OBJECTIVE: This study examines the comorbidity of DSM-III-R borderline personality disorder and the other axis II personality disorders. The extent and direction of overlap provides a measure of the clarity of its diagnostic boundaries and descriptive validity. METHOD: In 110 outpatients without concurrent major axis I conditions, axis II diagnoses were assessed in semistructured format and all DSM-III-R personality disorder criteria were rated. Multiple diagnoses were recorded. RESULTS: Twenty-two patients (20%) met criteria for borderline personality disorder; 18 (82%) had at least one additional personality disorder diagnosis. Using measures of frequencies and intercorrelation coefficients, the authors found that overlap was extensive and not confined to any one of the three designated axis II clusters. Factor analysis revealed 1) a group containing borderline personality disorder with paranoid, histrionic, narcissistic, antisocial, and passive-aggressive personality disorders and 2) another grouping of schizoid, schizotypal, avoidant, obsessive-compulsive, and self-defeating personality disorders. CONCLUSIONS: Borderline personality disorder appears to constitute a broad, heterogeneous category with unclear boundaries that embraces a general personality disorder concept. Both further refinement of the borderline personality disorder construct and investigation into alternative models to the DSM- III-R axis II classification system are suggested.Abstract Teaser