William K. Frankena, Ethics, second edition, 1973.


Meaning and Justification


Thus far, except for Chapter 1, we have been engaged in normative ethics, although we have also included a good bit of analysis and conceptual clarification, as well as some psychology. In other words, we have been endeavoring to arrive at acceptable principles of obligation and general judgments of value in the light of which to determine what is morally right, wrong, or obligatory, and what or who is morally good, bad, or responsible. As we saw in Chapter 1, however, ethics also includes another kind of inquiry called meta-ethics. Meta-ethics does not propound any moral principles or goals for action, except possibly by implication; as such it consists entirely of philosophical analysis. In fact, recent moral philosophy has concerned itself very largely with meta-ethical analysis; it has been primarily interested in clarification and understanding rather than in normative ethics, though it has included some discussion of punishment, civil disobedience, war, etc., and much debate about utilitarianism. For all that, what it has been doing is most mportant, since any reflective person should have some understanding of the neaning and justification of his ethical judgments, especially in this age vhen our general thinking about principles and values is said to be in a state of crisis. In any case, we ourselves must see what sort of justification, if any, can be claimed for the normative positions we have taken.

As usually conceived, meta-ethics asks the following questions.

  1. What is the meaning or definition of ethical terms or concepts like "right," "wrong," "good," "bad"? What is the nature, meaning, or function of judgments in which these and similar terms or concepts occur? What are the rules for the use of such terms and sentences?
  2. How are moral uses of such terms to be distinguished from nonmoral ones, moral judgments from other normative ones? What is the meaning of "moral" as contrasted with "nonmoral"?
  3. What is the analysis or meaning of related terms or concepts like "action," "conscience," "free will," "intention," "promising," "excusing," "motive" "responsibility," "reason," "voluntary"?
  4. Can ethical and value judgments be proved, justified, or shown valid? If so, how and in what sense? Or, what is the logic of moral reasoning and of reasoning about value?
Of these (1) and (4) are the more standard problems of meta-ethics; but (2) and (3) have been receiving much attention lately. We have touched a little on all of them, but will now concentrate on (1) and (4).

Of these two problems, it is (4) that is primary. What we mainly want to know is whether the moral and value judgments we accept are justified or not; and if so, on what grounds. Question (1) is not in itself important in the same way. Apart from conceptual understanding -- which is important to the pure philosopher -- we need to be concerned about the meaning or nature of ethical and value judgments only if this helps us to understand whether and how they may be justified, only if it helps us to know which of them are acceptable or valid. I shall therefore state and discuss the main answers to question (1) if and when they are relevant to the discussion of question (4). It is not easy to classify all of the different theories of the meaning of ethical and value terms and judgments, but they seem to fall under three general types: definist theories, intuitionism or non-naturalism, and noncognitive or nondescriptivist theories. I shall explain them as they become relevant.

For the purposes of such discussions as these, moral judgments and non-moral normative judgments are usually lumped together. This is a risky procedure, for it may be that rather different accounts must be given of the meaning and justification of the two kinds of judgments. Nevertheless, for convenience, we too shall adopt this procedure in our review of the various meta-ethical theories, and use the expression "ethical judgments" to cover all relevant normative and value judgments, not just moral ones.


One way of putting question (4) is to ask whether our basic ethical judgments can be justified in any objective way similar to those in which our factual judgments can be justified. It is, therefore, by a natural impulse that many philosophers have sought to show that certain ethical judgments are actually rooted in fact or, as it used to be put, in "the nature of things" as this is revealed either by empirical inquiry, by metaphysical construction, or by divine revelation. How else, they ask, could one possibly hope to justify them as against rival judgments? If our chosen ethical judgments are not based on fact, on the natures and relations of things, then they must be arbitrary and capricious or at best conventional and relative. One who follows this line of thought, however, seems to be committed to claiming that ethical judgments can be derived logically from factual ones, empirical or nonempirical. Opponents have therefore countered by contending that this cannot be done, since one cannot get an Ought out of an Is or a Value out of a Fact.

Now, we do sometimes seem to justify an ethical judgment by an appeal to fact. Thus, we say that a certain act is wrong because it injures someone, or that a certain painting is good because it has symmetry. However, it becomes clear on a moment's thought that our conclusion does not rest on our factual premise alone. In the first case, we are tacitly assuming that injurious acts are wrong, which is a moral principle; and in the second, that paintings with symmetry are good, which is a value judgment. In such cases, then, we are not justifying our original ethical judgment by reference to fact alone but also by reference to a more basic ethical premise. The question is whether our most basic ethical or value premises can be derived logically from factual ones alone.

This would mean that conclusions with terms like "ought" and "good" in them can be logically inferred from premises, none of which contain these terms; this simply cannot be done by the rules of ordinary inductive or deductive logic. To try to do so is essentially to argue that A is B, ∴ A is C, without introducing any premise connecting B and C. In this sense, those who insist that we cannot go from Is to Ought or from Fact to Value are correct. Such an inference is logically invalid unless there is a special third logic permitting us to do so. It has, in fact, been suggested by some recent writer that there is such a special logic sanctioning certain direct inferences from factual premises to conclusions about what is right or good, that is, an ethical logic with "rules of inference" like "If X is injurious, then X is wrong." But the theory and the rules of such a logic have not yet been satisfactorily worked out, and until they are we can hardly take this possibility seriously. In any case, it is hard to see how such a "rule of inference" differs in substance from the "premise" that injurious acts are wrong, or how its justification will be different.


There is, however, one possibility that must be taken seriously. This is the definist view that Ought can be defined in terms of Is, and Value in terms of Fact. For if such definitions are acceptable, then, by virtue of them, one can go logically from Is to Ought or from Fact to Value. For example, if "We ought to do. . ." means "We are required by society to do . . . ," then from "Society requires that we keep promises," it follows that we ought to keep promises. It will not do to reply, as some have, that no such definitions are possible since we cannot get an Ought out of an Is, for that is to beg the question. We must, therefore, take a closer look at definist theories.

