Bare Nouns and the Morphosyntax Of

The syntax and semantics of bare nominal expressions has been the subject of much research over the past few years. Two fundamental distinctions are generally assumed: bare nominals are generally restricted to plural and mass terms (unless they are proper names) and the "interesting" cases of bare nominals are those that appear in argument positions. We will call the first distinction the singular/plural asymmetry, and the second, the predicate/argument asymmetry. In the literature, two major proposals have emerged to account for bare nominals in argument position, one largely syntactic (Longobardi 1994, 1998), the other largely semantic (Chierchia 1998). Chierchia's proposal takes both asymmetries as part of the core data to be explained, while Longobardi takes both as background assumptions to an account of the syntactic and semantic differences between Romance and Germanic bare nominals based on their morpho-syntactic properties. In this paper we address some specific predictions made by Chierchia's system as it pertains to the distribution of bare singular count NPs (henceforth simply bare singulars). Chierchia makes two predictions about bare singulars in languages which have a morphological singular/plural distinction: they should never appear in argument position and they should freely appear in predicative position. Schmitt and Munn 1999 showed that Brazilian Portuguese disconfirms the first prediction. In this paper we argue that English in general disconfirms the second prediction because it does not allow bare singular predicates, even though the Romance languages, including Brazilian Portuguese, do. Our paper thus poses the following questions: