Event Detail

In speech and attitude reports, replacing a simple predicate with a complex predicate referring to the same property can change the range of readings available for a sentence, in a systematic way. For example, even though ‘tadpole’ and ‘juvenile frog’ refer to the same property, ‘I told the class that the animal was a juvenile frog’ seems to have interpretations that ‘I told the class that the animal was a tadpole’ lacks. Some have proposed explaining this phenomenon using an apparatus of “structured propositions”. But to avoid paradox, such systems must introduce systematic type-ambiguities that make trouble for quantificational reasoning like ‘They believe everything I told them; I told them that the animal is a juvenile frog; so they believe that the animal is a juvenile frog.’ I will develop an alternative approach in which predicates like ‘believe’ and ‘say’ take a univocally typed argument, which is built up using context-sensitive type-raising operations. This is compatible with coarse-grained views of the individuation of properties and other higher-order entities, providing a systematic solution to the problems of logical omniscience which threaten these views.