Words and Shape Similiarity Guide 13-month-olds' Inferences about Nonobvious Object Properties

S. Graham,Cari S. Kilbreath,A. Welder
Abstract:Words and Shape Similarity Guide 13-month-olds’ Inferences about Nonobvious Object Properties Susan A. Graham (grahams@ucalgary.ca) Department of Psychology, University of Calgary Calgary, AB T2N 1N4 Canada Cari S. Kilbreath (cskilbre@ucalgary.ca) Department of Psychology, University of Calgary Calgary, AB T2N 1N4 Canada Andrea N. Welder (anwelder@ucalgary.ca) Department of Psychology, University of Calgary Calgary, AB T2N 1N4 Canada Abstract We examined the influence of shape similarity and object labels on 13-month-old infants' inductive inferences. In two experiments, infants were presented with novel target objects with or without a nonobvious property, followed by test objects that varied in shape similarity to the target. When objects were not labeled, infants generalized the nonobvious property to test objects that were highly similar in shape (Expt. 1). When objects were labeled with novel nouns, infants generalized the nonobvious property to both high shape similarity and low shape similarity test objects (Expt. 2). These findings indicate that infants as young as 13 months of age expect those objects which share the same shape or the same label to possess the same nonobvious property. Introduction Inductive reasoning involves invoking the premise that things that are true for one member of a category (e.g., the blue ball bounces) will hold true for other members of the same category (e.g., therefore all balls bounce; Moore & Parker, 1989). The ability to reason inductively is an invaluable cognitive skill, allowing an individual to generalize knowledge to new instances and new situations. In recent years, a great deal of empirical attention has been devoted to examining preschoolers' inductive reasoning abilities, with particular focus on the nature of the categories that guide their inferences. In a typical inductive generalization task, preschoolers are taught a fact about a target object and then are asked whether that fact can be generalized to other test objects. Using this methodology, studies have demonstrated that by 2-1/2-years of age, children can reason inductively about object properties in remarkably sophisticated ways (e.g., Gelman, 1988; Gelman & Markman, 1986, 1987; Kalish & Gelman, 1992). For example, Gelman and Coley (1990) found that 2-1/2-year-olds will overlook perceptual similarity and generalize properties on the basis of shared underlying kind when the target and test objects are given the same count noun label. In recent years, researchers have begun to examine the development of inductive capabilities during the infancy period using the generalized imitation paradigm. In a typical task, an experimenter will first model a specific action on a target object. He or she will then hand infants test objects and observe whether or not they imitate the target action on the various objects. Studies using this paradigm indicate that infants as young as 9 months of age will draw inferences about nonobvious object properties based on knowledge gained during the experimental session (Baldwin, Markman, & Melartin, 1993). Furthermore, research suggests that both perceptual similarity and conceptual knowledge may play a role in guiding infants' inferences (Baldwin et al., 1993; Mandler & McDonough, 1996, 1998). In a recent series of studies (Welder & Graham, in press), we found that 18- month-old infants will rely on shared shape similarity to guide their inductive inferences about novel objects' nonobvious sound properties when no other information about category membership is available. More importantly, however, when infants were provided with information about conceptual category membership in the form of shared object labels, shape similarity was either attenuated in significance (in the case of novel labels) or disregarded (in the case of familiar labels). These findings indicate that 18-month-old infants can make inductive inferences about object properties based on a conceptual notion of object kind. Furthermore, these findings suggest that infants as young as 18 months of age recognize the conceptual information conveyed by object labels. That is, they recognize that noun labels supply information about underlying object kind and that members of the same kind share nonobvious properties. In the present studies, we pursued the investigation of infants' inductive abilities, with specific focus on the reasoning abilities of infants who are just beginning to acquire productive language. First, we examined whether 13-month-olds, like 18-month-olds, will rely on shared shape similarity to generalize nonobvious object properties, in the absence of other information about
What problem does this paper attempt to address?