--- In
cybalist@yahoogroups.com, Mate Kapoviæ <mkapovic@...> wrote:
>
> The problem is that you put far to much into "common
sense", "logic",
> "natural (?) expectations" and those often have nothing whatsoever
to with
> the process of linguistic change. That is exactly the reason why
one
> should first look at the empirical data before making general
conclusions
> such as: if *kW > p, then *gW > b. That indeed *does* happen often
but
> it's not the only possibility. Unparalleled changes have been
known to
> happen. It can hardly be called unusual.
>
> Mate
>
I agree that "shit happens" or at least "may happen" and that we
shouldn't let our expectations interfere with the formulation of
sound-laws etc.
On the Celtic side, however, it has already been mentioned that
there was a gap in the system, the fact that PIE *b was extremely
rare. The problem is of course in understanding why such gaps are
sometimes tolerated for century and sometimes eliminated right away.
There may also have been a further difference between *kW, *gWH on
the one side and *gW on the other. Going by the modern Celtic
languages, the unvoiced stops are aspirated. I don't know how far
back we can push the aspiration, but assuming that it is old, we
would have had a system *kHW, gHW, gW. Would this help?
Old aspiration of unvoiced stops might also explain *p > *ph > *f (>
*h > *Ø) better.
Anders