Syntactic location facilitates word segmentation by infants

Abstract

English learners begin to segment words and parse speech into prosodic units between 6 and 12 months. Recent experimental evidence suggests that segments are more fully realized at the edges of major syntactic constituents. For example, the [s] in ‘‘bass’’ will completely assimilate to [S] in the absence of a major syntactic boundary between the words ‘‘bass’’ and ‘‘shoppers.’’ In this study, we examine whether the location of words in a phrase (and hence the phonological strength of a word’s initial segment) affects infants’ segmentation. Eleven-month-olds were familiarized with stories containing the target words ‘‘acorn’’ and ‘‘shoppers,’’ using the head-turn preference procedure. For half of the infants, ‘‘shoppers’’ occurred after a major syntactic boundary, whereas ‘‘acorn’’ did not. For the other half, ‘‘acorn’’ occurred after a strong syntactic boundary, but ‘‘shoppers’’ did not. Infants in the first condition listened significantly longer to isolated versions of ‘‘shoppers’’ than to ‘‘acorn.’’ In contrast, infants in the second condition listened equally long to both familiar words. These findings suggest that words located at major syntactic boundaries are easier to segment. Our results also provide further evidence that infants’ segmentation of vowel-initial words is delayed compared to their segmentation of consonant-initial words.