1 Introduction
Delusions matter—both to the people who endorse them, and as a topic of investigation in philosophy, ethics, epistemology, psychology, and more. Lisa Bortolotti’s pioneering work in this area has played a significant role in convincing many philosophers of delusions’ theoretical and practical importance. Given this, I am grateful for the opportunity to comment on her recent book Why Delusions Matter (henceforth WDM), which builds on her previous contributions and adds several new and provocative ideas.
WDM has two overarching aims. The first, which is descriptive, is to analyse what we are thinking, feeling, and doing when we attribute delusions to others. Here, Bortolotti argues that we attribute delusions when we encounter beliefs that strike us as implausible and appear to function as unshakeable identity beliefs for the relevant speaker. The second, which is revisionist, is to give delusions a more positive reputation—to “challenge the idea that the negative connotations associated with delusional beliefs tell the full story” (Bortolotti, 2023, p. 34). “What if,” asks Bortolotti (2023, p. 34), “instead of being merely distortions of reality doomed to disrupt communication, the inevitable causes of triangulation breakdowns and other innumerable forms of harm, delusions were imperfect responses to an existing crisis?”
In this commentary, I will identify some concerns with this second revisionist goal. More specifically, I will first describe Bortolotti’s assumption that the concept of delusions should include popular beliefs that arise in non-clinical contexts (Section 2), and then draw attention to certain prominent examples of non-clinical delusions—namely, those that result from what I call “propagandistic cognition”—to challenge the claim that delusions in general deserve a more positive reputation (Section 3).
2 Non-clinical delusions
Much research on delusions assumes that it would be desirable to define the concept in ways that demarcate clinical delusions from non-clinical but epistemically irrational beliefs that prevail in the general population (see Murphy, 2013). For example, some definitions stipulate that delusions are distinguished not just by their alleged epistemic defects but by the fact they are idiosyncratic and hence not shared with others (e.g., DSM-5 2013, p. 819). Under the plausible and widespread assumption that the most spectacular forms of epistemic irrationality in the non-clinical population are collective, this definition excludes such beliefs as examples of genuine delusions.
One interesting and innovative aspect of Bortolotti’s project in WDM is to reject the assumption that delusions are an essentially clinical phenomenon. If, as she argues, delusions are attributed when beliefs strike interpreters as implausible and unshakeable identity beliefs, there is certainly no a priori reason why delusions should only arise in conditions such as schizophrenia, bipolar disorder, dementia, and so on. In fact, it is conceivable that delusions, so understood, could be a reliable marker of mental health and indeed mental flourishing under certain conditions. Moreover, as a matter of empirical fact, we do encounter many beliefs among psychologically healthy people that seem to possess the characteristics found in Bortolotti’s analysis of delusion attribution.
Nevertheless, one might worry about the utility of a concept that bundles together clinical delusions of the sort that arise in conditions such as schizophrenia with forms of epistemic irrationality that are widespread in the non-clinical population. If the concept of delusions becomes so broad that it picks out beliefs with very different aetiologies and consequences, reliable generalisation becomes challenging. This might be a problem for the revisionary aspect of WDM’s project, which seems to involve such a generalisation: namely, that delusions in general deserve a more positive reputation than they currently enjoy. For example, Bortolotti (2023, p. 12) argues that the “dismissal of the speaker’s perspective [in cases of delusion attribution] is something we can work harder to avoid as interpreters”. I agree that this is important when it comes to clinical delusions. However, it is not obvious that it is always or even typically the most appropriate course of action when it comes to non-clinical delusions.
3 The functions and harms of extraordinary popular delusions
Charles MacKay’s (1841) classic book “Extraordinary Popular Delusions and the Madness of Crowds” explored numerous examples of non-clinical delusions. One important case study concerned beliefs about witchcraft. Although MacKay’s focus was on witchcraft accusations in sixteenth- and seventeenth-century Europe, surprisingly similar beliefs about witchcraft are extremely common across diverse and independent cultural contexts, including within small-scale societies (Singh, 2021). It is understandable why MacKay included such beliefs in his book. They strike many outsiders of the relevant communities as bizarre. Moreover, they often seem to satisfy the conditions for delusionality that Bortolotti identifies in WDM.
Although the causes of beliefs about witchcraft are complex and diverse, anthropologists have long noted that they often emerge under conditions in which people are motivated to demonise other people, typically in the service of either social levelling (i.e., reducing a target’s status) or eliminating people viewed as a burden on the community (Singh, 2021). In this sense, witchcraft beliefs often seem to function as “demonizing narratives”, which emerge when “actors bent on eliminating rivals devise demonizing myths to justify their rivals’ mistreatment” (Singh, 2021, p. 3). As Pascal Boyer (2022) puts it,
“It helps to see witchcraft accusations as a form of stigmatization, providing a coordination point for coalitional alignment against a particular individual… People who have some interest in inflicting harm on a particular individual may use witchcraft accusations rather than a direct attack because the accusation makes it possible to recruit allies against the target, whilst maintaining one’s own reputation.”
