Abstract

Effects of Comparison and Explanation on Analogical Transfer Brian J. Edwards 1 ([email protected]), Joseph Jay Williams 2 ([email protected]), Dedre Gentner 1 ([email protected]), Tania Lombrozo 3 ([email protected]) Department of Psychology, Northwestern University, 2029 Sheridan Road – 102 Swift Hall, Evanston, IL 60208 USA Office of the Vice Provost for Online Learning and Graduate School of Education, Stanford University, Littlefield 253, 365 Lausen Street, Stanford, CA 94305 USA Department of Psychology, University of California, Berkeley, 3210 Tolman Hall, Berkeley, CA 94720 USA Abstract For example, Williams and Lombrozo (2010) found that participants who explained why a set of robots belonged to their respective categories were more likely than control participants to discover an abstract rule that could be used to accurately categorize all robots. The view that engaging in explanation can lead people to search for principles that apply to multiple cases – an implication of Williams and Lombrozo’s (2010) “subsumptive constraints” account – suggests synergy between comparison and explanation. Specifically, in order to determine whether a principle that applies to one case also applies to another case, comparing the two cases may be useful, or even necessary. Edwards, Williams, and Lombrozo (2013) investigated this possibility in a category learning task similar to that of Williams and Lombrozo (2010); participants who received explanation prompts reported doing significantly more comparison processing than participants who received control prompts. The findings that generating explanations can stimulate spontaneous comparison and that comparison may be a key mechanism by which generating explanations supports learning provide evidence for a relationship between comparison and explanation and raise the question of how these processes might work together to support learning. In one study, Gadgil, Nokes-Malach, and Chi (2012) found that comparing learner-generated and expert explanations helped participants detect errors in systems of relations between beliefs in addition to helping them notice errors in individual beliefs. More broadly, the Edwards et al. (2013) and Gadgil et al. (2012) studies suggest that combining comparison and explanation may yield even greater learning benefits than engaging in just one of these processes. In the present work, we investigate the relationship between comparison and explanation in analogical problem solving. We chose this domain for two reasons. First, analogical transfer is a hallmark of learning because it represents the ability to generalize knowledge from one’s past experience to situations with novel surface features but a shared underlying structure. Second, both comparison and explanation have been shown to play important roles in supporting analogical problem solving (Gick & Holyoak, 1983; Needham & Begg, 1991). In a seminal study, Gick and Holyoak (1983) found that students who compared two analogous stories were more likely to spontaneously transfer the solution principle to a Although comparison and explanation have typically been studied independently, recent work suggests connections between these processes. Three experiments investigated effects of comparison and explanation on analogical problem solving. In Experiment 1, explaining the solutions to two analogous stories increased spontaneous transfer to an analogical problem. In Experiment 2, explaining a single story promoted analogical transfer, but only after receiving a hint that may have facilitated comparison. In Experiment 3, irrelevant stories were interspersed among the two story analogs to block unprompted comparison; prompts to compare were effective, but prompts to explain were not. This pattern suggests that effects of explanation on analogical transfer may be greatest when combined with comparison. Keywords: Comparison; explanation; analogical transfer; problem solving; learning. Introduction Making comparisons and generating explanations can have robust effects on learning. The value of these processes in learning contexts, along with their centrality and ubiquity in everyday cognition, have inspired rich but largely separate literatures studying comparison and explanation (for reviews, see Gentner, 2010, on analogy and comparison; Lombrozo, 2012, on explanation). Comparison is the process of identifying similarities and differences between two cases. According to structure- mapping theory, comparison operates by a process of structural alignment based on finding common relational structure (Gentner, 1983; Gentner & Forbus, 2011; Gentner & Markman, 1997). As such, comparison is helpful for acquiring abstract relational schemas and for discovering deep relational commonalities between cases (e.g., Gick & Holyoak, 1983; Loewenstein, Thompson, & Gentner, 2003). Generating explanations (hereafter, “explanation”) consists of a number of related phenomena, including answering “why” and “how” questions. Explanation can support learning through a variety of mechanisms, including promoting deeper processing and helping learners detect gaps and inconsistencies in their knowledge (for a review, see Fonseca & Chi, 2011). Explanation can also affect learning by encouraging learners to seek broad patterns and unifying principles that apply to multiple cases (Williams & Lombrozo, 2010; Williams, Lombrozo, & Rehder, 2013).