According to such theories ethical terms can be defined in terms of non-ethical ones, and ethical sentences can be translated into nonethical ones of a factual kind. For example, R. B. Perry proposes such definitions as these:

"good" means "being an object of favorable interest (desire),"
"right" means "being conducive to harmonious happiness."1
For him, then, to say that X is good is simply another way of saying that it is an object of desire, and to say that Y is right is just another way of saying that it is conducive to harmonious happiness. A theologian might claim that "right" means "commanded by God"; according to him, then, saying that Y is right is merely a shorter way of saying that it is commanded by God. On all such views, ethical judgments are disguised assertions of fact of some kind. Those who say, as Perry does, that they are disguised assertions of empirical fact are called ethical naturalists, and those who regard them as disguised assertions of metaphysical or theological facts are called metaphysical moralists2 Many different theories of both kinds are possible, depending on the definitions proposed. In each case, moreover, the definition presented may be advanced as a reportive one, simply explicating what we ordinarily mean by the term being defined, or as a reforming proposal about what it should be used to mean. Perry's definitions are offered as reforming proposals, since he thinks our ordinary use of "good" and "right" is confused and vague. F. C. Sharp, on the other hand, offers the following as reportive definitions:
"good" means "desired upon reflection,''
"right" means "desired when looked at from an impersonal point of view."3

In offering definitions or translations of ethical terms and judgments, a definist also tells us how such judgments are to be justified. For example, when Perry tells us that "good" means "being an object of desire," he also tells us that we can test empirically whether X is good simply by determining whether it is desired or not. In general, on a naturalistic theory, ethical judgments can be justified by empirical inquiry just as ordinary and scientific factual statements can; and on any metaphysical theory, they can be justified by whatever methods one can use to justify metaphysical or theological propositions. Either way they are rooted in the nature of things.

Opponents of such theories, following G. E. Moore, accuse them of committing "the naturalistic fallacy," since they identify an ethical judgment with a factual one. To call this a fallacy, however, without first showing that it is a mistake, as is sometimes done, is simply to beg the question. The critics also claim, therefore, that all proposed definitions of "good" and "right" in nonethical terms can be shown to be mistaken by a very simple argument, sometimes referred to as the "open question" argument. Suppose that a definist holds that "good" or "right" means "having the property P," for example, "being desired" or "being conducive to the greatest general happiness." Then, the argument is that we may agree that something has P, and yet ask significantly, "But is it good?" or "Is it right?" That is, we can sensibly say, "This has P, but is it good (or right)?" But if the proposed definition were correct, then we could not say this sensibly for it would be equivalent to saying, "This has P. but has it P?" which would be silly. Likewise, one can say, "This has P but it is not good (or right)," without contradicting oneself, which could not be the case if the definition were correct. Therefore the definition cannot be correct.

To this argument stated in such a simple form, as it almost always is, a definist may make several replies.

  1. He may argue that the meaning of words like "good" and "right" in ordinary use is very unclear, so that when a clarifying definition of one of them is offered, it is almost certain not to retain all of what we vaguely associate with the term. Thus, the substitute cannot seem to be entirely the same as the original, and yet may turn out to be an acceptable definition.
  2. He may point out that the term being defined may have a number of different uses, as we saw in the case of "good." Then P may be correct as an account of one of its uses, even though one can still say, "This has P, but is it good?" For one can agree, say, that X is good intrinsically, and still ask sensibly if it is good extrinsically, morally, or on the whole.
  3. What we mean by some of our terms is often very hard to formulate, as Socrates and his interlocutors found. This means that one who doubts a certain formulation can always use the open question kind of an argument, but it does not mean that no definition can possibly be correct.
  4. A definist like Perry may reply that the open question argument does show that the proposed definitions are not accurate accounts of what we mean by "good" and "right" in ordinary discourse, but that it still may be desirable to adopt them, all things considered.
  5. A definist like Sharp, who thinks that his definitions do express what we actually mean, might even say that we cannot really ask significantly, "Is what we desire on reflection good?" or "Is what we approve when we take an impersonal point of view right?" His definitions are just plausible enough to give such a reply considerable force. In any case, however, although his critics may still be right, they will merely be begging the question if they rest their case on the open question argument.

The open question argument as usually stated, then, is insufficient to refute all definist theories. Its users almost never, in fact, make any serious effort to see what definists might say in reply or to consider their definitions seriously, as some of them certainly deserve to be. We cannot ourselves, however, try to consider separately all of the more plausible definitions which have been proposed. Even after studying them I find myself doubting that any pure definist theory, whether naturalistic or metaphysical, can be regarded as adequate as an account of what we do mean. For such a theory holds that an ethical judgment simply is an assertion of a fact -- that ethical terms constitute merely an alternative vocabulary for reporting facts. It may be that they should be reinterpreted so that this is the case. In actual usage, however, this seems clearly not to be so. When we are making merely factual assertions we are not thereby taking any pro or con attitude toward what we are talking about; we are not recommending it, prescribing it, or anything of the sort. But when we make an ethical judgment we are not neutral in this way; it would seem paradoxical if one were to say "X is good" or "Y is right" but be absolutely indifferent to its being sought or done by himself or anyone else. If he were indifferent in this way, we would take him to mean that it is generally regarded as good or right, but that he did not so regard it himself. We may be making or implying factual assertions in some of our ethical judgments -- when we say, "He was a good man," we do seem to imply that he was honest, kind, etc. -- but this is not all that we are doing.

It might be replied, by Perry for example, that we ought to redefine our ethical terms so that they merely constitute another vocabulary for reporting certain empirical or metaphysical facts (perhaps on the ground that then our ethical judgments could be justified on the basis of science or metaphysics). Then we would have to consider whether we really need such an alternative way of reporting those facts, and whether we can get along without a special vocabulary to do what we have been using our ethical terms to do -- which at least includes expressing pro or con attitudes, recommending, prescribing, evaluating, and so on.

