The Triviality Of The Debate Over "Is-Ought" And The Definition Of "Moral"

Peter Singer

"THE central problem in moral philosophy is commonly known as the is-ought problem." So runs the opening sentence of the introduction to a recent volume of readings on this issue.¹ Taken as a statement about the preoccupations of moral philosophers of the present century, we can accept this assertion. The problem of how statements of fact are related to moral judgments has dominated recent moral philosophy. Associated with this problem is another, which has also been given considerable attention - the question of how morality is to be defined. The two issues are linked, since some definitions of morality allow us to move from statements of fact to moral judgments, while others do not. In this article I shall take the two issues together, and try to show that they do not merit the amount of attention they have been given. I shall argue that the differences between the contending parties are terminological, and that there are various possible terminologies, none of which has, on balance, any great advantage over any other terminology. So instead of continuing to regard these issues as central, moral Philosophers could, I believe, "agree to disagree" about the "is-ought" problem, and about the definition of morality, provided only that everyone was careful to stipulate how he was using the term "moral" and was aware of the implications and limitations of the definition he was using. Moral philosophers could then move on to consider more important issues.

It has long been a commonplace in the debate about the definition of morality that moral terms are used in many different ways at different times and by different eople. The search for a definition, therefore, is not a search for the one true definition which expresses all that anyone has ever meant by the term. On the contrary, the search has been for the best definition, the definition that will express the most important or the most useful of the various meanings that moral terms have in ordinary speech. We are, then, to some extent at liberty to choose our definition from among the various definitions which could be given some justification from ordinary usage. On what basis should we make this choice? Provided a definition is not too much at odds with ordinary usage, so that it is still a definition of morality, and not of some newly invented concept, it seems to be generally agreed that the consequences for practical discourse of a particular definition are crucial. The kind of criticisms that are made of particular definitions bear this out. The emotivists, for example, were criticized for making morality a matter of personal whim, in which reason has no part to play, while they accused their opponents of restricting the possibilities of moral discourse by inserting their own values into the definition of morality. The real basis for choosing or rejecting a definition was, and is: how would that definition affect practical disputes? Will it allow sufficient scope, so that all dissenting opinions can be subject to discussion, without being ruled out from the start? Will it allow a sufficient role to reason, so that agreed principles of conduct can emerge, provided only that men are rational? The proponents of different definitions, some of which allow the move from "is" to "ought," and some of which do not, assume that it does make a difference which definition of morality is accepted. This assumption needs to be examined.

I shall begin by considering two possible views on the meaning of the moral "ought," and its relation to matters of fact. These two positions are at opposite ends of the spectrum of positions which can be taken on this issue.

The first position has often been called "subjectivism" but this is a misleading term and, extravagant as it may seem to invent yet another label, I shall call it "form-and-content neutralism," or, for brevity, just "neutralism," because the chief characteristic of this position is its complete neutrality about both the form and the content of moral principles. According to this position, there are no limits on the kind of principle which can be held as a moral principle. A moral principle can have any content whatsoever - that is, it is not restricted to a certain kind of subject matter, like suffering and happiness, or the satisfaction of wants, needs, or interests. "Clasp your hands three times every hour" might seem to have no connection with anything we value; nevertheless it could, on the neutralist view, be a moral principle. There are also no restrictions on the form a moral principle can take, so long as it is intelligible, not self-contradictory, and so on. By this I mean that to count as a moral principle, a principle does not have to satisfy any of the formal requirements that have sometimes been proposed, such as being able to be willed as a universal law, being acceptable to an impartial observer, being able to be formulated without the use of proper names. personal pronouns, or other singular terms. For example, the principle of pure egoism: "Everyone ought to do what is in my interest" fails the test of universalizability as propounded by R. M. Hare because it contains the singular term "my." According to the neutralist, however, this does not preclude it from being held as a moral principle.

