Title Young children's analogical reasoning across cultures: Similarities and differences

A cross-cultural comparison between U.S. and Hong Kong preschoolers examined factors responsible for young children's analogical reasoning errors. On a scene analogy task, both groups had adequate prerequisite knowledge of the key relations, were the same age, and showed similar baseline performance, yet Chinese children outperformed U.S. children on more relationally complex problems. Children from both groups were highly susceptible to choosing a perceptual or semantic distractor during reasoning when one was present. Taken together, these similarities and differences suggest that 1) cultural differences can facilitate better knowledge representations allowing for more efficient processing of relationally complex problems, and 2) that inhibitory control is an important factor in explaining the development of children’s analogical reasoning. Analogical Reasoning across Cultures 3 Young Children's Analogical Reasoning across Cultures: Similarities and Differences Analogical reasoning is a powerful mechanism in children's cognitive development. In an analogy, successful reasoners construct correspondences between two systems of relations to solve a problem (Gentner, 1983). For example, knowledge of using a stool to reach a toy can be used to figure out that a ladder could help reach a cookie jar. This skill enables children to draw on prior knowledge representations to make sense of new contexts and to build expertise by comparing and contrasting representations (see Goswami, 1992). Infants show analogical thinking and problem solving by a year and a half (Chen, Sanchez & Campbell, 1997); however, they do not approximate adults' levels of competence until adolescence (Halford, 1993; Richland, Morrison & Holyoak, 2006). Errors in young children's analogical reasoning are characterized by difficulty ignoring irrelevant object properties, like the difference in appearance between a ladder and a stool (e.g., Rattermann & Gentner, 1991; Richland et al, 2006). Also, young children have been shown to struggle with reasoning about multiple relationships at once (e.g., Halford, 1993; Richland et al, 2006). Several explanations have been posed to explain the development of analogy in children. The relational primacy theory holds that adequate knowledge about key relations is the main prerequisite for analogical reasoning (e.g., understanding "reaching" to solve the above example) (Goswami, 1992). Gentner and Rattermann (1991) argue, in addition, that until children have the adequate knowledge, they will not only fail to reason analogically, they will focus their answers on object properties and appearance (Relational Shift Hypothesis). This relational shift is not tied to the age of a child, but rather to knowledge, so that even an adult who is not knowledgeable Analogical Reasoning across Cultures 4 about an analogy task may tend to make mappings based on object properties, while s/he will shift to more relational mappings once adequate knowledge is acquired. Two aspects of executive resources have also been suggested as important for the development of analogical reasoning. Halford (1993) has suggested that the inability to process multiple relations in analogies may be due to limits in children’s working memory. Halford and colleagues found that children's developmental differences across a variety of tasks could be calculated as a function of the relational complexity, or the number of relationships that must be held in working memory simultaneously. Based on this model, children should be able to reliably solve tasks with a single level of relational complexity before approximately five years, after which two relations should be attainable. Likewise, we have previously posited that limits on inhibitory control may explain why young children solving analogy tasks sometimes still map correspondences based on object properties and general appearance correspondences in spite of understanding the relations and the analogy task (Richland, et al., 2006). The current paper used a cross-cultural approach to explore interactions among these factors and to investigate their explanatory power in a broader sample. Much of the prior research has been conducted with U.S. and Australian children, but adult and developmental research suggest that cultural experiences may impact the development of analogical reasoning. Relational Reasoning Across Cultures Cultural experiences may influence relational reasoning in several ways. Knowledge of cultural content may impact prerequisite knowledge of relations, and consequently, influence analogical reasoning on problem-solving tasks that rely on that cultural knowledge (Chen, Mo, & Honomichl, 2004). Analogical Reasoning across Cultures 5 In addition, culture may be related to normative patterns of relational reasoning and analogy production (Richland, Zur & Holyoak, 2007). Normative patterns for drawing relational inferences during problem solving can vary across cultures when content knowledge is comparable (see D'Andrade, Nisbett 2003). Chinese and Japanese reasoners may attend relatively more to relational patterns in visual representations and problems, while U.S. reasoners may attend relatively preferentially to object-based information (see Hansen, 1983; Nisbett, 2003). Such cultural variations have been demonstrated in visual scene interpretations, which are often used in analogical reasoning tasks. In one cross-cultural study, Chinese outperformed U.S. undergraduates in assessing covariation in presentations of arbitrary objects (e.g., judging whether schematic drawings of a coin and a lightbulb had been shown on the screen together). The Chinese students also showed greater attention to relations between the figures and background than U.S. participants, who demonstrated more attention to focal objects, or figure independence (Ji, Peng & Nisbett, 2000). Analyses of children's everyday experiences suggest these reasoning and attention patterns are part of children's socialized experience with relational inputs in schools (Richland, Zur & Holyoak, 2007) and at home with caregivers. Asian caregivers seem particularly interested in directing infants’ attention to how entities—human and otherwise—interact and relate to one another, including using action oriented language and referential verbs (e.g., swimming, ate, going to drive) (e.g., Korean: Au, Dapretto, & Song, 1994; Gopnik, Choi & Baumberger, 1996; Japanese: Fernald & Morikawa, 1993; Ogura, Dale, Yamashita, Murase, 2006; Chinese: (Mandarin) Tardif, et.al, 1997; Tardif, et.al, 1999; (Cantonese) Leung, 1998). The play of English speaking caretakers, who served as a comparison group in many of these Analogical Reasoning across Cultures 6 studies, is relatively object focused, including using more naming and non-referential verbs (e.g., 'looks like,' 'lookit, 'watch') (e.g., Goldfield, 1993; Gopnik, et al, 1996). The current paper uses a scene analogy task (Richland, et al., 2006) to explore the hypothesis that Chinese preschool children approach analogies differently than U.S. children based on their relatively greater experiences with relations. We specifically hypothesized that Chinese children might be able to processes analogies more efficiently by constructing higherlevel relational representations, thus making better use of comparable working-memory resources. The task used simple, common relations and a counterbalancing design that held necessary content knowledge constant across conditions, so any differences in performance should not be attributable to variations in prerequisite knowledge. We also sought to test our hypothesis (Richland, et al., 2006) that maturational limitations in inhibitory control explain children's low performance on analogies that include a strong object-similarity distractor. We predicted a common pattern across countries showing relatively lower performance on analogies with a distractor versus no-distractor despite comparable prerequisite knowledge. The relational shift hypothesis (Rattermann & Gentner, 1991), by contrast, would predict that children’s susceptibility to distraction would decrease as their relational knowledge increases. Thus Chinese children should be less susceptible to distraction