Realising the potential of general population research to reconceptualise the study of “delusions”: From normalising “psychosis” to defamiliarising “normality”

Challenging assumptions about the rarity of “delusional” belief: Epidemiological prevalence studies of the general population

In the first approach, researchers use structured diagnostic interviews or psychometric measures to study the prevalence of delusional beliefs in large general population samples. A good example is the study by van Os et al. (2000) where trained staff conducted interviews with 7,076 randomly selected adults in the Netherlands at their homes with some followed up by telephone interviews with psychiatrists. They reported that 3.3% of the sample met all the criteria for delusions whilst a further 8.7% had delusions that were not associated with distress and did not require intervention, much higher than what might be presumed by prevalence rates for diagnostic categories with which delusions are commonly associated, like schizophrenia.

Whilst van Os et al. (2000) focused primarily on the presence or absence of symptoms using diagnostic interviews, some studies use self-report methods like the Peters et al. Delusions Inventory (PDI; Peters et al., 2004) which assesses not only the presence of a belief but also the levels of conviction, preoccupation, and distress associated with it. This measure is informed by multidimensional approaches to the definition of delusion (e.g., Oltmanns, 1988) which attempt to address the deficiencies of standard psychiatric definitions. The PDI consists of items from the Present State Examination (Wing et al., 1974) rephrased to be less stigmatising and thus more acceptable to a general population sample—for example, “do you ever feel as if you are being persecuted in some way?”

Using a 21-item version of the PDI, Peters et al. (2004) compared a UK sample of 444 people from the general population with 33 inpatients with psychosis diagnoses whose beliefs were judged to be “delusional.” The mean number of PDI items endorsed by the inpatient group (11.9) was significantly higher than that of the general population sample (6.7) and the inpatient group had significantly higher scores on the measures of distress, preoccupation, and conviction. However, in terms of number of belief items endorsed, the nonclinical group’s scores were normally distributed, albeit skewed to the left, and there was considerable overlap between the two groups: 11% of the nonclinical participants had scores higher than the mean score of the clinical group. Peters et al. (2004) concluded that what differentiated the clinical and nonclinical groups was not necessarily the content of beliefs but the dimensions of distress, conviction, and preoccupation. Findings like these and those of van Os et al. (2000) provide empirical justification for the first three of the CBTp therapist assumptions noted in the introduction. Freeman’s (2006) review of nonclinical studies like these concluded that “delusional ideation, delusions, and clinically severe delusions” were “on a continuum, but its exact form requires further study” (p. 202).

However, in a review critiquing such studies from a phenomenological psychiatry viewpoint, Heilskov et al. (2020) argue that some self-report measures may result in “false positives” because they fail to capture what they regard as essential qualities of delusional beliefs like context and reasoning. Responding to similar arguments, Bentall (2015), has suggested that some psychiatric phenomenologists adopt an uncritical approach to the construct of schizophrenia and underemphasise the complexity and varied range of beliefs and experiences in the general population, for example, amongst those who have intense religious and spiritual experiences. This brings us to the second broad approach that has attempted to explore similarities and differences between the beliefs of people with these kinds of experiences and those of clinical samples.

Challenging assumptions that “delusional” beliefs are qualitatively different: Comparing clinical samples and nonclinical analogues along dimensions of belief

Whereas Peters et al. (2004) compared a clinical sample with a large general population sample, some studies have utilised an analogue methodology, comparing clinical groups with subgroups of the general population whose beliefs are presumed, on theoretical grounds, likely to share some phenomenological features of beliefs considered delusional. On the basis of previous research like the seminal study by Jackson and Fulford (1997), Peters et al. (1999) hypothesised that there might be similarities between “culturally idiosyncratic psychotic states and culturally validated mystical states” (p. 85). Using an earlier 40-item version of the PDI, they compared a clinical sample (described as “deluded, psychotic inpatients”) with four nonclinical samples: Christian, nonreligious, Hare Krishnas, and Druids—two New Religious Movements (NRMs). This term, from the sociology of religion, refers to groups which are “new” in that they emerged after the Second World War and “religious” in that they offer a religious or philosophical worldview or claim to provide the means by which some higher goal might be attained (Barker, 1996). Compared with the other three nonclinical samples, those in the NRMs scored higher on all the parameters except ratings of distress. The authors reported that, whilst the NRM groups did not endorse as many “florid” items on another scale as the clinical sample, they could not be differentiated from the clinical participants either on the total number of PDI items endorsed or on ratings of conviction though they were significantly less distressed and preoccupied by their beliefs. There were no differences between the Christian and nonreligious groups, suggesting that the presence of religious belief in itself could not explain the high ratings of the NRM samples. Studies like these appear to indicate that simply holding a belief that others find unusual is not inherently pathological (i.e., distressing or problematic)—evidence warranting the second and third of the normalising assumptions in CBTp noted in the introduction.

Challenging un-understandability: Applying “normal” psychological theories to explain “abnormal” beliefs

Researchers in this third tradition have attempted to understand the development and maintenance of ostensibly delusional beliefs by drawing on the conceptual frameworks and research methods used in studies of the general population within psychology’s subfields like perceptual, cognitive, and social psychology. Garety and Freeman’s (2013) review of cognitive behavioural research on delusions, for example, reviewed studies examining biases in probabilistic reasoning (e.g., the “jumping to conclusions” bias), externalising attributions, affective processes, and appraisal. Freeman (2006) concludes his review of nonclinical studies of “delusions and delusional ideation” by noting that “these experiences are not outside of understanding in terms of interactions of normal (albeit biased or dysfunctioning) processes” (p. 202) and he argues that, as a result, intervention with CBT is justified.

