Six-Month-Olds’ Detection of Clauses Embedded in Continuous Speech: Effects of Prosodic Well-Formedness
Thierry Nazzi
Institute of Human Development, University of California at Berkeley
Deborah G. Kemler Nelson
Department of Psychology, Swarthmore College
Peter W. Jusczyk
Department of Psychology and Cognitive Science, Johns Hopkins University
Ann Marie Jusczyk
Department of Psychology, Johns Hopkins University

Download this .exe file and launch in Windows Explorer. The sound files associated with the research in this article can then be played through Apple’s QuickTime media player.

Three experiments investigated the role of prosodic structure for infants’ recognition of embedded word sequences. Six-month-olds were familiarized with 2 versions of the same sequence, 1 corresponding to a well-formed prosodic unit and the other to a prosodically ill-formed sequence (although a successive word series). Next, infants heard 2 test passages. One included the well-formed unit, and the other included the ill-formed sequence. In Experiment 1, infants listened longer to the passage containing the well-formed unit, suggesting that such units, even when they are embedded, are better recognized. Experiments 2 and 3 showed that this better recognition does not depend on an acoustic match between the familiarized sequences and their later embeddings. This suggests that the advantage of the well-formed unit is at least partially due to infants’ use of prosody to parse continuous speech.