Abstract
Definiteness has often been assumed to play a role in syntax, most notably in relation to various "definiteness effects" and Case alternations (Belletti 1988, De Hoop 1992 and many others). The question whether this involves a semantic property which is relevant in syntax, or an independent syntactic representation of definiteness, remains to a large extent unanswered. This paper shows that, on the one hand, Hebrew provides independent evidence for assuming a definiteness feature in syntax; and on the other hand, this formal definiteness does not simply correlate with semantic definiteness, and there is no simple one-to-one mapping between the two kinds of definiteness. The second part of this paper focuses on the Hebrew object marker et, which appears only in front of DPs having the syntactic definiteness feature. I argue that et fulfills a requirement for structural Case which Hebrew verbs cannot assign, and that this requirement is related to the representation of definiteness as a formal feature and not to any semantic property. In this light I consider Belletti's (1988) theory of abstract Partitive, and show that Hebrew object marking seems to provide evidence against it. 1.0 Formal definiteness features Definiteness in natural language is usually seen as a semantic or pragmatic property of noun phrases. Over the years, however, definiteness has also been discussed in the syntactic literature as well. Definiteness FEATURES, taken as formal features that are marked on certain lexical entries and which play a role in syntactic processes, have often been either explicitly proposed or implicitly assumed (for Hebrew, see Hazout 1990, Siloni 1997, Borer 1998 and others). But as opposed to phi-features such as number and gender, whose morphological realizations in many languages are clear and which trigger purely syntactic agreement phenomena, the motivation for discussing definiteness is almost always semantic and not syntactic. As I will show in the first part of this paper, in Hebrew a definiteness feature can be motivated on syntactic grounds. But surprisingly, this formal feature does not always correlate with the semantic notion of definiteness or with any other known semantic property. The term 'feature' has been used in the linguistic literature in at least two ways which are quite different from each other; before arguing that Hebrew syntax makes use of a definiteness feature, we must first make it clear which sense of the term is meant. In its first use, the term 'feature' has been used to capture descriptive generalizations and to name natural sets of elements. So, for instance, to capture the generalization that nouns and adjectives share some properties, both are labeled (+N), as opposed to verbs and prepositions which are (-N). Similarly, in a binding-theoretic classification

This publication has 19 references indexed in Scilit: