The influence of bilingualism on statistical word learning: a registered report

Abstract While statistical word learning has been the focus of many studies on monolinguals, it has received little attention in bilinguals. The results of existing studies on statistical word learning in bilinguals are inconsistent, with some research reporting a bilingual advantage over monolinguals but others finding no difference between groups. Thus, our study investigated statistical learning using the cross-situational word learning (CSWL) paradigm between two groups: English–German bilinguals and English monolinguals. Participants (N = 201) learned 1 : 1 mappings (one word maps onto one object) and 1 : 2 mappings (one word maps onto two objects). In contrast to previous studies, we measured learning continuously and analysed trial-by-trial behaviour closely to understand fine-grained learning differences across language groups. Against our predictions, monolinguals generally outperformed bilinguals in word learning regardless of mapping type, and trial-by-trial analyses did not reveal differences in learning as expected. Moreover, language balancedness of bilinguals (operationalized via language entropy) did not account for differences in CSWL performance. These results suggest that any existing word learning advantages may be highly paradigm- and population-specific.