The role of prediction in novel vocabulary learning from context

When reading sentences, speakers generate predictions about upcoming words. Prediction can aid processing, as speakers have access to information before it is presented in a sentence. However, the predictability of a word in a sentence may support or hinder long-term retention. On the one hand, words presented in sentences that set up strong expectations for a particular word (e.g., she read the book) are better remembered as they confirm expectations. On the other hand, words presented in sentences that do not set up any expectations (e.g., she moved the book) are also remembered well due to their distinctiveness.
This project will explore the consequences of prediction for memory and learning in native (L1) and non-native (L2) speakers. Generating predictions can support word learning in L1 speakers.
However, although L2 speakers acquire new words more successfully than L1 speakers, prediction is slower and more effortful for L2 speakers. Thus, it remains unclear whether L1 and L2 speakers rely on similar processes for learning new vocabulary.
We will examine how prediction in sentence contexts can influence word learning in L1 and L2 speakers. Learning will be assessed by tracking speakers' eye movements and with tests of word
knowledge. We will also examine the relationship between word predictability and learning in a variety of predictable and less predictable sentences to test how the diversity of sentence contexts
influences learning. These findings will inform theories of language acquisition and bilingualism, with practical implications for foreign language learning and teaching