Event Details

In recent years, there has been a resurgence of interest in statistical approaches to language acquisition. This research is driven by two observations: (1) that distributional evidence for a given linguistic structure is sometimes available in the input and can be taken advantage of by learners; and (2) children's behavior in experimental conditions is often nonadult. In this talk, I use a combination of corpus analysis and experimentation with infants and children to show that learners often know things for which there is no evidence in the input and that nonadult behavior can sometimes provide the cleanest evidence for the representations that learners bring to the task of language acquisition.