Automaticity and prediction in non-native language comprehension

Abstract

Some evidence suggests that prediction is more limited in non-native language (L2) than native language (L1) comprehension. We evaluate the hypothesis that prediction is limited in L2 because prediction is largely non-automatic. We examine whether the subprocesses involved in prediction are unconscious, unintentional, efficient and uncontrollable (Bargh, 1994) to understand the extent to which prediction is automatic in L1 and L2. To unpack the subprocesses in prediction, we draw on Pickering and Garrod's (2013) proposal that people primarily use their production system for prediction, as well as a more automatic association-based mechanism. We conclude that at least some of the subprocesses in prediction are not fully automatic and suggest that these non-automatic processes can interfere with prediction in L2.