Does Ecological Rationality Solve the rationality debate (evolutionarily)?

Yuichi Amitani
University of British Columbia

     Full text: Not available
     Last modified: February 15, 2005
     Presentation date: 07/14/2005 2:00 PM in ROZH 107
     (View Schedule)

Abstract
It is well known that ordinary (or even learned) people often commit basic logical or probabilistic fallacies in psychological experiments like selection task. Some psychologists like Kahneman and Tversky think that those results show human thinking generally does not comply with logical and probabilistic laws and therefore is irrational.
Among psychologists dealing with those experiments and related debates on human rationality, Gerd Gigerenzer has made a unique contribution to the debate. He and his research group reinterpreted those psychological experiments in terms of theories of probability and evolution. Other evolutionary psychologists examined his hypothesis in more comprehensive contexts (e.g., Tooby and Cosmides 1996). From a series of experiments in which subjects solve Bayesian inference quizzes much better under frequentist representation, they argue that it is because a frequentist notion and representation of probability is adaptation.
This presentation is a critical examination from an evolutionary standpoint on Gigerenzer’s (and his followers') attempts to solve the debate. My examination is based on five points. Firstly, he does not seem to understand a principle of evolutionary biology that if there are no alternative traits, there cannot be any selection. Second, he does not seem to successfully distinguish a certain mode of representation of probability from an interpretation of probability. Thirdly, his scenario on how human beings acquired what he calls ecological rationality does not appear to have enough evidence to avoid the charge of a "just so story". Fourthly, Gigerenzer himself proposes an explanation of the performance on the Bayesian Inference which appeals only to computational simplicity. It is just because frequentist representation is computationally simpler that subjects perform better under this representation. This explanation does not need any specific adaptive mechanism which Gigerenzer assumes in another place. Finally, even though the methods of evolutionary psychologists are very productive in that they produce interesting empirical hypotheses which are confirmed as in this case, this defense is equally true of other theories which are rejected as scientific theories such as creationism.