Studies within this tradition, therefore, have challenged traditional psychiatric assumptions that “delusional” beliefs are “un-understandable” (Jaspers, 1963) or “empty speech acts” that are “not the symbolic expression of anything” (Berrios, 1991, p. 12). In contrast, Roberts (1991) reported that they were associated with measures of purpose and meaning in life and might offer a “preferred reality.” Researchers draw on a range of hypotheses derived from “normal” psychological theories to understand the meaning of beliefs, thus warranting the fourth normalising assumption noted in the introduction.

One line of research has examined associations between the content of ostensibly delusional beliefs and adverse life experiences. For example, Read et al. (2003) reported a link between documented abuse in childhood or adulthood and the thematic content of supposedly delusional statements. Persecutory beliefs have been associated with attachment-disrupting events (e.g., Bentall et al., 2014), childhood adversity, and experiences both of victimisation (e.g., Dickson et al., 2016) and discrimination—Janssen et al.’s (2003) large-scale general population study found that a higher level of “delusional ideation” was associated with experience of discrimination, particularly in two or more domains (e.g., gender, skin colour, or ethnicity, etc.).

The normalising programme has thus made a successful challenge to traditional psychiatric assumptions about beliefs labelled as delusional, breaking down the dichotomy between “abnormal” and “normal” belief; in James’ (1911) terms, familiarising the strange. However, as I argue in the next section, the potential of this programme to reconceptualise these beliefs has been hampered by a number of limitations, many of them similar to critiques of cognitive behavioural approaches (e.g., Boyle, 2002; Parker et al., 1995).

Medicalisation and pathologisation

Many studies within this programme frame the beliefs and experiences of general population participants within a medical conceptual framework, indicated by describing nonclinical samples as “subclinical,” “psychotic-like,” “schizotypal,” and “not in need of care.” Such terminology reifies and homogenises the participants in both clinical and nonclinical groups, implicitly undermining the programme’s challenge to the assumption of a qualitative difference between the groups. In addition, a psychiatric discourse of risk and vulnerability means that the beliefs and experiences of the general population can be transformed from aspects of human variation (e.g., Romme & Escher, 2012) into indicators of the “risk” of future “psychosis.” This can exacerbate the current problems of creeping medicalisation like the broadening of diagnostic categories, lowering of diagnostic thresholds, and “disease mongering” (the creation of new categories of disorder for problems of living previously understood in nonmedical terms). Whilst spectrum approaches like the continuum model hold out the possibility of an appreciation of the diversity of what we understand as “normal,” at the same time they can also enable increasing portions of the general population to be seen as “abnormal” and “at risk.” As McGuire (2017) notes, we move from asking “not ‘is disorder present or absent?’ but rather ‘how much disorder is present?’” (p. 407).

Decontextualisation

A second problem is the reliance—with notable exceptions like Heriot-Maitland et al. (2012)—on quantitative methods like diagnostic interview schedules and self-report measures. The reification of questionnaire responses as “beliefs” means that we are not given any contextual detail that might help us understand them or understand the relationship between belief claims and conduct nor do we know how the person themselves might verbalise their belief claim in everyday conversation. For example, if a NRM member scores highly on the PDI we do not know whether they looked for a group that shared their beliefs or whether the item resembles a key religious or spiritual belief of that community that they have adopted since joining it. With studies using diagnostic interview schedules we lack contextual information about how raters judge a person as being “in need of care.”

A problematic focus on the individual and the intrapsychic

A third problematic feature of the normalising programme is that it has tended to locate the cause of distress as lying within the individual—seen as a unitary rational subject (Harper, 1996; Parker et al., 1995)—with less of an emphasis on the person’s relational context. Yet, as Berkhout et al. (2019) note, even the “self-disorders” associated with psychosis diagnoses are decidedly relational. The variables investigated in the normalising programme are usually measures of intrapsychic (e.g., cognitive, affective, etc.) constructs or demographic data on individuals. NRMs are often simply seen as a source of “healthy” nonclinical participants but the belief community itself, the relationships within it and members’ relationships with people outside it (e.g., with family, friends, work colleagues, etc.) are not considered. Moreover, there has been little, if any, systematic empirical investigation of the dynamic responses, over time, of the social networks of members of a range of belief communities. In one of the few empirical studies of paranoia in the context of interpersonal relationships over time, Edwin Lemert (1962) drew on data gathered in a study of mental health service users, their families, and the courts system in the United States. He documented how problems developed, observing that families initially tolerated “paranoid” behaviour until the situation deteriorated, leading to a crisis point, after which those around the person began to respond differently to them. Cameron (1943) noted how a person’s paranoid preoccupations might cause them to react towards others in a manner that might lead others to exclude them, thus creating what he termed a paranoid “pseudo-community” (p. 233). In addition, Kaffman’s (1983) family therapy case studies examined the relational dynamics involving family members whose beliefs were seen as delusional. Yet, studies like Kaffman (1983) and Lemert (1962) were conducted decades ago and there has been little further longitudinal empirical investigation of social network responses. This is surprising since how those in a person’s social network become concerned about their beliefs is a crucial element in the process by which they may be referred for assessment of their mental health.

The neglect of power and the discursive and institutional context

As already noted, studies in the normalising programme often involve diagnostic interviews (e.g., prevalence studies) or utilise clinical samples where people have been given psychiatric diagnoses, thus, to some degree, implying the validity of diagnostic categories despite their inherent problems (e.g., Boyle, 2002). Of particular concern here is the lack of attention given to the interactional process of judging a belief to be delusional and the role of the diagnoser’s assumptions (Harper, 2011). Spitzer (1990) has proposed that such a judgement is prompted when a person makes “statements about external reality which are uttered like statements about a mental state, i.e. with subjective certainty and incorrigible by others” (p. 391). Palmer (2000) has drawn on the work of Wooffitt (1992), who observed that when ordinary people talk about paranormal experiences they tend to preface their reports with a description of a mundane activity, which might suggest to the listener that what the person was about to report was to be understood as strange and out of the ordinary. Palmer (2000) argued that people judged to hold delusional beliefs failed to do this. McCabe’s (2004) conversation analysis of psychiatric interactions concluded that a failure to persuade was key, arguing that her sample of service users could “recognize that others do not agree with their delusional claims and are not persuaded by the justification they provide for these claims” but that, despite this, “they do not revise their claims” (p. 738).

However, Spitzer and McCabe implicitly assume that people in the general population go about making “nondelusional” belief claims in a different manner whilst Palmer (2000) makes this assumption explicit but bases it on only one study of a nonclinical belief community. Yet there is evidence of incorrigibly held beliefs amongst a range of belief communities, such as political groups (Billig, 1991). Moreover, as Heise (1988) observes, we need to focus not only on the role of the person making belief claims but the other party in the interaction—the diagnoser or interviewer—and the asymmetrical power relations in psychiatric interviews where the professional has the socially sanctioned power to infer the beliefs of the service user and to act as an arbiter of belief plausibility. Indeed, Berkhout et al. (2019) argue that at least some apparently pathological phenomena may result from such power asymmetries.

In discourse analytic studies, Georgaca (2000, 2004) has examined transcripts of her own interviews with service users considered to hold delusional beliefs. She argues that her position as a research interviewer could be viewed as institutionally analogous to that of a diagnostic interviewer and that such institutional positions grant authority to question the belief claims of psychiatric service users and for them often to accept such questioning. She also observes that the judgement of delusion arises from a clash of competing interpretive frameworks, for example, for beliefs with religious content, between a secular medical framework and a religious one (Georgaca, 2004). She concludes that “delusions . . . employ culturally available discourses and discursive strategies for their construction” and “that they are argued and negotiated in speech with similar strategies to those used by ‘non-delusional’ individuals” (Georgaca, 2000, p. 227).

Diagnosis is often presented as an unproblematically objective categorisation but interviews with diagnosers can illuminate some of the discursive processes at work. Harper (1994), for instance, observed that psychiatrists and psychologists drew not only on an “empiricist” account but also a “contingent” account of diagnosis. For example, when explaining disagreement with their colleagues over the correct diagnosis for a service user, they might discuss their theoretical preferences and those of their colleagues. Unsurprisingly, given the subjectivity involved in diagnosis there is evidence that biases may affect judgements (Harper, 2011). In one of the few studies to examine diagnosers’ assumptions about beliefs, O’Connor and Vandenberg (2005) found that clinicians rated core beliefs of the U.S. Nation of Islam, but not Mormon and Catholic beliefs, as pathological. In another study, O’Connor and Vandenberg (2010) reported that psychology undergraduates appeared to evaluate whether beliefs were pathological on the basis of how similar they were to their own beliefs. Summarising the results of both studies, O’Connor and Vandenberg (2010) concluded that “ratings of psychopathology are based on the conventionality of the beliefs” (p. 181).

Neglecting the cultural context of belief

Many of the problems identified so far could be seen as reflecting particular cultural assumptions and this brings us to the fifth limitation: the neglect of the cultural context of beliefs. The assumption of a unitary rational subject is one example of this and Berkhout et al. (2019) argue that the Eurocentric “bounded, unified self of presumed ‘health’ and ‘normality’ is distinctly a product of specific gendered, racialized, classist, and ableist presumptions” (p. 460). The psychiatric terminology utilised in normalising programme research is steeped in Western dualistic conceptions like the distinction, adopted by the American Psychiatric Association’s (2013) Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5), between beliefs and sensory experiences. However, Jones and Luhrmann (2016) report that many of their interviewees gave accounts of “auditory hallucinations” that involved a blend of thought and sensory perception. Similarly, the role of the body is deemphasised although accounts of distress often seem to be referring to feelings (i.e., embodied affective states) rather than disembodied cognitions (Cromby & Harper, 2009). Affect (i.e., “distress”) is seen as a separable dimension from cognition in measures like the PDI. Following Jaspers (1963), a distinction is also made between the form and content of symptoms. In addition, DSM disorders are presented as culturally universal but potentially expressed differently in different cultural contexts and the DSM-5’s categories of delusion content (e.g., persecutory, grandiose, etc.) are treated similarly. In addition, researchers within the normalising programme have not examined how individual beliefs might be influenced by the changing cultural availability of particular ideas nor is there much consideration of how such changes might lead to changes in the cultural acceptability or “social currency” (Heise, 1988, p. 270) of beliefs, an issue that is likely to affect the extent to which beliefs are shared by others.