Given this strategic logic, one might be tempted to conclude that people do not sincerely believe witchcraft accusations when they advance them. This might be true in some cases. However, a significant body of evidence suggests that human beings are strongly biased towards beliefs they are motivated to spread to others (Butterworth et al., 2022; Hoffman & Yoeli, 2021; Melnikoff & Strohminger, 2020; Pinsof et al., 2023; Schwardmann & Weele, 2019; Williams, 2023a). That is, advocating for a belief often transforms us into true believers. This might be because sincerity makes us more persuasive propagandists (Hippel & Trivers, 2011), but it might also simply result from the fact that when we generate evidence and arguments to persuade others, we often end up persuading ourselves in the process (Hoffman & Yoeli, 2021).
Although there is still much that we do not know about this phenomenon, it seems plausible that people’s tendency to become convinced of beliefs they are motivated to advocate for—let us call it “propagandistic cognition”—plays an important role in many non-clinical delusions. For example, people often embrace and spread bizarre and completely unfounded demonising narratives when their goal is to eliminate rival groups in the context of pogroms and genocides (Horowitz, 2001; Mercier, 2020). Demonising narratives also emerge in many conspiracy beliefs, which often seem to function to demonise target groups and institutions (Cassam, 2019; Singh, 2021). Absurd Nazi conspiracy theories about the Jewish population in Europe provide one clear example, although by no means the only one. It is difficult to think of a clearer example of a demonising narrative than QAnon, which Bortolotti includes as a clear example of a non-clinical delusion in WDM. Moreover, conspiracy theories often serve propagandistic functions beyond demonisation. For example, Republican conspiracy theories about election fraud do not just demonise their political opponents; by calling into question the results of elections, they justify self-serving attempts to overturn the result.
In addition, propagandistic cognition plausibly plays a role in “system-justifying ideologies” (Williams, 2023b). For example, in contexts of group-based domination such as white supremacy or patriarchy, dominant groups are often motivated to embrace beliefs that justify their social position and legitimise the social arrangements from which they benefit. Thus, what Charles Mills (2017, p. 18) calls the “white delusion of racial superiority” might be rooted in the fact that white populations were historically motivated to propagate beliefs and narratives that rationalised their privileged social position. More generally, propagandistic cognition seems to play a powerful role whenever intergroup conflict exists, including in polarised political contexts, where rival political coalitions often propagate, embrace, and identify with implausible beliefs and narratives that promote and justify their coalitional interests (Pinsof et al., 2023; Williams, 2023a).
When non-clinical delusions result from propagandistic cognition, it is not obvious that they deserve a better reputation. This is not just because such beliefs are often socially and politically harmful. Bortolotti is attentive to societal and political harms of certain non-clinical delusions. It is also because they result from selfish, strategic goals. In analysing delusions, Bortolotti generally considers explanations in which they emerge from a combination of epistemic needs and intrapsychic needs—for example, needs for comforting feelings of control, certainty, and meaning. “Delusions,” she writes, “either let speakers see the world as they want the world to be; make speakers feel important and interesting; or give meaning to speakers’ lives, configuring exciting missions for them to accomplish” (Bortolotti, 2023, p. 9). I am sceptical of these intrapsychic explanations, both because evolution does not design belief-forming systems to maximise merely psychological benefits and because social-functionalist explanations of biased beliefs tend to be better supported by evidence than explanations that appeal to intrapsychic regulation (Mercier & Sperber, 2017; Tetlock, 2002; Williams, 2021). However, even if Bortolotti’s explanations are applicable in some cases, they do not seem applicable when non-clinical delusions result from propagandistic cognition. In such cases, the beliefs do not result from psychological “needs”; they result from strategic, social, and typically selfish goals, and it is not obvious that they deserve a more positive reputation. If anything, once we have a clearer understanding of their functions and causes, an even more negative reputation might be called for.
To be clear, I am not denying that some non-clinical delusions deserve a more positive reputation. For example, identity-defining religious beliefs that solve problems of social cooperation without imposing costs on others seem to satisfy Bortolotti’s criteria of delusionality, but their causes and effects often seem generally benevolent (Luhrmann, 2020; Raihani, 2021). The point is rather that non-clinical delusions are difficult to generalise about. Whether they deserve a more positive reputation must be decided on a case-by-case basis. In some cases—perhaps many—dismissing a speaker’s delusional perspective is probably not something we should try to avoid.