Spatial Intelligence and Learning Center
NSF Science of Learning Center
Analogy is an important strategy for reasoning and problem solving. It is also a powerful learning process. The process of comparison involves a structure-mapping, in which the two representations are aligned and further inferences are projected from one to the other (Gentner, 1983). Thus, one way that analogy facilitates learning is via projection of knowledge from a well-understood situation to another that is less familiar or more abstract. Another important form of analogical learning occurs when two situations are compared and their common relational structure is highlighted. There is abundant evidence that such highlighting renders the common structure more available for subsequent processing, including transfer to new contexts (Gentner, Loewenstein & Thompson, 2003; Gick & Holyoak, 1983). Comparison processing has been found to facilitate spatial learning (Kotovsky & Gentner, 1986; Loewenstein & Gentner, 2001; Vosmik & Presson, 2004), learning of noun (Gentner & Namy, 1999) and verb categories (Childers, 2005); mathematical insight (Zur, 2005) and scientific reasoning (Chen & Klahr, 1999; Kurtz et al, 2001). Benefits of comparison are apparent throughout development, from infancy (Baillargeon, 1991; Casasola, in preparation; Oakes & Ribar, in press) through adulthood (Catrambone & Holyoak, 1989; Gentner et al., 2003; Gick & Holyoak, 1983; Kurtz et al., 2001).
In SILC we will use analogy and comparison processing both as a source of insight into how children learn spatial knowledge and as a tool for facilitating children's spatial learning. For example, our work on maps in Strand 2 is informed by findings from analogical development that, early in learning, children's focus is on individual objects. Structure-mapping theory also suggests ways to promote the encoding of spatial relations. For example, we have found that children who carry out a comparison between two spatial models are then better able to map from one of those models to a new model, relative to a control group (Loewenstein & Gentner, 2001). This benefit was found even when the two initial models were highly similar to one another. This is important because young children often have difficulty with distant analogies. The finding that comparing highly similar (easily aligned) scenes can be informative offers a way to promote spatial learning even in very young children.
We will extend these methods to investigate the role of comparison in learning the relation between a map and the territory it represents. For example, can comparing two highly similar maps or scenes facilitate understanding the relationship between the map and the scene? We will also explore the use of progressive alignment--in which children begin with highly similar (easily alignable) comparisons and are gradually moved to dissimilar "far" comparisons--in teaching children the relation between a map and a spatial scene. We will also explore interactions between spatial analogies and symbolic representations. We have found that naming a spatial relation can improve children's ability to use that relation in an analogical mapping task (Gentner & Rattermann, 1981; Loewenstein & Gentner, in press). In addition to language interactions, we will also investigate how other kinds of external representations such as maps and spatial diagrams influence analogical processing. Finally, in all these arenas we will investigate the stability over delay and transfer distance, a crucial issue for effective education.
