«Sandra Waxman and Douglas Medin Northwestern University Experience and Cultural Models Matter: Placing Firm Limits on Anthropocentrism Abstract This ...»
Experience and Cultural Models Matter:
Placing Firm Limits on Childhood Anthropocentrism
Sandra Waxman and Douglas Medin
Experience and Cultural Models Matter:
Placing Firm Limits on Anthropocentrism
This paper builds on Hatano and Inagaki’s pioneering work on the role of experience
and cultural models in children's biological reasoning. We use a category-based induction
task to consider how experience and cultural models shape rural and urban children’s patterns of biological reasoning. We discuss the implications of these findings for developmental theory and educational practice.
Experience and Cultural Models Matter:
Placing Firm Limits on Anthropocentrism The concept ALIVE – a developmental view Considerable developmental research has been devoted to understanding the acquisition of biological concepts and reasoning. This work has roots in the Piagetian tradition, which claimed that children’s reasoning about core biological concepts, such as the concept, alive, was qualitatively different than that of adults (Piaget, 1954). In particular, young children’s responses in Piagetian interviews led to a widely-held view that they were animistic in their thinking, attributing living kind status to a broad range of non-living entities (e.g. clouds, bicycles). In other tasks, however, young children respond differently, tending to attribute living kind status to an overly-restricted set of entities (to animals, but not plants) (Hatano, Siegler, Richards, Inagaki, Stavy & Wax, 1993; Keil, 1983; Opfer & Siegler, 2004; Stavy & Wax, 1989). Another key finding in this arena concerns young children’s apparent anthropocentrism, in which they tend to rely heavily on humans as a privileged inductive base whenever they are asked to attribute biological properties (e.g., being alive, having a heart) to other animals. Clearly, questions concerning the nature of the biological concepts held by children and their relation to those held by adults are very much alive.
The pioneering contributions of Giyoo Hatano Giyoo Hatano’s fresh insights to this area of inquiry cannot be overestimated. In his pioneering work with his long-time collaborator Kayoko Inagaki, he has made lasting theoretical and educational contribitions regarding the biological concepts held by children. Our goal in this paper is to honor their contribution, pointing out the ways in which their insights into the role of experience and cultural models have shaped our cross-cultural developmental research program on biological reasoning.
Culture, experience and childhood anthropocentrism. Captivated by the issue of biological concepts and reasoning, domain-specificity, and developmental change, Hatano and Inagaki (Inagaki & Hatano, 1993, 1996, 2002; Hatano & Inagaki, 1994, 1999, 2000) launched a multi-faceted developmental research program. In one line of work, they considered how the cultural models espoused within a community shape children’s biological reasoning. These studies revealed that 5- to 8-year-old Japanese children understand many bodily processes in terms of vitalism -- a causal model that is pervasive in Japan and that relies on the distinctly biological concept of energy.1 In another facet of their ingenious program, Inagaki and Hatano (Inagaki, 1990;
Hatano & Inagaki, 1994, Inagaki & Hatano, 2002) focused on children’s direct experience and identified a potential source for the anthropocentric pattern we described earlier. Of course, children’s experiences with the biological world are varied, including not only the cultural models and beliefs espoused within the community, but also their habitual surroundings (e.g., rural vs. urban), their informal learning opportunities (e.g., direct hands-on experience, including farming, fishing, summer camp activities), and perhaps more remotely, their experiences in more formal learning environments and access to videos, books, and visits to the zoo (see Inagaki, 1990; Rosengren, Gelman, Kalish, & McCormick, 1991).
Hatano and Inagaki examined the role of experience by identifying urban Japanese children who had cared for and raised goldfish in their homes, and those who had not. When they then assessed these children’s attributions of biological properties, they found that children without goldfish-raising experience showed an anthropocentric pattern, but that those who had raised goldfish did not. This suggests that anthropocentrism may be mediated by children’s experience, a point to which we return below.
Still, the anthropocentric pattern produced by urban Japanese children who did not raise goldfish converged well Carey’s (1985) results with urban Boston children’s performance in a category-based induction task. In this task, an interviewer introduced children to a picture of a biological entity (either a human, dog, or bee) and taught them about an unfamiliar biological property (e.g., has an omentum) of that entity. Next, children were asked whether that novel property could also be generalized to other entities (including other humans, non-human animals, inanimate natural kinds (e.g., sun), and artifacts (e.g., garlic presses). Carey reported that 4- and 6-year-olds, unlike older children and adults, hold an anthropocentric conception of the biological world, treating human as a privileged inferential base.
She took this finding as evidence that young children’s knowledge pertaining to the biological world is organized around a model in which humans stand as the prototypical exemplar, and that development in this domain is characterized by pervasive conceptual change which is necessary to catapult children from this anthropocentric model to the more adult (Western science-oriented) model in which humans stand as one animal among many. (Medin and Waxman, in press, provide an extended discussion of alternative interpretations of these asymmetries in induction.) Limits on childhood anthropocentrism Importantly, however, the evidence from Hatano and Inagaki suggested that there may be limits on this anthropocentrism: This pattern may not be universal, but may instead be mediated by experience. This also challenged the view that development within the domain of biological knowledge necessarily entails radical conceptual change (Carey, 1985). In fact, it suggested an alternative: urban children’s propensity to view humans as a privileged inductive base may be driven by their relatively impoverished knowledge about animal kinds other than humans. If this interpretation is correct, then children with more knowledge and experience with the biological world should be less likely to reason anthropocentrically. This, of course, is exactly the pattern observed in urban Japanese goldfish-raising children.
