Conversational processes and causal explanation.

Abstract

Causal explanation takes place in and takes the form of conversation. Explanations are selected by questions and are thus governed by general rules of discourse. A conversational model of causal explanation is introduced that explicates social aspects of the explanation process by postulating that good explanations must be relevant to the focus of a why question, as well as being true. The notion of explanatory relevance enables an integration of the major models of the attribution process by showing that they use the same counterfactual logic but address different causal questions. The conversational perspective suggests a reinterpretation of many attributional biases, and also highlights the role of interpersonal goals in generating implicit questions, which in turn constrain explanations. Finally, the relevance of the conversational perspective for research on causal networks, the social context of explanation, and intrapsychic explanation is noted.Causal explanation is first and foremost a form of social interaction. One speaks of giving causal explanations, but not attributions, perceptions, comprehensions, categorizations, or memories. The verb to explain is a three-place predicate: Someone explains something to someone. Causal explanation takes the form of conversation and is thus subject to the rules of conversation.Causal explanation is therefore different from causal attribution, which does not involve an interpersonal exchange. In its wider sense, attribution describes a process whereby one ascribes some phenomenon to its origin. Thus, one attributes effects to causes, actions to intentions, and sayings to their authors. This sense of attribution is close to that which Heider (1958) had in mind when writing of how the sensation of light reflecting ofFa surface may be attributed to its distal source, the sun, or how an exam performance may be attributed to its source in a student's ability, effort, or luck. Thus, attributing something to someone is not the same as explaining something to that person.The difference between this sense of attribution and conversational explanation is nicely drawn out by the experience of U.S. Senator Joseph Biden. During his campaign for the 1988 Democratic Party nomination for president, Biden gave an impassioned campaign speech about social inequality that faithThis research was conducted while I was supported by the Cognitive Science committee of the University of Illinois at Urbana-Champaign. Portions of this article were read at the "Beyond Assertion" conference, Cleves, West Germany, September 3-5, 1987, and at the meeting of the Society for Experimental Social Psychology, Madison, Wisconsin, October 20-22, 1988.I would like to thank Ulf Bockenholt; Eirian Davies; Miles Hewstone; Robin Hogarth; Douglas Medin; Ben Slugoski; Richard Smith; the action editor, Nancy Cantor; and two anonymous referees for helpful comments on earlier drafts of this article and Curt Haugeveldt and Alan Strathman for help in preparation of the figures.Correspondence concerning this article...

Covariation in natural causal induction.

“…The causal attribution literature in social psychology has largely measured performance against an apparently different normative standard: namely, variations of Kelley's (1967Kelley's ( ,1973 influential proposal that people are "intuitive scientists" who use a mechanism of causal induction analogous to the analysis of variance (ANOVA) (Cheng & Novick, 1990a;Forsterling, 1989;Hewstone & Jaspars, 1987;Hilton, 1988Hilton, ,1990Jaspars, 1983;Orvis, Cunningham, & Kelley, 1975;Pruitt & Insko, 1980). This literature has, until recently, presented a chaotic picture of causal induction not unlike that in cognitive psychology: Causal induction sometimes conforms to the normative standard but often deviates from it.…”

“…A deviation from normative covariation that has received considerable attention in philosophy and related fields concerns the intuitive distinction people make between causes and enabling conditions (e.g., Einhorn & Hogarth, 1986;Hart & Honor6,1959/1985Hesslow, 1983Hesslow, ,1988Hilton, 1990;Kahneman & Miller, 1986;Mackie, 1965Mackie, ,1974Mill, 1843Mill, /1973Taylor, 1983;White, 1965). In response to the question "What caused the airplane to crash?"…”

“…Many models proposed in the literature on causes versus enabling conditions are deterministic (Hesslow, 1983(Hesslow, , 1988Hilton, 1990;Mackie, 1965Mackie, ,1974. It is not obvious how these models can be generalized to account for conjunctive probabilistic causes, even if the probabilistic generalization for simple causes is straightforward.…”

“…Thus, whereas a cause is always a condition assumed to be unknown to the hypothetical inquirer (otherwise there would be no reason for asking), an enabling condition is typically a condition assumed to be already known to the inquirer (Hilton, 1990;Mill, 1843Mill, /1973Turnbull, 1986;Turnbull &Slugoski, 1988;cf. Grice, 1975).…”

The abstinence violation effect in child molesters

“…It is now in the process of providing the social-cognitive underpinning of cognitive-behavioural therapy (Bradbury & Fincham, 1990;Fosterling, 1986;Hilton, 1990;Iacobucci & McGill, 1991;Kenardy, Evans, & Oei, 1990;Weiner, 1986).…”