It seems doubtful, then, that we can be satisfied with any pure definist theory of the meaning of moral and other value judgments. It also seems to me that such theories do not suffice to solve the problem of justification. If we accept a certain definition of "good," or "right," then, as we saw, we will know just how to justify judgments about what is good or right. But this means that the whole burden rests on the definition, and we may still ask how the definition is justified or why we should accept it. As far as I can see, when Perry tries to persuade us to accept his definition of "right," he is in effect persuading us to accept, as a basis for action, the ethical principle that what is conducive to harmonious happiness is right. He cannot establish his definition first and then show us that this principle is valid because it is true by definition. He cannot establish his definition unless he can convince us of the principle.

This may seem obvious, since Perry's definition is meant as a recommendation. But a definist who regards his definition as reportivc, and not reforming, would presumably rejoin by saying that his definition is justified simply by the fact that it expresses what we ordinarily mean, just as dictionary definitions are justified. It has been claimed that the notion of obligation as we know it was not present in Greek times and is due to the Judeo-Christian theology. It might be held, then, that "ought," as it is actually used in our moral discourse, means "commanded by God," and many people would accept this as an account of what they mean. If we ask such a reportive theological definist why we ought to do what God commands, he will probably answer, if he understands us to be asking for a justification and not for motivation, that we ought to do this because "ought'' simply means "commanded by God." But this, if true, would only show that his ethical principle had become enshrined in our moral discourse; it would not show why we should continue to give adherence to his principle, and this is the question. In other words, to advocate the adoption of or continued adherence to a definition of an ethical term seems to be tantamount to trying to justify the corresponding moral principle. Appealing to a definition in support of a principle is not a solution to the problem of justification, for the definition needs to be justified, and justifying it involves the same problems that justifying a principle does.

If this is true, then our basic ethical norms and values cannot be justified by grounding them in the nature of things in any strictly logical sense. For this can be done logically only if "right," "good," and "ought" can be defined in nonethical terms. Such definitions, however, turn out to be disguised ethical principles that cannot themselves be deduced logically from the nature of things. It follows that ethics does not depend logically on facts about man and the world, empirical or nonempirical, scientific or theological.

It still may be that there is some non-logical sense in which our basic norms and value judgments can be justified by appeal to the nature of things. We have already seen that ethical egoists seek to justify their theory of obligation by arguing that human nature is so constituted that each of us always pursues only his own good, and that Mill and other hedonists try to justify their theory of value by showing that human nature is so constituted as to desire nothing but pleasure or the means to pleasure. Neither the egoists nor the hedonists claim that their argument affords a strict logical proof. I have also indicated that such arguments nevertheless have a very considerable force, provided their premises are correct. But we saw reason to question the premises of the psychological arguments for egoism and hedonism, and hence must take them as inadequate. In any case, however, it is doubtful that one could find any similar "proofs" of principles like beneficence, justice, or utility. Many people hold that morality depends on religion or theology -- that ethical principles can be justified by appeal to theological premises and only by appeal to such premises. To those who hold this we must reply, in view of our argument, that this dependence cannot be a logical one. They may, of course, still maintain that morality is dependent on religion in some psychological way, for example, that no adequate motivation to be moral is possible without religion. This, I think, is true, if at all, only in a very qualified sense; however, even if religious beliefs and experiences are necessary for motivation, it does not follow that the justification of moral principles depends on such beliefs and experiences. Theologians may also contend that the law of love or beneficence can be rationally justified on theological grounds, even if it cannot rest on such grounds logically. They may argue, for instance, that if one fully believes or unquestionably experiences that God is love, then one must, if he is rational, conclude that he, too, should love. They may say that, although this conclusion does not follow logically, it would be unreasonable for one to draw any other or to refrain from drawing it. In this belief they may well be right; for all that I have said, I am inclined to think they are right. However, it does not follow that the principle of beneficence (let alone that of equality) depends on religion for its justification even in this non-logical sense. It may be that it can also be justified in some other way.


We must, then, give up the notion that our basic principles and values can be justified by being shown to rest logically on true propositions about man and the world. We may also have to admit or insist that they cannot be justified satisfactorily by any such psychological arguments as are used by egoists and hedonists. But now another familiar answer to the question of justification presents itself -- the view that our basic principles and value judgments are intuitive or self-evident and thus do not need to be justified by any kind of argument, logical or psychological, since they are self-justifying or, in Descartes's words, "clearly and distinctly true." This view was very strong until recently, and is held by many of the writers we have mentioned: Butler, Sidgwick, Rashdall, Moore, Prichard, Ross, Carritt, Hartmann, Ewing, and possibly even by Plato. It is sometimes called intuitionism, sometimes non-naturalism.

Intuitionism involves and depends on a certain theory about the meaning or nature of ethical judgments. Definist theories imply that ethical terms stand for properties of things, like being desired or being conducive to harmonious happiness, and that ethical judgments are simply statements ascribing these properties to things. Intuitionists agree to this, but deny that the properties referred to by words like "good" and "ought" are definable in nonethical terms. In fact, they insist that some of these properties are indefinable or simple and unanalyzable, as yellowness and pleasantness are. Sidgwick holds that "ought" stands for such a property, Moore that "good" does, and Ross that both do. These properties are not, therefore, unintelligible or unknown, anymore than pleasantness and yellowness are. But they are not natural or empirical properties as are pleasantness and yellowness. They are of a very different kind, being non-natural or nonempirical and, so to speak, normative rather than factual -- different in kind from all the properties dreamed of in the philosophies of the definists. According to this view, as for the definists, ethical judgments are true or false; but they are not factual and cannot be justified by empirical observation or metaphysical reasoning. The basic ones, particular or general, are self-evident and can only be known by intuition; this follows, it is maintained, from the fact that the properties involved are simple and non-natural.

On this view, ethical judgments may be and are said to be rooted in the natures and relations of things, but not in the sense that they can be derived from propositions about man and the world, as the views previously discussed hold. They are based on the natures and relations of things in the sense that it is self-evident that a thing of a certain nature is good, for example, that what is pleasant or harmonious is good in itself; or that a being of a certain nature ought to treat another being of a certain nature in a certain way, for example, that one man ought to be just, kind, and truthful toward another man.

