Individual differences in the acquisition of shared syntactic representations: a re-analysis of studies using an artificial language learning paradigm

A series of artificial language (AL) learning studies investigated the development of shared syntactic representations during early stages of second language (L2) acquisition. In this paradigm, the sharing of syntax is measured by means of structural priming: if structures are shared between two languages, a sentence in one language (e.g., the AL) should prime the structure of a sentence in another language (e.g., Dutch) and vice versa. According to Hartsuiker and Bernolet’s (Biling Lang Cogn 20(2):219–234, 2017. https://doi.org/10.1017/S1366728915000164) developmental theory, syntactic representations evolve gradually from item-specific to more abstract with increasing L2 proficiency. To test this hypothesis, we re-analyzed three earlier AL studies (Muylle et al. in Lang Learn 70(S2):194–220, 2020a. https://doi.org/10.1111/lang.12400; Biling Lang Cogn 24(4):767–778, 2021b. https://doi.org/10.1017/S1366728921000110; J Exp Psychol Learn Mem Cogn 47(9):1471–1493, 2021c. https://doi.org/10.1037/xlm0000997) while focusing on individual differences in AL proficiency during the first day of acquisition. We predicted that individuals with higher AL proficiency levels show more cross-linguistic priming than those with lower levels. AL proficiency was indeed related to the magnitude of structural priming, although the strongest evidence for modulation of priming by proficiency was obtained for item-specific priming. Additionally, we observed that working memory (WM) capacity and L1 proficiency correlated with AL proficiency and priming in general. Finally, WM capacity correlated with the magnitude of priming in ditransitive sentences, but not in transitive sentences, suggesting a larger role for WM in ditransitive versus transitive priming.