Medical professionals, including mental health professionals, largely agree that moral judgment should be kept out of clinical settings. The rationale is simple: moral judgment has the capacity to impair clinical judgment in ways that could harm the patient. However, when the patient is suffering from a "Cluster B" personality disorder, keeping moral judgment out of the clinic might appear impossible, not only in practice but also in theory. For the diagnostic criteria associated with these particular disorders (Antisocial, Borderline, Histrionic, Narcissistic) (...) are expressed in overtly moral language. I consider three proposals for dealing with this problem. The first is to eliminate the Cluster B disorders from the DSM on the grounds that they are moral, rather than mental, disorders. The second is to replace the morally laden language of the diagnostic criteria with morally neutral language. The third is to disambiguate the notion of moral judgment so as to respect the distinction between having morally disvalued traits and having moral responsibility for those traits. Sensitivity to this distinction enables the clinician, at least in theory, to employ morally laden diagnostic criteria without adopting the sort of morally judgmental (and potentially harmful) attitude that results from the tacit presumption of moral responsibility. I argue against the first two proposals and in favor of the third. In doing so, I appeal to Grice's distinction between conventional and conversational implicature. I close with a few brief remarks on the irony of retaining overtly moral language in an ostensibly medical manual for the diagnosis of mental disorders. (shrink)