Triggering Factivity: Prosodic Evidence for Syntactic Structure

It is well-known that, across languages, so-called “given” information (as opposed to “new” information) systematically correlates with lack of phonetic prominence, or deaccentuation (Halliday 1967, Taglicht 1982, Ladd 1996, Selkirk 1995, Schwarzschild 1999, among many others). The primary goal of this paper is to provide a minimalist syntactic account for a sub-case of this phenomenon, namely the correlation between “givenness” and / or “presupposition” of clausal complements (i.e., factivity, Kiparsky and Kiparsky 1970) on the one hand, and deaccentuation, on the other – a correlation that, as I will show, holds across several languages. In line with Chomsky’s (2004) “best case” scenario, according to which the components of the derivation of proceed cyclically in parallel, the central claim that I put forward in this paper is that (at least some) information which leads to certain prosodic outcomes, or spell-outs, is encoded in the (narrow) syntax. That is, (some) prosodically expressed information is part of the numeration (in the sense of an abstract morpheme, Embick and Noyer, to appear), or its analogue. Specifically, I contend that: (i) the [+presupposed] status of an embedded CP must be expressed; (ii) this is achieved syntactically (via functional structure); and (iii) the head hosting the relevant syntactic feature, which is instantiated overtly either by some expletive-like element (such as a clitic, a pronoun, a modal element, etc.), or simply by destressing (or deaccentuation), is a probe for the goal (here: CP) with an OCC feature, yielding the various patterns that are found and that are described in some detail in this paper. This paper is organized as follows. In section 2, I start out by presenting Kiparsky and Kiparsky’s (1970) construal of factivity, introduce some factivity triggers, and make a point as to why the distinction factive/non-factive may be thought of in terms of information structure. Then section 3 lays out a syntactic model of the correlation between factivity and deaccentuation, which crucially relies on Chomsky’s (2004) architectural best-case desideratum. Finally in section 4 I discuss certain extensions that push the research agenda introduced here further towards a more syntactic account of so-called information structure.