Prototypical Predicates Have Unmarked Phonology

Recent work recognizes that lexical category can be relevant for phonology, because phonological processes and phonotactics are sometimes category-sensitive (Smith 1997, 2001; Myers 2000; Bobaljik 2008; see also Cohen 1964; Chomsky & Halle 1968; Postal 1968; Kenstowicz & Kisseberth 1977). Moreover, there are strong cross-linguistic tendencies concerning the nature of phonological differences between categories. One such tendency is a hierarchy of phonological privilege (Smith 2011). Phonological privilege includes the ability to avoid neutralization and consequently to support a larger number of phonological contrasts (Beckman 1999). In optimality-theoretic terms, privilege reflects a comparatively high ranking for faithfulness constraints and a comparatively low ranking for markedness (well-formedness) constraints. Stated more generally, less phonological privilege for a given position or category means greater phonological unmarkedness for that position or category. The hierarchy of phonological privilege seen cross-linguistically in category-sensitive phonological patterns is N > A > V: nouns are most likely to be privileged, with adjectives next, and verbs being least privileged. Evidence is also emerging for lexical-category subclass effects in phonology. For example, proper nouns resist a syncope process that affects common nouns in Jordanian Arabic (Jaber 2011); for this process, proper nouns are privileged compared to common nouns. Another example is Itzaj Maya, in which transitive verbs undergo a phonological process of vowel deglottalization in a particular segmental context, but intransitive verbs do not (Hofling 2000:14), indicating less privilege for transitive verbs than for intransitives. I propose that such sub-category effects are evidence that the N > A > V hierarchy of phonological privilege is actually part of a potentially more finely-grained scale, which can be summarized as in (1):