Proceedings of the Annual Meeting of the Cognitive Science Society

Probabilistic reasoning has been proposed to underlie both word prediction in language comprehension and probability judgment in decision-making, yet the connection between the two cognitive processes remains unclear. This study investigates whether word predictions and probability judgments rely on shared probabilistic processes by examining unpacking effects. An offline probability judgment task and an online self-paced reading task were used to test probability shifts in each task, and further analyses examined the relationship between measures across tasks. Results show that atypical unpacking reliably led to lower probability judgments and slower reading times, whereas typical unpacking didn't consistently lead to higher probability judgments or faster reading times. Interestingly, item-level changes in probability judgments were correlated with changes in reading times only for atypical unpackings. These findings provide partial evidence that word prediction and probability judgment draw on a shared foundation of probabilistic reasoning.