Inflectional Dependencies: A Study of Complex Verbal Expressions in Ndebele

Abstract

This thesis is a study of the syntax of multi-verb constructions, in which a single clause contains more than one verbal element. Multi-verb constructions are of special interest in light of the hypothesis that clausal structure is headed by a verb. Implemented in terms of extended projections, this hypothesis states that clausal functional structure is the extended projection of a verb. A full understanding of multi-verbal constructions, in which there is no one-to-one verb–clause mapping, relies on refinements of the intimate relation between clause structure and the verbal category. Analyzing various types of multi-verb constructions in Ndebele, a Bantu language of Zimbabwe, I propose a theory multi-verbal syntax which makes explicit the ways in which a single extended projection may contain more than one verb. I propose a basic distinction within the verbal category into projecting verbs and non-projecting verbs. Projecting verbs have the ability to head a full clause and are what we typically think of as lexical verbs. Non-projecting verbs, on the other hand, are inherently unable to project full clausal structure. There are two types of non-projecting verbs in Ndebele. Aspectual auxiliaries are lexical items whose category feature has mixed properties: it has properties of both a lexical verb and some functional head in the clausal spine. The mixed properties are detected by different aspects of their syntax. The second type of non-projecting verb is the default auxiliary (be, in Ndebele), used in compound tenses. The default auxiliary is a verbal expletive, which becomes part of the derivation via last-resort selection. Any multi-verbal construction with more than one lexical (projecting) verb necessarily contains more than one extended projection. Additionally, the thesis is concerned with the role of two syntactic dependencies in the inflectional domain: agreement and selection. Both relations are involved in determining the morphological shape of multi-verb constructions by posing restrictions on the form, category and size of the main lexical verb. Depending on the light/auxiliary verb, it can be a participle (imperfective, past or future), have a subjunctive form (present or past), or be infinitival/clausal. It is argued that such cooccurrence patterns are largely predictable from clause structure and agreement. Selection is involved only in a small set of idiosyncratic cases of complement selection.