Inattention to epistemic injustice

A sixth and final limitation of studies in the normalising programme is that they have not investigated the experience of what philosopher Miranda Fricker (2007) has termed “epistemic injustice.” Fricker describes two forms of such injustice: (a) testimonial injustice, whereby “prejudice causes a hearer to give a deflated level of credibility to a speaker’s word” (p. 1), and (b) hermeneutical injustice, which “occurs at a prior stage, when a gap in collective interpretive resources puts someone at an unfair disadvantage when it comes to making sense of their social experience” (p. 1).

Sanati and Kyratsous (2015) note that, because of prejudicial stereotypes about people with psychiatric diagnoses, a service user may find their credibility is questioned. An example of such testimonial injustice can be seen in cases of the “Martha Mitchell effect” (Maher, 1988) where clinicians, because of such stereotypes, judge a belief as delusional when it is not. Maher named such cases after the late wife of President Nixon’s Attorney General (John Mitchell). Because she gave journalists information about the Watergate affair, the Nixon White House attempted to undermine her by labelling her as unreliable due to alcoholism and mental health problems (Neyfakh, 2017). However, the information she gave journalists was confirmed as being true in the subsequent investigations. Yet, as Sanati and Kyratsous (2015) observe, the veracity of all a service user’s statements, not simply those related to a supposedly delusional belief, may be questioned. Harper (2011) suggests that, in diagnosing paranoia, professional judgements of the plausibility and credibility of accounts may be influenced by the service user’s status in relation to social categories of identity such as gender, race, culture, and social class and that there is the potential for a double injustice since women, Black people, and those from a working-class background not only are more likely to experience structural discrimination but more likely, in general, to have their credibility questioned. Sanati and Kyratsous (2015) suggest that hermeneutical injustice may arise when a belief is seen as delusional because the service user cannot find the language to adequately convey their concerns in some way or because mental health professionals lack the conceptual resources to understand what is being communicated. Hermeneutical injustice will be exacerbated by power asymmetries in the process of diagnosis and also when there is an inequality of access either to nonpathologising explanations for anomalous experiences or to the kinds of belief communities where such accounts might be found—they may, for example, be more available to those with higher levels of education.

These limitations have prevented the normalising programme from fully realising the potential of general population research to reconceptualise how beliefs labelled as delusional are understood. In James’ (1911) terms, these normalising studies could be characterised as being a programme of familiarising the strange. However, by not fully investigating the diversity of belief in the general population, this programme has failed to view the “familiar as if it were strange” (p. 7). In the next section, I make a case for the importance of social norms in the process of judging beliefs as delusional and propose the study of a range of unconventional belief communities. In the final section, I suggest ways in which such studies could address each of these six problems.

The importance of social norms in judging beliefs as delusional

The DSM-5 includes this glossary definition of delusion:

A false belief based on incorrect inference about external reality that is firmly held despite what almost everyone else believes and despite what constitutes incontrovertible and obvious proof or evidence to the contrary. The belief is not ordinarily accepted by other members of the person’s culture or subculture (i.e., it is not an article of religious faith). (American Psychiatric Association, 2013, p. 819)

Psychiatric diagnoses are normative and we can see that this definition implies several assumptions about “nondelusional” beliefs. For example, it is assumed that they will be: of a propositional nature, true and based on correct inference about external reality and that each of these can be straightforwardly determined. There is an extensive literature debating the validity of this definition on conceptual and empirical grounds (e.g., Boyle, 2002; Oltmanns, 1988; Spitzer, 1990) but here I will focus on its empiricist assumptions. Maher (1992) has argued that the judgement that a belief is delusional is “typically made by a clinician on the basis of ‘common sense,’ and not on the basis of a systematic evaluation of empirical data” (p. 261). This claim is consistent with observational studies of assessment interviews showing that psychiatrists frequently elicit the content of beliefs, less frequently challenge the content or explore alternative explanations, but do not appear to examine the evidence for beliefs in any systematic manner (Zangrilli et al., 2014). Perhaps one of the reasons why such examination does not occur is that, as Georgaca (2000) has argued, ostensibly delusional beliefs “are statements about the self and the world whose truth and falsity cannot be definitely settled in speech” (p. 227). In practice, it seems that whether a person’s beliefs are shared by others and whether they have been able to persuade others of their beliefs are key concerns (Boyle, 2002; Heise, 1988). As Oltmanns (1988) puts it, “the balance of evidence for and against a belief is such that other people consider it completely incredible [emphasis added]” (p. 5).

General population research as a corrective to the “clinician’s illusion”

One of the central challenges of psychiatry is that although its central constructs, like delusions, are defined in normative terms, clinicians and researchers lack comparative data on the general population since the vast majority of their work is with people who have been referred to mental health services. As a result, they can fall prey to the “clinician’s illusion” (Cohen & Cohen, 1984) whereby they overgeneralise from their clinical experience to the general population. Thus, on the basis of their clinical experience seeing service users whose beliefs are associated with distress and disruption, they might presume that such beliefs are inherently pathological. I would argue that the DSM-5 definition (American Psychiatric Association, 2013) is influenced by this bias since it was not formulated on the basis of empirical data about “nondelusional” beliefs in the general population. For example, many beliefs widely endorsed in the general population may lack persuasive empirical evidence: a Gallup survey of 1002 U.S. adults reported that 73% believed in at least 1 of 10 paranormal experiences with 41% believing in extrasensory perception and 37% believing that houses could be haunted (Moore, 2005). There is, therefore, a danger that those holding beliefs considered delusional are being held to a higher standard of rationality—a hyper-rationality—than are ordinary people.

