Published in: Thinking & Reasoning, 1 (4), 1995, 324–333. www.tandf.co.uk/journals/pp/13546783.html Reprint 14/1997 by the Max Planck Institute for Psychological Research. The Taming of Content: Some Thoughts About Domains and Modules Gerd Gigerenzer Max Planck Institute for Psychological Research, Munich, Germany Commentary on “Pragmatic Reasoning With a Point of View” by Keith J. Holyoak and Patricia W. Cheng Peter Wason invented the Selection Task in 1966. Thirty years and many, many experiments later, two results are evident for me. First, the view that sound reasoning can be reduced to prop- ositional logic (or first order logic) is myopic. Human thought operates in more dimensions than entailment and contradiction (Strawson, 1952). We need to work out how the mind infers the meaning of a conditional statement from the content of the Ps and Qs and the social context in which the statement is uttered, rather than exclaiming “Cognitive illusion! Hurrah! Error!” when- ever human reasoning cannot be reduced to propositional logic. Second, the hope that these “errors” or their flip-side, the “facilitation” of logical reasoning, would be the royal road to dis- covering the laws of human reasoning did not materialise. This hope was fueled by the (mislead- ing) analogy with visual illusions (Gigerenzer, 1991). What were seen as “errors” were attributed to deeper cognitive deficits such as confirmation biases, to crude heuristics such as availability, or simply to “the subjects’ incorrigible conviction that they are right when they are, in fact, wrong” (Wason, 1983, p. 356). Unfortunately, this programme of research has brought little progress on the theoretical front. In the last decade, a few adventurous researchers freed themselves from the straitjacket of propositional logic and looked at dimensions of reasoning beyond entailment and contradiction. The content rather than the logical structure of the conditional statement moved into the fore- ground. As early as 1972, Wason and Johnson-Laird pointed out (p. 245) that, contrary to their expectation, “content is crucial” to reasoning and that “any general theory of human reasoning must include an important semantic component.” But neither they nor the others who studied human reasoning at that time found a way to include semantics in a theory of reasoning. Instead, content remained but a decorative element in reasoning problems—which either might “confuse” subjects or “facilitate” their logical reasoning. In the 1980s, Patricia Cheng (Cheng & Holyoak, 1985; Cheng, Holyoak, Nisbett, & Oliver, 1986) and Leda Cosmides (1989; Cosmides & Tooby, 1989, 1992) dared to introduce the content of the Ps and Qs into their theories. Cheng and her collaborators postulated a set of pragmatic reasoning schemas, such as permission and obligation schemas. Cosmides and Tooby postulated Darwinian algorithms, with social contracts and threats as examples. Content, the terra incognita where no established researcher on reasoning dared or cared to venture, became a legitimate topic of study. Cheng and her colleagues made a significant first proposal: permission and obligation schemas, each defined by four production rules. This was an important move away from propositional logic, but the production rules still resembled the four rules of the truth table, with Ps and Qs replaced by “actions” and “preconditions.” Cosmides and Tooby made a bolder and theoretically richer leap,