Inspection Times, Relevance, and Reasoning: A Reply to Roberts

Roberts (this issue) has criticized the methodology of a particular study (Evans, 1996) offered in support of the ``heuristic± analytic’ ’ (H± A) account of reasoning on the Wason selection task. It is important to realize that the H± A account (originally published by Evans, 1984) has been applied much more widely than on the selection task (see especially, Evans, 1989), and that even on that task it is supported by data other than those of Evans (1996), as we shall see later. T he theory was a development of ideas described initially by Wason and Evans (1975) as a ``dual process theory’ ’ of reasoning. In recently developing the ideas further, Evans and Over (1996) have suggested a return to this earlier title, but talk now of ``implicit’ ’ and `̀ explicit’ ’ processes. Evans (1984, 1989) used the terms ``heuristic’ ’ and ``analytic’ ’ processes. T he heuristic processes were described as preconscious and as resulting in the representation of selected problem information, together with associated knowledge retrieved from longterm memory as psychologically relevant. Analytic (logical) reasoning processes could only be applied to such selected representations. T hus Evans (1989) offered a general account of reasoning biases as caused by the failure or heuristic processes, on occasion, to select the logically relevant information for further processing. In the special case of the Wason selection task, it was argued, people simply choose those cards that appear relevant, so that their choices are entirely heuristically determined. T his position has been somewhat modi® ed in recent accounts in the light of verbal protocol analysis (Evans, 1995, 1996, Evans & Over, 1996). I now argue that analytic reasoning does occur on the selection task, but that it generally serves only to rationalize choices and not to alter them. It is important to realize that the H± A account does generally envisage that analytic reasoning will occur and affect behaviour on reasoning tasks and that the theory does not depend upon the Wason selection task being an exception. It simply appears that it is an exception, and I have as yet to see any evidence that choices on this task involve more than heuristicÐ or, if you like, intuitiveÐ judgements of the relevance of the cards. Hence, when I introduced the inspection time methodology (Evans, 1996), I predicted that people would think only about the cards they ended up choosing. By instructing participants to point the computer’ s mouse at images that represented the four cards, I was THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 1998, 51A (4), 811± 814