[PHIL-WiP] Mike McCourt - Polysemy, semantics, and psychology
[PHIL-WiP] Mike McCourt - Polysemy, semantics, and psychology
It’s been observed that polysemy raises a challenge for truth-conditional semantics (Chomsky 1996). Some respond to this challenge by abandoning truth-conditional semantics and acknowledging that sentences do not determine anything truth-evaluable, not even relative to an index of evaluation (Pietroski 2005). An alternative to truth-conditional semantics would instead map a sentence to something underspecific, such as an instruction to fetch and combine concepts (Pietroski 2019). It has also been argued that such an underspecific semantics is especially well-suited to a mentalist approach to linguistic theory (Harris 2020), on which semantics is the study of a component of individual human psychology (Chomsky 1964). However, there are at least two strategies for accommodating polysemy in a mentalist semantic theory without abandoning truth conditions: a minimalist semantics on which words are mapped to highly general concepts (Borg 2004), and an overspecified semantics on which words are mapped to hybrid concepts (Vicente 2018). For mentalists, semantic theory is answerable, at least in principle, to evidence from experimental investigation of sentence comprehension (cf. Carston 2012). In this talk, guided by currently available psycholinguistic methods and data, I articulate some specific predictions that might differentiate these three options (minimalism versus underspecificity versus overspecificity). My goal here is not to defend one view over the others, but to spell out predictions that might be tested in order for such a defense to be successful.