It is a suggestive feature of English and other languages that an indicative sentence such as 'Premarital sex is wrong' can be described not only as an expression of the belief that premarital sex is wrong, but also as an expression of disapproval of premarital sex. Disapproval is plausibly regarded as an attitude that is distinct from belief, in that it does not have truth conditions. What sort of attitude, then, should we take 'Premarital sex is wrong' to express: disapproval, belief, or perhaps both? One group of contemporary philosophers advocates the first option. They hold that evaluative claims serve essentially to express positive and negative attitudes that are more like desires than beliefs, and that cannot be said to be true or false — at least in the robust way in which claims about the ages of trees (for example) can be true or false. Call these philosophers 'expressivists.' 2 Seemingly opposed to expressivists are those who hold [End Page 459]that evaluative claims express beliefs, and can be true or false. Call these philosophers 'realists.' 3 A third option has recently been suggested by David Copp, according to which moral claims express both moral beliefs and a certain kind of approval. 4 Copp calls this option 'realist-expressivism,' but for reasons of terminological convenience we can call it 'the hybrid view.'
In arguing against expressivists, the analogy between color and value has seemed useful to many realists. This is true despite debates that go on amongst color-theorists regarding both the ontological status of colors and the objectivity of color claims. 5 For if a realist could manage to show that the claim that capital punishment is wrong is just as reasonably regarded as objectively true as the claim that snow is white, he might well regard his work as done. He need pay little regard to the philosopher who subsequently asks how reasonable it is to regard it as objectively true that snow is white. Against this sort of defense of realism, some philosophers have explicitly attacked the analogy between color and value. 6 The purpose of this paper is to suggest that the analogy can actually serve to reconcile expressivists and realists, at least for some very central normative terms, yielding a view that shares some features with Copp's hybrid view. 7 However, unlike Copp's view, the view developed here makes the relevant evaluative discourse much more like normal descriptive discourse. Moreover, while it is essential to expressivism that a sincere and competent speaker who makes an evaluative claim has a certain pro- or con-attitude, and while Copp's view involves the idea that something needs explaining whenever such a speaker does [End Page 460]not have the relevant attitude, the proposal advocated here does not have this strong implication. 8 Rather, many evaluative claims made in the absence of the relevant pro or con attitude need be no more exceptional than color claims based on testimony or memory.
I Agreement, Ostension and Reference
We human beings are so constituted that that we have more or less the same phenomenological responses to variously colored objects seen under the same circumstances. Even those who worry about inverted spectra have to concede that human beings make virtually the same similarity judgments regarding the colors of different objects. 9 And this agreement in similarity judgments is all that is needed for a community of language speakers to make use of ostensive methods to teach language learners the meanings of common color words such as 'green' and 'orange.' For it is this agreement that allows for the following two things. First, we can be pretty sure that the overwhelming majority of language learners will go on in the same way as we do after the requisite number of ostensive uses have been made. Second, those few humans who are not set up to go on in the same way (color-blind people, for example) will be corrected when they go on in the wrong way. Perhaps they will never, therefore, be able to apply color words correctly simply by looking at objects. But they will learn that there are acceptable and unacceptable way...