Are Positive Illusions Epistemically Innocent?

A belief is epistemically rational if it is well supported by evidence and responsive to counter-evidence. But do epistemically rational beliefs contribute to our psychological wellbeing? Some believe that epistemic rationality contributes to psychological wellbeing, and that epistemic irrationality is often responsible for psychological distress (for a version of the traditional view, see Healthy Personality, by Jourard and Landsman, 1980). Others believe that psychological wellbeing requires epistemic irrationality, and that there are circumstances in which epistemic rationality is responsible for psychological distress (for a version of the trade-off view, see e.g. Positive Illusions, by Taylor, 1989).

The traditional view tells us that people who are psychologically healthy have cognitions that are constrained by evidence and are accurate, that is, they track how things actually are. Their memory reports are reliable, their beliefs well-grounded, and their predictions realistic. People who experience psychological distress have cognitions that are not constrained by evidence and are inaccurate, that is, they do not track how things actually are. Their memory reports are distorted, their beliefs ill-grounded, and their predictions unrealistic. This view obviously has some implications for the goals of psychological therapy: psychological health is enhanced when epistemic rationality is restored. 

The trade-off view suggests that epistemic irrationality is often not inimical, but conducive to psychological wellbeing. People who are psychologically healthy are unreasonably optimistic when they form beliefs about their skills and talents, when they assess their capacity to control external events, and when they predict their future. Interestingly, people who are affected by low mood do not share such an inflated conception of their skills and talents, they do not overestimate their capacity to control external events, and they predict their future more realistically than people without low mood. This view also has implications for the goals of psychological therapy: psychological wellbeing is improved when the right kind of distortion (such as a doxastic bias towards self-enhancement) is introduced or re-instated. 

I am offering caricatures of views that have been defended in the psychological literature, and that have been influential in driving psychological research, and in shaping our understanding of the relationship between epistemic rationality and psychological wellbeing. Both views should be challenged in the light of recent empirical evidence. The traditional view cannot accommodate the ‘depressive realism’ effect, which proves to be a robust phenomenon across a number of contingency tasks. The trade-off view can easily explain the phenomenon of depressive realism, but is silent about forms of psychological distress different from those manifesting with low moods. Moreover, in the recent psychological literature on optimism there is also a suggestion that is not immediately compatible with the trade-off view: some positive illusions have significant psychological costs as well as benefits (see Bortolotti and Antrobus 2014).

The traditional view offers an oversimplified account of the relationship between epistemic rationality and psychological wellbeing, potentially leading to an unjustified stigmatisation of poor mental health as a manifestation of epistemic irrationality. But in the trade-off view epistemic rationality and psychological wellbeing are presented as incompatible, and this conclusion does not reflect the complexity of the empirical data. 

A more balanced view acknowledges that some epistemic benefits can be mediated by psychological ones, and some psychological benefits can be mediated by epistemic ones. It also recognises that a belief can be at the same time psychologically beneficial and epistemically irrational, or psychologically harmful and epistemically rational. We can use the notion of epistemic innocence as a way of telling a richer story about the epistemic evaluation of self-enhancing beliefs.

An epistemically innocent cognition is one which has epistemic costs for an agent, but also epistemic benefits that the agent would not be able to attain by adopting less epistemically costly cognitions. The notion of epistemic innocence enables us to show that even epistemically irrational beliefs can have epistemic benefits, and that these benefits should be taken into account when evaluating such cognitions. Some of the epistemic benefits of a cognition are mediated by its psychological benefits, and some are independent of those. The beliefs and predictions due to positive illusions are typically ill-grounded, and resistant to counter-evidence. But two epistemically relevant considerations apply to them. 

First, they enhance socialisation and thus increase the chances of getting feedback from peers, and, second, they support agency, including the opportunity to pursue and achieve epistemic goals. Combined, such roles are likely to promote the capacity to acquire, retain, and use relevant information and the exercise of some intellectual virtues. Moreover, it is not obvious that people can always avoid positive illusions merely by regulating their epistemic practices, and thus the epistemic irrationality of their self-enhancing beliefs and optimistic predictions can be difficult and sometimes even impossible to avoid. This suggests that people are not always epistemically responsible for their unrealistically optimistic beliefs.

The idea behind epistemic innocence is to approach the epistemic evaluation of epistemically irrational beliefs in a way that is sensitive to both their costs and benefits. This strategy is promising when applied to positive illusions.