Is third language learning influenced by working memory, implicit learning and inhibitory control?

Abstract

The present study examines how individual differences in executive functions, specifically working memory, implicit learning, and inhibitory control, influence the early learning of morphosyntax in a third language (L3). Thirty-three adult participants (17 from Spain and 16 from Norway) were exposed to an artificial language that included grammatical gender agreement. During this initial learning phase, participants completed standardized executive function tasks and a grammaticality judgment task while their brain responses were recorded using EEG.

The present study examines how individual differences in executive functions, specifically working memory, implicit learning, and inhibitory control, influence the early learning of morphosyntax in a third language (L3). Thirty-three adult participants (17 from Spain and 16 from Norway) were exposed to an artificial language that included grammatical gender agreement. During this initial learning phase, participants completed standardized executive function tasks and a grammaticality judgment task while their brain responses were recorded using EEG.

At the behavioral level, participants showed no significant difference in accuracy between grammatical and gender violation conditions. Only working memory was significantly associated with grammaticality judgment accuracy, with higher working memory scores predicting better performance. Implicit learning and inhibitory control did not show reliable behavioral effects.

At the behavioral level, participants showed no significant difference in accuracy between grammatical and gender violation conditions. Of the executive functions tested, only working memory was significantly associated with grammaticality judgment accuracy, with higher working memory scores predicting better performance. Implicit learning and inhibitory control did not show reliable behavioral effects.

ERP analyses revealed condition-specific effects that were modulated by individual differences. In the 200-500 ms window, an interaction emerged between condition and inhibitory control, suggesting that participants with weaker inhibitory control exhibited reduced neural sensitivity to violations. Marginal trends also pointed to possible effects of working memory and implicit learning in this early window. In the 300-600 ms window, a significant interaction was found between condition and implicit learning, indicating that participants with stronger implicit learning abilities showed greater neural differentiation between grammatical and violation trials. There were no significant findings in the 400–900 ms window. Multilingual language background, as measured by the Multilingual Language Diversity Score, did not significantly influence neural responses in any model.

Keywords third language learning, executive functions, working memory, implicit learning, inhibitory control, artificial language, ERP, morphosyntax