Symptom-based models, typically operationalized through diagnostic interview, and trait models, typically operationalized via questionnaire inventories, reflect historically competing conceptions of personality disorder (PD). DSM-5 includes models of both types, in Sections II and III, respectively. In this study, we sought to synthesize these alternative conceptualizations by fitting bifactor models to data for both Section II PD symptoms (assessed using the SCID-II interview protocol) and dimensional traits for the six PDs retained in Section III (assessed using the Personality Inventory for DSM-5). Bifactor models fit the data effectively for all six PDs, and trait and symptom indicators both loaded appreciably on general factors reflecting cross-domain PD constructs. These results provide the basis for a principled, quantitative synthesis of categorical/interview and dimensional/self-report approaches to operationalizing and studying PDs, with considerable implications for diagnosis, research, and practice.