Re: [tied] Re: Hittite preterites

On Thu, 24 Apr 2003, Sergejus Tarasovas wrote:

> --- In cybalist@yahoogroups.com, Jens Elmegaard Rasmussen <jer@...>
> wrote:
>
> > So 'to be' is not enough? I would appeal to the augment also to
> explain
> > the long circumflex vowels of Lithuanian e:~jo 'went',
> e:~me: 'took' (e:
> > being the dotted e) which would be regular from *e-H1ey-, *e-H1em-
> + prt.
> > morphemes.
>
> Why do we have an acute in <díeveris> < *deh2iwé:r then? Because of a
> syllabification like [deh2j.wé:r] rather than [de.h2i.wé:r]? Because
> *i even in its vocalic (rather than glidic) incarnation still keeps
> some immanent qualities *e doesn't possess? Or do you reconstruct
> *daih2wé:r? If so, what would be the source for *a? It's not flanked
> by uvulars in that case.

I don't think any of us knows how much coarticulation there was in PIE
phonetics. Without thinking too much about that point, I have myself on an
earlier occasion reached the clarification that a sequence -VHyCV- was
regularly metathesized to -VyHCV-. The basic reason is the working of
Hirt's Law retracting the accent in such sequences. Hirt's law is then
always triggered by a syllable-final asyllabic laryngeal, and it only
retracts the ictus from the immediately following syllable. I therefore
simply accepted the vocalism as /a/, not by coloration, but
underlyingly. Today I am less certain that this is necessarily right.
Many scholars make very majestic statements about IE syllable boundaries.
I don't know where they get that knowledge from, unless they simply
invent it.

Jens