Event-related potential studies of cross-situational word learning in four-year-old children

• Published in: Journal of experimental child psychology 2022-10, Vol.222, p.105468-105468, Article 105468
• Main Authors: Mangardich, Haykaz, Sabbagh, Mark A.
• Format: Article
• Language: English
• Subjects: Cross-situational learning; Event-related potentials; Familiar object labeling; N400; Semantic representations; Word learning
• ISSN: 0022-0965; 1096-0457
• DOI: 10.1016/j.jecp.2022.105468

•4-year-old children show an N400 effect following cross-situational word learning.•This effect disappears in cross-situational task without familiar word pretraining.•Pretraining on the structure of the task may be critical for semantic word learning. Young children learn word-referent links in cross-situational learning paradigms despite uncertainty as to a given word’s correct referent on individual exposures. However, the semantic status of these word-referent links is unknown. Here, we used a novel event-related potential (ERP) testing approach to investigate whether children showed electrophysiological signatures of semantic representations following a typical cross-situational learning paradigm. In Study 1, 20 4-year-old children (mean age: 51.35 months) learned eight new word-object pairs in a cross-situational learning task which included familiar object labeling before training. Then, they completed an ERP task which involved seeing pictures of the novel objects and then hearing words that were either congruent or incongruent with the word-referent association that was trained. Results showed a robust N400 ERP effect, thereby providing evidence for having acquired a semantic representation during the cross-situational learning paradigm. Children also showed good performance on a comprehension test approximately 20 min after training. In a second study, a different sample of 20 4-year-old children (mean age: 52 months) completed the same cross-situational task but without the initial familiar word pretraining. Here, we did not see the N400 effect we saw in Study 1, but children still showed good performance on the delayed comprehension test. We conclude that rapid semantic encoding can occur during cross-situational learning, but that pretraining on the structure of the paradigm may be critical for facilitating that encoding. We speculate on the mechanisms by which pretraining might have effects and discuss the implications of these findings for early word learning.