[PS-3.24]Are bilinguals better at statistical learning?

Onnis, L. 1 , Chun, W. E. 1 & Esposito, G. 1, 2

1 Nanyang Technological University
2 University of Trento

Using multiple languages may confer distinct social and cognitive advantages, such as sustained attention and switching between tasks effectively. However, there is relatively little research on whether bilingualism confers an advantage in implicit statistical learning, a core cognitive ability underlying language. We hypothesized that the lifetime exposure to two different sets of linguistic regularities may have a positive influence on implicit language learning of NOVEL statistical regularities. We tested young adults (n=57) of varying degrees of bilingual experience on a challenging artificial grammar learning task. This involved simultaneous learning of two miniature languages containing different sets of statistical regularities. Following a 12-min training session we obtained accuracy scores on a grammaticality test for each language as a measure of learning. We found that participants learned each grammar significantly better than chance and both grammars equally well. In addition and crucially, a gradient bilingual dominance index obtained from the Bilingual Language Profile questionnaire (https://sites.la.utexas.edu/bilingual/) predicted accuracy scores for both artificial grammars in a generalized linear model. To our knowledge, this is the first study demonstrating an advantage in learning novel statistical relations in more balanced bilinguals.