Impact and evaluation of research in the humanities has been the subject of several recent publications (e.g. Benneworth et al. 2016, Ochsner et al. 2016, Severinson 2017). A common theme is the need for a re-evaluation of how the humanities are evaluated, as well as for research on the effects of evaluation practices on the humanities (Ochsner et al. 2016, p. 9). Another important idea is that this re-evaluation should consider the nature of the public value of the humanities (Benneworth et al. 2016, especially ch. 7). The dominant model of research evaluation has as its focus societal economic benefits of research, a focus humanities scholars have mostly been uncomfortable with (Benneworth et al., ch. 2). One of the main tools of research evaluation, bibliometric analysis, is problematic – to say the least – for the humanities (Ochsner et al. 2016). As the policy and practice of research evaluations influences the way research is conducted (van Leeuwen 2016, Severinson 2017) there is a potential danger that the inherent value of the humanities will be eroded and their value to society diminished through evaluation. This problem will be addressed from two sides in this paper. The a theory of societal capacities in Benneworth et al. (2016) and virtue epistemology (particularly as developed by Linda Zagzebski).