Developmental Science 5:2 (2002), pp 219 –232 © Blackwell Publishers Ltd. 2002, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA. Blackwell Publishers Ltd PAPER Relevant properties How children know the relevant properties for generalizing object names Susan S. Jones and Linda B. Smith Department of Psychology, Indiana University, USA Abstract Young children’s novel word extensions indicate that their animal categories, like those of adults, are characterized by multiple similarities among instances; whereas their artifact categories, again like those of adults, are characterized more simply by com- monalities among instances in shape. Three experiments shed light on the nature and development of a mechanism that enables children to organize novel lexical categories differently for different kinds of objects. Experiment 1 shows that, by adult judgments, animals and artifacts present different category organizations. Experiment 2 shows relations between both age and the number of nouns young children have acquired, and children’s kind-specific generalizations of newly learned nouns. Experiment 3 is a training study in which even younger children show an ability to learn and then generalize highly abstract relations between different contextual cues and different category structures; and importantly, to learn more than one set of such relations at a time. Together, these three findings indicate one way in which children are able to rapidly and accurately form lexical categories that parallel those of adults in their language community. In introducing a series of papers by eminent development- alists, R. Gelman (1990) pointed to attention as a key problem. She asked ‘ How is it that our young attend to inputs that will support the development of concepts they share with their elders?’ (p. 3). The problem of selecting the right information for categorization is known as the problem of feature selection. As Murphy and Medin (1985) stated, selecting the right features for constructing a new category seems to require already knowing something about that category. For example, color matters more for whether something is a pea than for whether it is a ball. How then when we see some novel object – a potential pea or ball – do we know whether to attend to color? In the present research, we address the problem of feature selection as it applies to young children’s learning of object names. Two classes of theoretical solution to the problem of feature selection have been offered. One is that conceptual knowledge – a naïve theory about categories – directs attention (e.g. S. Gelman & Coley, 1991; Medin & Ortony, 1989; Quine, 1969; Soja, Carey & Spelke, 1991). Knowledge that an object is, for example, to be eaten versus to be bounced may direct attention to certain per- ceptible properties and away from others. The second solution to the feature selection problem is that learned associations among perceptual features and contexts auto- matically shift attention among perceptual properties (e.g. Jones & Smith, 1993; Jones, Smith & Landau, 1991; Krushke, 1992; Nosofsky, 1986; Smith & Jones, 1993; Smith, Jones & Landau, 1996). For example, the surface textures and curved shapes of fruits may be cues for atten- tion to color (Macario, 1991), or as we consider here, the property of eyes may be a cue for attention to texture (Jones et al., 1991). Although both kinds of knowledge – concepts and associations among perceptual features – are likely to play a role, in this paper we focus specific- ally on learned associations among perceptual properties. The phenomenon A number of researchers – for example, Jones et al. (1991), S. Gelman (1988), Keil (1994) and Markman (1989) – have all suggested that animal categories are richly structured by being based on multiple similarities, whereas common artifact categories are more simply structured and based principally on shape. Jones et al. (1991) tested this idea in a noun generalization task. They showed 3-year-old children novel laboratory Address for correspondence: Susan S. Jones, Department of Psychology, Psychology Building, 1101 East 10th Street, Bloomington, Indiana 47405- 7007, USA; e-mail: jones1@indiana.edu