Evolutionary perspectives on adaptive misbelief are not new (Byrne & Kurland Reference Byrne and Kurland2001; Haselton & Buss Reference Haselton and Buss2000; Trivers Reference Trivers1985; Reference Trivers, LeCroy and Moller2000; see also, Gigerenzer & Brighton Reference Gigerenzer and Brighton2009; Gigerenzer et al. Reference Gigerenzer, Todd, the and Research1999). What is new, however, is the precise analysis of the conditions of adaptive misbelief presented in the target article. Unfortunately, the target article's impact is limited by its reliance on controversial “better than average” effects and the relatively non-restrictive nature of the proposed sufficient conditions. Here, we briefly document these concerns and discuss some relevant phenomena in individual differences research. Ultimately, we argue that adaptive misbeliefs are likely much more widespread than is implied.

McKay & Dennett (M&D) suggest that adaptive misbeliefs are reflected in better-than-average and similar overconfidence type effects. However, there are concerns about the stability, universality, and reality of such illusions (Gigerenzer et al. Reference Gigerenzer, Hertwig, Hoffrage, Sedlmeier, Plott and Smith2008; Larrick et al. Reference Larrick, Burson and Soll2007; Moore & Healy Reference Moore and Healy2008; see also, Juslin & Olsson Reference Juslin and Olsson1997; Juslin et al. Reference Juslin, Winman and Olsson2000). To illustrate, when most people report that they are better than average drivers they are not wrong or biased. Instead, data indicate that only a very small number of people are responsible for the vast majority of motor vehicle accidents. Thus, driving ability is not normally distributed and so most people are technically correct when they believe they are better than average drivers. This kind of example is not uncommon. Better-than-average and overconfidence type effects are often complicated by statistical artifacts and non-ecological task contexts (Gigerenzer et al. Reference Gigerenzer, Todd, the and Research1999; Krueger & Mueller Reference Krueger and Mueller2002).

More problematic than the quality of the proposed evidence, however, are the following set of alleged sufficient conditions offered for systematic adaptive misbelief: (a) the belief is the result of “design” (where design is appropriately defined); (b) the belief misrepresents information to the possessor of the belief; (c) the misrepresentation of information is beneficial to the possessor of the belief (sect. 2, para. 5); and (d) the misbelief is systematic (sect. 4). If these conditions are only sufficient, then in contrast to what is implied, M&D have not captured a unique way in which misbelief can be adaptive. Rather, they have only pointed out one of many possible ways. This worry results in interpretative issues with M&D's general argument.

Assuming that many beliefs could satisfy (a), it is unclear what degree of misrepresentation or benefit is sufficient for a belief to satisfy conditions (b) and (c). According to M&D, a misbelief is one that is “false,” or “to some degree departs from actuality,” or “to some extent wide of the mark” (sect. 1, para. 1). These comments indicate that any belief that departs from reality in any way satisfies condition (b). The only way a belief could fail to satisfy (b) is if the content of the belief does not even in part misrepresent reality. If that is correct, then it is likely many (if not most) of our beliefs satisfy condition (b) (something M&D realize, sect. 1). It is also unclear how and in what ways the misbelief must be beneficial in order to satisfy condition (c). We can grant that positive illusions may be adaptively beneficial to the possessors of those beliefs in a number of profoundly interesting ways. But again, it is a very modest and easily satisfied condition if the misbelief only needs to provide some adaptive benefit to the possessor.

Condition (d) also is satisfiable in a number of ways. M&D appear to endorse a “one size fits all” model of misbelief that would be adaptive for whoever holds such misbeliefs (condition [d]). But there is more than one way that misbeliefs can be systematic. For instance, there can be misbeliefs that are systematically related to stable individual differences among groups of people. There is evidence that personality traits (e.g., the Big Five) are related to individual differences in beliefs about the nature of the world (Langston & Sykes Reference Langston and Sykes1997) and to fundamental philosophical beliefs regarding moral objectivism, compatibilism, and intentional action (Cokely & Feltz Reference Cokely and Feltz2009a; Reference Cokely and Feltz2009b; Feltz & Cokely Reference Feltz, Cokely, Love, McRae and Sloutsky2008; Reference Feltz and Cokely2009).

To take just one example, those who are neurotic are likely to think that the world is dangerous. Those who are not neurotic tend not to have this belief (particularly so for extraverts and those who are agreeable) (Langston & Sykes Reference Langston and Sykes1997, p. 154). On the face of it, these are contrary beliefs. So, either neurotic individuals have a misbelief or non-neurotic individuals have a misbelief – and perhaps both have misbeliefs. Evidence also indicates that some personality types are related to beneficial life outcomes and that personality traits are partially genetic in origin (Bouchard Reference Bouchard1994). Hence, it appears that at least some systematic individual differences in beliefs are likely to be excellent candidates to satisfy (a)–(d).

Given that it is likely that quite a few of our beliefs satisfy (a)–(d), M&D underestimate the number of misbeliefs that are adaptive. Moreover, it may be that individual differences in misbeliefs are adaptive for both the specific misbelieving actor and for other non-misbelieving members of their group. That is, differences in belief might enable more effective allocation of limited resources in groups, benefiting both accurate and misbelievers alike (Wolf et al. Reference Wolf, van Doorn, Leimar and Weissing2007). In summary, we argue that although the proposed parameters offered by M&D do provide substantive increases to theoretical specification, they do not support bold claims such as “the exchange rate with truth is likely to be fair in most circumstances” (sect. 15, final para.). It is possible that adaptive misbeliefs are in the minority; however, this has yet to be adequately evaluated and does not follow from the evidence or argument provided. In contrast, we suspect that there are many relatively unexplored opportunities for theoretical and translational progress at these frontiers (e.g., the modeling of decisions and design of better choice environments; Johnson & Goldstein Reference Johnson and Goldstein2003; Todd & Gigerenzer Reference Todd and Gigerenzer2007; Weber & Johnson Reference Weber and Johnson2009).