We can find three broad types of positive illusions, following Taylor and Brown (1994): (1) excessively positive self-appraisals; (2) the belief that one has greater control over events than it is actually the case; (3) more rosy views of the future than statistics can warrant. It is important that positive illusions are regarded as mild distortions of reality and do not involve "denying the obvious": most researchers interested in positive illusions are keen to distinguish positive illusions from cases of self-deception or from defence mechanisms. Taylor and Brown have shown not only that positive illusions are widespread in non-clinical populations, but that there are strong links between certain forms of positive illusion and the promotion of mental health (in terms of creativity and productivity), and physical health (in terms of prolonged longevity).
A classic example of positive illusions is the better-than-most effect. People tend to find themselves warmer, kinder, more sincere, etc., than the average person and these "self-aggrandising" views are correlated with higher achievement. Similar results have been found with young children overestimating their capacities. They develop better language, problem-solving, or motor skills.
People affected by serious illness who believe that they are coping better than other patients are found to experience reduced stress. Another classic example concerns illusions of control. In a lottery situation, people who have been assigned random tickets prefer to swap their tickets with tickets they choose themselves, even if this does not impact on their chances to win (Langer 1975). People's belief that they can change alter external circumstances for the better, when the circumstances are such that they can be to an extent changed, contributes to better adjustment in context of trauma or chronic illness.
Finally, unrealistic optimism is the phenomenon by which "people anticipate that their future will be brighter than can reasonably be justified on statistical grounds" (Taylor and Brown 1994). People who are optimistic in this way are more creative, and cope better with stressful situations. Taylor et al. (1992) studied men who had tested seropositive for HIV and found that they were more optimistic about not acquiring AIDS than men who knew they were seronegative. This (illusory) optimism was correlated to health-promoting behaviour and use of positive coping techniques.
Now, this is all good, but can positive illusions be also epistemically beneficial? In the endless source of wisdom that is McKay and Dennett (2009), we find some suggestions about epistemic features of positive illusions. First, the authors argue that positive illusions are both psychologically and biologically adaptive, because they do not just make people feel better about themselves, but can improve physical health too, and have evolutionarily relevant benefits. One example is parenting. The better-than-average effect applies not just to oneself but also to one's romantic partner, and seeing one's partner as better than the average partner in intelligence, attractiveness, etc. (Gagné and Lydon 2004) is good for the relationship and (maybe less romantically) matters to the efficient "completion of species-specific parental duties". If the relationship lasts, the offspring receives better care and more protection, so she is more likely to survive and flourish. Similarly, people have inflated views of their own children (Wenger and Fowers 2008), which makes parents more attached to their children, and more willing to dedicate time and resources to them.
More interesting from our point of view, if it is true that unrealistically positive views of, and a sense of control over, one's medical history increase health and longevity (Taylor et al. 2003), then, as McKay and Dennett notice, the optimism of positive illusions is "not so unrealistic after all". The belief that one will be healthy and live longer is no longer unjustified, because it acts as a self-fulfilling prophecy and it becomes true. McKay and Dennett conclude that: "such doxastic departures from reality are not culpable but entirely forgivable". In the light of their epistemic features, maybe we could even say that some positive illusions do not really count as departures from reality at all.