|THE DEMAND FOR JUSTIFICATION IN ETHICS
ABSTRACT. The common belief that the epistemic credentials of ethics are quite questionable, and therefore in need of special justification, is an illusion made possible by the logical gap between reason and belief. This gap manifests itself sometimes even outside ethics. In ethics its manifestations are common, because of the practical nature of ethics. The attempt to cover it up takes the form of exorbitant demands for justification and often leads to espousing noncognitivism.
In The Methods of Ethics Henry Sidgwick asked why is ethics commonly taken to include discussion of the nature of the moral faculty, while mathematics and physics are not usually taken to include discussion, respectively, of the mathematical faculty and the faculty of sense- perception. Part of the explanation he proposed was that, unlike mathematics and physics, ethics has a practical aim, and that while we cannot help believing what we see to be true we can help doing what we see to he right. He also suggested that it is partly because of this fact that people commonly ask, Why should I do what I see to be right?, but not Why should I believe what I see to he true? But the main explanation Sidgwiek offered was that there is a great diversity of methods and principles inherent in ordinary moral thought.i
Sidgwick went on to describe the fact needing explanation as “the persistent unsatisfied demand for an ultimate reason,” for a demonstration of “the ultimate reasonableness of conduct.”ii Decades later, Prichard described it as the demand for a proof or at least justification that we ought to do what in our non-reflective consciousness we have unquestioningly thought, sensed, that we ought to do.iii Prichard’s answer was, roughly, that such a proof or justification is not possible, but his explanation of why nevertheless we look for one was similar to Sidgwick’s. We may not accept his answer, or the particular ethical theory in which it is embedded, but I believe his explanation, like Sidgwick’s, was generally correct. There is a hint of that explanation also in Kant. However, our understanding of it must go considerably beyond what either Siclgwick or Prichard or Kant said. I shall refer mainly to Sidgwick, but this is not a paper about him. As on many other topics, Sidgwick provides us with an especially convenient point of departure.
Of course, philosophers ask general epistemological questions also about mathematics, to say nothing of physics, but for a different reason. They do not ordinarily doubt that most mathematical judgments they accept express unquestionable knowledge; ordinarily they are only curious about the nature of this knowledge, about how they may accommodate it in their epistemological theories, and especially about whether it can be given a realist interpretation. And philosophers often have this sort of reason also for raising the corresponding questions about ethics. But here I shall not be concerned, except in passing, with such technical philosophical reasons, important though they are, just as Sidgwick was not. I shall be concerned with a reason peculiar to ethics, which is no less familiar in ordinary, nonphilosophical ethical thought. In ethics the demand for justification is made largely because there is an impression of uncertainty infecting all ethical judgments, and then of course it becomes a crucial question for ethics whether this impression corresponds to the facts. Why is there this impression of uncertainty? I believe an answer roughly like Sidgwick’s is the right one. But to explain it, I must make a number of preliminary and, unavoidably, very sketchy remarks.
I suggest that in general we do not find ethics to he inherently less firmly grounded than mathematics and physics. It’s just that we can be so easily led, by others or by ourselves, to think that it is. In all three (if we understand them broadly as Sidgwick did and as in this paper we also should), we employ perception or immediate awareness (often called ‘intuition), we employ intuitive induction, deduction, Humean induction; we appeal to systematic considerations, for example, to coherence and explanatory power; and we marshal the sort of arguments that Mill thought could determine the intellect but not constitute an actual proof, for example, appeals to analogies.iv In all three, by such means, though to widely varying degrees because of the differences in subject matter, we come to declare certain propositions justified, probable, known to be true, perhaps even self-evident. I speak here not of what epistemologists or moral philosophers, or mathematicians, or physicists would ordinarily say, though many do say it. I speak of what the ordinary person would say to questions such as “Is lying wrong?’, “Are three and two five?,” “Does water evaporate when boiled?” Much of the power of realism and cognitivism in ethics is due to this fact about ordinary thought. What needs to be understood, however, is why we can be so easily led to question even some of our immediate so-called ‘intuitions’ about the subject matter of ethics, but not those about the subject matter of mathematics or physics.