Of course, there must be some way in which even the neutralist distinguishes a man's moral principles from other principles which he may hold. For the neutralist, a man's moral principles are the principles, whatever they may be, which that man takes to be overriding. This is made true by definition. In support of the definition, thy neutralist can refer to usages like "They gave up everything for Art; Art was their morality" or: "His morality was just egoism, for he cared about nothing but himself."

The neutralist view, then, is that whether a principle is a moral principle for a particular person is determined solely by whether that person allows the principle to override any other principles which he may hold. Any principle at all is capable of being a moral principle for a person, if that person should take it as overriding.²

The strength of this account of morality is that it provides a very close logical connection between the moral principles a man holds and the way he acts. It is often said that we can tell what a man's moral principles are by observing how he acts. Certainly we feel uneasy in saying of a man who owns three cars that his moral principles dictate that every rich man should give whatever he can to the poor. The neutralist is able to explain why, if man acts on the basis of a coherent set of principles at all, he will act in accordance with his moral principles. If a man recognizes that a certain action is prescribed by his overriding principles, he surely will do that action, if he can. There are, of course, instances in which a man does not do what is in accord with his own overriding principles, but in these cases we look for some explanation, such as succumbing to temptation, kleptomania, addiction, or whatever the case may be. This is the sort of explanation that we look for when a man does not act rationally, and it is clear that behavior that has to be explained in this way does not count against saying that the man who acts against his overriding principles has all the reason that he could possibly have for doing the opposite.

The significance of the sort of tie between moral principles and action which the neutralist account provides can be seen in a situation of the following sort: Jack is trying to persuade Bill to do something which he, Jack, thinks would be a morally good thing to do - say, give money to famine relief. If Jack can convince Bill that giving the money is in accordance with Bill's own moral principles, then, so long as Bill continues to act on his own overriding principles, that is, so long as he continues to make rational decisions about what to do, and carries out these decisions, he will give money to famine relief. There is no step between accepting the view that an action is in accord with his moral principles and the decision to do that action.

On the other hand, the neutralist position has a less welcome consequence, which can also be illustrated by the above example. Whatever facts Jack mentions in his attempt to convince Bill that giving to famine relief is demanded by Bill's own moral principles, Bill can always retort that no moral conclusions follow from these facts. Jack might, for instance, refer to the immense amount of suffering that will be relieved by a substantial donation, and compare this with the relatively trivial difference in pleasure that Bill will get if he uses the money to buy a Mercedes rather than a Morris. Bill could reply that according to his moral principles, suffering matters only when experienced by himself, his family, or friends. Jack would have no effective counter to this reply - he could not argue, for example, that Bill would not be prepared to accept his own moral principles if he were one of the victims of a famine. Bill could reject this argument as simply irrelevant, since there is no reason why a moral principle should conform to such a requirement. This the neutralist must accept. So the only role which reason has to play in the choice of moral principles is that role which it has in all areas of life, associated with correct beliefs about what is the case, what ends are attainable, the best means to them, and so on. When any principle at all can be a moral principle, there can be no special kind of reasoning applicable to moral principles.

At the other end of the meta-ethical spectrum there is the kind of position which R. M. Hare has termed "descriptivism." It is often called "naturalism," but since the intuitionist account of morality cannot really be called naturalist, but may, for our purposes, be grouped together with naturalist theories, I shall use Hare's term. Descriptivism is the direct opposite of neutralism in that for a principle to be a moral principle, as the descriptivist defines the term, it must satisfy criteria of both form and content. Thus, to give just one example of the many possible forms of descriptivism, it might be said that moral judgments are logically tied to suffering and happiness, impartially assessed. In other words, a judgment is not a moral judgment unless it is somehow connected to suffering and happiness, and a judgment is also not a moral judgment unless it is an impartial judgment, in the sense that it does not arbitrarily place more importance on the suffering and happiness of a particular person or group of persons than on the suffering and happiness of any others.