There are a number of reasons why intuitionism, for almost two centuries the standard view among moral philosophers, now finds few supporters. First of all, it raises several ontological and epistemological questions. An intuitionist must believe in simple indefinable properties, properties that are of a peculiar non-natural or normative sort, a priori or nonempirical concepts, intuition, and self-evident or synthetic necessary propositions. All of these beliefs are hard to defend. Do our ethical terms point to distinct and indefinable properties? It is not easy to be sure, and many philosophers cannot find such properties in their experience. It is also very difficult to understand what a non-natural property is like, and intuitionists have not been very satisfying on this point. Moreover, it is very difficult to defend the belief in a priori concepts and self-evident truths in ethics, now that mathematicians have generally given up the belief that there are such concepts and truths in their field.

Intuitionism is also not easy to square with prevailing theories in psychology and anthropology, even if we do not regard relativism as proved by them, a point we will take up later. An enriched view of the meanings of meaning and of the functions and uses of language likewise casts doubt on the view that ethical judgments are primarily property-ascribing assertions, as intuitionists, like definists, believe.

Intuitionism may still be true in spite of such considerations. But there are two arguments against it that many have regarded as decisive. Both are used by noncognitivists or nondescrigtivists and, interestingly enough, the first is similar to the open question argument used against definists by intuitionists themselves. Let us suppose, it is said, that there are such brave non-natural and indefinable properties as the intuitionists talk about. Let us also suppose that act A has one of these properties, P. Then one can admit that A has P and still sensibly ask, "But why should I do A?" One could not do this if "I should do A" means "A has P"; hence it does not mean "A has P" as intuitionists think.

I do not find this argument convincing. "Why should I do A?" is an ambiguous question. One who asks it may be asking, "What motives are there for my doing A?" or he may be asking, "Am I really morally obligated to do A?" That is, he may be asking for motivation or he may be asking for justification. Now, of course, one can admit that A has P and still sensibly ask, "What motives are there for my doing that which has P?" But this, an intuitionist may say, is irrelevant, since he is proposing a theory of justification and not a theory of motivation, although he is also ready to provide a theory of motivation at the proper time. Therefore, the question is whether one can admit that A has P and still ask sensibly, "Ought I really to do A?" Here we must remember that the intuitionist holds that "I ought to do A" means "A has P" or, in other words, that P is the property of obligatoriness. Hence, he can answer the argument in its relevant form by saying that if "I ought to do A" does mean "A has P," then one cannot sensibly say, "A has P but ought I to do it?" His critic may still insist that he can sensibly say this, but not if he first admits that "I ought to do A" means "A has P." For him simply to assert that it does not mean "A has P" is to beg the question; however, his argument does not prove his conclusion, but assumes it. If there is a property of obligatoriness, as the intuitionist holds, then one cannot sensibly admit that A has this property and ask, "But is it obligatory?"

The second argument, which comes from Hume, is used against many kinds of definism as well as intuitionism, and has to do with motivation rather than justification. It begins with an insistence that ethical judgments are in themselves motivating or "practical" in the sense that, if one accepts such a judgment, he must have some motivation for acting according to it. It then contends that, if an ethical judgment merely ascribes a property, P, to something, then, whether P is natural or non-natural, one can accept the judgment and still have no motivation to act one way rather than another.

Intuitionists (and definists) also have a possible answer to this argument. They can maintain that we are so constituted that, if we recognize X to be right or good (i.e., that X has P), this will generate a pro attitude toward X in us, either by itself or by way of an innate desire for what has P. One may, of course, question their psychological claims, but one must at least give good reasons for thinking these are false before one takes this argument as final.

On the whole, however, intuitionism strikes me as implausible even if it has not been disproved. As was indicated earlier, ethical judgments do not seem to be mere property-ascribing statements, natural or non-natural; they express favorable or unfavorable attitudes (and do not merely generate them), recommend, prescribe, and the like. Of course, one could maintain that they do this and also ascribe simple non-natural properties to actions and things, but such a view still involves one in the difficulties mentioned a moment ago. The main point to be made now is that the belief in self-evident ethical truths, and all that goes with it, is so difficult to defend that it seems best to look for some other answer to the problem of justification.


The third general type of theory of the meaning or nature of ethical judgments has no very satisfactory label. However, it has been called noncognitivist or nondescriptivist because, as against both definists and intuitionists, it holds that ethical judgments are not assertions or statements ascribing properties to (or denying them of) actions, persons, or things, and insists that they have a very different "logic," meaning, or use. It embraces a wide variety of views, some more and others much less extreme.

1. The most extreme of these are a number of views that deny ethical judgments, or at least the most basic ones, to be capable of any kind of rational or objectively valid justification. On one such view -- that of A. J. Ayer -- they are simply expressions of emotion much like ejaculations. Saying that killing is wrong is like saying, "Killing, boo!" It says nothing true or false and cannot be justified in any rational way. Rudolf Carnap once took a similar view, except that he interpreted "Killing is wrong" as a command, "Do not kill," rather than as an ejaculation. Bertrand Russell held that moral judgments merely express a certain kind of wish. Many existentialists likewise regard basic ethical judgments, particular or general, as arbitrary commitments or decisions for which no justification can be given.

I should point out here that such irrationalistic views about ethical judgments are not held only by atheistic positivists and existentialists. They are also held by at least some religious existentialists and by other theologians. For example, a theologian who maintains that the basic principles of ethics are divine commands is taking a position much like Carnap's. If he adds that God's commands are arbitrary and cannot be justified rationally, then his position is no less extreme. If he holds that God's commands are, at least in principle, rationally defensible, then his position is like the less extreme ones to be described.

