In most areas of medicine, the physician's primary task is to diagnose the patient's presenting problem by correctly identifying the underlying pathology causing that problem. This goal is achieved by correlating the patient's symptoms with objective evidence of the presence of disease derived from the physical examination and test results. The quality of diagnostic reasoning is evaluated through the accuracy of the diagnosis made—that is, good reasoning leads to the right diagnosis. Because there are no physical examination signs or test results that indicate the underlying pathology of most mental illnesses, it does not seem that, in reasoning toward diagnosis, the psychiatrist can get the 'right answer' about what is causing the patient's problems. Instead, psychiatric diagnostic reasoning is more like sketching a roadmap that will enable the understanding of those problems. In this article, we argue that psychiatrists must use intersubjective strategies of knowing to diagnose their patients. First, we discuss some of the main approaches to diagnostic reasoning in medicine, pointing out that these models do not capture intersubjective strategies of knowing. Second, we argue that intersubjective knowing is necessary for psychiatric diagnostic reasoning, including diagnosing that is strictly criteriological. Finally, we contend that because intersubjective knowing can yield diagnostically relevant evidence, it ought to be included in models of psychiatric diagnostic reasoning.