Best footnote ever?
Right up there, anyway. From Nevins & Wagner 2001, “Can Correspondence Fix Fixed Segmentism?”
M&P claim that avoidance phenomena are metalinguistic. As Bert Vaux has mentioned, this is quite unsatisfying: any phenomena that manipulates phonological elements in a systematic manner should be accounted for in a theory of phonology. In fact, OT enthusiasts could even be creative and suggest that [Smidt Smidt] is ungrammatical due to some variant of MorphDis, in which FixSeg and the reduplicant are indistinguishable. Whatever.