Re: *gwh in Gmc.

The rule stated by Miguel sounds strange to me, too. For example,
there is reconstructed IE. *gwhermos giving Skt. gharma, Gk. thermos,
Lat. formus, but we see Goth. warms, Eng. warm. There must be plenty
of counterexamples.

--- In cybalist@egroups.com, Miguel Carrasquer Vidal <mcv@...> wrote:
> In Collinge's "The Laws of Indo-European", under "Siever's Law II"
> (i.e. the development of PIE *gwh), reference is made to a work by
> Seebold 1967 ("Die Vertretung idg. *gwh- in Germanischen", KZ
> 81.104-133). In Collinge's summary table, it is stated that */gwh/
> gives /b/ in initial position (except before /u/). What is this
> about?