I outline how the DSM and ICD fail to describe important aspects of individual people but there are still advantages to employing abstract and idealised diagnoses. Descriptions of individual people are less abstract and idealised than psychiatric diagnoses but knowledge of individual people is still conceptually constituted. We will need very different types of concepts to understand individual people compared to concepts used to formulate psychiatric diagnoses. Finally, I suggest that metaphysics and science do not give us good reason to see psychiatric diagnoses as part of identity.