2. C. L. Stevenson's form of the emotive theory is somewhat less extreme than Ayer's. He argues that ethical judgments express the speaker's attitudes and evoke, or seek to evoke, similar attitudes in the hearer. But he realizes that to a very considerable extent our attitudes are based on our beliefs, and so can be reasoned about. For example, I may favor a certain course of action because I believe it has or will have certain results. I will then advance the fact that it has these results as an argument in its favor. But you may argue that it does not have these results, and if you can show this, my attitude may change and I may withdraw my judgment that the course of action in question is right or good. In a sense, you have refuted me. But, of course, this is only because of an underlying attitude on my part of being in favor of certain results rather than others. Stevenson goes on to suggest that our most basic attitudes, and the ethical judgments in which we express them, may not be rooted in beliefs of any kind, in which case they cannot be reasoned about in any way. He is open-minded about this, however, and allows a good deal of room for a kind of argument and reasoning.

3. More recently, from a number of Oxford philosophers and others, we have had still less extreme views. They refuse to regard ethical judgments as mere expressions or evocations of feeling or attitude, as mere commands, or as arbitrary decisions or commitments. Rather, they regard them as evaluations, recommendations, prescriptions, and the like; and they stress the fact that such judgments imply that we are willing to generalize or universalize them and are ready to reason about them, points with which we have agreed. That is, they point out that when we say of something that it is good or right, we imply that there are reasons for our judgment which are not purely persuasive and private in their cogency. They are even ready to say that such a judgment may be called true or false, though it is very different from "X is yellow" or "Y is to the left of Z." For them ethical judgments are essentially reasoned acts of evaluating, recommending, and prescribing.

The arguments for such theories -- the open question argument against definists and the two arguments against intuitionists -- we have found to be less conclusive than they are thought to be. To my mind, nevertheless, these theories, or rather the least extreme of them, are on the right track. The kind of account the latter give of the meaning and nature of ethical judgments is acceptable as far as it goes. Such judgments do not simply say that something has or does not have a certain property. Neither are they mere expressions of emotion, will, or decision. They do more than just express or indicate the speaker's attitudes. They evaluate, instruct, recommend, prescribe, advise, and so on; and they claim or imply that what they do is rationally, justified or justifiable, which mere expressions of emotion and commands do not do. The more extreme views, therefore, are mistaken as a description of the nature of ethical judgments. Moreover, it is not necessary to agree with them that such judgments cannot be justified in any important sense. They generally assume that if such judgments are not self-evident and cannot be proved inductively or deductively on the basis of empirical or nonempirical facts, as we have seen to be the case, then it follows that they are purely arbitrary. But this does not follow. It may be that this conception of rational justification is too narrow, as I have already intimated in discussing psychological egoism and hedonism. Mill may be right when he says, near the end of Chapter I of Utilitarianism,

We are not . . . to infer that [the acceptance or rejection of an ethical first principle] must depend on blind impulse, or arbitrary choice. There is a larger meaning of the word "proof," in which this question is . . . amenable to it . . .The subject is within the cognizance of the rational faculty; and neither does that faculty deal with it solely in the way of jntuition. Considerations may be presented capable of determining the intellect either to give or withhold its assent . . .
Here, Mill is with the less extreme of the recent nondescriptive theories, as against the definists, the intuitionists, and the more extreme nondescriptivists. All of these share the conception of justification as consisting either in self-evidence or in inductive or deductive proof. Only the definists and intuitionists believe that ethical judgments can be justified in one or the other of these ways, while positivists and existentialists deny that ethical judgments can be justified at all. Mill and the less extreme recent philosophers, on the other hand, agree with intuitionists and definists that they can be justified in some rational sense or in some "larger meaning of the word 'proof.'," though they have different and various views about the nature of such justification.

At this point, it may help to notice that even such things as "mere" expressions of feeling and commands may be justified or unjustified, rational or irrational. Suppose that A is angry at B, believing B to have insulted him. C may be able to show A that his anger is unjustified, since B has not actually insulted him at all. If A simply goes on being angry, although he no longer has any reason, we should regard his anger as quite irrational. Again, if an officer commands a private to close the door, believing it to be open when it is not, it is reasonable for the private to answer, "But, sir, the door is closed," and it would be quite irrational if the officer were seriously to command the private to close it anyway. Emotions and commands, generally at least, have a background of beliefs and are justified or unjustified, rational or irrational, depending on whether these beliefs themselves are so.


In my opinion, even the less extreme of recent nondescriptivist theories have not gone far enough. They have been too ready to admit a kind of basic relativism after all. They insist that ethical judgments imply the presence of, or at least the possibility of giving, reasons which justify them. But they almost invariably allow or even insist that the validity of these reasons is ultimately relative, either to the individual or to his culture, and, therefore, conflicting basic judgments may both be justified or justifiable. Now, it may be that in the end one must agree with this view, but most recent discussions entirely neglect a fact about ethical judgments on which Ewing has long insisted, namely, that they make or somehow imply a claim to be objectively and rationally justified or valid. In other words, an ethical judgment claims that it will stand up under scrutiny by oneself and others in the light of the most careful thinking and the best knowledge, and that rival judgments will not stand up under such scrutiny. Hume makes the point nicely, though only for moral judgments: \

The notion of morals implies some sentiment common to all mankind, which recommends the same object to general approbation. . . . When a man denominates another his enemy, his rival, his antagonist, his adversary, he is understood to speak the language of self-love, and to express sentiments, peculiar to himself, and arising from his particular circumstance and situation. But when he bestows on any man the epithets of vicious or odious or depraved, he then speaks another language, and expresses sentiments, in which he expects all his audience are to concur with him. He must here. . . depart from his private and particular situation, and must choose a point of view, common to himself with others. . . .4
And, he must claim, Hume might have added, that anyone else who takes this point of view and from it reviews the relevant facts will come to the same conclusion. In fact, he goes on to suggest that precisely because we need or want a language in which to express, not just sentiments peculiar to ourselves but sentiments in which we expect all men are to concur with us, another language in which we may claim that our sentiments are justified and valid, we had to
. . . invent a peculiar set of terms, in order to express those universal sentiments of censure or approbation. . . . Virtue and vice become then known; morals are recognized; certain general ideas are framed of human conduct and behavior. . . .

