<p>According to a traditional distinction in moral philosophy, the deontic and the evaluative constitute two wholly separate families of normative categories, with properties and relations such as ‘ought’, ‘ought not’, ‘required’ and ‘forbidden’ belonging to the first group, and ‘good’, ‘bad’ and ‘better than’ belonging to the second. Selim Berker offers a systematic defence of this distinction by proposing five logical features which all deontic categories have but evaluative categories lack, or vice versa. In this paper, I show how each of Berker’s five features can in fact be found among categories across the two families. Along the way I acknowledge some differences that survive scrutiny. The upshot is that the deontic and the evaluative families are much more closely interrelated than is often assumed. I then offer a new account of the relationship which can make sense of these results: the similarities and differences in the logical structures of certain deontic and evaluative categories can be understood in terms of the specific ways these categories guide our reasoning about normative matters.</p>