Gender exponence and apparent polarity in a class of Omani Mehri plurals

Abstract

A specific class of Mehri broken plurals raises two questions: why does Gender exponence appear to be possible in two different places? Does the fact that Gender exponence in one site is the mirror image of what appears in the other imply polarity? In parallel to these two questions, the issue of what templates really are is raised. Assuming that (i) templates are direct reflections of deep morphosyntactic architectures, and (ii) Mehri roots do not spread freely to repair ill-formed phonological structures, makes it possible to identify one site in the templates characterizing these nouns as the canonical site for Gender exponence and the other as artefactual. When the differential status of the two sites is established, the polarity apparently inherent in the system can be shown to reduce to an OCP effect.