ISCA Workshop on Plasticity in Speech Perception (PSP2005)
Senate House, London, UK
A considerable amount of research has focused on infants' attention to speech typically directed to them by adult caretakers (infant directed speech or IDS). Although IDS differs from adult-directed speech (ADS) in many ways, primary focus has been on pitch characteristics; IDS is typically higher in absolute pitch, and more variable in pitch (exaggerated pitch excursions; wider pitch ranges). However, IDS is also typically slower in temporal form, largely because many of its voiced segments are stretched over time. Little attention has been paid to this characteristic of IDS and its affect on infant attention.
In these experiments, we took IDS-Normal utterances and either attenuated segmental durations (i.e., decreased voiced segments, producing IDS-Fast) or augmented segmental durations (i.e., increased voiced segments, producing IDS-Slow); in both cases, the relative pitch characteristics remained unchanged. We then tested groups of 4- and 8-month-old infants in a serial attention protocol with either IDS-Normal vs. IDS-Fast, or IDS-Normal vs. IDS-Slow recordings. Two measures of attention were recorded: fixation of a visual target and heart rate activity. Visual fixation was recorded because looking at the target (a colorful bullseye pattern) was contingently yoked to hearing the recordings; that is, infants heard one of two speech patterns as long as they looked at the target. Heart rate activity was simultaneously recorded because increases in heart period reflect heart rate (HR) deceleration, an indicator of sustained attention. We expected no differences in average looking times or HR as a function of speech type if segmental duration does not influence infants' attention to IDS.
In the first experiment, both 4-month-olds (n = 19) and 8-month-olds (n = 18) heard ID-Normal and IDFast speech. The results showed that the younger infants looked significantly longer to ID-Normal; older infants showed no difference in visual attention. Interestingly, both ages showed greater degrees of HR deceleration (sustained attention) to ID-Normal. In the second experiment, both 4- month-olds (n = 16) and 8-month-olds (n = 20) heard ID-Normal and ID-Slow speech; the results showed equal visual attention across groups, but increased sustained attention (HR deceleration) to ID-Slow in the 4-month-olds.
These data suggest that the temporal features of IDS affect infants' attention, with longer segmental durations (slower speech) increasing attention more in younger than older infants. Attending more to slower speech may reflect younger infants' increased interest in highly emotive speech, but may also act to increase certain aspects of language learning such as native phonemes.
Bibliographic reference. Panneton, Robin / McIlreavy, Megan (2005): "Developmental differences in infants' attention to utterance duration", In PSP2005, 52-55.