| Summary: | This dissertation explores the semantics and syntax of switch-reference (SR). It makes novel generalizations about the phenomenon based on two empirical sources: A broad, cross-linguistic survey of descriptive reports, and semantic fieldwork that narrowly targets the Kiowa language of Oklahoma. It shows that previous attempts at formalizing switch-reference cannot work, and offers a new theory of switch-reference that derives the facts through effects that emerge from the interaction between the syntax and the semantics. The empirical investigation results in four major findings: First, SR is introduced by its own head, instead of being parasitic to T° or C°. Second, switch-reference can track Austinian topic situations. Third, it must track topic situations when it is found with coordination, and it cannot do so with intensional embedded clauses. Finally, generalizations or theories based solely on the syntax are not able to account for these facts. These findings are explained by analyzing switch-reference as a pronominal head in the extended verbal projection of the embedded clause. This head introduces a relation of identity or non-identity between two arguments. One of these is in the dominant clause, the other is the highest indexed constituent in the sister of the SR head. The arguments are selected indirectly, through binding structures that are interpreted as lambda-abstraction. The clausemate argument is bound by the SR head; the properties of feature valuation derive the height constraint. The pronoun introduced by the SR head is bound by the connective. Binding by the connective results in the interpretation of the SR-marked clause as a property. This property is then ascribed to an argument in the dominant clause. This theory accounts for the generalizations, and makes fruitful predictions about other aspects of switch-reference, notably when it tracks non-referential subjects. This dissertation improves our understanding of switch-reference, of situation semantics, and of reference-tracking in general. It ties reference-tracking to contextual restriction by use of topic situations, which are anaphoric pronouns used to restrict sentential interpretation. It provides the first solid evidence of morphology sensitive to situations. In addition, the theory of switch-reference proposed here relies on independently-motivated mechanisms in the grammar. This reliance links switch-reference to other mechanisms of co-reference from inside an embedded clause, and finds a solid place for switch-reference in linguistic theory.
|
|---|
