The validity of claims made about 'schizophrenia' (that it is a scientific concept and that it refers to a syndrome) was assessed in three analyses: of the writings of those said first to have described schizophrenia; of the development of rules for inferring schizophrenia and of 'genetic' research said to be central to the concept's theoretical network. Four major conclusions were drawn: 1. That there was no evidence to support the original introduction of 'schizophrenia', but good evidence that some of the population from which it was derived were suffering from a later-identified neurological disorder. 2. That the rules for inferring schizophrenia have been developed in a manner quite different from the development of concepts in the empirical sciences and in medicine. 3. That there is no evidence that the rules set out in DSM-111 refer either to a syndrome or to any pattern of phenomena and 4. that 'genetic' research has been seriously misrepresented in secondary sources and does not support 'schizophrenia'. A number of factors were discussed as possibly important in explaining the continued use and influence of ‘schizophrenia'. These included the use of popular but fallacious types of argument to defend the concept, the functions it apparently serves for psychiatry and the public, the perceived primacy of biological or dispositional explanations of behaviour and the habits of 'seeing' patterns in unrelated phenomena, of inferring before describing, of reifying constructs and of confusing observation and inference. Finally, the implications of abandoning 'schizophrenia' were discussed and the weak foundations of the distinctions between 'normal, and abnormal' behaviour emphasised. An alternative framework, derived from the experimental analysis of behaviour was suggested and illustrated, both for the analysis of bizarre behaviour and of the conditions under which it is seen as symptomatic of schizophrenia.