Franco, A. , San Anton, M. E. , Destrebecqz, A. & Cleeremans, A.
Université Libre de Bruxelles
Prior research has shown that learners possess powerful statistical learning abilities to solve the word-to-world mapping problem. However, little is known about people’s capacity to use statistical information to map a new word onto an already known object. Here, we explore monolingual and bilingual participants’ ability to map two different words onto the same object. To do so, we adapted the cross-situational learning paradigm proposed by Smith & Yu (2008). Participants were first exposed to twelve word-referent pairs (Language A). They were then exposed to another set of twelve word-referent pairs (Language B) composed of the same objects but now associated to new words. On each trial, two spoken words were presented with the two related objects along with 0, 2 or 4 distractors. The test phase consisted in an adaptation of the Process Dissociation Procedure (Jacoby, 1991) so as to test people’s ability to consciously control their knowledge. We found that monolingual and bilingual participants learn both languages in the 0 distractors condition. However, in the 2 and 4 distractors condition, we found that while both groups learn Language A equally well, only bilinguals were able to learn Language B.