Validity is one of the most debated constructs in our field; debates abound about what is legitimate and what is not, and the word continues to be used in ways that are explicitly disavowed by current practice guidelines. The resultant tensions have not been well characterized, yet their existence suggests that different uses may maintain some value for the user that needs to be better understood. We conducted an empirical form of Discourse Analysis to document the multiple ways in which validity is described, understood, and used in the health professions education field. We created and analyzed an archive of texts identified from multiple sources, including formal databases such as PubMED, ERIC and PsycINFO as well as the authors' personal assessment libraries. An iterative analytic process was used to identify, discuss, and characterize emerging discourses about validity. Three discourses of validity were identified. Validity as a "test characteristic" is underpinned by the notion that validity is an intrinsic property of a tool and could, therefore, be seen as content and context independent. Validity as an argument-based "evidentiary-chain" emphasizes the importance of supporting the interpretation of assessment results with ongoing analysis such that validity does not belong to the tool/instrument itself. The emphasis is on process-based validation (emphasizing the journey instead of the goal). Validity as a "social imperative" foregrounds the consequences of assessment at the individual and societal levels, be they positive or negative. The existence of different discourses may explain--in part--results observed in recent systematic reviews that highlighted discrepancies and tensions between recommendations for practice and the validation practices that are actually adopted and reported. Some of these practices, despite contravening accepted validation "guidelines", may nevertheless respond to different and somewhat unarticulated needs within health professional education.