The study of general population samples can provide a corrective to this bias and Boyle (2013) suggests four ways in which such research can be useful. First, experiences and behaviour seen as associated with “psychosis” can be understood in the context of “normal” responses to “abnormal” events like the range of adversities that people can face in their lives (e.g., childhood abuse, bullying, etc.). Second, she suggests there is a need to identify links between apparently “psychotic” experiences and other “normal” experiences distributed in the general population (i.e., that they can be understood by drawing on similar principles and theories). For example, Bentall (2015) argues that “psychotic” experiences should not be compared with ordinary life experiences but, rather, with “the exceptional experiences of ordinary people” like “sudden life changes such as religious conversion” (p. 114). Third, researchers could investigate the wide range of functions (e.g., personal, interpersonal, social, and cultural) that “psychotic” experiences and beliefs might fulfil in the same way that other experiences and beliefs fulfil functions. Last, she suggests it is important to explore links between the content of “psychotic” experiences and beliefs and a range of life experiences, for example, with adverse experiences. To some degree, the normalising research programme has made forays into these areas but, as we saw in the previous section, it has been hampered by a number of limitations. However, building on Boyle’s (2002) arguments, if normative judgements about beliefs are key, but these judgements might be subject to bias because of a focus on clinical groups, then one way forward is to develop a systematic programme of research on groups in the general population whose beliefs may be seen as implausible because they breach a range of socially established belief norms.

Examining the role of social norms by researching unconventional belief communities

We saw earlier how normalising programme studies identified subgroups within the general population whose beliefs were seen, on theoretical grounds, to share some features with beliefs labelled as delusional, with the most studied groups being religious and spiritual belief communities like NRMs. The alternative programme proposed here is different in two respects. First, the identification of groups should be focused on the breaching of social norms. Second, building on studies like O’Connor and Vandenberg (2010), the range of groups studied should be expanded to include members of a range of communities whose shared beliefs are seen as unconventional, which I will refer to as unconventional belief communities (UBCs). The Cambridge Dictionary defines “unconventional” as “different from what is usual or from the way most people do things” (Cambridge University Press, n.d.). “Unconventional” and related synonyms like “unusual,” capture the way in which the belief claims of members of UBCs are likely to challenge the norms of conventional society in a number of different ways. For example, they might appear strange, difficult to understand or explain, different from what is normal or expected, out of the ordinary, and either not conforming with or deviating from what is seen as normal. A focus on the unconventionality of belief opens up the range of belief communities that could be studied including subgroups within mainstream religious faiths, like evangelical Christians who might hear the voice of God (e.g., Luhrmann, 2012), groups studied within the fields of parapsychology and anomalistic psychology (e.g., C. D. Murray & Wooffitt, 2010) such as mediums within the Spiritualist church (e.g., Roxburgh & Roe, 2014), and people claiming to have been abducted by extraterrestrials (e.g., Clancy, 2005) as well as those endorsing a range of political beliefs, including conspiracy beliefs (Harambam, 2020). UBCs could be identified on theoretical grounds by selecting beliefs and belief communities that might breach a particular norm, for example, those implied by definitions of delusions or in descriptions of categories of content.

Although quantitative research can be useful in investigating patterns in larger groups, there is also a need for more qualitative work (e.g., Heriot-Maitland et al., 2012; C. D. Murray & Wooffitt, 2010; Roxburgh & Roe, 2014), including ethnographic studies (e.g., Harambam, 2020; Luhrmann, 2012). However, although there will be significant overlaps between ordinary ethnographic work on these topics and this alternative programme, the questions focused on by researchers might be those that would lead to a richer understanding of the kinds of distressing and disruptive beliefs that prompt referral to mental health services. As well as asking some different questions, they might also draw on different conceptual resources (e.g., Critical Disability, Postcolonial, and Mad Studies perspectives).

In the next and final part of this argument, alternative research questions and study designs are proposed for the investigation of lived belief narratives in context through a programme of research focused on UBCs in the general population.

A demedicalising and nonpathologising approach

It is essential that researchers studying UBCs do so by using a nonmedical conceptual framework and terminology. For religious and spiritual UBCs there are well-established conceptual frameworks within the fields of theology, sociology, and anthropology (e.g., Dein et al., 2012; Luhrmann, 2012). Within psychology, researchers have suggested alternative terminology like “anomalous experience,” “nonordinary experience,” or “exceptional experience” (Lynn, 2017). Since different research traditions utilise different concepts, terminology, and methods researchers should include detailed and contextualised descriptions of beliefs. In addition, UBC members may be understandably wary that mental health researchers might pathologise them or seek to “de-bunk” their beliefs and thus it will be important to approach these communities in a respectful manner. Mental health researchers, including psychiatric survivor researchers, could work collaboratively with members of these communities, theologians, anthropologists, and sociologists, seeking to understand the community’s beliefs in a more ethnographic manner, within the milieu of the particular group. Rather than simply contrasting their beliefs and experiences with psychiatric categories, a focus on social norms might provide a less pathologising way of framing research questions. Beginning with the assumption that unconventional beliefs and experiences are not inherently problematic might enable us to understand why some people’s beliefs become so overwhelming and distressing or so disruptive that they cause them or others concern.