The fundamental answer, I suggest, is the first one Sidgwick in effect gave: we have a stake in ethics that we do not have in mathematics or in physics. Our inclinations, whether egoistic or altruistic or neither, often conflict with our moral intuitions, with our putative awareness of moral truth, or with the conclusions of our reasoning from these intuitions, and sometimes even overpower them. And since by nature we are truth-respecting beings, it is natural for us to want to suppose that this is so because of some inadequacy of moral reason, that the intuitions conflicting with our inclinations are spurious, that they do not constitute genuine knowledge. Of course, there need not be such a conflict and some of our inclinations (e.g., those involved in what Moore called personal affection) may have the highest value, perhaps, pace Kant, even moral value.
But while this may explain why we are tempted to question our ethical judgments, we need to explain also why we succumb to that temptation so easily. This may be thought of as the purpose of the second part of Sidgwick’s explanation, in which he appealed to the diversity of methods and principles present in ordinary ethical thought. Though I believe that the first part, namely, his appeal to the practical nature of ethics, suffices as an explanation of both the temptation and our succumbing to it, to understand it properly we must briefly consider the second part first. But my account of it will he rather different from Sidgwick’s and eventually I will disagree with him on a crucial issue.
There is indeed great diversity in the ethical judgments we make, both in subject matter and in epistemic status, a diversity absent from mathematics and physics, at least as they arc known to ordinary thought. There are abstract singular judgments, some of which may deserve to he called axioms of ethics, such as, perhaps, “Friendship is a good,” Pain is bad,”” Lying is wrong.” If we find ourselves disagreeing about them, this is almost certainly due to a misunderstanding. It maybe a misunderstanding of what is meant by the words “friendship,” “pain,” and “lying,” a misunderstanding that occurs because the notions they express, unlike those of mathematics and physics, are quite imprecise. (And they remain imprecise partly because, unlike mathematicians and physicists, modern moral philosophers have generally not been doing their job; they have been talking more about ethics than about the subject matter of ethics, they have been doing more meta-ethics than ethics.) Or the disagreement may be due to a misunderstanding not of individual words but of the grammar of the statements. This sort of misunderstanding typically takes the form of objections such as “What if the pain was deserved?” or “What if lying is the only way to prevent disaster?” But in saying, e.g., that pain is bad, we are saying, of course, that pain as such, in itself, is bad, not that some particular instances of it, though as such also bad in themselves, may not have other characteristics that are good, or that they may not be essential elements in more complex states that are good, or that they may not have good consequences, and therefore that they may not be good on the whole. The distinction is simple, easy to understand, and once made the misunderstanding is almost certain to be eliminated. The Platonism implicit in it would still he rejected by most philosophers, but it seems quite natural to common sense.
A second class of ethical judgments concern the rightness or wrongness of particular actions. And even if we are not consequentialists, we certainly regard the consequences of an action as very much relevant to the question whether the action ought to be done. But we seldom, if ever, know enough about the consequences of our actions, and disagreement about them is possible, indeed often unavoidable. This, however, is a fact about the inherent limitations of human knowledge in general, not about any inherent epistemic problems of ethics. The qualification it requires in any cognitivist position in ethics is readily acknowledged by common sense. And this is why, though ordinary ethical thought, or moral common sense, is unqualifiedly realist, it is not unqualifiedly cognitivist, in the literal sense of this term suggested by its etymology. Indeed, realism and cognitivism, and so irrealism and noncognitivism, ought not to he identified, as they usually are. For example, a utilitarian is a realist about right action, for he takes for granted that it is a fact about reality that a certain action would or would not have the best possible consequences, hut he may well be a noncognitivist in the literal sense of this term by denying, as Sidgwick came close to doing, that we can ever know whether it would or would not. And a moral philosopher, such as John Raw may be a cognitivist if he accepts a coherence theory of justitication hut an irrealist if he also accepts a coherence theory of truth, with respect to ethical judgments.
A third class of ethical judgments concern praiseworthiness and blameworthiness, the morality of actions, of traits of character, and of agents. Here there are two special kinds of uncertainty, one the uncertainty inherent in psychological judgments in general, the other that engendered by legitimate but essentially metaphysical (though quite familiar to common sense) doubts about the notions of responsibility and freedom of choice, which seem to be presupposed by the ethical judgments in this third class. But, once again, these are kinds of uncertainty rooted in facts and considerations external to ethics.