This kind of an account of our normative discourse strikes me as eminently wise. It is a language in which we may express our sentiments -- approvals, disapprovals, evaluations, recommendations, advice, instructions, prescriptions -- and put them out into the public arena for rational scrutiny and discussion, claiming that they will hold up under such scrutiny and discussion and that all our audience will concur with us if they will also choose the same common point of view. That this is so is indicated by the fact that if A makes an ethical judgment about X and then, upon being challenged by B, says, "Well, at least I'm in favor of X," we think he has backed down. He has shifted from the language of public dialogue to that of mere self-revelation. This view recognizes the claim to objective validity on which intuitionists and definists alike insist, but it also recognizes the force of much recent criticism of such views.


Against any such view it will be argued, of course, that this claim to be objectively and rationally justified or valid, in the sense of holding up against all rivals through an impartial and informed examination, is simply mistaken and must be given up. This is the contention of the relativist and we must consider it now, although we can do so only briefly.

Actually, we must distinguish at least three forms of relativism. First, there is what may be called descriptive relativism. When careful, it does not say merely that the ethical judgments of different people and societies are different. For this would be true even if people and societies agreed in their basic ethical judgments and differed only in their derivative ones. What careful descriptive relativism says is that the basic ethical beliefs of different people and societies are different and even conflicting. I stress this because the fact that in some primitive societies children believe they should put their parents to death before they get old, whereas we do not, does not prove descriptive relativism. These primitive peoples may believe this because they think their parents will be better off in the hereafter if they enter it while still able-bodied; if this is the case, their ethics and ours are alike in that they rest on the precept that children should do the best they can for their parents. The divergence, then, would be in factual, rather than in ethical, beliefs.

Second, there is meta-ethical relativism, which is the view we must consider. It holds that, in the case of basic ethical judgments, there is no objectively valid, rational way of justifying one against another; consequently, two conflicting basic judgments may be equally valid.

The third form of relativism is normative relativism. While descriptive relativism makes an anthropological or sociological assertion and meta-ethical relativism a meta-ethical one, this form of relativism puts forward a normative principle: what is right or good for one individual or society is not right or good for another, even if the situations involved are similar, meaning not merely that what is thought right or good by one is not thought right or good by another (this is just descriptive relativism over again), but that what is really right or good in the one case is not so in another. Such a normative principle seems to violate the requirements of consistency and universalization mentioned earlier. We need not consider it here, except to point out that it cannot be justified by appeal to either of the other forms of relativism and does not follow from them. One can be a relativist of either of the other sorts without believing that the same kind of conduct is right for one person or group and wrong for another. One can, for example, believe that everyone ought to treat people equally, though recognizing that not everyone admits this and holding that one's belief cannot be justified.

Our question is about the second kind of relativism. The usual argument used to establish it rests on descriptive relativism. Now, descriptive relativism has not been incontrovertibly established. Some cultural anthropologists and social psychologists have even questioned its truth, for example, Ralph Linton and S. E. Asch. However, to prove meta-ethical relativism one must prove more than descriptive relativism. One must also prove that people's basic ethical judgments would differ and conflict even if they were fully enlightened and shared all the same factual beliefs. It is not enough to show that people's basic ethical judgments are different, for such differences might all be due to differences and incompletenesses in their factual beliefs, as in the example of the primitive societies used previously. In this case, it would still be possible to hold that some basic ethical judgments can be justified as valid to everyone, in principle at least, if not in practice.

It is, however, extremely difficult to show that people's basic ethical judgments would still be different even if they were fully enlightened, conceptually clear, shared the same factual beliefs, and were taking the same point of view. To show this, one would have to find clear cases in which all of these conditions are fulfilled and people still differ. Cultural anthropologists do not show us such cases; in all of their cases, there are differences in conceptual understanding and factual belief. Even when one takes two people in the same culture, one cannot be sure that all of the necessary conditions are fulfilled. I conclude, therefore, that meta-ethical relativism has not been proved and, hence, that we need not, in our ethical judgments, give up the claim that they are objectively valid in the sense that they will be sustained by a review by all those who are free, clear-headed, fully informed, and who take the point of view in question.


We now have the beginnings of a theory of the meaning and justification of ethical judgments. To go any farther, we must distinguish moral judgments proper from nonmoral normative judgments and say something separately about the justification of each. How can we distinguish moral from other normative judgments? Not by the words used in them, for words like "good" and "right" all have nonmoral as well as moral uses. By the feelings that accompany them? The difficulty in this proposal is that it is hard to tell which feelings are moral except by seeing what judgments they go with. It is often thought that moral judgments are simply whatever judgments we regard as overriding all other normative judgments in case of conflict, but then aesthetic or prudential judgments become moral ones if we take them to have priority over others, which seems paradoxical. It seems to me that what makes some normative judgments moral, some aesthetic, and some prudential is the fact that different points of view are taken in the three cases, and that the point of view taken is indicated by the kinds of reasons that are given. Consider three judgments:

  1. I say that you ought to do X and give as the reason the fact that X will help you succeed in business;
  2. I say you should do Y and cite as the reason the fact that Y will produce a striking contrast of colors; and
  3. I say you should do Z and give as the reason the fact that Z will keep a promise or help someone.
Here the reason I give reveals the point of view I am taking and the kind of judgment I am making.

