I recently made a post to Twitter to explain why the simplest kind of moral subjectivism necessarily collapses into emotivism or moral non-cognitivism. The basic argument is that one cannot coherently reduce “P is wrong” to a mere opinion or belief that “P is wrong” because, opinions and beliefs are propositional attitudes towards a proposition, in this case, the proposition “P is wrong,” which, if it is proposition, must necessarily have a truth-value. If it does not have a truth value, it is senseless to claim that the analysis of such judgements is to say that someone has a view about what its truth value is.
This point is made very cogently by philosopher David S. Oderberg in his magisterial Moral Theory: A Nonconsequentialist Approach, which I highly recommend to all my readers, as well as its companion volume Applied Ethics: A Nonconsequentialist Approach.
I here reproduce Oderberg’s analysis from Chapter 1 of his book, specially about the semantic objection to moral relativism—a case originally made by philosopher Peter Geach, whose work I also highly recommend to anyone interested:
Perhaps the most widespread form of relativism, again deriving from the philosophy of David Hume, is what I shall call personal relativism, more usually called subjectivism. The central claim of personal relativism is that the truth or falsity (truth value) of moral statements varies from person to person, since morality is merely a matter of opinion. Now there are various ways in which subjectivists have elaborated this basic thought, developing more or less sophisticated semantic theories linking moral judgements with statements of opinion. It is impossible to look at them all, but since the sorts of objection I will raise can be applied in modified form to different versions, let us take just one kind of subjectivist theory. It is one of the more simple varieties, and while many philosophers would say it was too simple, it also happens to be the sort of subjectivism that the vast majority of students of moral philosophy believe; and it is an approach that many will continue to believe even after they have finished studying philosophy!
According to this version of subjectivism, there is no objective truth to the statement, for instance, ‘Child abuse is wrong’: all that a person is entitled to claim is something equivalent to ‘I disapprove of child abuse.’ Instead of saying ‘I disapprove of child abuse’, Alan may say ‘Child abuse is wrong for me’, or ‘Child abuse is wrong from my subjective viewpoint’, but he is not then allowed to say ‘Child abuse is wrong, pure and simple’, since it might be right from Brian’s subjective viewpoint—he will say ‘Child abuse is right for me, though it is wrong for Alan, who personally disapproves of it.’ Generally speaking, moral judgements can never be considered apart from the question of who makes them. A moral judgement, ‘X is wrong’, made by a person P, can only be assessed for truth or falsity by relativizing it to P: the subjectivist says that ‘X is wrong’, uttered by P, is equivalent in meaning to ‘I disapprove of X’ uttered by P. If an observer were to report on P’s opinion, he would say, ‘X is wrong for P, or as far as P in concerned; in other words, P disapproves of it.’ But the observer can still say, ‘However, I personally approve of it, so “X is wrong” is not true for me.’
For the subjectivist, to claim that there is a fact about the morality of child abuse, which transcends mere personal opinion, is a philosophical mistake. Certainly, there are facts about what is wrong for Alan, right for Brian, and so on. These facts are genuine—they are reports of the opinions (or ‘sentiments’ to use Hume’s term) of individual moral judges—but since each judge makes law only for himself, he cannot impose his view of things on others. For the subjectivist, once the facts are in concerning the moral opinions of those engaged in a disagreement, there is no room for further argument. More accurately, there might be room for argument over other facts: Alan might claim ‘I approve of child abuse’ because he does not know the psychological damage it does to children. Had he known, he would have claimed ‘I disapprove of child abuse.’; and another person might change Alan’s mind by pointing out the relevant facts. But what the personal relativist holds is that as long as there is no dispute over the facts, two people can make opposing claims about the morality of a certain action or type of behavior with no room left for rational dispute. That have, as it were, reached bedrock.
As was said, the version of subjectivism just outlined is a simple one and all sorts of refinements can be added. Still, it is the view held by very many philosophy students, not to say quite a few philosophers (and certainly vast numbers of the general population), and should be assessed in that light. Further, as was also noted, the general kinds of observations that can be raised against it apply to the more sophisticated versions. We can only consider a few devastating objections here, but it should be noted that the validity of any one on its own is enough to refute subjectivism, whatever the strength of the others. Given the weight of all the objections, however, it is surprising that personal relativism is so widely held.
First there is the semantic problem: A proposition of the form ‘Doing X is wrong’ uttered by P (for some action or type of behavior X and some person P) is, according to the personal relativist, supposed to mean no more nor less than ‘P disapproves of doing X’: the latter statement is claimed to give the meaning or analysis of the former. But ‘P disapproves of doing X’ cannot, on this analysis, be equivalent to ‘P believes that doing X is wrong’, since ‘Doing X is wrong’ is precisely what the relativist seeks to give the meaning of; in which case the analysis would be circular. On the other hand, the relativist might again analyze the embedded sentence ‘Doing X is wrong’ in ‘P believes that doing X is wrong’ as ‘P believes that doing X is wrong’, and so on, for every embedded occurrence of ‘Doing X is wrong’, thus ending up with an infinite regress: ‘P believes that P believes that P believes . . . that doing X is wrong.’ This, of course, would he no analysis at all, being both infinite and leaving a proposition of the form ‘Doing X is wrong’ unanalyzed at every stage.
Such an obvious difficulty might make one wonder that any relativist should support such a way of trying to analyze ‘Doing X is wrong’; but if he is committed to the idea that morality is a matter of opinion or personal belief, it seems that he tacitly invokes just such a pseudo-analysis. The only other route the relativist can take is to assert that ‘P disapproves of doing X’ needs no further gloss: it is a brute statement of disapproval that does not itself invoke the concept of wrongness (or rightness, goodness and the like). But then personal relativism collapses into emotivism, the theory that moral statements are just expressions of feeling or emotion and only appear to have the form of judgements that can he true or false. Emotivism is a different theory from relativism, however, and more will he said about it in the next section. Unless the personal relativist can give an analysis of disapproval that is neither circular, nor infinitely regressive, nor collapses his theory into emotivism, he is in severe difficulty; and it is hard to see just what such an analysis would look like.
David S. Oderberg, Moral Theory: A Nonconsequentialist Approach; Blackwell: Oxford, 2000; pp. 16-18.
In case it adds something, I will also append the write-up I made for Twitter: