In recent years, considerable attention has been devoted to the mismatches that often exist between a word's morphological structure and its semantics. Here, I argue that the full range of mismatches can only be resolved within the framework of a paradigm-based theory of morphology in which morphological rules are formulated as operations on morphological expressions and in which the formal relations between the root of a paradigm and the words in that paradigm are defined by a set of paradigm functions. I demonstrate that this kind of framework affords a straightforward resolution of certain morphosemantic mismatches from Breton which nevertheless present insuperable difficulties for structure-based approaches to ‘bracketing paradoxes’ such as that proposed in Pesetsky 1985. The paradigm-based approach is shown to apply with equal success in the domains of inflection and derivation. Moreover, paradigm functions are shown to be more adequate for the resolution of certain kinds of mismatches than either the head operations of Hoeksema 1985 or the process of analogical back-formation proposed by Spencer 1988.
Stored only when both papers are in the corpus — 109 citations exist upstream.