Now let us take up the justification of nonmoral normative judgments. We are interested primarily in judgments of intrinsic value such as were discussed in the previous chapter, for such judgments are relevant to ethics because, through the principle of beneficence, the question of what is good or bad comes to bear on the question of what is right or wrong. Besides, if we know how to justify judgments of intrinsic value, we will know how to justify judgments of extrinsic and inherent value, for judgments of the latter sorts presuppose judgments of the former. It is true, as we have already seen, that we cannot prove basic judgments of intrinsic value in any strict sense of proof, but this fact does not mean that we cannot justify them or reasonably claim them to be justified. But how can we do this? By taking what I shall call the evaluative point of view as such, unqualified by any such adjective as "aesthetic," "moral," or "prudential," and then trying to see what judgment we are led to make when we do so, considering the thing in question wholly on the basis of its intrinsic character, not its consequences or conditions. What is it to take the nonmorally evaluative point of view? It is to be free, informed, clear-headed, impartial, willing to universalize; in general, it is to be "calm" and "cool," as Butler would say, in one's consideration of such items as pleasure, knowledge, and love, for the question is simply what it is rational to choose. This is what we tried to do in Chapter 5. If one considers an item in this reflective way and comes out in favor of it, one is rationally justified in judging it to be intrinsically good, even if one cannot prove one's judgment. In doing so, one claims that everyone else who does likewise will concur; and one's judgment is really justified if this claim is correct, which, of course, one can never know for certain. If others who also claim to be calm and cool do not concur, one must reconsider to see if both sides are really taking the evaluative point of view, considering only intrinsic features, clearly understanding one another, and so on. More one cannot do and, if disagreement persists, one may still claim to be right (i.e., that others will concur eventually if. . .); but one must be open-minded and tolerant. In fact, we saw in Chapter 5 that one may have to admit a certain relativity in the ranking of things listed as intrinsically good, although possibly not in the listing itself.

What about the justification of moral judgments? Already in Chapters 2 and 3 we have, in effect, said something about the justification of judgments of right, wrong, and obligation. We argued that a particular judgment essentially entails a general one, so that one cannot regard a particular judgment as justified unless one is also willing to accept the entailed general one, and vice versa. This is true whether we are speaking of judgments of actual or of prima facie duty. We have also seen that judgments of actual duty, whether particular judgments or rules, cannot simply be deduced from the basic principles of beneficence and justice, even with the help of factual premises, since these must be taken as prima facie principles and may conflict on occasion. Thus, we have two questions: first, how can we justify judgments of actual duty, general or particular, and second, how can we justify basic principles of prima facie duty ? The same answer, however, will do for both. It seems fair to assume that it will also do for the question of justifying judgments of moral value.

First, we must take the moral point of view, as Hume indicated, not that of self-love or aesthetic judgment, nor the more general point of view involved in judgments of intrinsic value. We must also be free, impartial, willing to universalize, conceptually clear, and informed about all possibly relevant facts. Then we are justified in judging that a certain act or kind of action is right, wrong, or obligatory, and in claiming that our judgment is objectively valid, at least as long as no one who is doing likewise disagrees. Our judgment or principle is really justified if it holds up under sustained scrutiny of this sort from the moral point of view on the part of everyone. Suppose we encounter someone who claims to be doing this but comes to a different conclusion. Then we must do our best, through reconsideration and discussion, to see if one of us is failing to meet the conditions in some way. If we can detect no failing on either side and still disagree, we may and I think still must each claim to be correct, for the conditions never are perfectly fulfilled by both of us and one of us may turn out to be mistaken after all. If what was said about relativism is true, we cannot both be correct. But both of us must be open-minded and tolerant if we are to go on living within the moral institution of life and not resort to force or other immoral or nonmoral devices.

If this line of thought is acceptable, then we may say that a basic moral judgment, principle, or code is justified or "true" if it is or will be agreed to by everyone who takes the moral point of view and is clearheaded and logical and knows all that is relevant about himself, mankind, and the universe. Are our own principles of beneficence and justice justified or "true" in this sense? The argument in Chapters 2 and 3 was essentially an attempt to take the moral point of view and from it to review various normative theories and arrive at one of our own. Our principles have not been proved, but perhaps it may be claimed that they will be concurred in by those who try to do likewise. This claim was implicitly made in presenting them. Whether the claim is true or not must wait upon the scrutiny of others.

The fact that moral judgments claim a consensus on the part of others does not mean that the individual thinker must bow to the judgment of the majority in his society. He is not claiming an actual consensus, he is claiming that in the end -- which never comes or comes only on the Day of Judgment -- his position will be concurred in by those who freely and clear-headedly review the relevant facts from the moral point of view. In other words, he is claiming an ideal consensus that transcends majorities and actual societies. One's society and its code and institutions may be wrong. Here enters the autonomy of the moral agent -- he must take the moral point of view and must claim an eventual consensus with others who do so, but he must judge for himself. He may be mistaken, but, like Luther, he cannot do otherwise. Similar remarks hold for one who makes nonmoral judgments.


What is the moral point of view? This is a crucial question for the view we have suggested. It is also one on which there has been much controversy lately. According to one theory, one is taking the moral point of view if and only if one is willing to universalize one's maxims. Kant would probably accept this if he were alive. But I pointed out that one may be willing to universalize from a prudential point of view; and also that what one is willing to universalize is not necessarily a moral rule. Other such formal characterizations of the moral point of view have been proposed. A more plausible characterization to my mind, however, is that of Kurt Baier. He holds that one is taking the moral point of view if one is not being egoistic, one is doing things on principle, one is willing to universalize one's principles, and in doing so one considers the good of everyone alike.5

Hume thought that the moral point of view was that of sympathy, and it seems to me he was on the right wavelength. I have already argued that the point of view involved in a judgment can be identified by the kind of reason that is given for the judgment when it is made or if it is challenged. Then the moral point of view can be identified by determining what sorts of facts are reasons for moral judgments or moral reasons. Roughly following Hume, I now want to suggest that moral reasons consist of facts about what actions, dispositions, and persons do to the lives of sentient beings, including beings other than the agent in question, and that the moral point of view is that which is concerned about such facts. My own position, then, is that one is taking the moral point of view if and only if

  1. one is making normative judgments about actions, desires, dispositions, intentions, motives, persons, or traits of character;
  2. one is willing to universalize one's judgments;
  3. one's reasons for one's judgments consist of facts about what the things judged do to the lives of sentient beings in terms of promoting or distributing nonmoral good and evil; and
  4. when the judgment is about oneself or one's own actions, one's reasons include such facts about what one's own actions and dispositions do to the lives of other sentient beings as such, if others are affected.
One has a morality or moral action-guide only if and insofar as one makes normative judgments from this point of view and is guided by them.


