Rob:
> The perfect reduplication seems more linked to the
> other form of durative reduplication, [...]
The perfect and durative are two different aspects.
The durative is _ongoing_; the perfect is a _state_
reached by an action. I don't see any connection
at all aside from the superficial connection with
reduplication.
I'd say that the e-grade reduplicated duratives
are simply taken from the perfect reduplication
by later analogy. The original form before Schwa
Diffusion would be *C&CéC- and seperate from a
perfect in *CeCáC-.
> I wonder if, in fact, the sigmatic aorist qualifies
> as a "root extension".
That's basically what I'm saying, but unlike the
extension in *-g- which would come recently from a
particle, the sigmatic ending would always have been
a suffix of some sort, inherited from Proto-Steppe.
In my theory, we'd never see **bar-sa-mi because
the theme is strictly *-as- in MIE. So we'd find 1ps
*ber-as-am which is *bHe:rsm after Syncope. That's
all we'd see from this because the stem is *ber-as-
and not **ber-asa-. It makes a difference because
of QAR.
The g-extension is evidently different. This would
be because it is originally the particle *ge,
remember? The different syllabic shape of the
resulting suffix in comparison to the original
aorist alone can then explain the stem alternations.
> I'm not as sure as you are about this, but suffice
> it to say that if IE had denominal verbs late in
> the game, it surely had some earlier on as well.
Yes, it's neverending :)
= gLeN
__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com