Relevance in Real-Time Interpretation of Utterances

This article proposes a new evaluation of relevance in the context of relevance theory (RT). Although RT has made substantial contributions to communication science, there are few studies that employ an experimental setting to examine RT in the real-time interpretation of utterances because the evaluation of relevance has remained inadequate for experimental examination. In the original RT, relevance is evaluated based on cognitive effects and processing costs, which are mutually dependent—although conceptually distinct—in the real-time interpretation of utterances. In fact, the mutual dependence of cognitive effects and processing costs is the primary reason why RT is difficult to investigate experimentally. This article proposes cognitive effects as a sigmoid function of processing costs and relevance as the ratio of cognitive effects to the processing costs required. This new evaluation of relevance is shown to have favorable theoretical consequences regarding both the convergence of the computation of utterance interpretation and the disambiguation of potentially ambiguous utterances. Furthermore, this new evaluation of relevance enhances the possibilities of examining RT experimentally.