Another problem that remains has been mentioned before. Why should we be moral? Why should we take part in the moral institution of life? Why should we adopt the moral point of view? We have already seen that the question, "Why should . . .?" is ambiguous, and may be a request either for motivation or for justification. Here, then, one may be asking for

  1. the motives for doing what is morally right,
  2. a justification for doing what is morally right,
  3. motivation for adopting the moral point of view and otherwise subscribing to the moral institution of life, or
  4. a justification of morality and the moral point of view.
It is easy to see the form an answer to a request for (1) and (3) must take; it will consist in pointing out the various prudential and non-prudential, motives for doing what is right or for participating in the moral institution of life. Most of these are familiar or readily thought of and need not be detailed here. A request for (2) might be taken as a request for a moral justification for doing what is right. Then, the answer is that doing what is morally right does not need a justification, since the justification has already been given in showing that it is right. On this interpretation, a request for (2) is like asking, "Why morally ought I to do what is morally right?" A request for (2) may also, however, be meant as a demand for a nonmoral justification of doing what is morally right; then, the answer to it will be like the answer to a request for (4). For a request for (4), being a request for reasons for subscribing to the moral way of thinking, judging and living, must be a request for a nonmoral justification of morality. What will this be like?

There seem to be two questions here. First, why should society adopt such an institution as morality? Why should it foster such a system for the guidance of conduct in addition to convention, law, and prudence? To this the answer seems clear. The conditions of a satisfactory human life for people living in groups could hardly obtain otherwise. The alternatives would seem to be either a state of nature in which all or most of us would be worse off than we are, even if Hobbes is wrong in thinking that life in such a state would be "solitary, poor, nasty, brutish, and short"; or a leviathan civil state more totalitarian than any yet dreamed of, one in which the laws would cover all aspects of life and every possible deviation by the individual would be closed off by an effective threat of force.

The other question has to do with the nonmoral reasons (not just motives) there are for an individual's adopting the moral way of thinking and living. To some extent, the answer has just been given, but only to some extent. For on reading the last paragraph an individual might say, "Yes. This shows that society requires morality and even that it is to my advantage to have others adopt the moral way of life. But it does not show that I should adopt it, and certainly not that I should always act according to it. And it is no use arguing on moral grounds that I should. I want a nonmoral justification for thinking I should." Now, if this means that he wants to be shown that it is always to his advantage -- that is, that his life will invariably be better or, at least, not worse in the prudential sense of better and worse -- if he thoroughly adopts the moral way of life, then I doubt that his demand can always be met. Through the use of various familiar arguments, one can show that the moral way of life is likely to be to his advantage, but it must be admitted in all honesty that one who takes the moral road may be called upon to make a sacrifice and, hence, may not have as good a life in the nonmoral sense as he would otherwise have had.

The point made at the end of Chapter 5 must be recalled here, namely, that morally good or right action is one kind of excellent, activity and hence is a prime candidate for election as part of any good life, especially since it is a kind of excellent activity of which all normal people are capable. It does seem to me that this is an important consideration in the answer to our present question. Even if we add it to the usual arguments, however, we still do not have a conclusive proof that every individual should, in the nonmoral sense under discussion, always do the morally excellent thing. For, as far as I can see, from a prudential point of view, some individuals might have nonmorally better lives if they sometimes did what is not morally excellent, for example in cases in which a considerable self-sacrifice is morally required. A TV speaker once said of his subject, "He was too good for his good," and it seems to me that this may sometimes be true.

It does not follow that one cannot justify the ways of morality to an individual, although it may follow that one cannot justify morality to some individuals. For nonmoral justification is not necessarily egoistic or prudential. If A asks B why he, A, should be moral, B may reply by asking A to try to decide in a rational way what kind of a life he wishes to live or what kind of a person he wishes to be. That is, B may ask A what way of life A would choose if he were to choose rationally, or in other words, freely, impartially, and in full knowledge of what it is like to live the various alternative ways of life, including the moral one. B may then be able to convince A, when he is calm and cool in this way, that the way of life he prefers, all things considered, includes the moral way of life. If so, then he has justified the moral way of life to A. A may even, when he considers matters in such a way, prefer a life that includes self-sacrifice on his part.

Of course, A may refuse to be rational, calm, and cool. He may retort, "But why should I be rational?" However, if this was his posture in originally asking for justification, he had no business asking for it. For one can only ask for justification if one is willing to be rational. One cannot consistently ask for reasons unless one is ready to accept reasons of some sort. Even in asking, "Why should I be rational?" one is implicitly committing oneself to rationality, for such a commitment is part of the connotation of the word "should."

What kind of a life A would choose if he were fully rational and knew all about himself and the world will, of course, depend on what sort of a person he is (and people are different), but if psychological egoism is not true of any of us, it may always be that A would then choose a way of life that would be moral. As Bertrand Russell once wrote:

We have wishes which are not purely personal. . . The sort of life that most of us admire is one which is guided by large, impersonal desires. . . Our desires are, in fact, more general and less purely selfish than many moralists imagine. . .6

Perhaps A has yet one more question: Is society justified in demanding that I adopt the moral way of life, and in blaming and censuring me if I do not?" But this is a moral question; and A can hardly expect it to be allowed that society is justified in doing this to A only if it can show that doing so is to A's advantage. However, if A is asking whether society is morally justified in requiring of him at least a certain minimal subscription to the moral institution of life, then the answer surely is that society sometimes is justified in this, as Socrates argued in the Crito. But society must be careful here. For it is itself morally required to respect the individual's autonomy and liberty, and in general to treat him justly; and it must remember that morality is made to minister to the good lives of individuals and not to interfere with them any more than is necessary. Morality is made for man, not man for morality.


1 Realms of Value (Cambridge, Mass.: Harvard University Press, 1954), pp. 3, 107, 109. See selections in Frankena and Granrose, Chap. VI.

2 Most writers today use "naturalism" to cover all kinds of definism,

3 Ethics (New York: The Century Co., 1928), pp. 109, 410-11.

4 An Enquiry into the Principles of Morals, pp. 113 -- 14.

5 The MoralPoint of View (New York: Random House, 1965), Chap. 5.

6 Religion and Science (New York: Henry Holt and Co., 1935), pp. 252-54.