It is counterintuitive only if you assume
that pre-Germanic accent was like the strong expiratory stress typically found
in the historical Germanic languages. This is unlikely to have been the case, if
you reflect that there were no vowel reductions in the millennia between PIE and
the sub-historical stages of Proto-Germanic. My guess is that Proto-Germanic
already began to develop demarcative (root-initial) expiratory stress
while retaining the inherited pre-Germanic "pitch stress" (phonological
prominence realised as a high tone). A root-initial voiceless fricative did not
undergo Vernerian voicing because it was in the onset of
a _stressed_ syllable. What happened word-medially or word-finally was
conditioned by the pitch contour rather than expiratory strength: a preceding
low tone favoured fricative voicing.
In the unmarked case, tonal and expiratory
prominence occur together as cues for primary stress, so the further
evolution of the Germanic stress system consisted in aligning _all_ kinds of
phonological prominence (including the full range of vowel contrasts) with the
root-initial syllable.
Piotr
----- Original Message -----
Sent: Sunday, December 16, 2001 10:21 PM
Subject: Re: [tied] Grimm and Verner
- Which brings us to "why Verner's law?". On the face of it, the
law
makes no sense: we have <mó:þar> vs. <fadár> >
<fádar>, which is
counter-intuitive if we compare it to phenomena in
English such as
<metal> ~ <metallic> (['metl] or ['medl] or
['meRl] vs. [m&'thælIk])
or <basil> ~ <basilic> (['bæzIl] vs.
[b&'sIlIk] (or [b&'zIlIk])).
It seems strange that a consonant
immediately before the stressed
vowel should have been
lenited.