Exploiting the biological and psychological plausibility of parallel distributed processing, a neural network model of cortical processing is proposed in order to provide a framework for understanding pathological processes in schizophrenia. Major psychosocial and neurobiological factors are assigned to appropriate parameters in the model. Behaviour of the network is explored through simulation, in particular interaction between the temperature and memory loading parameters is studied. The emergent network behaviour provides a model for complex interaction between dopaminergic activity, cortical and hippocampal pathology, and psychosocial factors in the production of psychotic symptoms. This integrative model provides an account of a wide range of observations in schizophrenia. It also produces testable predictions concerning cognitive performance and the course of schizophrenia.