How non-epistemic values can be epistemically beneficial in scientific classification

Abstract

The boundaries of social categories are frequently altered to serve normative projects, such as social reform. Griffiths and Khalidi argue that the value-driven modification of categories diminishes the epistemic value of social categories. I argue that concerns over value-modified categories are an endorsement of problematic assumptions of the value-free ideal of science. Contrary to those concerns, non-epistemic value considerations can increase the epistemic success of a scientific category. For example, the early history of the category infantile autism shows how non-epistemic value considerations can contribute to delimiting and establishing infantile autism as a distinct category in mainstream psychiatry. In the case of infantile autism, non-epistemic considerations have led to a new interpretation of existing data, the expansion of research to include biology, and the creation of diagnostic criteria that further contribute to collecting relevant data. Given this case study, we see that non-epistemic considerations may not be epistemically detrimental but can be epistemically beneficial in scientific classification.