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

On 2007-07-10 21:35, stlatos wrote:

> It doesn't seem likely to me: PIE had words ending in *-etos,
> *-elos, *-etlos, *-enos, and even *-enwos. My take on their relation
> and meaning doesn't matter; it seems certain they existed. One case
> ending became an infinitive, *-enxa(i), among others; they mixed in
> Sanskrit and Greek dialects to various degrees.
>
> The Greek inf. seem to show *+en added to nouns: *bheros >>
> *bheresen; *bhermn. >> *bhermnen > *bhermen by n-n dissimlation. The
> -ai marker was added to -men (and many others in dialects) after it
> was abstracted from the inf.

I was actually thinking of parallel cases elsewhere, e.g. the
Celtiberian infinitive -unei, which seems to be related to neuter
abstracts in *-m(e)n, the original proterokinetic dative *-men-ei being
replaced by *-mn-ei > -wnei after a vowel.

BTW, Celtic has plenty of *-m(e)n abstracts, but the only sure example
of *-wr./*-w(e)n- there is *arwar/n- 'grain' < *h2arh3-wr. It would be
interesting to examine the distribution of *-wer/n- vs. *-men- in IE to
see if it correlates with the presence of *h3 at the end of the root.

Piotr