In our research, we have been inspired by Hatano and Inagaki to pursue this possibility experimentally. We are involved in a cross-cultural developmental series of studies on biological knowledge and reasoning spanning several different communities (Anggoro et al., 2005; Ross et al., 2003, Medin & Atran, 2004; Waxman, Medin & Ross, in press). In this paper, we focus on urban vs rural children being raised in the US. The logic underlying this urban-rural comparison is straightforward: with regard to experience, rural children are the counterpart of Inagaki and Hatano’s goldfish-raisers, except that they likely have even more extensive experience with a wider variety of biological kinds. Thus far, our results with rural children offer no evidence for an anthropocentric pattern, providing support for the view that children’s experience with a biological kind affects the strength of that kind as an inductive base (Atran et al., 2001;
Ross, et al., 2003; Medin & Waxman, in press). However, the performance of our urban Chicago preschool-aged children throw a fly in the ointment: they too have failed to show the anthropocentric pattern (Angorro, et al, 2005).
Why might this be the case? Perhaps this discrepancy in the performance of urban children in our work and Carey’s and Hatano and Inagaki’s reflects, at least in part, a methodological difference in experimental design. In these latter versions, information about a biological kind is gathered at one point in time, and children participate in the category-based induction task several days later. Carey’s version involved explicit teaching. Children were introduced to a base-object (either a human, dog, or bee) and taught extensively about a novel property of that object (e.g. “has an omentum”). In Inagaki and Hatano’s designs, there was no explicit teaching: they simply measured the cognitive consequences of intensive experience raising goldfish. We refer to procedures like these, in which training occurs prior to the category-based induction test, as “spontaneous generalization” tasks. In contrast, in our lab we have relied upon what we might call “prompted generalization”: We teach children that some biological entity (e.g., a human or a dog) has some biological property (e.g., an omentum) and then immediately observe children’s inductive inferences from this base to a variety of other kinds.
Investigating anthropocentrism in U.S. children from rural and urban communities To examine the consequences of this methodological difference, we designed a close conceptual replication of Carey’s procedure. On Day 1, children were introduced to a novel property and were taught that this was a property true of either humans, dogs, or bees. On Day 2, children were given an opportunity to project that property to a range of kinds in a spontaneous generalization task. To examine the role of experience, we worked with urban and rural children. If the discrepant results for young urban children reflect primarily methodological differences, then in the current task, then urban children should reveal an anthropocentric pattern, treating humans as a stronger inductive base than either of the non-human animals (dog, bee). Moreover, if experience with biological kinds shapes the trajectory of children’s biological reasoning, then rural children should not treat humans as a privileged inductive base.
We tested 203 4- to 7-year-old children from public schools in Shawano, Wisconsin (rural) and Chicago, Il (urban). Shawano County is replete with farmland, small forest plots, and lakes and rivers. Hunting, fishing, and water recreation are popular activities for adults and children. For children in Chicago, recruited from a racially and ethnically diverse public magnet school, direct contact with animals is generally more limited to visits to the zoo, caring for pets, and noticing the native animals (squirrels, pigeons) that live in urban areas. Table 1 shows the number of children in each community and age-group.
Children were interviewed individually in a quiet place in their school on two different days. On the first day, they were introduced to an object (either a human, a dog or a squirrel) and taught that a novel property (e.g., “…has an omentum”) applied to that kind of object. Children then answered a few questions about their pets and outdoor activities. A day or two later, children completed the category-based induction task, involving color photographs of 16 target items, including a person, dog, bear, aardvark, bee, fly, eagle, toucan, trout, angel fish, maple tree, dandelion, sun, rock, computer, and pencil. For each photograph, children were asked, “Do Xs have an omentum?”2 We computed each child’s tendency to generalize the novel property from their designated base to the targets.
The results, depicted in Figure 1, indicated that the anthropocentric pattern was apparent only in the youngest urban children, and only weakly so at that. In every other group, non-human animal bases (both dogs and bees) were on at least equal footing with humans in their inductive strength. Thus, in this ‘spontaneous’ version of the categorybased induction task, we more or less replicated Carey’s and Inagaki and Hatano’s anthropocentric pattern -- but only with our youngest urban children. We did not observe an anthropocentric pattern in any other population – including the young rural children.
This pattern of results suggests that although procedural differences may be instrumental in the expression of an anthropocentric pattern in young urban children, a thorough developmental account will need to go beyond task and age considerations alone. Perhaps most importantly, the current results underscore the importance of considering the role of experience in the development of core biological concepts and reasoning. This focus on experience brings us full circle back to the insights provided by Hatano (Hatano & Inagaki, 1994; Inagaki & Hatano, 2002).
(Re)conceptualizing ‘experience’ and development.
First, how can we best conceptualize the role of experience in children’s biological reasoning? Perhaps children across the world’s communities undergo a universal developmental trajectory. Perhaps as Carey suggested, this trajectory involves radical conceptual change, in which a new domain (folkbiology) emerges from a previously-established domain (folkpsychology). In her view, it is this developmentallyprior domain of folkpsychology that undergirds the anthropocentric pattern, in which humans are seen as the prototype. On this view, perhaps rural children fail to show an anthropocentric pattern simply because their richer experience with the natural world permits them to move more rapidly along this universal trajectory. If this is correct, then we should find evidence of an anthropocentric pattern in rural children, but at a stillyounger age. We are currently pursuing this possibility.
There is, however, an alternative interpretation. Perhaps there is a universal developmental trajectory, but one that does not necessarily include an early anthropocentric period. Perhaps the anthropocentric pattern is itself culturally-inflected, reflecting urban children’s sensitivity to an anthropocentric cultural model that is passed along within the discourse of their communities.
Parental input to young children. This interpretation becomes plausible when considered in conjunction with recent evidence concerning parental input to young children. First, in their analysis of conversations between mothers and their 2-year-old
children, Gelman, Chesnick, and Waxman (2005) observed an intriguing phenomenon: