Sub-segmental detail in early lexical representations

In previous studies of phonological sensitivity, toddlers have failed to differentiate mispronunciations of varying severity. We provide evidence of more sophisticated phonological knowledge. Nineteen-month-olds were presented with displays consisting of one familiar and one unfamiliar object. In Experiment 1, names of familiar objects were pronounced correctly or had onset mispronunciations of one, two, or three phonetic features. Under these referential conditions, subjects demonstrated linearly graded sensitivity to the degree of mismatch. In Experiment 2, mispronunciations involved one-feature place, voice, or manner changes; in Experiment 3, subjects heard three types of two-feature onset mispronunciations. Within each of these two experiments, responses were similar to the three types of mispronunciations. Moreover, the pattern across Experiments 2 and 3 confirmed the graded sensitivity observed in Experiment 1. These results converge to suggest that 19-month-olds’ representations of familiar words are quite mature and that lexical processing in toddlers (as in adults) is affected by sub-segmental detail.