The strength of the descriptivist view is that once the definition of morality is accepted, watertight reasoning from statements of fact to moral conclusions is possible. This means that (to continue with the example used earlier) from the fact that Bill's money will reduce suffering and increase happiness to a greater extent if given to famine relief than if spent on a Mercedes, Jack can argue that Bin ought, morally, to give the money to famine relief rather than buy the more expensive car. If the descriptivist is right in tying morality to suffering and happiness, impartially assessed, Bill has no way of resisting this argument, for the conclusion follows deductively from the definition of morality and the facts of the case. Of course, most descriptivist views are not as straightforward as the example I have given, and so the argument may be more complicated. Nevertheless, to the extent that the descriptivist gives definite form and content to his definition of morality, he is able to show that reason has a prominent role in moral argument, and that moral arguments are objective.

Unfortunately the significance of this conclusion is reduced by the fact that descriptivism, strong where neutralism is weak, is also weak where neutralism is strong. To show that an action is required by a moral principle does not, if the descriptivist view is accepted, have the consequences it would have if moral principles were necessarily overriding. We are not, on the descriptivist view, free to form our own opinion about what is and what is not a moral principle; but we are free to refuse to concern ourselves about moral principles. Bill has to grant that if morality is tied to suffering and happiness, it follows that he is morally obliged to give to famine relief, but he may say that if that is what morality is about, he is not interested in acting according to moral principles. The descriptivist cannot tie morality to action, as the neutralist did, because he has tied it to form and content. So morality may become irrelevant to the practical problem of what to do.

We have, then, two quite opposed views about the meaning of moral judgments and their relations to statements of fact. These two views differ as much as any of the contending views that have been put forward in the "is-ought" controversy. Yet on the issue of how statements of fact are connected with reasons for acting in general (and not just moral reasons for acting) neutralism and descriptivism do not differ at all. To go from the statement of fact: "Giving money to famine relief will reduce suffering and increase happiness to a greater extent than spending the money on a more expensive car" to the practical conclusion of giving the money away is neither easier nor more difficult if we adopt one position rather than the other. The arguments Which we might use are, in fact, substantially the same in either case, although the way we express them may differ. Thus if a person accepts, on the basis of an argument from a descriptivist definition of morality, that morally he ought to give to famine relief, but asks what reason there is for taking any notice of morality, we may answer by appealing to the feelings of sympathy and benevolence which, in common with most of mankind, he probably has to some extent. We may talk of the fulfillment and real happiness that can come through knowing that one has done what one can to make the world a little better, and contrast this with the disappointments and ultimate sense of futility which are likely to come from a self-centered existence devoted to nothing but selfish concerns. We might mention the value of friendship between open people who respect each other, a kind of friendship impossible for the narrowly egoistical man or woman. These are just some of the considerations we might mention, and they may or may not be valid reasons for leading a life which the descriptivist would say was morally good. Whether these are valid reasons is not my concern here; it might depend on the person to whom they are addressed. My point is that the neutralist could use exactly the same reasons in an attempt to persuade the man whose overriding, that is, moral, principles take no account of the happiness or suffering of people other than himself, his family, and friends, to widen his area of concern, and so, perhaps, to adopt principles which would involve giving to famine relief.³

I hope it is now clear that the issue that really matters, that is of practical significance, is how statements of fact are connected with reasons for acting, and not how statements of fact are connected with moral judgments. The latter question is encompassed by the former. To hold, as the neutralist does, that action follows from moral judgments but moral judgments do not follow from facts, is to place morality close to the "action" side of the "fact-action" or "reason-action" gap, while to hold, as the descriptivist does, that moral judgments follow from facts but action does not follow from moral judgments, is to place morality on the opposite side. The dispute between the neutralist and the descriptivist, therefore, is a dispute about where, within a limited framework, morality shall be placed. Since nothing of any practical significance hangs on the placing of this term within this framework - the prospects for going from facts to action are the same in either case - the dispute is merely terminological.⁴

I said earlier that neutralism and descriptivism are at opposite ends of the spectrum of positions which can be taken over the "is-ought" issue. If this is true, it would mean that other positions differ from each other, and from the two extreme positions, even less than the two just discussed, so that if the difference between neutralism and descriptivism is of no real importance, the differences between any other positions will be no more significant. Yet it might be thought that a middle position between neutralism and descriptivism could combine the strengths and avoid the weaknesses of the extreme positions - that is, that it could preserve the tie between action and morality as well as the tie between fact and morality. If this could be done, such a position would differ significantly from both neutralism and descriptivism, for it would eliminate the gap between facts and action.

