On Opining: Modal Verbs, Dispositions, Free Choice, and Negation
This dissertation is concerned with the semantics of a specific set of intensional verbs, those that are used to report a subject's standpoint on a given possibility. Among these verbs are permit, promise, offer, guarantee, demand, insist on, recommend, suggest, encourage, and a handful of others. When the objects of these verbs are disjunctive, we find the kind of free choice effects previously observed with possibility and necessity modals. Based on whether the verbs pattern like may or like must with respect to these inferences, we separate the verbs into two classes, which we call Class I (may-like), and Class II (must-like). This behavior suggests that at the level of interpretation, these verbs contain quantifiers over possible worlds—an existential one in the case of Class I, and a universal one for the members of Class II. However, motivated by an unexpected range of readings found when sentences built with these verbs are negated, an investigation reveals that the members of Class I and II are more than just modal. They also appear to be accomplishment verbs that describe external events. As a result, we give a semantic analysis of these verbs that casts them as complex creatures, describing external events in which subjects indicate their modal opinions.
Taking the verbs to be reporters of external events, we then need to explain why some of the negative sentences built with the Class I/II verbs appear to describe internal cognitive states. The solution to this involves two elements: first, we appeal to a version of the habitual operator that can deliver dispositions that are not necessarily established by repetitive action. Second, after noticing that all habitual sentences have extra, unexpectedly strong readings with negation, we enlarge the scope of the phenomenon previously called Neg-raising, and show how an existing pragmatic account for this (that of Romoli (2013)) can be modified to deal with the broader array of extra strong negative readings.
Along the way, we will account for why dispositions described by habitual Class I/II predicates seem to have different establishment requirements than those described by similar accomplishment verbs. We also address how the performativity of these verbs follows from the semantics proposed.