Cognitive Development What children infer from social categories

Children hold the belief that social categories have essences. We investigatedwhat kinds of properties children feel licensed to infer about a person based on social category membership. Seventy-two 4–6-year-oldswere introduced tonovel social categories defined as having one internal – psychological or biological – and one external – behavioral or physical – property. For half of the participants, the internal propertywas described as causing the external one; for the others, no causal relationship between properties was mentioned. Children were asked to choose as a novel exemplar of a category one with only the internal or only the external property. Children inferred thatexemplarshadapsychologicalproperty irrespectiveof causal status, but they inferred the presence of a biological property only when described as causal. Children did not draw systematic inferences regarding either of the two external properties. These findings indicate that children treat psychological and causal properties as central – and perhaps essential – to human kinds. © 2011 Elsevier Inc. All rights reserved. Extensive evidence coming from the anthropological (Gil-White, 2001), aswell as the social (Keller, 2005) and developmental (Astuti, Solomon, & Carey, 2004) psychology literature, suggests that adults and children around the world hold essentialist beliefs about certain human groups. Namely, people believe that certain social group distinctions mark fundamental differences between group members, and that these differences are inherent, permanent, exclusive, and causally predictive of group members’ characteristics. Thus, studies have found that children believe membership in racial categories to be established at birth and stable throughout life (Hirschfeld, 1996), that gender innately disposes people towards certain behavioral preferences (Taylor, 1996), that certain social categories reflect objectively truthful classes of the world (Rhodes & Gelman, 2009), and that ethnic categories ∗ Corresponding author at: Department of Psychology and Gonda Brain Research Center, Bar-Ilan University, Ramat-Gan 52900, Israel. Tel.: +972 3 5318630; fax: +972 3 7384106. E-mail address: [email protected] (G. Diesendruck). 0885-2014/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.cogdev.2010.11.001 G. Diesendruck, E. Eldror / Cognitive Development 26 (2011) 118–126 119 are inductively powerful (Birnbaum, Deeb, Segall, Ben-Elyiahu, & Diesendruck, 2010; Diesendruck & haLevi, 2006). Although these lines of work provide evidence for the existence in children of a belief that social categories have essences, very little attention has been devoted to what children consider to be the essence of social categories. In a sense, this state of affairs was envisioned in Medin and Ortony’s (1989) original definition of psychological essentialism as a “placeholder belief”; namely, a belief that categories have essences, without commitment to a particular instantiation of essences. This stipulation notwithstanding, in the present study we attempt to provide indirect clues as to what children believe constitutes the essential properties of human kinds. In fact, in the domain of natural kinds, similar attempts have been made. One line of this work has focused on the “causal status” of properties. For instance, studies have found that, in determining an animal’s categorymembership, children treat properties described as causal to the definition of an animalkindasmore important thanpropertiesdescribedaseffects (Ahn,Gelman,Amsterlaw,Hohenstein, & Kalish, 2000; Opfer & Bulloch, 2007). A second line of work has attempted to identify the “content domain” of properties that children find as definitional of animal categories. Dating back to Keil’s (1989) “transformation” studies, it has been found that from a young age, and in different cultures, children believe that internal, non-obvious, biological properties define animal category membership (Diesendruck, 2001; Newman & Keil, 2008). In a recent study, Meunier and Cordier (2009) integrated these two lines of work and found that 5-year-olds relied on causal properties for categorizing animals only when these were described as internal, as opposed to surface, properties of the animals. Altogether, these findings about the nature of essentialist beliefs about animals have supported the claim that children’s categorization of animals is governed by a naïve biological theory (Atran, 1990; Gelman, 2003). In the present study, rather than asking which properties children take as definitional of category membership, we ask the complementary question: Given an exemplar’s categorymembership, which properties do children infer the exemplar must have? As Gelman (2003) noted, “when given the category label [for a category in an essentialized domain], you can be certain that the instances have the category essence” (p. 44). Following this reasoning, we were interested in what children believed they could infer about a person, based on knowledge of that person’s social category membership. If children believe that a certain property is central to categorymembership, and they are given a choice between an exemplar that lacks that property vs. another one that has it, children should select the latter as a member of the category. This strategy is similar to the one pioneered by Gelman and Markman (1986) in their studies on category-based induction in the domain of living kinds and later extended to the domain of human kinds (Diesendruck & haLevi, 2006; Heyman & Gelman, 2000). A crucial difference between these previous