Representation vs Derivation
Joaquim Brandão de Carvalho
Abstract: Plurilinear representations and constraint-based derivation have been the tools used in the last 40 years by autosegmental approaches and OT respectively in order to get rid of SPE rule-based serialism. It will be argued here that the border between representation and derivation depends on another key division in phonological theory: in phonology proper, there are no such things as “allophonic processes” involving feature propagation or delinking; only by virtue of allomorphy, that is of phonology/morphosyntax interaction, may an object be said to change into a different one. Thus, representational theories better describe core phonology, i.e. the internal workings of phonology seen as an autonomous module of grammar, while constraint-based models are more appropriate to deal with phonology’s upper interfaces.
Carvalho, Joaquim (Brandão de). 2019. “Representation vs Derivation: The Case for a modular view of phonology”. Manuscript available in Radical: A Journal of Phonology.