Prosodic conditioning of the null complementizer in English

Abstract

This paper presents a novel account of the distribution of the null complementizer (C) in English, arguing that the licensing of null C is a case of prosodically conditioned allomorphy (Carstairs 1988, 1990; Paster 2006; Inkelas 2014). The account expands on prior proposals (Bošković & Lasnik 2003, An 2007a-b), observing that environments which prohibit null C all show an obligatory Intonational Phrase (IP) boundary preceding C. In contrast, an obligatory IP boundary does not appear in environments which allow null C. I argue that null C is an allomorph associated with a specific prosodic environment: “medial” position within IP, (…C…). Overt C, on the other hand can appear in “initial” position within IP, (C…). This synchronic account is supported by a diachronic account, tracing how null C arose from phonologically weak overt forms of C (/ðæt/ → /(ð)æt/, /(ð)ət/, /-t/ → Ø). Crucially, the distribution of these weak forms is itself prosodically conditioned by principles which allow phonological weakening in medial prosodic positions but prevent weakening in initial position, a result of “domain-initial strengthening” (Pierrehumbert & Talkin 1992; Fougeron & Keating 1997; Beckman 1998; Keating et al. 2004; White & Turk 2010).