Re: [tied] xW>v>w (was: -tlo- vs -tro-)

--- In cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...> wrote:

>
> On 2007-07-09 21:29, stlatos wrote:
>
> > Not at all; that was just one example:
> >
> > *dovixY+ > L duim
> >
> > *dovenai > G inf. dou^nai (Cyp to-ve-na-i); Skt da:váne:
> >
> > *dove:+ etc > Lith pret. daviau~
> >
> > And as I said, analogy adding xW to w that itself actually came from
> > xW > E know, glow; Lith do:vanà 'gift'.
>
> This word is trickier than that. With all other verb roots in Gk. the
> infinitive has <-men(ai)>

Not Attic-Ionic and Arcado-Cypriot -(e)nai, which is the evidence
I'm using.

> which may mean that <dou^nai> (<dowenai>) has
> a suffix that was originally an allomorph of *-men-, namely *-wen-, and
> so the lexem in question would be *doh3-wen-/*d&3-wen-. Jens has a
> theory about a pre-PIE change of *-h3m- > *-h3w- whose scattered

remains

> can be seen in PIE morphology. I wonder if this analysis couldn't be
> extended to cases like *pr.h3-wó- ~ *pr.h3-mó- 'first', *sloih3-wo-
> 'sloe', and the like.

Why should this be a PIE change for this ev.? There are many
changes of m>w in various IE languages.