Voice hearing (VH) is often regarded as pathognomic for schizophrenia. The purpose of this article is to review and integrate historical, clinical, epidemiological, and phenomenological evidence in order to suggest that VH may be more appropriately understood as a dissociative rather than a psychotic phenomenon. First, we discuss the lifetime prevalence of VH in the general population, which is estimated to range between 1% and 16% for adult nonclinical populations and 2% and 41% in healthy adolescent samples. Second, we demonstrate how the ubiquity of VH phenomenology, including variables like voice location, content, and frequency, limits its diagnostic and prognostic utility for differentiating psychotic from trauma-spectrum and nonclinical populations. Finally, we report on the empirical associations between VH, measures of dissociation, and trauma particularly (though not exclusively) childhood sexual abuse. There are 2 main conclusions from this review. First, we argue that available evidence suggests that VH experiences, including those in the context of psychotic disorders, can be most appropriately understood as dissociated or disowned components of the self (or self-other relationships) that result from trauma, loss, or other interpersonal stressors. Second, we provide a rationale for clinicians to use psychotherapeutic methods for integrating life events as precipitating and/or maintaining factors for distressing voices. Potential mechanisms for the relationship between trauma, dissociation, VH, and clinical diagnosis are described, including the relevance of literature from the field of attachment in providing a diathesis for dissociation. Suggestions for future research are also discussed. (Contains 4 footnotes, 5 tables, and 1 figure.)