Re: bidet

--- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy <bhrihstlobhrouzghdhroy@...> wrote:

>
> Kluge's Law is in fact being rehabilitated by Leiden School; Stokes'
> formulation must be precised by Zupitza's one.

I do not find convincing examples of *-dn-' > Celt. *-dd- etc. in Zupitza's paper. Many of the etymologies are vexed, and the unvexed ones cannot be referred unambiguously to protoforms with accent following a media-nasal cluster.

As a working hypothesis, a "Stokes-MacBain Law":

*'-tn- > *-tt-
*'-dn- > *-dd-
*-tn-' > *-tt-
*-dn-' > *-tt-

etc.

> PIE root constraints cannot inhibit the proposal of a law, because
> they are themselves speculative (the very prohibited sequences *deg-
> and *ged- are normal roots in Pokorny and require always controversial
> explanations).

Fine, and I believe in *kneigWH-, but where else is *teg(^)H-?

> Gaulish accent is in turn matter of discussion; Greek renderings, at
> face value (for what they can say, that is admittedly little), do
> exhibit oxytony, often in accordance with PIE etyma.

We do have some modern place-names.

> *paþa can (not: must) anyway have a good Germanic etymology, why
> should a borrowing be preferred? I hope not for the tale of the lack
> of PIE */b/, because this would be a circular argument

PIE initial */b/ is rare, but where else is *bet-?

DGK