Understanding lived beliefs in context

An important aspect of a UBC programme of research will be to provide not only rich but contextualised descriptions of people’s beliefs. These descriptions might include information on the meaning of the belief to the person and how it is integrated into their life both inside and outside of the community as well as the relationship between belief claims and the belief system shared within the community. Thus Harambam’s (2020) study of conspiracy beliefs considers the influence of personal biographies and the conspiracy milieu, as well as broader cultural shifts in wider society. In other words, a UBC programme should focus on beliefs as they are lived. In addition, as Boyle (2002, 2013) suggests, attention could be paid to the range of functions that beliefs serve. For example, both paranoid and conspiracy beliefs are associated with a range of adversities like powerlessness and victimisation (e.g., Freeman & Bentall, 2017) and Cromby and Harper (2009) have described how such adversities, pattered by social inequality, can lead to an embodied suspicious or paranoid stance toward the world. The British Psychological Society’s Power Threat Meaning Framework (Johnstone & Boyle, 2018) hypothesises that beliefs and experiences regarded as “symptoms” within a medical framework could, instead, be seen as adaptive “threat responses,” coping strategies developed as a response to the threats posed by a range of adversities resulting from the unequal operations of power in society. An alternative programme could examine belief-related threat responses in the general population, investigating links between them and life adversities as well as a range of functions. Such studies could examine particular UBCs whose members are likely to have experienced adversities and also groups within the population likely to have experienced certain adversities rather than UBCs per se. For example, Grier and Cobbs (1968/1992) argued that it was necessary for African Americans to develop a “‘healthy’ cultural paranoia” (p. 161) in order to survive living in a racist society. However, such research also needs to attend to the ways in which such intelligible and adaptive responses to social inequality have been historically pathologised. Metzl (2010), for example, has documented the way in which African American men, expressing legitimate concerns about racism have, historically, been seen as “hostile” and “aggressive” and given schizophrenia diagnoses.

It is also possible to examine the belief dimensions of conviction, preoccupation, and distress in a more contextualised manner (Boyle, 2002; Georgaca, 2000). Boyle (2002) has hypothesised that the “conviction” with which an idiosyncratic belief is held may be associated with whether a person, because of their life circumstances, is threatened by the “loss” of a belief and that there might be less preoccupation when a belief is not rejected by others but shared by them and integrated into a person’s day-to-day life. Peters et al. (1999) found high ratings of conviction amongst their NRM participants but less preoccupation, and one possibility is that belief conviction was rated highly because the PDI items endorsed were core beliefs of the NRM. Was there less preoccupation because the beliefs were shared within the NRM and integrated into their lives? More ethnographic research with UBC members might shed light on such questions. Boyle (2002) also hypothesises that a person holding idiosyncratic beliefs might be distressed not only because the content concerns threat from others but also, if the content is positive, because others respond negatively, for example, not being persuaded that the belief is true. Peters et al. (1999) did not find high ratings of distress amongst their NRM sample and, again, qualitative research with UBC members might help us understand distress in a more contextualised manner.

Studying the interpersonal context

The study of the interpersonal context of belief would be greatly helped by more research into how a wide range of beliefs in the general population are negotiated within families and other social networks (e.g., Billig, 1991). This would provide a comparative basis for the study of UBC members’ social networks. Researchers could examine whether UBC members discuss their beliefs with people outside the belief community and, if so, how they respond. In Roxburgh and Evenden’s (2016) study of people with anomalous experiences, they reported that others often responded negatively: “whenever I’ve ever spoken about my experiences, people have, think that I’m barmy, that I’m a bit mad” (p. 215). Roxburgh and Evenden observed that, as a result, they were dissuaded from talking to other people about them. It is possible, therefore, that many UBC members, aware of how their beliefs might be seen by others, do not discuss them or, at least, are careful about who they share them with. Consequently, it is possible that they do not, therefore, cause others concern and thus avoid coming to the attention of mental health services.

Lemert’s (1962) study suggests that families may have a high threshold of tolerance for unconventional beliefs and it would be useful to explore whether this is the case across a wide range of UBCs. Do, for instance, family members and friends of UBC members disagree with their beliefs? Or do they tolerate, accept, or even agree with them? What influences these processes and how do they evolve over time? How often do family members and others become concerned by the beliefs of UBC members and what prompts such concerns? For example, if there are concerns, is it the person’s beliefs per se or their impact on the person’s life and others which give rise to them? And, for those considered “eccentric,” who may not have significant social networks, have they managed to avoid coming into contact with mental health services and, if so, how?

Interpersonal relationships within the belief community could be another topic of interest. A range of UBCs could be studied including those that are more structured (e.g., with a defined belief system, a consistent membership, and perhaps regular in-person meetings) and those that are more of a network, perhaps only existing online with little direct contact except through discussion forums. Members of faith groups often report the importance of fellowship and social support from fellow believers (Mankowski & Rappaport, 2000) and it is likely that many UBC members will have similar experiences. Moreover, within some UBCs, certain beliefs and experiences might be validated and valorised rather than pathologised. Thus, within the Spiritualist movement clairaudient mediums are seen as having a gift (Roxburgh & Roe, 2014). Researchers could examine the way in which group members are given such valued social roles and how they are experienced—for example, what is it like both for those feeling they are able to contact deceased loved ones through a medium and those who are seen as having such a gift?

