Although recent research shows that infants are sensitive to structural regularities in linguistic input, little is known about the degree to which these sensitivities reflect general-purpose mechanisms (e.g., association) versus innate language-specific expectations about the kinds of structure that will occur. The objective of the proposed research is to determine the nature of some of the learning mechanisms available to infants. This research will address questions pertaining to: 1) the architecture of infant language learning mechanisms; 2) kinds of input learned; and 3) developmental sequencing of such learning. Learning is assessed by exposing infants to auditory strings generated by an artificial grammar. Infants participate in training and testing phases using the head-turn preference procedure. After initial exposure to the grammar, infants are tested to see if they can discriminate new grammatical strings from ungrammatical ones. In Year 1, 9- and 12- month old infants will be trained on finite-state grammar, then tested to see if their discriminations are due to an associative learning mechanism or rules involving salient cues in linguistic input. Year 2 investigates the role played by grammatical morphemes in learning phrase categories in a phrase structure grammar.
