Singh's cultural evolutionary theory of shamanism provides a nuanced explanation for the recurrence of shamanistic practices across diverse cultures. Nonetheless, although his stated aim is to build a comprehensive theory of shamanism, his theory has a lacuna: It does not offer a satisfying account of who becomes a shaman. Singh mentions that it may be easier for individuals with “psychological peculiarities” to become shamans, but he says little about what these peculiarities entail or why they are important. In this commentary, we propose that continuum models of psychosis could illuminate the psychological profile of individuals who are most likely to pursue the shaman vocation.

The notion that shamanism is associated with psychosis – a serious loss of contact with reality typically involving hallucinations and delusions – has long been contentious. One of the earliest proposals in this vein was made by Silverman (Reference Silverman1967), who argued that shamanism is a manifestation of schizophrenia (the primary psychiatric diagnosis associated with psychosis). Silverman claimed that shamanism and schizophrenia have closely related cognitive, experiential, and behavioral profiles. Indeed, he suggested that the only substantial difference between shamans and patients with schizophrenia is that shamans attain a greater level of cultural acceptance for their aberrant behaviors. However, Silverman's theory was built on psychoanalytic foundations that have subsequently fallen out of favor. More recently, Polimeni and colleagues (Polimeni Reference Polimeni2012; Polimeni & Reiss Reference Polimeni and Reiss2002) have drawn on contemporary cognitive and evolutionary accounts of religion to develop a theory that likewise links shamanism and schizophrenia. Polimeni (Reference Polimeni2012) argues that, “people with a … biological propensity towards psychosis were almost always placed in the role of shaman” (p. 168) and “the anthropological literature on shamanism reveals repeated patterns of psychotic experiences – commonly auditory hallucinations – occurring in young people soon to be shamans” (p. 154).

Proposals that shamanism is closely related to schizophrenia have met with sustained criticism (Luhrmann Reference Luhrmann2011). Noll (Reference Noll1983) provided a particularly influential critique. He argued that there are several important differences between shamanic trance and psychotic episodes: Shamanic trances tend to have a voluntary onset and offset, are not overly negative, are consistent with cultural expectations about appropriate behavior for shamans, and serve community needs by procuring information and blessings from the spirit world. By contrast, episodes of psychosis are typically involuntary and highly distressing, are interpreted as “madness” by the community, and do not serve any obvious social function. We contend that it is particularly relevant that understandings of shamanic trance and “madness” are culturally negotiated. As we have argued elsewhere, scholars should be very cautious about pathologizing religious beliefs; just because it is notoriously difficult to identify a precise dividing line between religious beliefs and delusions does not mean that they are the same thing (Ross & McKay Reference Ross and McKay2017

We suggest that there is a middle path between equating shamanism with schizophrenia and disavowing any relationship whatsoever. Converging lines of evidence suggest that the positive symptoms of psychosis (particularly delusions and hallucinations) lie at the extreme end of a continuum of psychosis-like phenomena in the general population (Linscott & van Os Reference Linscott and van Os2013; van Os et al. Reference van Os, Hanssen, Bijl and Ravelli2000; van Os & Reininghaus Reference van Os and Reininghaus2016). We propose that individuals nearer the psychosis end of this psychosis continuum are more likely to become shamans than people who are not. There may be at least two reasons for this. First, frequent and vivid psychosis-like experiences could be interpreted as evidence of “special gifts,” which could signal that shamanic training would be appropriate. Second, the process of shamanic training could intensify and refine psychosis-like experiences such that they are increasingly under voluntary control and are more consistent with cultural expectations about shamanism. Studies of prayer training in evangelical Christians provide evidence for each of these proposals. Imagination-based prayer practice (as opposed to conventional Bible study) increases the vividness of mental imagery, the frequency of the use of imagery, and the frequency of unusual sensory experiences associated with feeling God's presence. Individual differences in sensory absorption are positively associated with the magnitude of some of these effects (Luhrmann et al. Reference Luhrmann, Nusbaum and Thisted2010; Reference Luhrmann, Nusbaum and Thisted2013). On a related note, we suggest that shamanic training may cultivate an “intuitive cognitive style” (Stanovich Reference Stanovich2011), which has been linked to the endorsement of paranormal explanations for anomalous experiences (Ross et al. Reference Ross, Hartig and McKay2017) and supernatural beliefs (Pennycook et al. Reference Pennycook, Ross, Koehler and Fugelsang2016) in contemporary Western populations.

We are not aware of any existing individual differences in research on shamans in traditional societies that could be used to test our proposal directly. Quantitative evidence for a relationship between shamanism and a psychosis continuum, however, comes from research on a shaman-like role in the West: the psychic. Some people in Western societies have experiences of contact with the supernatural. Much like shamans in traditional societies, such individuals may undertake training to cultivate these experiences and make a vocation of contacting the spirit world to help members of their communities. A recent study of clairaudient psychics (people who report hearing voices from other realms) revealed that this group had hallucinatory experiences that share phenomenological characteristics with patients diagnosed with psychotic disorders (Powers et al. Reference Powers, Kelley and Corlett2017a). Nonetheless, there were important differences. In particular, psychics had more control over the onset and offset of their voices, and their voices were less distressing – a pattern that looks remarkably similar to differences between shamanic trance and psychotic episodes identified by Noll (Reference Noll1983).

In this commentary, we have proposed that Singh's cultural evolutionary theory of shamanism could be enhanced by drawing from continuum theories of psychosis to make predictions about who takes up this particular vocation. Nonetheless, caution is warranted. As we have noted, there appear to be important phenomenological differences between clinical hallucinations and non-clinical hallucination-like experiences. Consequently, a more nuanced, multidimensional model that treats psychosis as a complex constellation of phenomena rather than a particular thing likely will be needed to understand the causal pathways underlying psychosis in clinical populations and psychosis-like experiences in the general population (Luhrmann Reference Luhrmann2017) – and, we would add, shamanism.