Studying discourse, power, and pathologisation

Studying how members of a range of UBCs draw on different discursive strategies to make and defend their belief claims could help explore the questions raised by Georgaca, Heise, McCabe, Palmer, and Spitzer. For example, Bennett and Bennett’s (2000) qualitative study showed that bereaved people who experience a sense of the continuing presence of their deceased loved one drew on both “materialist” and “supernaturalist” discourses in describing their experiences, sometimes within the same sentence. Similarly, in a study of people with psychosis diagnoses, Jones et al. (2016) found that their participants “weighed and debated competing secular and supernatural explanations, often juxtaposing and blending different explanatory frameworks” (p. 488). Although Jones et al. (2016) hypothesised that this negotiation might be heightened for those with psychosis diagnoses, Bennett and Bennett’s (2000) study suggests this might be a more common experience and so further study of this amongst members of UBCs would be useful. Moreover, Bennett and Bennett (2000) reported that people were often reluctant to espouse a supernaturalist discourse, particularly with strangers, “for fear of ridicule” (p. 141), illustrating the powerful role of social norms in influencing whether beliefs are disclosed, as noted in the previous section.

Obviously, since UBC members are not located in mental health institutions, they are not subject to the same power asymmetries as service users, like the socially sanctioned power of a professional to label their beliefs as delusional. However, lay people, too, make insanity ascriptions as O’Connor and Vandenberg’s (2010) study of undergraduates shows. Indeed, as Coulter (1973) suggests, such lay ascriptions may form the basis for the more technical judgements made by professionals. Lay judgements could, therefore, be a focus for researchers in the UBC research programme and it is noteworthy that the terms used to refer to those with unconventional beliefs (e.g., “weirdo,” “nut,” “crank,” etc.) are similar to the kinds of derogatory terms for mental health service users. Do others consider UBC members’ beliefs as simply strange and unconventional or do they view them as pathological and indicative of psychiatric problems? What might influence such judgements? To what extent are people outside of UBCs persuaded by their belief claims and by what criteria do they judge the plausibility of such claims? Do they, for example, make judgements about falsity, basis in reality, and the extent to which beliefs are viewed as culturally shared? In a Q methodology study of how a general population sample judged the plausibility of conspiracy beliefs, Lauren Daniel and I found that, whilst a dominant account focused on conventional scientific criteria for evaluating evidence and the quality of reasoning, heuristics like expert consensus were also important and some participants seemed to consider it important to assess the credibility of information sources (Daniel & Harper, 2020).

As we have seen, a key social norm is the extent to which a belief is shared by others and this could be investigated in different ways. For example, one could research UBCs with memberships of different sizes. One could also study a range of UBC beliefs: those that are shared by a larger proportion of the general population and those shared by smaller proportions as indicated by general population survey data. For example, in relation to beliefs about extraterrestrials, Patry and Pelletier (2001) reported that 48% of their Canadian student sample believed in UFOs but only 25% believed in alien abductions. An even rarer belief is “that shape-shifting reptilian people control our world by taking on human form and gaining political power to manipulate our societies,” which was endorsed by only 4% in a U.S. survey (Public Policy Polling, 2013, p. 3). Interviewing people holding this “reptilian” belief, propounded by conspiracy author David Icke (Harambam, 2020), would provide an insight into what it is like to hold a belief shared by very few. Those who report having been abducted by aliens may be an even smaller group and quantitative studies like Clancy (2005) suggest a number of potential avenues for qualitative research.

Coulter (1973) suggested that beliefs might be regarded as unusual but still be seen as legitimate and socially supported if they were categorised as religious beliefs or superstitions, so researchers could examine the differential social support for a range of UBCs assessed, for example, via opinion polls and media reporting. UBCs may be represented in a range of ways in the media (e.g., as harmless and even entertaining or, on the other hand, as dangerous). Thus, whilst some religious and spiritual beliefs might receive the legitimacy Coulter suggests, some NRMs are labelled as dangerous “cults.” Similarly, conspiracy beliefs vary in how much support they attract. For example, 55% of British adults polled in 2015 agreed with the statement, “the Government is deliberately hiding the truth about how many immigrants really live in this country” (Drochon, 2018, p. 339). In contrast, reptilian beliefs are endorsed by fewer people. Thus, some conspiracy beliefs more than others may be regarded as “stigmatized knowledge” because those endorsing them do so “despite the marginalization of those claims by the institutions that conventionally distinguish between knowledge and error” (Barkun, 2013, p. 26). Many conspiracy believers are aware of how negatively they may be perceived and often reject the label “conspiracy theorist” because they view such labelling as an attempt to delegitimise their beliefs (Harambam, 2020).

Coulter (1973) observed that, outside of socially supported beliefs, other beliefs might be seen as falling within a category of “bizarre idiosyncrasy” (p. 136) and Roberts (1991) notes that, whereas political and religious belief systems may be shared with fellow believers, “the person committed to a delusional world view is strikingly alone” (p. 21). But the social category of “eccentric” can be ascribed to people whose views are so idiosyncratic they do not belong to an identifiable community. For example, Weeks and James (1997) interviewed over a thousand people they defined as eccentrics because they were “nonconforming,” “happily obsessed with one or more hobbyhorses,” and “convinced that he [sic] is right and that the rest of the world is out of step” (p. 18). What might prompt a judgement that a person is not simply eccentric but mentally ill? Researchers could investigate whether those considered eccentric have been labelled in such a way, whether they have tried and failed to persuade others of their beliefs, or whether they have simply not been concerned to discuss them with others.