The difficulty which must be faced by any attempt to combine the advantages of neutralism and descriptivism without their disadvantages should be obvious from what has already been said. If we insist that the principles a person acts upon are, by definition, his moral principles, we must recognize that people can act on all sorts of principles without committing any logical error. How can we claim a necessary connection between a person's moral principles and the way he acts, while simultaneously denying that some conceivable principles of action can be moral principles?

Let us consider one way in which it might be thought possible to deny the status "moral principle" to at least some principles of action, while preserving the tie between the way a person acts and what we regard as his moral principles. It might be thought that one can maintain that moral principles are, by definition, prescriptive, so that to assent to a moral principle is to commit oneself to acting upon it when it is appropriate to do so, and at the same time maintain that, while a moral principle can have any content whatsoever, it must satisfy the formal requirement of universalizability.⁵ Does this position overcome the difficulty I mentioned?

First, what advantage does such a position have over ordinary neutralism? It will be remembered that the position I have been referring to as neutralism was neutral as to both the form and the content a moral principle could have. The position we are now considering is neutral as to content, but not as to form. This means that there are some principles on which people might act which are not moral principles. This restriction on the form a moral principle can take makes it possible to develop a more effective form of moral argument than was possible with form-and-content neutralism. To illustrate, using the same example as before: Jack can ask Bill how he would feel about other people indulging in luxuries while he starved to death. So long as Bill continues to claim that he acts on moral principles, he cannot deny the relevance of this question, for if he cannot prescribe universally that people who feel inclined to spend their money on luxuries should do so, even when others (including, perhaps, Bill himself) are starving, then he cannot defend his purchase of the Mercedes on moral grounds. This argument is not quite as watertight as the direct argument from a descriptivist definition of "moral," for it is at least logically possible that Bill will reply that, according to his moral principles, suffering, whether his own or anyone else's, is not very important. He might say that the ideal of "standing on one's own two feet," and not being dependent on anyone else, is more important than suffering, or there may be some other ideal for which he would be prepared to suffer, and thinks others should, if necessary, suffer for too. Nevertheless, as a matter of fact, few people are prepared to starve to death for an ideal, and so the universalization argument does provide a means of linking moral conclusions with statements about suffering, happiness, and other matters which may, with some reservations, be termed "factual."

The composite position, then, does allow a more powerful form of moral argument than is possible for the neutralist. Does it avoid the disadvantage of descriptivism? It would seem that it does, for by defining moral principles so that they are necessarily prescriptive, it seems to make it logically Impossible for anyone to say that, while he can see that from the moral point of view he ought to do a particular act, he is not interested in acting according to moral principles. So this position appears to have advantages over both the positions we considered earlier.

Unfortunately, this impression is not sustained on closer examination. To see this, we only have to ask what a defender of the composite position would say about a person whose overriding principle of action is non-universalizable - for instance a person who acts on the principle of pure egoism, mentioned earlier. The only thing that can be said about such a person, consistent with the composite position, is that he does not hold any moral principles at all, for there are no universalizable principles which he holds prescriptively. The restriction on the form a moral principle can take is incompatible with the view that whatever principles of conduct a person espouses are his moral principles. So, on this account of morality, it is possible to "opt out" of acting on moral principles, and thus there is still a gap between action and morality. Certainly, to opt out of morality defined in this way is not quite the same as opting out of morality as defined by a descriptivist. To opt out of morality as defined by a typical descriptivist one would have to deny that one is concerned, say, about suffering and happiness, impartially assessed. To opt out of universal prescriptive morality, one would have to deny that one is concerned to act on judgments which one is prepared to universalize. Still, it is possible to do this, and there are no doubt many people whose overriding principles are selfish ones, which they would not be prepared to universalize. If we were to try to persuade these people to act only on judgments which they are prepared to universalize, we should have to use arguments the same as or similar to those already mentioned, which the descriptivist could use in attempting to persuade people to act on moral principles as he defines them, and which the neutralist could use to get an egoist to take the interests of others into account.

Does the middle position have any advantage over descriptivism, then? It may be - depending, I think, on just what account is given of the notion of universalizability - that there are some people who are prepared to act on universalizable judgments, but not on judgments which are tied to suffering and happiness, or whatever other content the descriptivist gives to his definition of morality. For these people, on the middle position, but not on the descriptivist position, there still is a logical tie between action and morality. But is this really a gain? It is doubtful if it is. For it is precisely these people - people who are prepared to universalize their judgments but do not concern themselves with the suffering or happiness of others - who will hold ideals which allow them to remain unmoved by the "how would you like it if you were in that position?" argument. To see this requires only a little reflection. The person who, because he holds the ideal of "standing on one's own two feet" is able to resist the universalization argument and buy his Mercedes while others starve, is just the sort of person who would not be prepared to act on moral principles, if "moral" were defined in terms of suffering and happiness, but would be prepared to act on moral principles, defined so that they must be universalizable but can have any content. A person who does not hold ideals which he considers more important than suffering and happiness, and is prepared to act only on universalizable principles, could be moved by the universalization argument, but would also, presumably, be prepared to act on moral principles defined as a descriptivist might define them, in terms of suffering and happiness.

So the middle position does not seem to have any advantage over the other two positions. The only advantage it seemed to have over descriptivism was that there are some people for whom it can show a logical tie between action and morality, while descriptivism cannot, but it turns out to be just these people for whom the middle position cannot provide a tie between reason, or fact, and morality. Nor does the middle position have any real advantage over neutralism, for, as we have seen, to get an egoist to act according to moral, that is, universalizable, principles, an advocate of the middle position would have to use the same arguments that a neutralist can use to get an egoist to take into account considerations other than his own interests.

I have not, of course, considered every possible account of the meaning of moral concepts, nor every way of either overcoming or preserving the "is-ought" gap. Nevertheless, I think that the three positions I have examined are illustrative of the kinds of account that can be offered and have been offered in recent years. I have tried to show, by examining these positions, that there are limits to what any account of morality can do. No definition of morality can bridge the gap between facts and action. Nor does any one definition of morality have any important overall advantages as against the other plausible definitions that have been suggested. It follows that the disputes over the definition of morality and over the "is-ought" problem are disputes over words which raise no really significant issues.

In conclusion, I should perhaps add that I am not denying that lack of clarity about the meaning of words is an important source of error, both in philosophy and in practical argument. Confusion is likely to occur unless both parties to a discussion are sure about what the use of a word implies. It is not difficult to see how, if one did not make the way in which one was using the term "moral" clear, errors could happen. It might be thought that one could argue that moral principles must be concerned with suffering and happiness, impartially considered, and then maintain that if a person is to act on a coherent set of principles at all, he must act on moral principles, since moral principles are necessarily overriding. This would be a particularly obvious slide from the descriptivist to the neutralist definition of morality, but such moves can be better disguised. It is therefore necessary, before embarking on a discussion of morality, to make quite clear in what sense one is using terms like "moral judgment," and what follows and what does not follow from such a use of the term. This is an essential preliminary; but it is only a preliminary. My complaint is that what should be regarded as something to be got out of the way in the introduction to a work of moral philosophy has become the subject-matter of almost the whole of moral philosophy in the English-speaking world.

University College, Oxford.

⁵ This position has been advocated by R. M. Hare in The Language of Morals (Oxford, 1952) and Freedom and Reason (Oxford, 1963). See also P. H. Nowell-Smith, Ethics (Baltimore, 1954), pp. 307-309.

The Triviality Of The Debate Over "Is-Ought" And The Definition Of "Moral" has been published in American Philosophical Quarterly, Volume 10, Number 1, January 1973.