Learning Through Prediction: A Case of Verb Bias Learning

Linguistic prediction, which emerges from experience, is a pervasive process in language comprehension. However, how prediction develops as learning unfolds and how it drives the learning process remains unclear. This study examines three key questions: (1) whether learning is associated with growth in prediction, (2) whether stronger prediction errors are associated with greater learning, and (3) whether linguistic prediction skills are stable across tasks. Our results revealed that learners who successfully updated their verb biases showed greater growth in anticipatory looking patterns upon hearing the verb in ambiguous test sentences. Greater prediction errors operationalized by the mismatch between learners’ initial verb bias and the subsequent training predicted greater learning. Finally, individuals who predicted more in a separate comprehension task showed greater changes in anticipatory looking behavior and ambiguity resolution in the verb bias learning task. Taken together, these results provide empirical support for a prediction-based language learning framework.