Studying the relationship between individual and cultural narratives

In order to explore the relationship between beliefs and culture it may be helpful to draw on a conceptual framework that avoids individualistic cognitivist terms like “belief.” In the folklore research tradition, the term “belief narrative” is used to refer to the way in which “individuals and communities verbalise and communicate” about “mysterious phenomena” or “supernatural experiences” as well as the “various interpretations and beliefs connected with such experiences” (Hiiemäe, 2016, p. 8). The narrative metaphor has been used by researchers working at individual, interactional, ideological, and cultural levels of analysis (M. Murray, 2000) and by scholars from different psychotherapeutic traditions, within the survivor movement and across cultures. It could thus provide a way of integrating both different levels of analysis and different theoretical traditions (e.g., Angus & McLeod, 2004). Narratives can be viewed as not inherently pathological but as potentially adaptive whilst viewing them as embodied and lived can avoid unhelpful distinctions between belief and experience.

A broadly narrative perspective also opens up questions about the availability and acceptability of the broader cultural narratives that shape our beliefs. For example, Jones et al. (2016) discuss the ways in which the rise of a postsecular cultural shift is associated with an “embrace of a complex understanding of what is real that neither accepts nor rejects the supernatural” (p. 488). They also note the availability of new virtual communities and Bell et al. (2006) and Shullenberger (2017) have described the flourishing online subculture ³ devoted to beliefs about “mind control” with members claiming they are “targeted Individuals” who are victims of “gang-stalking.” These beliefs are thematically similar both to earlier notions of “influencing machines” (Jay, 2012; Shullenberger, 2017) and to the DSM-5 description of delusions “of being controlled” (American Psychiatric Association, 2013). Jones et al. (2016) suggest that virtual communities may enable people, who might otherwise be socially isolated, to find others with similar experiences across national boundaries, but may also expose them to sceptics, which may heighten their awareness of cultural tensions and doubt.

Sheridan et al. (2020) estimate the prevalence of “gang-stalking” belief as up to 0.17% of adult men and 0.66% of adult women. Researchers could investigate the way in which “targeted individual” belief narratives relate to broader cultural narratives. Shullenberger (2017) argues that they engage with public concerns about government and commercial surveillance technologies, journalistic and congressional investigations into research by intelligence agencies, as well as popular culture. For example, in a recent U.S. survey, 15% endorsed the belief that “media or the government adds secret mind-controlling technology to television broadcast signals” (Public Policy Polling, 2013, p. 3). The internet is a key vector for the circulation of such narratives and Shullenberger (2017) notes that, whereas previously, information in the media might pass through a filter of interpretation by experts (e.g., mental health professionals), digital media flatten culturally sanctioned hierarchies of expertise. They have also facilitated a process of “mainstreaming” by which conspiratorial explanatory narratives have moved from the cultural margins to the mainstream (Barkun, 2013). The hypertextuality of the web means that a website on one topic can easily link one to other conspiratorial topics, and recommendation algorithms serve a similar function for other digital media (e.g., YouTube videos), leading to the increased cultural availability of previously marginal narratives. Popular culture, such as The X-Files TV series (1993–2002, original series), also contributes to this mainstreaming, meaning that conspiratorial explanations have become destigmatised over time (Barkun, 2013). The increasing cultural availability of previously marginal narratives is also facilitated by broader cultural changes. Thus Aupers (2012) links the rise of both conspiratorial beliefs and New Age spiritual beliefs with a search for meaning in response to growing societal epistemological and ontological insecurity. The relationship between the availability and cultural acceptability of beliefs merits further study. Is acceptability simply an effect of availability? For example, Jenkins (2000) notes new religious groups have gradually become accepted over time and thus “the cult oddities of the 1920s became the religious orthodoxies of the 1990s” (p. 231).

Addressing testimonial and hermeneutical injustice

As with the issue of pathologisation, whilst a UBC research programme would not necessarily directly focus on the experiences of service users, it could still shed light on epistemic injustice. For example, to what extent do UBC members experience testimonial injustice and how is it influenced by UBC members’ social locations (e.g., gender, social class, ethnicity, etc.)? Moreover, UBCs vary in their demographic composition and the influence of this on perceptions of credibility and pathology could be examined. O’Connor and Vandenberg (2005) expressed concern that clinicians rated the core beliefs of the U.S. Nation of Islam as pathological because this is an exclusively African American organisation. Thus, experiences of testimonial injustice could be explored with members of UBCs comprised exclusively or predominantly of people subject to discrimination (e.g., women; working class people; and people from a range of ethnic, sexual, and other minorities, etc.) and compared with UBCs whose members come from more privileged backgrounds.

Researchers could investigate hermeneutical injustice by exploring how people encounter UBCs and whether awareness of—and access to—them is influenced by their economic, social, and cultural capital. Might it also be the case that some beliefs are adopted because people have been deprived access to other kinds of explanatory narratives because of inequalities of educational opportunity? For example, might conspiracy beliefs, associated as they are with a range of adverse circumstances and lower levels of education (Freeman & Bentall, 2017), as well as a country’s level of inequality (Drochon, 2018), function as an attempt to explain injustice in a context where people have not had access to other explanatory frameworks that could have rendered opaque and abstract political and structural forces more intelligible?