A fourth class of ethical judgments, often overlapping with the second, that of judgments about the rightness or wrongness of actions, but containing also evaluations of matters that are not actions, may be described as putative remote theorems supposedly derived from the axioms of ethics but in fact only in conjunction with various nonethical propositions. They constitute what Moore called casuistry and we call applied ethics, the importance of which philosophers in recent years have once again come to recognize. Examples would be ethical judgments about sexual attitudes and modes of behavior, conditions such as irreversible coma, everyday political problems, certain medical procedures, and so on. They are highly specific in content, and presuppose a great deal of knowledge of nonethical facts, including consequences of actions. We often disagree about them, but it is plausible to hold that the disagreement is either about the relevant nonethical facts or about the validity of the derivation from the axioms and the nonethical facts to which we appeal. For example, we may find it self-evident that life, being alive, is a great good, but does it follow that being in irreversible coma is a good? Or that abortion is wrong? The existence of such quandaries has no tendency to show that ethics is not a cognitive discipline.
These remarks arc intended to suggest that the diversity of methods and principles in ordinary ethical thought is quite benign, quite compatible with unqualified realism and with a suitably qualified cognitivism. The uncertainties and quandaries present in it cast doubt on its realism and qualified cognitivism no more than the uncertainties and quandaries present in sophisticated branches of engineering, e.g., those used in the space program, cast doubt on physics. t believe that its realism and qualified cognitivism are generally correct. The usual philosophical arguments to the contrary are hardly compelling. They are not my concern here, but to make what is my concern clear I must briefly refer to them.
One is Hume’s essentially phenomenological claim that, to use his example, he could not find the vice in a case of murder. But a more sophisticated phenomenology, such as that of Husserl, Seheler, and Hartmann, would not be impressed by that claim, which rested on an extraordinarily impoverished conception of consciousness.
Another, related argument is that ethical properties, such as goodness and badness, are very mysterious; indeed, Thomas Nagel describes the badness of pain with just that word, and so he is unwilling to allow that there is such a property as the badness of pain, preferring instead to speak of “the fact that there is reason for anyone capable of viewing the world objectively to want it [the pain] to stop,”v a fact he presumably finds crystal clear. But an adequate metaphysics would include a more sophisticated theory of properties, or if you wish of universals, both generic and specific, and will hold that if there is such a property as goodness or badness, it would almost certainly be a generic property, not a specific property like a shade of yellow. Then the mystery some find in it would be dispelled when it is pointed out that if, for example, a person is good because she is kind, her goodness is not a property merely additional to her kindness, even if supervenient upon it (whatever, if anything, this may amount to beyond its purely formal characterization)vi but rather is a genus of which kindness is a species. If so, goodness would be no more mysterious a property than, say, color and shape as such, as generic properties, are mysterious.
A third argument is that science does not and need not appeal to ethical properties, or that ethical properties serve no explanatory role of the sort familiar in science. But only a crude, still unargued, commitment to “the scientific image of the world” could force us to infer from this fact, if indeed it is a fact, that there are no ethical properties. Notoriously, in this respect even mathematics seems to be rather like ethics, as the worries of causal theorists of knowledge about it show, though recently there have been attempts to dispel these worries, Of course, in the broader, deeper, and more proper sense of “explain,” namely, to render plain, to make intelligible, ethical properties do have an explanatory role. They render intelligible our moral consciousness, in particular its intentional character, and thus the moral fabric of our life, In this sense they are far more explanatory than the entities postulated by even the best confirmed physical theories.
A fourth argument is that disagreement about ethical matters is widespread and unresolvable. But, if what I have already said about where and why such disagreement occurs is correct, its existence, such as it is, is hardly a reason for abandoning cognitivism. And the rest of what I shall say may be thought of as a further explanation of it, one that also is entirely compatible with cognitivism.
However, my purpose here is not to evaluate these arguments or to defend realism and cognitivism. I have tried to do so in detail elsewhere.vii They are technical philosophical arguments, and whatever their merits, they are hardly the reason for the peculiar fact about ethics with which I am concerned. My purpose here is to understand why Sidgwick’s question, with which I began, arises even in ordinary ethical thought, despite its unqualified realism and its qualified cognitivism. We are in a position now to consider the first part of Sidgwick’s answer to that question.
