An editorial in the September 2010 edition of the British Journal of Psychiatry challenges our ideas about the nature of hallucinations and delusions as a sign of illness or madness. Population-based studies using both self-report and interview surveys show that the prevalence of psychotic symptoms is far greater than had been previously considered. 
Hallucinatory experiences are as old as humankind and until the nineteenth century these experiences were generally attributed to mystical or divine sources such as gods or demons, whereas in modern times, hallucinations have generally been regarded as signs of illness.
Population-based studies using both self-report and interview surveys show that the prevalence of psychotic symptoms is far greater than had been previously considered, with a recent meta-analysis suggesting a prevalence rate of 5-8% in the general population (Schizophrenia has a prevalence of 0.8-1% in the general population).
Psychosis exerts strong negative fitness effects. Despite this apparent disadvantage, schizophrenia maintains a relatively stable prevalence worldwide. Several theories have been formulated to explain the persistence of psychosis in the human population.
Burns proposed that schizophrenia is a ‘costly by-product’ in the evolution of complex social cognition and Nesse developed this idea further in terms of ‘cliff-edge’ fitness, whereby certain traits may increase fitness up to a critical threshold, but beyond this point, fitness falls precipitously. For instance, strong tendencies to use metarepresentation and theory of mind can increase the ability to predict other people’s behaviours and discern their intentions, but, as Nesse explains, ‘it is only one step further, over the cliff’s edge of psychotic cognition… to finding secret meanings and evidence for conspiracies in other people’s most casual gesture.’
Recently, Dodgson & Gordon have proposed that certain types of hallucinations could be viewed as evolutionary by-products of a cognitive system designed to detect threat since, from a survival perspective, it is much worse to fail to recognise a threat such as the sound of an approaching predator than to mistakenly believe that a predator is approaching when it is not. Evolution may therefore favour a selective skew towards propagation of genes that promote false positives over false negatives, thus resulting in ‘hypervigilance hallucinations’ in the population. In social or pack animals, such as humans, hypervigilance is not necessary for every member, but the presence of this trait in at least some members stands to benefit the entire group. Such a trade-off allows the persistence of advantageous traits even in the presence of increased risk of disorder.
Higher than expected prevalences have been reported for depression and anxiety as well. Prospective studies are showing that up to half (and perhaps even a majority) of the population can expect one or more episodes of depression over their lifetime. Taking an evolutionary perspective, one could speculate that the tendency for unhappiness and dissatisfaction is intrinsic to the human species (man as the ‘unhappy ape’) and this is one of the features that has been a contributory factor in the success of our species, driving Homo sapiens onwards, even to the surface of the moon, in search of new terrain and challenges.
Evolutionary theories lend support for the idea of a continuum approach to the diagnosis of psychosis and provide some clues as to why the traditional Kraepelinian or categorical approaches to psychosis are beginning to prove problematic.
Van Os and Hansen interviewed 7076 men and women aged 18-64years with the Composite International Diagnostic Interview (CIDI). Of the 1237 individuals with any type of positive psychosis rating (sample prevalence 17.5%), only 26 (2.1%) had a DSM-III-R diagnosis of non-affective psychosis. 
(1) Ian K, Jenner JA, Cannon M: Psychotic symptoms in the general population – an evolutionary perspective., Br J Psychiatry. 2010 Sep;197:167-9)
(2) van Os J, Hanssen M, Bijl RV, Ravelli: Strauss (1969) revisited: a psychosis continuum in the general population?, A (Schizophr Res. 2000 Sep 29;45(1-2):11-20.)