we saw that a rule which appeared to underapply internally in reduplicated forms was a cyclic rule, which we do not expect to apply morpheme-internally in non-derived environments. Simply by identifying the nature of the rule in question we could predict its behaviour with respect to reduplicated form. Similarly, I shall show that the remaining class of phonological processes which either over- or underapply to reduplicated forms share property which predicts their over- and underapplication. Crucially, these processes do not apply to all morphemes meeting their structural description. For each morpheme, one must learn separately whether or not the process will apply (although there may be some generalizations that can be expressed regarding the classes of morphemes which undergo or do not undergo the process). The processes also have morphological rather than purely phonological environments. Aronoff [Aronoff, Mark, 1976, “Word Formation in Generative Grammar”, Linguistic Inquiry Monograph 1, MIT Press, Cambridge] has called processes meeting the above criteria "allomorphy rules".
- Marantz (2004), a pag.345