verner

----- Original Message -----

From: Miguel Carrasquer Vidal

To: cybalist@yahoogroups.com

Sent: Monday, December 17, 2001 11:48 PM

Subject: Re: [tied] Grimm and Verner

>>The internal periodisation of Grimm's Law is somewhat uncertain, and my favoured (but, needless to say, tentative) [solution] is different from yours. I do not think it is profitable to attempt to connect Grimm's Law with the Armenian consonant shift.

> So any connection with Celtic (see my other message) would not be
profitable either, I suppose.

 

Contrariwise, there could be at least an areal connection, if not a genetic one. As a matter of fact, I think that the spirantisation of the breathy series may be a very old process, _perhaps_ a shared apomorphy of "Northwestern IE" (with *dH > *D merging with *d in Celtic).

> ... But for a while then, PGmc would have had a consonantal system:

/f/ /B/  -
/T/ /D/ /d/
/x/ /G/ /g/

> I don't think that's very realistic.  And neither (though slightly
better) is:

/f/ /B/  -
/T/ /D/ /t/
/x/ /G/ /k/

 

I'd actually vote for the second system (without any gaps -- see below), or rather for the instant reinterpretation of the *d series as unspecified for phonation type as soon as the spirantisation of the remaining stops had become phonologised.

> If there is a voice contrast, it's more likely to be in the stops
alone than in the fricatives alone (speaking as an Hispano-Hollandish
person, but I believe it tends to be true in general).

 

The fact that something "is more likely" or "tends to be true" does not mean that counterexamples should be dismissed. Who says that reconstructed Proto-Germanic must represent the cross-linguistic average? The {T, D, t} triad is not my invention but the standard Germanic reconstruction, and it looks solid enough on comparative grounds. Theoretical considerations (Boersma 1989) do not rule such a system out; they merely suggest that it would usually evolve into the more stable {T, d, t} configuration, which is just what the bulk of Germanic did in due time.

 

There would have been no labial gap in the stop series. First, the spirantisation of voiceless stops failed in sC clusters, where the second element merged with the reflexes of the PIE "plain" [or maybe not so plain] voiced series. Secondly, there were at least a handful of lexical items like the "apple" or "lip" words, which means that PGmc. *p < *b was never non-existent, though it may have been pretty rare.

Piotr