Copredication in Context: A Predictive Processing Approach

Area of Science:

• Cognitive Psychology
• Linguistics
• Computational Semantics

Background:

• Copredication statements, where a nominal denotes two distinct entities simultaneously (e.g., "The book is heavy and informative"), pose challenges for standard truth-conditional semantics.
• Existing research has not fully addressed the cognitive mechanisms underlying judgments of copredication felicitousness or the reasons for varying degrees of felicitousness.

Purpose of the Study:

• To propose a cognitive-psychological model explaining linguistic intuitions about copredication statements.
• To investigate the word representations and cognitive mechanisms responsible for judgments on the felicitousness of copredication.
• To explore the reasons behind the varying degrees of felicitousness in similar copredication statements.

Main Methods:

• Development of a cognitive-computational model of copredication.
• Application of the predictive processing framework to model linguistic intuitions.
• Analysis of how higher-order priors (discourse context, world knowledge) influence expectations and felicitousness judgments.

Main Results:

• The proposed model accounts for judgments on the felicitousness of copredication statements.
• Asymmetries in felicitousness judgments are explained by a set of expectations influenced by discourse context and world knowledge.
• The model offers a novel perspective on semantic interpretation and cognitive mechanisms.

Conclusions:

• Linguistic intuitions about copredication are rooted in cognitive mechanisms within the predictive processing framework.
• World knowledge and discourse context play a crucial role in shaping expectations and determining the felicitousness of copredication statements.
• The study provides a computational account for semantic phenomena previously considered problematic for traditional theories.