ABSTRACT

The thesis of this book has been that patterns of individual differences, if interpreted in creative ways, can help to shed light on the various reasons why descriptive accounts of human behavior depart from normative models. This thesis has been demonstrated with empirical results that ranged widely over the tasks in the heuristics and biases literature. The data patterns have been varied. In this chapter, I intend to step back from the minutiae of individual tasks and draw more general conclusions. The chapter begins with a brief consideration of performance errors and computational limitations. Considered next are the implications of algorithmic-level limitations for the problem of applying the right normative model to performance. Finally, it is argued that, taken collectively, the results summarized in this volume warrant another broad conclusion. Specifically, after ruling out algorithmic limitations (which makes the normative model not prescriptive for some individuals) there were reliable deviations from normative responding that tended to covary across tasks. This reliable covariation, as well as relationships involving thinking dispositions, suggest that debates about rational thought might benefit from a reversal of figure and ground. Traditionally these debates have foregrounded issues of competence and backgrounded issues of performance. The results discussed in this volume suggest an inversion of these priorities-an inversion that has some unexpected consequences for the debate between the Panglossians, Apologists, and Meliorists. The chapter concludes with some caveats/qualifications and with some speculations about the pretheoretical views that determine study design and data interpretation in these debates.