A Formal Model of Plausibility Monitoring in Language Comprehension

Recent work in psychology provided evidence that plausibility monitoring is a routine component of language comprehension by showing that reactions of test persons were delayed when, e.g., a positive response was required for an implausible target word. These experimental results raise the crucial question of whether, and how, the role of plausibility assessments for the processes inherent to language comprehension can be made more precise. In this paper, we show that formal approaches to plausibility from the field of knowledge representation can explain the observed phenomena in a satisfactory way. In particular, we argue that the delays in response time are caused by belief revision processes which are necessary to overcome the mismatch between plausible context (or background resp. world) knowledge and implausible target words.