ASA 129th Meeting - Washington, DC - 1995 May 30 .. Jun 06

1aSC12. The cocktail party effect in infants.

Peter W. Jusczyk

Rochelle Newman

Dept. of Psych., SUNY at Buffalo, Buffalo, NY 14260

Most research on speech recognition by infants occurs in quiet laboratory rooms with no outside distractions. But the real world may not resemble the laboratory: Much speech directed to infants occurs with competing acoustic signals. To learn language, infants need to follow their care-givers' speech even under less than ideal listening conditions. Infants' abilities to follow a female speaker's voice when a male voice was talking simultaneously were examined. In pilot studies, infants presented with two competing streams of fluent speech showed no evidence of recognizing words spoken by the target voice. In the present study, infants heard a target voice repeating isolated words, while the second voice spoke fluently at either of three different loudness levels. Subsequently, infants heard passages produced by the target voice using either the familiar words, or novel words. Infants listened longer to the familiar words when the target voice was 10 or 5 dB louder, but not when the distractor was equally loud. This suggests that infants could separate the two voices, provided the target voice was at least 5 dB louder. Hence, infants have some capacity to extract information from speech even in the face of competing acoustic signals. [Work supported by NICHD to PWJ and an NSF Graduate Fellowship to RN.]