Preschoolers' sensitivity to talker information in online recognition

It is unknown what degree of acoustic detail children pay attention to when learning language. We understand that children register sound differences important for word identification; however, we do not know if differences in acoustic structure that result from differences in dialect (accent) or talker are meaningful to children. In this study, we conducted various experiments to determine which acoustical differences were most salient for preschoolers. By using this method, we predicted that there would be variation in ability to perceive different types of sound. In a word-learning paradigm we have previously found that pre-school children (n=24) learned the names of two creatures (e.g. keeb) after n=16 repetitions of each name (M=74%; two-choice task). Further, these children were much lower in accuracy when the name was heard with a vowel change (kib), suggesting they pay attention to phonemic changes similar to those found in accent variation. In the current study, we used this learning paradigm but, instead of learning different names for creatures, children learned to associate particular voices with creatures. Experiment 1 used two voices matched on gender, accent and age, following previous research. We found that preschoolers (n=24) could distinguish the speakers with varying degrees of accuracy (M=62%), with a possible developmental effect. In follow-up experiments, the task remained challenging for children (n=16; M=56%) even when same-gender speakers differ in intonation, but children (n=16) performed at ceiling for different-gender speakers (M=92%). Comparing these studies that recognizing voices may have a longer developmental trajectory than recognizing speech sounds.