Modals in the scope of attitudes: a corpus study of attitude-modal combinations in Mandarin

Abstract

This dissertation explores how attitude verbs constrain the interpretation of modal expressions in their complements, with the focus on configurations involving embedded concord modals (E-CM’s, e.g. John suspects the culprit might be Dan). The goal of this dissertation is two-fold. The first is to identify E-CM’s and the sub-variations among them. The second is to explain how each subtype of E-CM’s arises compositionally based on the lexical semantics of the attitudes and modals.

I use original data from Mandarin Chinese. Efforts are made to semi-automatically retrieve attitude verbs from a parsed corpus (Penn Chinese Treebank 7.0). The selected verbs are surveyed for their abilities to license modals of different types and forces in their complements. Two diagnostic tests are then applied to the licit attitude-modal combinations in search of constructions with E-CM’s.

I argue that E-CM’s fall into two major sub-classes: Real Concord and Pseudo Concord. Real Concord requires semantic equivalence between a sentence with the scheme a Attitude Modal p and its counterpart without the modal, a Attitude p. Pseudo Concord arises when one of a Attitude Modal p and a Attitude p entails the other, and the entailed one implicates the stronger statement. The generalization reached is that pure circumstantial modals can only exemplify Pseudo Concord; epistemics are concord under representative verbs; priority modals are in concord with jussive and factive-emotive verbs.

It is proposed that Real Concord is from multiple sources. For attitudes with quantificational semantics (Hintikka 1962), it is a result of domain binding and vacuous quantification (Yalcin 2007). The modal retrieves the domain anaphorically from the embedding attitude. Due to domain binding, if the modal and the attitude verb have the same modal force, the quantification contributed by the attitude verb becomes trivial, and thus concord effect arises. For attitudes with neo-Davidsonian semantics (Kratzer 2013), Real Concord happens because the modal expression is an overt counterpart of the modality feature heading the complement clause. Pseudo Concord is partially explained by pragmatic mechanism. The Principle of Informativeness (Atlas & Levinson 1981) is adopted to derive the strengthening implicatures underlying the Pseudo Concord constructions.