Alice the Camel Has No Humps: Why We Should Only Be Quasi-Expressivists

Abstract: Quasi-expressivism assigns to conversational pragmatics those features of discourse that expressivism-proper assigns to meaning. This talk argues that we should only be quasi-expressivists, about moral/normative discourse but also any other vocabulary (e.g. 'true', 'probable', 'tasty') that substantially emulates representational terms in its embedding behavior, because expressivism's "Frege-Geach" or embedding problem remains unsolvable despite expressivists' best efforts. My case takes the form of a dilemma. Expressivists can either opt for a single-attitude strategy, or a dual-attitude strategy. I'll argue that the problem for single-attitude expressivism is not (as others have suggested) its commitment to indefinitely many distinct attitude-types, but rather the kinds of some of these required attitude-types. Whereas dual-attitude strategies escape the Frege-Geach problems, I argue that they achieve this only by abandoning the central insights motivating expressivism, which are better captured by quasi-expressivism.