The article presents the major findings from a research project on the social and symbolic function of evaluations of education. The aim and perspective of the research project was to formulate a social theory about evaluations. The project was documented in the author's dissertation (Nørholm 2008a). The original article Danish article consisted of an edited version of the oral presentation of the project, published in Danish as Nørholm (2008b). The analyses draw upon theories by Durkheim (art or practical practice, practical theory, science or theoretical theory respectively), Bourdieu (the notions of habitus, capital, field and the theories from Reproduction), Ulf P. Lundgren (a scientific theory of evaluations) and Staf Callewaert (pedagogy as science). The empirical material consists of eight evaluation reports regarding the educations/formal training programmes for the (semi-)professions within a medical field, of different text about evaluations, and of dominant examples of Danish normal evaluation research. The analysed evaluations were conducted by representatives at all hierarchical levels, and was implicitly relating to evaluation research performed by a group of researchers/experts allocated seemingly external to the representatives. In a typically meritocratic vision and according to a New- Public-Management model, the evaluations and the evaluation research presuppose that in contemporary society, performances e.g. in school are only legitimate, they only exist, when they are evaluated. Any other performance or effort is illegitimate, non-existent. The premise being that the directions formulated in the purposive evaluation research are realized. The social and symbolic function of evaluations of educations is constructed as a societal tool for social sorting, reproducing an already existing arbitrary social order. Normal evaluation research is constructed as a symbolically necessary part of the scientific object evaluation of education. Evaluations of education, an accompanying normal evaluation research and the evaluated educations are constructed as parts of a societal tool for social sorting, for reproducing an already existing arbitrary social order. The legitimacy of the evaluations is borrowed from a research constituting a nonautonomous part of a political-administrative field, the field of power, despite 2 the fact that it is administratively allocated externally related to the evaluations and to the evaluand.