In my opinion the Armenian consonant shift
and the less securely established Thracian shift are independent from, and
probably much older than Grimm's Law (or at least than its final stages). I
don't see how a common Latin/Greek shift could be defended -- the
developments in Greek and Latin are only partly convergent. The typology of
possible systems resulting from {dH, t, d[~]} is limited and pretty well
exhausted by the attested IE developments; little wonder that different branches
sometimes followed similar evolutionary paths.
I'm sympathetic to the view that a fortis
pronunciation of the voiceless series (I'll use the notation {t}) was
characteristic of "Western IE" (Celtic, Italic and Germanic). One possibility is
that both {dH} and {t} were pronounced with a spread glottis (as opposed to
{d}). A spread glottis produces aspiration-like effects, and the resulting sound
can easily change into a continuant. In Italic, spirantisation affected only the
{dH} series (e.g. *dH > *D [*-D-/*T-] > *f-/*-d-/*-b-), and the {t} : {d}
contrast was stabilised as [-/+ voice] rather than [+/- spread]. In
Germanic, spirantisation ultimately affected the {dH, t} subsystem in most
positions, while {d} ended up with unmarked phonation. In Celtic, there was at
least allophonic spirantisation of both {dH} and {t} (leading, amont other
things, to the erosion of *p > *f > *h > zero), whereas {d} merged
with {dH} as {[d/D]}.
Piotr
----- Original Message -----
Sent: Tuesday, December 18, 2001 4:11 AM
Subject: Re: [tied] Grimm and Verner
Hmm.. My focus was more on the aspiration of the voiceless series
(*t
> *th), and, for "Western-IE" in general (including Armenian and
some
of the Balkan languages, but excluding Balto-Slavic, Indo-Iranian
and
some other Balkan languages) a common (areal) plosive system /th/
/t/
/d/ (with different origins for it, to be sure, in Latin-Greek
vs.
Celtic-Germanic-Armenian).