A Cognitively Grounded Bayesian Framework for Misinformation Susceptibility
For researchers studying misinformation and cognitive science, this work offers a cognitively grounded model that explains susceptibility patterns, though it is incremental as it extends existing theory with known bounded rationality concepts.
This paper introduces Bounded Pragmatic Listener (BPL), a Bayesian framework extending Rational Speech Act theory with cognitive bounds to model misinformation susceptibility. BPL achieves competitive veracity classification on LIAR and MultiFC benchmarks and provides experimental support for the depth-mismatch paradox.
In this (work in progress) paper, we present Bounded Pragmatic Listener (or BPL), a cognitively grounded Bayesian framework for modelling susceptibility to information disorder. BPL extends Rational Speech Act theory with three cognitively motivated bounds derived from the bounded rationality literature with a) a recursion depth bound (that emphasises working memory limits);b) a prior compression parameter (which is oriented at capturing information bottleneck); and c) an availability sample size (that operationalises importance sampling with saliency-weighted proposals). This allows us to test predictions about misinformation susceptibility, annotator disagreement, and the differential vulnerability to mis-, dis-, and mal-information as defined in the Information Disorder framework. We validate BPL on the LIAR and MultiFC benchmarks showcasing competitive veracity classification and experimental support for the depth-mismatch paradox.