Background: While statistically robust, the association between psychosis and violence remains causally unexplained. Objective: To provide an overview of possible causal pathways between risk factors and violence in psychosis. Methods: A structured narrative review of relevant studies published between 1990 and 2013, found via online databases and bibliographies. Both reviews and empirical studies were included. No restrictions were applied to language, study design, sample characteristics and measurement of psychosis and violence. Case reports and studies about self-harm were excluded. A final sample of 69 studies was used. Findings: The lack of knowledge regarding the causal relationship between psychosis and violence is partially due to methodological aspects of research. These aspects include study design, sampling, operationalization and confounding variables. Moreover, violence is the potential outcome of several interrelated risk factors: demographics, social factors, persecutory delusions, command hallucinations, comorbid antisocial personality pathology, substance use, inadequate insight, treatment non-adherence and physiological factors. Forty-one possible causal pathways between these risk factors and violence are presented. Conclusions: This study stimulates research by providing a theoretical framework, avenues for future investigation and methodological recommendations. Understanding violence in psychosis enhances its prevention and treatment, decreases stigma associated with psychosis and improves the patient's legal position.