In their recent editorial, Kelleher et al ^{Reference Kelleher, Jenner and Cannon1} emphasised the importance of evolutionary theory for explaining the persistence of psychotic symptoms, depression and anxiety in humans. The authors did not mention the difference between proximate and ultimate explanations, in other words between ‘how’ and ‘why’ explanations, ^{Reference Mayr2} and this could make their argument for using evolutionary theory in psychiatric research more specific. In the development of treatments one needs an explanation at the proximate level, whereas the ultimate level can be necessary for generating hypotheses.

In evolutionary-based research the challenge is to find not which behaviour is beneficial now, but which behaviour has been advantageous for the procreation of ancestors in the past. This is the ultimate-level explanation. We know very little about our human ancestors and hypotheses can easily become ‘just-so’ stories with limited predictive value. Therefore rigorous testing at the how level is required. ^{Reference DeBruine3} Furthermore, there are complicating factors such as cliff-edged fitness, ^{Reference Nesse4} whereby a limited number of traits is beneficial but too many are detrimental.

The possible theories for psychosis or schizophrenia mentioned by Kelleher et al vary enormously. It might have something to do with language development, complex social cognition, hypervigilance or with something completely different. However, all these theories need to be further developed to generate hypotheses at the how level, for example how language/hypervigilance/social cognition skills differ in humans with genes associated with schizophrenia or in family members of people with schizophrenia. The aim is to explain psychotic disorders at the proximate level, because that is needed to find the best possible treatment.