Let me turn, finally, to Wilhelm’s constructive effort to defend non-cognitivism. Non-cognitivists claim that moral judgments are not like ordinary assertions; they cannot be assessed as true or false, but rather merely express the speakers’ attitudes and commendations. If so, I ask, why do we express moral judgments as assertions? This is a familiar challenge. In my discussion, I merely I point out that they can be backed up by empirical data. Wilhelm has a two-part reply. First, he observes that, for non-cognitivists, the primary function of moral discourse is to persuade. C. L. Stevenson, for example, says that “x is bad” does not just mean “boo to x!”; it also means “say boo to x as well!”. Second, Wilhelm makes the original and plausible suggestion that this persuasive function is most effective when it is covert. People, he notes, don’t like to be manipulated. If I explicitly exhort you to say “boo!” you may resist, because no one likes being told what to do. But if I present my attitude in the form on an assertion, you might causally take it on board, as you would if I were presenting an ordinary statement of fact.
I think Wilhelm’s proposal deserves serious exploration. Cognitivists can respond in two ways. First, they can try to show that moral discourse often occurs in contexts that don’t aim at persuasion. This might seem implausible. After all, why should we bother engaging in moral discourse if we don’t intend to persuade anyone? On closer analysis, however, it does seem that much of our moral discourse involves preaching to the choir. In political debates, for example, left wing pundits and right wing pundits engage in a lot of moral discourse, but they never seem to persuade each other. This raises the intriguing possibility that moral judgments are not primarily in the business of persuasion. An alternative possibility is that we make moral judgments to assert our identity, or express solidarity with like-minded individuals. Empirical tests might be designed to compare the persuasion model and the self-expression model.
Cognitivists might also try to resist Wilhelm’s conjecture that people do not like to be manipulated by consulting research on explicit persuasion. In defense of Wilhelm’s conjecture, there is literature suggesting that people sometimes resist explicit persuasion (e.g., Petty & Cacioppo 1979). On the other hand, resistance does not occur in all contexts. Indeed, in a consumer product context, Reinhard et al. (2006) found that, when a person is regarded as likeable (or attractive!), they become even more persuasive when they make their intent to persuade explicit. Similarly, in studies of college drinking behavior, Neighbors et al. (2008) found that injunctive norms (which explicitly reference attitudes) are effective when and only when they are expressed by members of the students’ social groups. Further work could test the effects of explicit injunctions in the moral domain.
I should underscore that I think more testing is required to settle these debates. Wilhem’s explanation for surface discourse remains viable, and we can make progress on these issues by devising new ways to test it. These are manifestly empirical issues. While I wager with the cognitivists, I grant that the case is far from closed.