11-08-03 00:16, Richard Wordingham wrote:
> I've still not found any examples of tonal features affecting the
> phonation of plosives.
Whereas _stress_ may affect phonation in a Vernerian way, see Eng.
<é[ks]ercise> vs. <e[gz]ért>, <an[gz]íety> vs. <án[kS]ious>, <lú[kS]ury>
vs. <lu[gZ]úrious>. The voicing of original /s/ and /tS/ after an
unstressed syllable (as in <matches>, <gallows>, <Norwich>) and in
function words (<was>, <is>, <as>, <his>) is also Verneroid.
I'd formulate the voicing rules for Germanic fricatives in the following
way:
(1) An originally voiceless fricative followed by another obstruent
assimilates to it with regard to phonation.
(2) A cluster of voiceless fricatives always stays voiceless.
(3 = Verner's Law) Elsewhere, fricatives stay fortis (voiceless) when
preceded by a stressed syllable or the initial boundary of a lexical
root; otherwise they become lenis (voiced).
Rule 3 implies the rightward spread of some laryngeal feature that (a)
can serve as a phonetic correlate of stress, (b) can be employed to mark
the beginning of a word, and (c) can inhibit spontaneous voicing. The
feature [stiff vocal folds], as proposed by Page (1998), looks quite
plausible.
*Ga-' (*< xan-') owes its voicing to being a bound proclitic unrelatable
to any lexical morpheme. Note also the behaviour of fossilised compounds
in which the second component may lose its phonological separateness and
be affected by Verner's Law. My personal favourite is Ger. Messer < OHG
mezzirahs < *mati-saxsa- 'food-knife'.
Piotr