Beyond Words and Phrases: A Unified Theory of Predicate Composition

Dissertation Abstract

One of the ongoing debates in linguistic theory has centered around the notion of "word". Much research has been carried out in generative linguistics to distinguish the component of word-formation and the module responsible for the formation of phrases in order to characterize the relation between morphology and syntax. This dissertation investigates cross-linguistic phenomena that lie at the interface of these two components and proposes to cpature the distinct properties of verbal constructions within a computational model that consists of parallel nominal and verbal domains and which provides uniform principles of predicate composition.

By studying complex predicates, such as causatives in Eastern Armenian and Japanese and light verb constructions in Persian, this dissertation isolates the primitive atoms used to encode meaning in the syntactic code and argues for a single computational domain responsible for the formation of predicates. In the model proposed, syntactic principles are used to form predicates by combining the primitive grammatical features. The notion of word is thus defined as a level in the syntactic structure and the distinction between a "word" and a "phrase" is characterized by the structural complexity of the constituents involved in the formation of a particular predicate. The dissertation further argues that the distinct morphological, syntactic and semantic properties of words and phrases can be captured straightforwardly from the resulting configuration and the interface conditions. In the computational model developed, the lexicon consists of a list of the primitive atoms of meaning and of unordered sets of associated atoms. These lexical entries, combined with the spell-out node to the PF component, determine the parameters for language variation and can derive the structure-meaning mismatches observed in verbal predicates.

A close examination of the projection of arguments in Easterna Armenian and Finnish leads to the development of parallel nominal and verbal primitive features that account for the direct correlation between the assignment of object case, word order and semantic interpretation. In addition, an investigation of the nominal elements in Persian complex predicates provides a three-way distinction between predicate modifiers, non-specific objects and specific arguments, based on the syntactic configuration in the verb phrase.

The dissertation formalizes a restrictive theory of the computational system within the framework of Parallel Domains \cite{vergnaud:01}, whereby all predicate composition is achieved by combining primitive assemblies of feature sets that are used to encode meaning in natural language. This combinatorial system allows us to capture the compositionality of verb phrase formation and the direct correlation between syntactic configuration and semantic interpretation without positing richly annotated lexical entries. In addition, the dissertation provides a computational model of predicate composition that is able to derive the distinct properties of ``words'' and ``phrases'' within a single component of grammar and its interface conditions, thus providing an insight into the interaction between morphology and syntax.