In my review of Sipt�r and T�rkenczy's Phonology of Hungarian,
I made the following incorrect statement:
This somewhat negative view is supported by the fact that,
according to S+T, /kp/ is actually found in ten monomorphemic words
(page 129: Table 17).
There are actually no examples of -kp- monomorphemic clusters--
the ten examples are of -pk-. I think the other arguments against
positing a constraint to explain this gap still stand, however,
and welcome further feedback.