Development of Category-Based Reasoning: Results from a Longitudinal Study

Development of Category-Based Reasoning: Results from a Longitudinal Study Karrie E. Godwin (kegodwin@andrew.cmu.edu) Anna V. Fisher, (fisher49@andrew.cmu.edu) Bryan J. Matlen (bmatlen@cmu.edu) Carnegie Mellon University, Department of Psychology, 5000 Forbes Avenue, Pittsburgh, PA 15213 USA Abstract Godwin, et al., in press; Fisher, et al., 2011; Fisher 2010; Fisher & Sloutsky, 2005). However, it remains unclear what leads to the development of spontaneous category-based reasoning during the preschool years. Two classes of explanations have been put forth to explain changes in various areas of cognitive development, namely Representational Change and Executive Function development. We briefly discuss both explanations below. Prior research on the development of category-based reasoning indicates a protracted developmental course of this ability as well as a high degree of individual variability. However, the sources of this individual variability as well as the sources of developmental change remain unclear. The present study aimed to examine these issues, with a focus on the role of representational change and executive function development. Across two time points spaced approximately 7 months apart, children’s category-based reasoning was assessed along with a battery of executive function and representational change measures. Results replicated prior work in that only a small proportion of children exhibited spontaneous category-based reasoning at Time1, and this proportion increased with development. In addition, both executive function and representational change were found to predict the development of category-based reasoning. Representational Change Representational change is “reorganization of existing knowledge or a difference in the utilization of information, rather than the acquisition of new information” (Nelson, 1977, p. 109). Representational change has been implicated as an explanatory factor in several areas of cognitive development, including analogical reasoning (e.g., Gentner et al., 1995), problem solving (e.g., Karmiloff-Smith, 1984), and numerical development (e.g., Opfer & Siegler, 2007). With regards to semantic development, there are several compelling sources of evidence pointing to representational change in the multidimensional scaling literature (e.g., Howard & Howard, 1977), free association studies (e.g., Brown & Berko, 1960), and development of semantic priming (e.g., McCauley, Weil, & Sperber, 1976). Furthermore, different approaches to modeling semantic cognition suggest that early conceptual organization is fairly undifferentiated (such that penguin, trout, and alligator may start out as belonging to the same cluster) with greater differentiation emerging with development (Kemp & Tenenbaum 2008; Rogers & McClelland 2004). At present, there is no direct empirical evidence testing these predictions, although Carey’s (1985) seminal work is largely consistent with these developmental profiles. Keywords: Category-based reasoning; inductive reasoning Introduction Category-based reasoning is central to mature cognition and underlies much of our learning and functioning in the world (e.g., Osherson et al., 1990; Sloman, 1993). Despite early reports that even very young children spontaneously engage in category-based reasoning (e.g., Gelman & Markman, 1986; Gelman & Coley, 1990; Welder & Graham, 2001), recent evidence suggests that development of category-based reasoning follows a relatively protracted developmental course (e.g., Badger & Shapiro, 2012; Godwin, Matlen, & Fisher, in press; Fisher, Matlen, & Godwin, 2011; Fisher, 2010; Fisher & Sloutsky, 2005). One of the hallmarks of category-based reasoning is one’s ability to make inferences based on the knowledge that two (or more) items belong to similar kinds in the absence of supporting perceptual information. For example, if one is shown a picture of a rock, a sponge, and another rock and asked to predict which two items have properties in common, one could rely on perceptual similarity to make an inference. Similarly, if the pictures are ambiguous (or not presented) and labels are used to indicate category membership, one could base their inference on matching labels (e.g., rock–rock), not necessarily because one understands that labels refer to kinds, but because the labels are perceptually identical (Sloutsky & Fisher, 2004). However, one’s ability to rely on semantically-similar labels (e.g., rock-stone) to make inferences is commonly interpreted as an index of category-based reasoning (e.g., Gelman & Markman, 1986). Several studies have documented that the ability to spontaneously engage in category-based reasoning appears between 4 and 6 years of age (e.g., Badger & Shapiro, 2012; Executive Functions Executive Functions (EF) are psychological processes thought to control other (typically, higher-order) psychological processes such as planning, reasoning, and problem-solving. Most researchers distinguish the following EF processes: set shifting, active maintenance of representations (sometimes referred to as working memory), and inhibitory control (Bunge et al., 2002; Carlson et al., The EF system is traditionally associated with prefrontal cortex, which is believed to be one of the slowest brain regions to mature (e.g., Diamond, 2002). Development of EF has been implicated in developmental accounts of category learning (Sloutsky, 2010), and there is evidence that representation maintenance and inhibitory control play a role in the development of analogical reasoning (Morrison