TALKER VARIABILITY AND EARLY WORD REPRESENTATIONS

点击翻译按钮获取中文摘要

DESCRIPTION (Applicant's abstract): Recent research in infant speech perception indicates that infants learn a great deal about the sound organization of their language during the first year of life and that they use this information as a indicator of the syntactic organization of the language. Current diagnostic methods for identifying Specific Language Impairments (SLI) are only applicable with children 3 to 4 years of age. A better demarcation of the critical linguistic landmarks during the period between 6- and 18-months of age will help in developing diagnostic tools for SLI that can be used at a much earlier developmental stage. These experiments will investigate the conditions in which infants form representations of words that are generalizable across different voices. Infants will be familiarized with works presented with one or several voices and then tested on their recognition of these words presented in a different voice by examining their orientation times to passages containing the familiarized words compared to passages that do not. The first part of the project will investigate how well infants can generalize words across specific voice- property differences (i.e., voice gender, average pitch, pitch contour). In Part Two infants are familiarized with words using several voices to explore the possibility that word representations might be more or less robust and generalizable when formed with multiple talkers than with a single talker. Part Three investigates the possibility that experience with a talker's voice improves infants' ability to analyze and encode the speech from that talker. Specifically, infants will be pre-exposed to the voice that will present the passages before the familiarization phase to see if word recognition performance is improved.