This paper is forthcoming in Philosophy and Phenomenological Research

The realism-irrealism debate in metaethics has mostly transpired with representation as its fundamental semantic notion. Realists say ethical claims get their meaning from what they represent, and some of them correctly represent the world. Irrealists accept that representation is the default semantic notion but insist that ethical claims do not represent. My aim in this paper is to begin to explore what happens to the realism-irrealism debate in metaethics when we take inference rather than representation as our master concept in philosophical semantics. More specifically, I want to consider the fortunes of the most prominent form of irrealism—expressivism—and urge that a new form of this position, which takes the distinction between theoretical and practical reasoning (rather than the distinction between representational and nonrepresentational mental states) as basic, has the resources to address one of the main objections threatening contemporary versions of the view.