Functional Architecture and the Distribution of Subject Properties

Abstract

It has long been assumed that subjects have special properties cross-linguistically. Depending on basic assumptions, these have been attributed to the property of “subjecthood” itself or to “certain epistemologically prior notions’. But this assumption (underlying both positions) appears to be flawed. There is ample evidence, cross-linguistically, that no unified notion of subject exists. In this paper, we propose a syntactic account for the unified behavior of subjects in certain types of languages, showing why it is that subjects in certain other languages fail to display these properties. We will first show that languages such as English and French include a syntactic requirement that all subjects be DPs—importantly, this induces a DP node dominating non-NP subjects. However, this is not universal, as data from Bulgarian, Russian, and some V-initial languages show. This makes any semantic account of subject properties untenable. Instead, we suggest that this clustering of DP properties is attributable to the [D]-feature, as proposed in Chomsky 1995. English-type languages are `D-feature prominent’ while languages without this cluster of subject properties are `V-feature prominent’. We further propose that this is correlated with clausal architecture—V prominent languages are those in which TP dominates all Agr projections while in D prominent languages an Agr projection dominates T.

We are extremely grateful to the following individuals for their questions, comments, and suggestions: Samuel bayer, Si Belasco, Odutan Bode, Nicholas Sobin, three anonymous reviewers for NLLT, participants athe 1999 LSA Institute Workshop on Grammatical Functions in Transformational Syntax, and audiences at ESCOL 1999 and WECOL 1999. Additionally, we are indebted to the following native speakers for extending their time and energy on behalf of our research: Hanaa Dornik, Elena Gavruseva, Denis Kopyl, Olga Petrova, Eric Roman, Elena Schmitt, and Mila Tasseva-Kurktchieva. Previous versions of parts of this paper have been presented and/or appeared as Davies & Dubinsky 1999a, b, 2000b.