Whether language input acts simply as a trigger for innately stored information or provides infants with information that supports active learning of language is highly controversial. One method of addressing this controversy is to investigate whether young infants' ability to remember and recognize relevant sections of speech such as sounds, syllables and words is differentially supported by two types of input. This project will compare infants' memory for sound sequences heard in either the adult-directed or infant-directed speech register. To begin acquiring the sound categories and words of a language, infants must be able to remember previously heard sound sequences and group together productions that are acoustically distinct but categorically equivalent. To investigate infant memory for sound sequences in these two speech registers, 6.5 month-old infants will be divided into two groups. Each infant will hear a familiarization passage containing target CVC sequences. Sentences in the familiarization passage will be produced with either infant-directed or adult-directed prosody. After familiarization, infants will be tested on the familiar target CVC sequences and a set of unfamiliar CVC foils, produced in the same prosody to which they were familiarized. Testing will be conducted using the Headturn Preference Procedure. If sequences heard in infant-directed speech are remembered better than those heard in adult-directed speech, infants who were familiarized with infant-directed speech will orient longer to the familiar, target CVC sequences. In the first experiment, the test CVC targets will be acoustically identical to those heard during familiarization. In the second experiment, test CVC targets will be acoustically distinct tokens of the same sequence, produced either by the same speaker or by a different speaker. Evidence that infant ability to recognize CVC sequences varies with input type would provide strong evidence that young infants are actively learning to categorize speech tokens: an important early step to acquiring language. If the early stages of language acquisition are better supported by one input type than the other, this will have important implications for theories of language acquisition. Both normal populations and populations that are at risk for language delay or disorder may benefit from early exposure to input that optimally supports infants' ability to segment sound sequences from running speech, remember them, and accurately compare memories with new input.