Being often torn between what we know we ought to do and what we want to do, we naturally tend to demand reasons for doing the former. But if we do know what we ought to do, then we have sufficient reasons, perhaps all the reasons that there could be, for believing that we ought to do it, and so Sidgwick’s question “Why should 1 do what I see to be right?” seems to contain its own answer, namely, “Because I see it to be right,” though for reasons I shall note later he ignores this fact. And let us suppose we have also taken stock of all of our inclinations for not doing what we see to be right. The resolution of the conflict then calls not for additional reasons or considerations, since ex hypothesi there aren’t any, even if we exploit the ambiguity of the promiscuous word ‘reason” between what Hutcheson called justifying and exciting reasons. Rather, as Prichard pointed out,viii the resolution of the conflict calls simply for decision. Yet, being by nature truth-respecting creatures, we want our decisions to be based, as we say, on the facts, and feeling the power of our contrary inclinations we are tempted to question whether after all we do really know what we ought to do and, if we generalize the question, which as truth-respecting creatures we may also do, to suppose that there is a special epistemological problem about ethics. And then we either proceed to try to resolve it or cut the knot by adopting some noncognitivist or irrealist position. Kant wrote: “Man feels in himself a powerful counterweight to all the commands of duty presented to him by reason as so worthy of esteem--the counterweight of his needs and inclinations. From this there arises a natural dialectic--that is, a disposition to quibble with these strict laws of duty, to throw doubt on their validity or at least on their purity and strictness...”ix
What makes this “natural dialectic” possible is that while a valid argument is necessarily truth-preserving, it is not necessarily belief-preserving, that there is a logical gap between argument and belief, and in general (thus including judgments made not on grounds of argument but on grounds of self-evidence) between reason and belief. However, contrary to what Sidgwick implies, this is not a fact limited to ethics: it is a fact about human nature that there is a duality of reason and belief but this duality casts no legitimate doubt on the pronouncements of reason. The most familiar and least controversial example of this fact is the relative inefficacy of argument on political and religious beliefs. Some conservatives eventually become liberals, and some liberals eventually become conservatives, but hardly as a result of argument. And some sophomores can find nothing wrong with the premises and validity of a simple version of the cosmological argument for the existence of God and yet resolutely refuse to believe that its conclusion is true, presumably because they consider themselves too enlightened to do so, though they cheerfully accept a causal argument for the existence of the external world which has similar structure but virtually unintelligible premises about our “sensations.”
Far less common, yet hardly unfamiliar, examples concern precisely what Sidgwick called physics and mathematics. In a recent article, a psychiatrist tells us of a person who is ‘unable to stop washing,” because he is “haunted by the notion that he “was dirty--in spite of the contrary evidence of his senses,” and of another person who constantly checks the door to make sure that it is locked. The psychiatrist I am quoting calls such a person “an ultimate skeptic who cannot credit his sense data or his attempts to refute the obsession by means of logic. The sufferer cannot accept reassuring information, he is unable to believe.”x Elsewhere, our psychiatrist writes: ‘This is a disease that may be thought of as skepticism gone wild.”xi (The French aptly call it folie de doute). An example involving mathematics is described as follows by a patient. “I have stayed till midnight at my laboratory compelled to check my computer’s simplest calculations by hand. The work is unpublished because I can never be certain that the numbers were averaged correctly.”xii An example of failure to believe the conclusion of a simple deductive argument, even after accepting its premises and acknowledging its validity, might be the often cited case of a parent who refuses to believe that her son is dead, but not because of any doubts about the evidence or about its logical sufficiency.
We don’t need to go to psychiatry for such examples. It is the trenchancy of the descriptions of some of them that led me to include the above quotations. Such cases are hardly unknown to us, and what they exemplify may be familiar to us, I hope in a very mild form, from our own experience--especially if we are episteniologists. Nevertheless, though in mathematical and empirical ordinary thought such cases do occur, they are relatively rare and we are inclined to regard them as cases of madness - or, if you imagine this to he more illuminating, follow the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM III) and call it Obsessive-Compulsive Disorder (OCD). In moral ordinary thought, however, they are common, and we regard people who tend to ignore the dictates of reason not as mad or even as unusual but just as immoral.
