McKay & Dennett (M&D) provide a useful analysis of the evolution of misbelief, making a number of important distinctions, including one between misbeliefs that are tolerable byproducts of evolved psychological adaptations and those that would have been adaptive in and of themselves. A reasonable primary hypothesis is that selection has shaped the human mind to form true beliefs about the world. The ultimate criterion of evolutionary selection, as M&D rightly point out, is reproductive success, not the accurate detection or preservation of truth. We, and others, have argued that selection has favored psychological adaptations that do not always maximize truthful beliefs; these adaptations instead can result in misbeliefs (e.g., Haselton & Buss Reference Haselton and Buss2000; Haselton & Nettle Reference Haselton and Nettle2006).
Humans appear to possess cognitive biases which lead to systematic misbeliefs and require scientific explanation. These include the positive illusions that compel us to have a rosy outlook on the future (Taylor & Brown Reference Taylor and Brown1988), sex-linked biases such as men's tendency to overestimate women's sexual interest (e.g., Abbey Reference Abbey1982), and perceptual biases such as auditory looming, the tendency to overestimate the proximity to self of approaching objects compared to receding objects that are in fact equally distant (Neuhoff 2001). We articulated error management theory (EMT; Haselton & Buss Reference Haselton and Buss2000) as a theory to explain how evolution could lead to adaptive biases, some of which entail misbeliefs. Many problems of judgment under conditions of uncertainty can be framed as having two possible errors – false positive and false negative errors. According to EMT, in forming judgments under uncertainty, if there were recurrent asymmetries in the costs of these errors over evolutionary history, selection should produce a system that errs in the less costly direction. For example, for men estimating a woman's sexual interest, we hypothesized that the reproductively more costly error would have been to underestimate her interest and miss a reproductive opportunity. Thus, EMT predicts that men possess an adaptive bias toward overestimating women's sexual interest.
M&D affirm the logic of EMT, but argue that selection can solve adaptive problems of the sort explained by EMT in ways other than creating misbeliefs. They argue that humans do not need to possess biased beliefs if biased actions can accomplish the same ends while preserving true beliefs. We agree entirely with this point. It is possible, for example, that selection could design an adaptation in which men acted as if a larger number of women were sexually interested in them than actually were, in order for them not to miss a potential sexual opportunity, while not truly believing that those women are sexually interested. Similarly, it might be possible for selection to fashion an adaptation in which people act as though more people harbor homicidal intent than they actually do, in order to avoid the costly cases in which people actually do harbor such thoughts, without actually believing that those individual do harbor homicidal intent.
Just because selection can solve these adaptive problems without misbelief does not mean that selection has solved these problems without misbelief. The argument that selection could craft an adaptation for thermoregulation other than sweat glands (e.g., dogs thermoregulate through evaporation from a protruding tongue) is not an argument that selection has not fashioned sweat glands in humans.
Ultimately, the question of whether misbeliefs are part of the design of EMT biases is an open issue that must be decided on a case-by-case basis with empirical research. However, we suggest that there are no compelling reasons to discount the possibility that misbeliefs, including functional misbeliefs, are part of the evolved design of EMT biases. Consider the male sexual overperception bias. A misbelief that a woman is sexually interested could facilitate access to sexual opportunities in at least three ways. First, it could provide the motivational impetus for courtship behavior. Second, it could allay a man's anxiety about being rejected, eliminating a common cognitive barrier to initiating courtship (Kugeares Reference Kugeares2002). If it turns out that his belief was indeed incorrect, it is not terribly costly for him to revise his beliefs about a particular woman after being rebuffed (e.g., “I thought she was sending me sexual signals, but it turns out I was wrong”). Third, a man's misbelief, by motivating attraction tactics or elevating confidence, could transform a woman who is initially sexually uninterested in him into one who is sexually interested – an outcome showing that the initial misbelief itself can sometimes provide functional benefits. Hence, the EMT-generated misbelief can, in principle, solve the adaptive problem of maximizing sexual opportunities more effectively than an adaptation lacking the misbelief design feature.
Although we advanced the theory to explain cognitive biases, the core logic of EMT is neutral in predicting where in the perception-belief-action chain selection will shape a bias. All that is required is that, ultimately, humans behave so that they minimize the more costly of the two errors in question, even if this cost minimization ends up producing a larger number of overall errors. To discover where in this chain a bias exists must be empirically adjudicated. On the basis of the existing empirical evidence, however, we suggest that biasing action is unlikely to be the sole outcome of selection in which there has been recurrent cost asymmetries associated with errors.
M&D's analysis will stimulate empirical research about particular EMT biases. Some biases may be instances of biased action without involving misbelief. Others may entail misbeliefs. A subset of these may be cases in which the misbelief is not simply a tolerable byproduct of an adaptively biased cognitive system but is itself adaptive. M&D make a compelling argument that positive illusions qualify as adaptive misbeliefs because they positively affect an individual's fitness by motivating striving for favorable outcomes. We suggest that that some EMT biases, such as the male sexual overperception bias, also can motivate adaptive action through misbeliefs by providing motivational impetus for action, overcoming inhibitions associated with action, and transforming the psychological states of others in ways beneficial to the holder of misbeliefs.
