A Conspiracy Argument for Optimality Theory: Emakhuwa Dialectology

This paper [to be published in the proceedings of the University of Pennsylvania Colloquium in Linguistics for 1999] argues that while the “conspiracy argument” (originally developed in Kisseberth (1970)) clearly supports the Optimality Theory model of phonology, it is not the case that conspiracies are just an attribute of the synchronic grammars of a single language. Conspiracy arguments can be replicated in a variety of phonological domains (acquisition, speech pathology, language change, dialectology). The present paper examines tonal phenomena in three dialects of Emakhuwa (a Bantu language spoken in northern Mozambique and neighboring regions of Tanzania and Malawi) and argues that rule-based generative phonology fails to capture the conspiratorial relationships obtaining in these dialects, wheras OT (in its Optimal Domains Theory implementation) succeeds in interpreting dialectal relationships as differences in the ranking of universal constraints.