Why might young children be especially anthropocentric? One factor may be that they are presumably exposed more to humans than to other biological kinds. Another idea is that children are reluctant to generalize from any base without extensive knowledge about that biological kind. In support of this notion and as discussed in Chapter 4, Inagaki (1990) examined generalization of biological properties by children in Tokyo, some of whom had extensive experience raising goldfish. She found that children who had no pets showed a familiar anthropocentric pattern of generalization, whereas children raising goldfish showed two generalization gradients—one around humans and one around goldfish (e.g., they generalized from goldfish to turtles).
If intimate experience with biological kinds governs patterns of generalization, then rural children may not show anthropocentrism at all. Ross, Medin, Coley, and Atran (2003) examined inductive generalizations from different bases among urban children, rural European American children, and rural American Indian children using a procedure similar but not identical to that employed by Carey (1985).
For both groups of rural children, human was not a better base for generalization than a nonhuman mammal. Young urban children showed broad and relatively undifferentiated generalization. Ross et al. (2003) also found evidence for an alternative strategy for generalization. Older rural European American and American Indian children of all ages sometimes generalized in terms of ecological or causal relations. For example, when they were told that bees have “andro” inside them, they might reply that bears also have andro inside them, justifying their judgments by saying that andro might be transmitted to bears when bees sting them, or that andro might also be in honey (which bears eat).
Other results suggest that evidence for anthropocentrism in young children depends on the details of tasks and procedures (Waxman and Medin, 2007) but that it is seen only in young urban children. Although anthropocentrism may reflect a lack of intimate experience with the biological world, it may also reflect an anthropocentric cultural model, as seen, for example, in Disney movies and in the way urban pets are often treated (e.g., dogs are typically seen as part of the family).
Related work reinforces the idea that urban (as differentiated from rural) environments influence the development of children’s biology. For example, Coley and associates (Coley, Vitkin, Seaton, and Yopchick, 2005) have examined taxonomic and ecological generalization as a function of age and experience. Rather than dichotomizing children as urban versus rural, Coley used the continuous measure of population density. He found that taxonomic generalization shows little, if any, variation as a function of age or population density (see Waxman and Medin, 2007, for similar results using a different paradigm) but that ecological generalization increased systematically with age and decreased systematically with population density. In addition, the distinction between properties that may be distributed by ecological agents