Abstract
Thick concepts provide a focal point for several important issues in ethical theory. Separatists argue that the descriptive and evaluative elements of a thick concept can be separated out. Non-separatists deny this and claim that there are no descriptive boundaries delimiting a thick concept. A common strategy for both camps in the debate has been an appeal to armchair intuitions of various everyday thick concepts. My alternative strategy consists in a closer study of the professional practice of risk analysis. As a well-developed practice, it provides substantial material for analysis. Moreover, its central concepts of risk and safety are typically seen as scientific concepts fitting the separatist analysis. Still, I argue that there are several evaluative aspects in risk and safety ascription that are hard to account for on a separatist analysis. I consider three separatist strategies, and conclude that they all fail. The result is a corroboration of the general non-separatist thesis put forward by theorists such as John McDowell and Bernard Williams.