Perhaps the most discussed case in ethics is that of the egoist. It occupied much of Sidgwick’s attention and perhaps was part of what he had in mind when pointing out the special epistemic demands we make on ethics and then explaining them by appealing to the diversity of methods and principles present in ordinary ethical thought. I think moral philosophers take it more seriously than it deserves to be taken, hut it is very familiar and provides us with a particularly suitable example. Sidgwick asserted that the principle of benevolence, roughly, that one ought to promote the general good, which for him was universal happiness, is one of the three maxims of ethics he found self-evident, comparable in self-evidence to mathematical propositions (pp. 383, 507). Yet he also wrote, with obvious a that even if one does admit the self- evidence of this principle one “may still hold that his own happiness is an end which it is irrational for him to sacrifice for any other” (p. 498), even though earlier he had held that to describe an action as rational, or reasonable, is to say that it ought to be done (p. 23; see also the entry for section 1 of chapter III in the analytical table of contents, p. xxiv), that is, he had explicated the notion of a rational action in terms of the notion of an action that ought to be done, the exact opposite of current fashion. So Sidgwick found it self-evident that one ought to promote the general happiness, i.e., that this is the rational thing to do, and yet claimed that the egoist might justifiably hold that this is irrational. The inconsistency seems obvious. Why did Sidgwick fall into it?xiii
The answer, I suggest, can he found in Sidgwick’s writing of one’s special concern for one’s own happiness, of its being important to one, of the alleged fact that “it is contrary to Common Sense to deny that the distinction between any one individual and any other is ultimate and fundamental, and that consequently “I” am concerned with the quality of my existence as an individual in a sense, fundamentally important, in which I am not concerned with the quality of the existence of other individuals” (p. 498). As Moore was to exclaim later, What does all this mean?xiv Surely not that the mere fact that one has special concern for something, that the thing is important to one, is eo ipso relevant to the question whether the thing can be rationally desired and pursued. (This may he Bernard Williams’s view, but it was not Sidgwick’s). And how does the mere numerical difference between oneself and another generate a qualitative difference between one’s moral relation to oneself and one’s moral relation to the other? Egoism maybe thought of as the ultimate policy of discrimination, namely, discrimination against everyone else, infinitely irrational when compared with political discrimination on grounds of race or sex, for race and sex at least are qualitative differences, irrelevant though they are. The egoist does not appeal even to irrelevant differences, since it is no part of his view that he has special qualities justifying his policy of discrimination. Surely what can be legitimately meant by Sidgwick in this passage is only that the self-evidence of the principle of benevolence may, indeed usually does, fail to extinguish selfish inclination when in conflict with it, and that such inclination is common, indeed natural. We do not have here a paradox, or a dualism, of practical reason, which could be given only a theological resolution, as Sidgwick seemed to believe, but just an especially vivid and very familiar example of the logical gap between reason and belief. Of course, one’s inclinations need not he in conflict with reason; they could he directed toward great goods, including one’s own happiness, the pursuit of which may be quite compatible with the dictates of reason.
Tile chief attraction of ethical egoism, I suggest, is psychological, not logical, or metaphysical, or epistemological, or dialectical, except in the case of philosophers who defend it for systematic reasons, e.g., because they accept psychological egoism, or some evolutionary hypothesis, or some Freudian theory, or possibly metaphysical solipsism. Solipsism seems to provide by far the most interesting reason; it is a logically impeccable (though trivially so) basis for egoism, and is certainly itself quite defensible on epistemological grounds. But, for obvious reasons, actual defenses of it are even harder to find than straightforward defenses of egoism. At any rate, I shall not he concerned with such philosophers in this paper. The truth is that all of us are to a great extent selfish, myself and those who may agree with what I say in this paper no less (possibly more) so than those who would not, and so we have a motive for falling into bad faith by believing that our selfishness is justified by reason. Or, discovering that no such justification is possible, we fall deeper into bad faith by asking the nonegoist to justify his position and by insisting that he do so in ways that clearly are impossible, e.g., to show us a property of goodness or badness, of rightness or wrongness, that is visible to us as colors are, or to ground his position in, of all things, physics! And when this demand is, of course, not met, we may sink even deeper into had faith. We may conclude that ethical judgments cannot be justified at all because they lack cognitive content and merely express personal attitudes or interests, Of course, we need not go that far. Instead of falling into bad faith, we sometimes resolve the conflict by making excuses, in the ordinary sense of this term. (“I couldn’t help ii. The money was too much to pass up”). But ethics affects us much too deeply and broadly for mere excuses to he very effective.
