On 2006-08-31 09:53, P&G wrote:
>> [Sean:] What about Greek ephthitai as the perfect of 'decay'?
>
> (a) There are other examples of non-reduplicated perfects in Greek, e.g.:
> erxatai = are shut in < *werg-, pluperfect erxato, and with augment
> eerxato
> heimai = I am clothed with < *wes-mai, 2nd p.sing hes-so, pluperfect
> hesso, hestai; participle heimenos < *wes-menos
> amphi-akhuia crying around (not < amphi-*wiagh, but from amphi-wakh)
> dekhatai await pluperfect edegme:n, participle degmenos
> (The examples are in Monro)
These conspicuously lack the o-grade and show a lot of formal
convergence with aorists (athem. epHtHito, etc. -- even the augment has
been copied). I't quite clear that the inherited aorist and perfect
influenced each other in Greek (see Dunkel's 2004 article on Gk.
ka-aorists and ka-perfects). In those cases where Vedic and Greek forms
agree, reduplication is always present (with the usual lonely exception
of *woid-e, which must be of PIE date).
> (b) How could we tell the difference between an inherited obligatory
> reduplication and independent regularisation of reduplication in languages
> that show it?
That's why we need a general theory of IE verb-stems explaining more
than just isolated fragments. See the controversies over such bold
proposals as Jasanoff's theory of o-presents, Rasmussen's intensive
de-reduplication theory, etc. The competition is far from over, and
since I've broken my crystal ball I can't predict the outcome. At the
moment I put my money on the agreement between Vedic and Greek, but if I
find something better, I may change my mind yet.
> (c) Why discount the evidence of Latin, where no -o- stem perfect shows
> reduplication?
The Latin perfect is a syncretic category. I suppose you really mean
that reduplicated perfects usually show no o-grade, don't you? That's
explicable as the commonplace generalisation of the weak-grade
allomorph. Actually, there are some reduplicated perfects with the
o-grade in Latin, but they are formed from o-grade causatives/iteratives
(spondeo:/spopondi: ~ spepondi:, tondeo:/totondi:) and may reflect (or
at least have been influenced by) old reduplicated aorists.
Piotr