McKay & Dennett (M&D) address a wide variety of misbeliefs and whittle the range down to conclude that one type of misbelief – positive illusions – remains as viable evidence for the existence of adaptive misbeliefs. We concur with M&D's line of reasoning and propose an additional form of positive illusion that may serve as an example of an adaptive misbelief – overconfidence in the veracity and generalisability of one's theories. This positive illusion is common across a range of domains in adult reasoning (e.g., Klahr & Dunbar Reference Klahr and Dunbar1988; Rozinblit & Keil Reference Rozinblit and Keil2002), but we will focus on its prevalence in theory formation in the developing mind as an incidence that is naturally occurring, pervasive, and seemingly robust. M&D are clear about the qualities that a possible candidate for adaptive misbelief must have: It must be a belief, it must arise in the normal course of the doxastic system's proper functioning and, most important, it must confer adaptive advantage on the individual. We address each of these points in turn.
Karmiloff-Smith (Reference Karmiloff-Smith1992), in her outline of the representational redescription model, suggests that all children go through three phases in theory formation. Briefly, in Phase 1 children collect data from the world, treating each experience as an independent event with little or no generalisation between occurrences. In Phase 2 they consolidate independent representations into a unified theory, rejecting contrary external evidence while the theory is strengthened. And in Phase 3 they test the theory on a range of external examples, adjusting and broadening it to account for a variety of anomalies. The theoretical entrenchment exhibited in Phase 2 can result in errors and inflexibilities not evident in Phases 1 and 3 that lead to the characteristic U-shaped curve of behavioural success on a variety of tasks (e.g., Karmiloff-Smith Reference Karmiloff-Smith1986; Newport Reference Newport and Collins1981). The classic demonstration of this developmental pattern is Karmiloff-Smith and Inhelder's (1974) block-balancing task. When asked to balance a series of blocks, some of which had been covertly weighted such that their balancing point was off-centre, 4- and 8-year-olds consistently passed while 6-year-olds consistently failed. Karmiloff-Smith suggests that the 4-year-olds succeed by treating each block as an independent task but the 6-year-olds have a theory that blocks balance in their symmetric centre and they overgeneralise this to apply to all blocks, even when this strategy consistently fails. Eight-year-olds, by contrast, hold the same theory but are flexible enough to also take into account the extra dimension, asymmetric weight, and adapt their strategy.
To what extent can children's experience of theoretical entrenchment in Phase 2 be referred to as a “belief”? M&D dismiss aliefs and judicious psychological biases as candidates for adaptive misbelief because, they argue, the consumers don't really believe the bias, they just respond as though they were in danger or in case danger might be lurking. Conversely, in many ways children's overconfident belief in their theory is akin to M&D's description of delusions as examples of misbeliefs par excellence in that they are “held with strong conviction regardless of counterevidence and despite the efforts of others to dissuade the deluded individual” (sect. 4, para. 2). To test whether 3-year-olds had a perseverative theory that all objects fall straight down, Hood (Reference Hood1995) presented an array in which objects fell down a curved tube to a displaced location. All 3-year-olds searched directly below the dropping point even if there was only one tube (and thus no physical connection between the dropping point and the favoured target), repeatedly in the face of counterevidence (up to 20 consecutive trials) and persistently regardless of how many times the experimenter explained the role of the tubes to them. This persistence implies that children really believe that their theory is correct. M&D dismiss delusions as examples of adaptive misbelief because they arise from an improperly functioning doxastic system. Conversely, Phase 2 theoretical perseveration occurs in all children across a range of domains and seems to be a built-in feature of a properly functioning theory-formation mechanism. Indeed, children often make up observables in support of their theory when the perceptual experience lets them down (e.g., Baker et al. Reference Baker, Murray and Hood2009; Massey & Gelman Reference Massey and Gelman1988).
Last, it is necessary that a proposed adaptive misbelief should convey adaptive advantage to the individual. Overconfidence in one's theories conveys adaptive advantage insofar as it enables them to creatively simplify a problem by ignoring some of the complicating factors. “[I]t seems possible for the child to experience surprise and question his theory only if the prediction he makes emanates from an already powerful theory expressed in action” (Karmiloff-Smith & Inhelder Reference Karmiloff-Smith and Inhelder1974, p. 209). Thus, Phase 2 enables children to unify representations into coherent (but overgeneralised) theories that in turn lead to new, broader theories and greater behavioural mastery. Second, overconfidence in one's theories sustains and enhances health in an everyday sense by decreasing exposure to cognitive dissonance, which has been shown to lead to feelings of anxiety and stress (Aronson Reference Aronson and Berkowitz1969), which in turn result in negative physiological effects. Consequently, overconfidence in one's theories may also result in exaggerated feeling of control, a positive illusion that M&D list as adaptive in its own right.
Thus, overconfidence in the veracity and generalisability of one's theory fits the criteria laid out by M&D as necessary to be considered as an adaptive misbelief. Children certainly believe that they are right; this belief is systematic and misinforms the organism as a whole, occurs for all children across a range of microdomains, and persists into adulthood. Therefore, it can be considered a naturally occurring feature of a properly functioning doxastic system. It can also be construed as adaptive in leading the individual to undertake adaptive actions and by enhancing health and fitness. In children, this tendency is evident not only in subjective self-evaluation, but also in objective theories about how the world works that, in turn, guide their behaviour. A phase in which this is especially prominent occurs across a variety of microdomains and may be a fundamental and important feature of properly functioning theory-building doxastic systems.
