Combining neurophysiology and formal semantics and pragmatics: the case of the N400

Abstract

We present an outline of how experimental data from neurolinguistics related to one particular ERP-component, the N400, can be analyzed in a probabilistic extension of In cremental Dynamics with frames and situation models. We show that none of the semantic and/or pragmatic properties proposed in the neurolinguistic literature alone can explain the whole range of data. Our own approach is similar to that of Werning et al. in taking the pragmatic dimension seriously by incorporating both the perspective of the speaker and the listener using RSA using a Bayesian model. Probabilities are calculated by using both semantic information which is based on an information ordering on situation models and discourse information which is based on a linking relation between discourse referents.