Title: Do Perceptual Complexity and Object Familiarity Matter for Novel Word Extension
Abstract: Do Perceptual Complexity and Object Familiarity Matter for Novel Word Extension? Catherine M. Sandhofer ([email protected]) Department of Psychology, 1101 E. 10th Street Bloomington, IN 47404 USA Linda B. Smith ([email protected]) Department of Psychology, 1101 E. 10th Street Bloomington, IN 47404 USA Abstract This paper examines the relationship between shape complexity and familiarity in extending novel adjectives. Previous research has suggested that familiarity with an object's basic level label determines the likelihood that a novel adjective will be extended to new instances. The present results do not support that conclusion. Instead the results suggest that given an adjectival syntactic frame children are likely to extend novel words to other objects of the same material when the objects are simple in shape. This result suggests that the perceptual properties of objects and the lexical form class cues are integral to understanding how children come to learn new words. Introduction How do children come to extend words to new instances? This question is at the heart of research understanding language development partially because much of language learning presumably takes place using ostensive definition: children learn a label for one object, event, or property and are able to extend that label to new instances. The task used to study this is the novel word extension task. In this task a child is shown an exemplar and the exemplar is labeled. The child is then given other objects that match the exemplar on different dimensions and the child is asked to select the one that also has the same label. Although much of the research in this area tends to focus on how children extend novel count nouns, other grammatical classes have been studied as well. For example, previous work has suggested that when encountering novel adjectives, children are likely to extend the novel adjective to other objects of the same material only if the objects are familiar to them. (Hall, Waxman, & Hurwitz, 1993). Thus, young children can extend the novel adjective “plush” to other objects of the same material if the plush objects are familiar to the child e.g. a shoe, but not if the objects are unfamiliar to the child, e.g. a widget, even when the children are provided with an adjectival syntactic frame. This finding has been interpreted as evidence that children are biased to expect a novel word to refer to a kind of object. By this account children should extend novel words to other objects of the same shape if the object is unfamiliar to them and children should extend novel words to objects sharing some other property when the object is familiar to them. However, Landau, Smith, and Jones (1992) and Smith, Jones, and Landau (1992) have shown that young children can generalize novel adjectives to other objects that match in material. These results are seemingly at odds with Hall, Waxman, & Hurwitz (1993) because the objects presented to children in these studies were unfamiliar objects and thus by Hall et al’s proposal children should initially interpret the novel words as referring to objects of the same shape or object kind. In addition, other research has shown that children take the specific perceptual properties of objects into account when extending novel words. For example, several researchers (Soja, 1992; Dickinson, 1988; but see Markman & Wachel, 1988 for contradictory findings) have demonstrated that children extend novel nouns to solid objects with the same shape, but children extend novel nouns to non-solid substances with the same material as an exemplar. Further, Imai and Gentner (1997) have shown that children as young as two years of age generalize simple objects, complex objects, and substances differently. Based on these previous results, we propose an additional constraint that may guide whether children are likely to extend a novel word to objects that match an exemplar in shape vs. objects that match in other properties. We call this the perceptual complexity hypothesis. By this hypothesis, complex objects, objects like tractors with multiple parts may be labeled by more possible words than simple objects, like ball. This may foster attention to shape if these other labels point to properties (wheels, smokestack, engine) are themselves correlated with shape. Thus, we predict that attention to shape should be a stronger pull when the objects are complex than when they are simple. This idea is supported by previous findings by Imai and Gentner (1997) that show when Japanese 2 year olds
Publication Year: 2001
Publication Date: 2001-01-01
Language: en
Type: article
Access and Citation
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot