McKay & Dennett (M&D) argue that misbeliefs are adaptive only “in certain rarefied contexts” (sect. 15, last para.). Only certain “positive illusions” about one's own capacities are adaptive. In the following, I argue that the authors seriously underestimate the prevalence and importance of adaptive self-directed misbeliefs in human populations. One of the most important tasks facing a human being is making oneself as interpretable as possible to one's fellows (Cash Reference Cash2008). Behaving consistently with self-interpretations afforded by one's culture, even if they are inaccurate, is an important means by which human beings accomplish this task. Religious beliefs constitute an important class of such culturally afforded self-interpretations.

It is surprising that M&D ignore certain prominent kinds of adaptive, self-directed misbelief suggested by Dennett's own corpus. According to Dennett (Reference Dennett1991), the widespread belief that human cognition is under the control of a conscious self is, in many respects, false – a “user illusion” (p. 216) – yet adaptive, enabling cognitive “self-control” (pp. 417–18). Elsewhere, Dennett suggests that all available evidence often fails to determine which of two competing interpretations of a person's behavior is correct: There may be no fact of the matter whether a famous art critic really admires his son's mediocre art or is actually a victim of self-deception (Dennett Reference Dennett1978, pp. 39–49). If Dennett is right, then, to the extent that our self-interpretations presume to settle such interpretive indeterminacies, our self-interpretations are often false. If such self-directed misbeliefs can nonetheless help our survival prospects, then they can be adaptive.

Here is how this might work. We often form beliefs about what we want and think based on the categories of interpretation afforded by our language and culture. Such self-interpretations are often false because, even if Dennett is wrong that all the evidence fails to select between competing interpretations, in everyday contexts, we often do not have the time to consult all relevant evidence before settling on one. Do I really think an author I am reading is pretentious, or is this self-deception concealing my envy of her masterful prose? None of the subjectively or objectively available data to which I realistically have access need settle this question. Indeed, the available data might also fail to rule out some third, incompatible and, perhaps, linguistically inexpressible, self-interpretation. Yet, context often requires that, even in the absence of determinative evidence, we express self-interpretations employing categories familiar to likely interactants. We also have a standing desire to behave consistently with our publicly expressed self-interpretations (Carruthers Reference Carruthers2009, p. 127). So, once we pick among such underdetermined self-interpretations, we tend to act consistently with our choice. Why is this adaptive? It makes us easier to interpret by our likely interactants, enabling complex coordination and cooperation.

There is strong evidence for this mechanism from pathological cases, such as split-brain patients. It is possible to induce behavior in such subjects in ways that escape their awareness (Carruthers Reference Carruthers2009, p. 126; Gazzaniga Reference Gazzaniga and Gazzaniga1995, p. 1393). For example, it is possible to induce standing in a sitting split-brain subject by flashing the word “walk” in her left visual field, to which her left, complex-language-encoding hemisphere has no access. The right hemisphere knows enough language to trigger compliance with the one-word command. When such subjects are asked why they stand up, they immediately confabulate a self-interpretation, for example, “I'm going into the house to get a Coke,” and proceed to comply with this completely false self-interpretation. Presumably, when the subject's left hemisphere processes the question, it cannot access the correct answer, and so, quickly confabulates a culturally afforded self-interpretation, upon which the subject proceeds to act. My suggestion is that, in dynamic, temporally constrained, quotidian contexts, everyone's access to their true motivations is incredibly limited, so our self-interpretations are usually false – based on quick and dirty motivation attributions afforded by our language and culture. Thanks to the desire to act consistently with one's self-interpretations, however, such self-directed misbeliefs can dramatically facilitate interpretation and coordination among similarly enculturated interactants.

If this proposal is on the right track, then it is relatively easy to see why many religious misbeliefs are adaptive, contrary to the “by-product” theory of the evolution of religious belief endorsed in the target article. The reason is that religious interpretive frameworks are widespread in many populations – that is, many cultures afford religious self-interpretations. Thus, for many human beings, the only interpretive tools available in quotidian, interactive contexts are saturated with religious self-conceptions. So the mechanism described above ensures that many human beings develop religious, self-directed misbeliefs that are adaptive. Relative to some populations, coordination is facilitated by self-interpretations such as “I am inspired by the Holy Spirit!” This is not a defense of “belief in belief” (Dennett Reference Dennett2006a). Much as the capacity to learn English is adaptive only relative to an environment that includes other English speakers, the disposition to self-interpret in religious terms is adaptive only relative to a social environment consisting of interpreters that find such interpretations intuitive. Religious self-interpretation might be adaptive relative to the kinds of interpretive frameworks that happen to have dominated human populations historically. This does not mean that it is adaptive relative to any kind of interpretive framework, or that only religious interpretive frameworks are possible for human populations.

If I am right, adaptive, self-directed misbeliefs are much more prevalent and important than M&D acknowledge. In order to coordinate with our fellows, we need to make ourselves as easily interpretable by them as possible. But, usually, we have very poor access to our true motivations in dynamic, quotidian contexts. One mechanism for dealing with this involves the cultural affordance of ready-made self-interpretations with which most members in a population are familiar. We interpret our motivations using such culturally afforded resources, even though such self-interpretations are likely false. Then we act in ways that confirm such self-interpretations, making interpretation by and coordination with others familiar with such interpretations much easier. As Dennett aptly puts it, we are “creatures of our own attempts to make sense of ourselves” (Dennett Reference Dennett1987, p. 91).