From: Miguel Carrasquer
Message: 40267
Date: 2005-09-21
>And is it typologically normal or at least attested anywhere for q > k?It's surprisingly hard to find attested cases. The most
> X, x) and q > ? (as in Arabic).There are a few North Caucasian languages (e.g. Khinalug)
>And in any case why did IE *q become *k in all cases (satem, centum, Luwian)Because it happened only once, _before_ the breakup of PIE.