Subjectivity and the Expressive – Reportive Speech Acts

Predicates of taste (e.g. fun), as well as interjections (e.g. wow), form a class of inherently subjective expression-types, consequently resisting straightforward treatment within truth-conditional semantics. While subjective expression-types semantically require an experiencer-argument, because that argument can potentially be set to anyone at all semanticists seem unable to explain how its value is interpreted systematically. I argue that interpreting the correct value depends directly on the speech act performed: Expressive or Reportive. One can only possibly express one’s own mental states, yet may report on anyone’s, including oneself. Rules of sentence choice thus predict distinct linguistic structures to be employed when expressing: as the experiencer-argument may become overt (fun for me) or remain covert (fun), when performing an expressive speech act it remains covert – it is otiose to refer to oneself when expressing. Assuming interjections are normally used expressively, I predict the experiencer-argument to remain covert when predicates of taste follow; e.g. “Wow! It’s fun!” is more likely than “??Wow! It’s fun for me!” Google yields many such distributional contrasts robustly. Importantly, interjections are not semantically expressive, as some semanticists assume; rather, expressing (and reporting) denote categories of use, not structure – things speakers do in uttering tokens of subjective expression-types.