How each prosodic boundary cue matters: Evidence from German infants

Previous studies have revealed that infants aged six to ten months are able to use the acoustic correlates of major prosodic boundaries, that is, pitch change, preboundary lengthening, and pause, for the segmentation of the continuous speech signal. Moreover, investigations with American-English- and Dutch-learning infants suggest that processing prosodic boundary markings involves a weighting of these cues. This weighting seems to develop with increasing exposure to the native language and to underlie crosslinguistic variation. In the following, we report the results of four experiments using the headturn preference procedure to explore the perception of prosodic boundary cues in German infants. We presented eight-month-old infants with a sequence of names in two different prosodic groupings, with or without boundary markers. Infants discriminated both sequences when the boundary was marked by all three cues (Experiment 1) and when it was marked by a pitch change and preboundary lengthening in combination (Experiment 2). The presence of a pitch change (Experiment 3) or preboundary lengthening (Experiment 4) as single cues did not lead to a successful discrimination. Our results indicate that pause is not a necessary cue for German infants. Pitch and preboundary lengthening in combination, but not as single cues, are sufficient. Hence, by eight months infants only rely on a convergence of boundary markers. Comparisons with adults’ performance on the same stimulus materials suggest that the pattern observed with the eight-month-olds is already consistent with that of adults. We discuss our findings with respect to crosslinguistic variation and the development of a language-specific prosodic cue weighting