studies and the present one is that the former focused on the kinds of similarity relations that children privilege for drawing inferences (e.g., perceptual vs. taxonomic similarity), the goal of those studies being to uncover whether and which categories children essentialize. For instance, Diesendruck and haLevi asked children to infer the novel psychological property of a character based on ethnic vs. gender status or ethnic vs. religiosity status. In other words, the properties to be inferred were always psychological, and what varied was the category membership of the target characters. The present study, instead, targets the kinds of properties children feel licensed to infer based on social categorymembership. Specifically,we investigatewhether children’s inferences varydepending on the causal status and the content domain of the to-be-inferred properties. In this respect, the design and logic are similar to thoseusedbyKalish and Lawson (2008). In their study, childrenwere told about the category membership of a character and were asked to select which property the character was likely to have. As Kalish and Lawson describe, this method allows one to determine how central to the representation of social categories different types of properties are. They found that already by kindergarten age, children view “obligation”-related properties (what one has to do) as more central to social categories than “preference”-related properties (what one likes to do). The present study extends this question to additional types of properties and property relations. To investigate the role of both property domain and causal status, we adapted the method developed by Ahn et al. (2000). They presented 7–9-year-olds with pictures of a novel animal category, told children about properties descriptive of the category, and asked children to decide which novel 120 G. Diesendruck, E. Eldror / Cognitive Development 26 (2011) 118–126 exemplar was a member of the category: one missing property X or another missing property Y. The crucial manipulation was that for some children property X was defined as causing the other properties (i.e., Y was an effect-property in this condition), while for others X and Y were simply listed as two of the properties of the animal category. Ahn et al. found that children weremore likely to choose as another exemplar of a target category, an animal missing a non-causal property than one missing a causal property. We used a simplified version of this task with younger children, asking children to infer which property a category exemplar had rather than which category a particular combination of properties defined. Also, instead of targeting animal categories with pertinent properties, we presented children social categorieswith additional kinds of properties. Guided by theoretical positions in the relevant literature, most studies of animal categorization have presented childrenwith the possibility of drawing inferences about internal/biological, visual/superficial, or behavioral properties. In light of theoretical debates regarding the domain of human kinds, however, an additional type of property seemed crucial – social/psychological properties. Researchers debate whether social categorization is governed by domain-specific conceptual schemes or by analogy to a natural domain. Some argue that children construe social categories in social/psychological terms by kindergarten age (Heyman & Gelman, 1998; Hirschfeld, 1996; Kalish, 2002). Thus, they should treat social/psychological properties of human kinds as proxies for essences. Consequently, provided the definition of a social category, and being informed about the category membership of a novel exemplar, children should feel licensed to draw inferences primarily about the exemplar’s psychological properties. Others argue that concepts of social categories derive from beliefs about animal kinds (Atran, 1990). In this case, children should treat internal/biological properties as central to social categories. Additionally, children might hold a “performative” view of human kinds, i.e., you are what you do, leading them to draw inferences about behavioral properties based on category membership (Astuti et al., 2004). Finally, if children’s categories are defined by perceptual similarity, they should feel licensed to draw inferences primarily about physical features. In sum, the present study extends Ahn et al.’s (2000) study to address a number of questions. First, we target social rather than animal categories, allowing us to assess whether causal status impacts children’s category-based inferences in that domain as well, and which property domain children most strongly believe is determined by social category membership. Second, and related, we add another content domain of properties to those typically included in studies on animal categorization – social/psychological properties. Doing so permits evaluation of the broad conceptual theories that underlie children’s social categories. Third, we simplified the procedure slightly for kindergarten children, to examine category-based reasoning at the age at which social essentialism becomes evident. Moreover, given debates regarding the age at which children develop firm conceptual distinctions between naïve psychological and biological understandings (Carey, 1995; Gelman, 2003), kindergarteners seemed relevant and potentially different from the 7–9-year-olds examined by Ahn et al. In fact, in their studies on social-category based property inferences, Kalish and Lawson (2008) found differences in the response patterns of children from these two age groups.