Weiss, D. . , Bulgarelli, F. & Poepsel, T.
The Pennsylvania State University
Over the past two decades, the field of language acquisition has been profoundly influenced by the discovery that infant and adult learners are capable of statistical learning. However, the majority of research in this field has restricted the task of the learner to acquiring a single, uniform input. That is to say, the learner can extract the same information from the input at any point during familiarization. Given that real world input is far more variable, research in my lab has begun to explore how learners contend with variability in the speech stream that could potentially signal the presence of a second causal structure. What are the conditions under which learners come to form multiple representations to accommodate multiple statistical inputs? In this talk, I will summarize the findings from several studies by our group that indicate important roles for factors such as contextual cues, the number of switches between languages, amount of exposure, and phonological overlap. In addition, I will also briefly touch on the possibility that, as a consequence of their language experience, bilingual learners might approach this problem with a different perspective relative to monolinguals, noting that our findings suggest this may be highly task-dependent.