Re: Family terms [was: Kluge's Law in Italic?]

--- In cybalist@yahoogroups.com, "stlatos" <stlatos@...> wrote:

>
>
>
> --- In cybalist@yahoogroups.com, "dgkilday57" <dgkilday57@> wrote:
> >
>
> > Anyhow, Olsen says that Lat. <stabulum> "appears to have replaced *sta:bulum", which she compares to Skt. <sthá:tram>. I can find no such Skt. form in Uhlenbeck's dictionary. Regardless, let us suppose for argument's sake that *sta:bulum was the inherited Lat. form, shortened perhaps by contamination with <stabilis>. Then the PIE protoform had full grade, and if I understand Olsen's theory correctly, *stéh2-tlom was preaspirated to *stéh2tHlom without laryngeal absorption, leading to Proto-Italic *sta:Tlom and the assumed Lat. *sta:bulum. (Surface-truthers are free to substitute -ah2- for -eh2-, to-mah-toes for to-may-toes, whatever.) We must then find some other way to explain Lat. <Sta:tius> and Osc. <Staatiis>, since Olsen's theory would demand *Sta:dius and *Staafiis if they were derived from *steh2-tH-.
> >
> > Lat. <su:bula> 'awl' appears to have the zero-grade of <su:tus> 'sewn', continuing PIE *sjuh{x}-. However, since <su:bula> is first attested with Seneca, it might conceivably be a neologism formed after <su:tus> on the model of <fa:bula> 'speech, story' against <fa:tus> 'having spoken', and cannot by itself refute Olsen's model. It is generally agreed that the participle <fa:tus> has replaced short-vowel *fatus (cf. Grk. <phatós>), which survives in the specialized sense <fatum> 'decree, fate' in Plautus, later <fa:tum>. I presume that the Olsenian protoform of <fa:bula> is full-grade PIE *bHéh2-tleh2- > *bHéh2tHleh2- > PItc *fa:Tla:-.
> >
> > Lacking sufficient snippets, I do not know how Olsen explains Grk. <génethlon>, <genéthle:> 'lineage, descent'.
>
>
>
> It's almost pointless to argue about the exact theory of someone who isn't part of the discussion and could have changed any part of it in the years since. I'd say it was opt., someone else could say all x()t>H and some analogy operated in many IE, etc.
>
>
> The important thing is there is no ev. for *-dHlo- anywhere, especially where tH and dH survived as distinct sounds (Skt).
>

The last time I argued w Piotr about this, I think he said it could have happened when PIE x wasn't syllabic between C:

*
gYenxYtlo+
gYenxYtHlo+
gYen-xY-tHlo+

which seems unlikely and can not explain other ev. showing the change was after PIE started breaking up, opt., not apart from h2 (xW), etc. The Latin ev. of *tlo>klo etc. isn't relevant since analogy created new words w tlo not flo after all V later, not just o: .