What happened in Romance was not "satemisation" but someting much
more trivial -- palatalisation of velars before front vowels. The
Satemic shift does not depend on the phonetic environment (*k^ >
palatal even before back vowels and consonants), and so is a
different type of process, and Romance parallels are not very
enlightening.
In satem languages, *k and *kW simply fell together so early that
their development is the same in all satemic branches (except perhaps
in Albanian, but the evidence is inconclusive) -- that is why my
table has one column for both.
Piotr
--- In cybalist@..., tgpedersen@... wrote:
> So it hinges on the existence of a stage where this merger hadn't
> taken place? Romance, in its "satemisation" also had kW -> k. Does
> the relative ocurrence of these two phenomena in Romance tell us
> anything of possible temporal sequences of them in the IE
> satemisation?
> BTW would you care to elucidate on the fate of k and kW in the
table
> you already made?
>
> Torsten