The evaluation of research in human and social sciences has been subject to intense criticism in recent years. Several authors argue that research in these fields is radically incommensurable and should not be subject to any kind of measurement. Based on Bourdieu and Foucault, these contributions suggest that research evaluation is a subtle device for controlling ideas and reducing the potential for dissent and radical critical work on society, under an apparently objective and neutral scientific methodology. The paper takes issue with these reconstructions and suggests an alternative theoretical path. First, commensuration is integral part of modernity. The effort to make social reality more measurable has historically been part of a process of emancipation, trying to reduce the power of traditional, implicit and opaque sources of knowledge. Second, scientific communities can reflexively find an agreement on what is quality of research and which are the criteria for its identification in practice. Third, while there is no question that research is incommensurable by nature, yet it is possible to transform purely qualitative judgments by experts into measures that can be subject to inter-subjective comparisons. Based on several streams of literature in sociology, the paper suggests that the tensions induced by evaluation are potentially beneficial to human and social sciences.