New Perspective for Verb Learning Hanako Yoshida (yoshida@uh.edu) Department of Psychology, University of Houston, 126 Heyne Building Houston, TX 77204-5022 USA Linda B. Smith (smith4@indiana.edu) Psychological and Brain Sciences and Program in Cognitive Science, Indiana University Bloomington, IN 47405 USA Brian Weisinger (brian.weisinger@hotmail.com) Department of Psychology, University of Houston, 126 Heyne Building Houston, TX 77204-5022 USA Abstract When technology eventually allows us to zoom in to the point of view of children, will it provide us with any informative information about what the early word-learning environment really looks like? The present study uses a child-centered view that simulates infants’ in-the-moment view to test whether such a newly obtained perspective may provide us with new insight into early verb learning. We present adult participants with two perspectives taken from child-parent play sessions in which parents were teaching the child a set of nouns and verbs. One was the child-centered view, and the other was a typical third-person view, and both were used without audio. The participants’ task was to identify what words the parents taught by looking at the video clips. The results support higher accuracy in identifying nouns than verbs, and they provide a new finding about adults identifying verbs better than nouns when viewing the child-centered view. Keywords: verb learning; word learning context; child- centered view. Introduction One of the most crucial questions in language learning is how children map words to meanings. Needless to say, there are a number of factors contributing to the process of word learning, including social cues such as eye gaze and gesture, prosody, language structure, input frequency, pragmatics, and much more. However, a complete explanation may require more than a characterization of the learning environment because learners come with biases. Moreover, learners actively engage in their world, and in so doing they distort the regularities and carve the input in systematic ways. This means that one cannot really consider the input separately from the learners’ own actions because the learner selects and creates the input. Although the study of human cognitive development has focused on the “input” as divorced from the learner’s activity, the growing interest in embodiment suggests the need to take a closer look. One recent study that motivated this present study investigated how early learners create their own visual input by studying the first-person view during toy play (Yoshida & Smith, in press). This study used a small camera attached to small children’s foreheads and thus documented how this view may differ from what can be captured from a third- person view (commonly used for standard observation studies). In the present paper, we show how this child- centered view might change our understanding of early word learning, and particularly the relevant information for learning nouns and verbs. Perceptual Accessibility and Word Learning Gentner (1978) pointed out that there are perceptual differences in the accessibility of meanins among different word categories that determine how easily the words can be learned. In particular, she suggests that nouns dominate verbs in early productive vocabulary (see also Gentner 1982; Gentner & Boroditsky, 2001) because noun meanings are more available from the perceptual input. The natural partitions part of their hypothesis further proposes that animates and objects present cohesive perceptual properties easily separable from complex scenes. These cohesive perceptual bundles are most likely lexicalized as nouns, and as a result, they argue, young word learners universally learn nouns easily (see also Gillette et al., 1999). Verbs, in contrast, being about relations, are not easily specified by the perceptual input. Human Simulation Paradigm (HSP) The effect of perceptual accessibility was made dramatically apparent in a study of adults’ ability to interpret the referents of real child-directed utterances of parents (Snedeker & Gleitman, 2000). The evidence comes from what is known as the Human Simulation Paradigm (HSP). In these studies, adults are asked to guess a set of words given visual information of child-mother play. The evidence shows that their rate of successful guessing differs depending upon the word type (Snedeker & Gleitman, 2000; see Piccin & Waxman, 2007, for a similar result with children; see also Snedeker, Gleitman, & Brent ,1999; Gillette, Gleitman, Gleitman, and Lederer, 1999; and Snedeker, 2000). In brief, they are better at guessing nouns from such scenes than verbs. 2215