The term 'schizophrenia' refers to a group of disorders that have been described in every human culture. Two apparently well established findings have corroborated the need for an evolutionary explanation of these disorders: (1) cross-culturally stable incidence rates and (2) decreased fecundity of the affected individuals. The rationale behind this relates to the evolutionary paradox that susceptibility genes for schizophrenia are obviously preserved in the human genepool, despite fundamental reproductive disadvantages associated with the disorders. Some researchers have therefore proposed that a compensatory advantage must exist in people who are carriers of these genes or in their first-degree relatives. Such advantages were hypothesised to be outside the brain (e.g. greater resistance against toxins or infectious diseases), or within the social domain (e.g. schizotypal shamans, creativity). More specifically, T.J. Crow has suggested an evolutionary theory of schizophrenia that relates the disorders to an extreme of variation of hemispheric specialisation and the evolution of language due to a single gene mutation located on homologous regions of the sex chromosomes. None of the evolutionary scenarios does, however, fully account for the diversity of the symptomatology, nor does any one hypothesis acknowledge the objection that the mere prevalence of a disorder must not be confused with adaptation. In the present article, I therefore discuss the evolutionary hypotheses of schizophrenia, arguing that a symptom-based approach to psychotic disorders in evolutionary perspective may improve upon the existing models of schizophrenia.