Questions and quantifiers : an inquiry in natural language metaphysics
|Authors:||Shimelman, Wendy Aviva|
|Abstract:||This dissertation looks at quantified questions (“QQs”) and, specifically, at Pair-List readings, Weak Islands and Intervention Effects (“PLRs, ” “WIs” and “IEs”). In many cases, we observe that there are constraints on the interpretation of QQs that are irreducibly linguistic. Although no logical or pragmatic constraint precludes understanding, say, Who did few girls see? as a request for a pair-list, that interpretation is unavailable. The “big question” that structures this project is: What ontological presuppositions best support an explicit semantics for natural language? We contribute to the growing body of evidence that natural language quantifies over at least two basic domains, viz., individuals and events. We develop a somewhat novel method of testing speakers’ intuitions with regard to the “answerhood conditions” of questions. We found, first, that some quantified interrogatives can be ambiguous between an object- and an event-related reading. The truth conditions of the PL response to a quantified question are those of its event-related reading. The truth conditions of responses to non-quantified in-situ questions are also those of the event-related reading; the truth conditions of responses to raised quantified questions are those of the object-related reading. We hypothesize that the interpretive possibilities in QQs are a function of the values of two parameters: the domain being quantified over (events or individuals, E or D) and the type of the subject phrase (referential or quantificational, < d> or < < dt> t> ). A PL reading results from the combination [E, < d> ]. WIs are the product of the “combination” *[E, < < dt> t> ]. We hypothesize that there are no extant examples of this combination in natural language – a type mismatch precludes it. Both the combinations [D, < d> ] and [D, < < dt> t> ] yield object-related readings. IEs result from the combination of the combination [D, < < dt> t> ] and an intervener, “WH-AGENT.” We develop two proposals for the semantics of QQs. The first is algebraic and makes use of “alternative sets.” The second is dynamic. Here, the question operator, Q, is treated like an adverbial quantifier of the ilk of usually or always. This approach permits a more uniform treatment of the semantics of nominal and verbal expressions and of declaratives and interrogatives.|
|Document Type:||Thèse de doctorat|
|Open Access Date:||13 April 2018|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.