The Pragmatics of Explanation

The Pragmatics of Explanation Seth Chin-Parker (chinparkers@denison.edu) Department of Psychology, Denison University Granville, OH 43023 USA Alexandra Bradner (bradnera@denison.edu) Department of Philosophy, Denison University Granville, OH 43023 USA of focusing our attention on one of these approaches exclusively, we instead explore a fourth approach, the pragmatic approach, as proposed by Van Fraassen (1988). This approach subsumes the other three, for explanation is seen as a selection process (Bradner, 2005) – at any time there are multiple candidate explanations that could be referenced, but we choose the most appropriate for a given situation by working through two stages. First, the topic of the explanandum is determined by identifying the contrast class for the explanation, defining the explanandum. For instance, the answer to “Why is she shivering – as opposed to why is he shivering?” is different than the answer to “Why is she shivering – as opposed to why is she dancing?” After the contrast class has been identified, relevance relations motivate the content of the explanation. The relevance relations for “Why is she shivering?” would possibly include knowledge of human physiology or knowledge of the temperature outside. Van Fraassen proposes that background knowledge does a good deal of the work in constraining the possible explanations at each of these stages. Earlier work in artificial intelligence (e.g. Leake, 1991) has explored the implications of this view, but there has not been empirical study of how background knowledge affects the selection of the contrast class or the relevance relations. Recently, there has been active study of what parameters influence when causal explanations (e.g. Glymour, 1998) and functional explanations (e.g. Chaigneau, Barsalou, & Sloman, 2004; Lombrozo & Carey, 2006) are warranted. In the current study, we extend this work by asking participants to provide an explanation for an explanandum when either a causal or functional explanation is licensed. By manipulating the prior experiences of the participants, we are then able to assess whether the pragmatic theory is correct in predicting that the participants’ explanations are determined by available background. Although there are different proposals as to how people come to appreciate the presence of causal relations (see Shanks, Holyoak, & Medin, 1996, for one overview), there is consensus that cause is often invoked during explanation (Keil, 2006). We expected that participants would readily generate causal explanations for events that were physically and temporally contiguous and that followed basic beliefs associated with naive physics – e.g. the movement of object A as object B moves into a space previously held by object A can be explained by object B causing object A to move. Abstract The pragmatic theory of explanation (Van Fraassen, 1988) proposes that background knowledge constrains the explanatory process. Although this is a reasonable hypothesis and research has shown the importance of background knowledge when evaluating explanations, there has been no empirical study of how the background constrains the generation of explanations. In our study, participants viewed one of two sets of preliminary movie clips of some novel items engaged in a series of actions and then all were asked to explain the same final clip. Between conditions, we varied whether the events in the preliminary clips completed a system. In the systematic condition, a greater proportion of functional explanations were generated for the final clip compared to the non-systematic condition. Interestingly, despite the difference in the types of explanations generated, the participants showed high agreement in the evaluation of explanations provided by the experimenters. Keywords: explanation; cognition; function; pragmatics. Explanation holds a special place in the cognitive sciences (Lombrozo, 2006). Philosophers, psychologists, members of the artificial intelligence community, and computer scientists have all studied various facets of explanation. The allure to cognitive scientists is obvious: An explanation embodies an individual’s ability to express the understanding that she has acquired for events that occur in the world. If she is able to explain why peaches are fuzzy or why a person shivers, it indicates that she has found meaning in the relationship between those simple facts (a peach is covered in fuzz) or events (the person is shivering) and her more general knowledge of the world. The explanation goes beyond describing the event or recognizing associations, e.g. it is cold and that person is shivering, to providing connections between the explanandum (what is to be explained) and the explanans (what does the explaining). Philosophical inquiry into the nature of explanation has a long and rich history. At the risk of over-simplifying this body of work, we propose that there have been three main approaches to what constitutes an explanation. Much work has illuminated our understanding of causal explanation (e.g. Salmon, 1998), functional or teleological explanation (e.g. Cummins, 1975; Wright, 1973), and the unification approach to explanation (e.g. Kitcher, 1981). In the current study, we are interested in the cognitive underpinnings of explanation as opposed to the explication of an explanation itself, so instead