What happens, I suggest, would not he unlike an investor’s becoming a noncognitivist regarding mathematics because of the disastrous, life-shattering results of his calculation of his assets after a stock market crash. It’s just that we would regard such a case as one of madness though probably temporary, presumably because it would be so extraordinary. And it would be extraordinary because a mathematical conclusion one strongly dislikes would be an isolated case even in one’s own life; one has many other uses for mathematics that involve no conflict between reason and inclination. There is a clear sense in which mathematics does not directly have a practical aim, as ethics does. Had it been otherwise, we might have made unreasonable demands for justification on mathematics as well.
Sidgwick tried to explain why people commonly ask, ‘Why should I do what I see to be right?,’ even though, as I have said, the question seems to contain its own answer (“Because I see it to be right,) a fact he ignored presumably because of his concern over egoism, his belief that egoism is not irrational.xv But what I have done instead is to try to explain the apparently very different fact that people often question, indeed deny, what they see to be right. As I shall point out shortly, Sidgwick’s question was still on the right track, but only if it is not cheapened by being confused with the question of weakness of the will (to which, strictly speaking, it is irrelevant), or with the question supposedly asked by the egoist, whether doing what one sees to be right would be in one’s self-interest (which has an obvious answer, acknowledged by Sidgwick: sometimes it would, sometimes it would not, an answer even Plato’s philosophers acknowledged when they chose to return to the cave and serve as kings). Seduction may succeed because passion literally overpowers moral conviction, yet without shaking it in the least. But it may succeed also because moral conviction is shaken, begins to waver, becomes corroded by “rationalization,” perhaps is entirely destroyed. The former case is true weakness of the will, which is made possible by the logical gap between belief and action, which often is also a gap between belief and feeling, since action often involves feeling. The latter case is the very different phenomenon with which this paper is concerned; it is made possible by the logical gap between reason and belief.xvi (Incidentally, the gap between belief and action, or feeling, accounts, at least in part, for the undoubted fact that the cognitivist is in general not more likely to be moral than the noncognitivist: the former’s moral failures are generally explained by weakness of the will, though sometimes by the falsity of his moral beliefs, the latter’s moral successes are sufficiently explained by the inability of belief to destroy or even weaken deeply ingrained ways of acting and feeling).
And the question before us should not he confused with that supposedly asked by the egoist, for besides the obviousness of the answer to it I mentioned earlier, egoists should not ask it at all since they see nothing to be right unless it is in their self-interest. Moreover, it is not only with selfish inclination that reason may conflict. Altruistic behavior (e.g., a foolish self-sacrifice prompted by sexual passion) and behavior not directed toward people at all (e.g., devotion to an unworthy artistic goal) are also sometimes contrary to reason, perhaps by violating Sidgwick’s principle of prudence,xvii and occasion self-deceiving questioning of its dictates.
To belong in the context where it occurs (the beginning of The Methods of Ethics), to have the generality evidently attributed to it by him, and to he essentially epistemological, as it obviously was taken by him to be, Sidgwick’s question must be understood rather as asking, at least in part, “Why should I believe what I see to be right?” And this is merely a special case of the question ‘Why should I believe what I see to he true,” just the question Sidgwiek claimed people never ask. Whether they do or do not is beside the point, though we have seen reason to think that sometimes they may, at least by implication, when they say that they do not, can not, believe what they see to be true. What is to the point is that the question, though it too seems to contain its own answer, makes sense by in effect acknowledging the gap between reason and belief and thus drawing attention to the deeper issue before us.
This is why my explanation rested on pointing out that there is a general gap between reason and belief, that there are cases even outside ethics in which this gap is evident. It’s just that such eases, for example, refusing to accept the result of a calculation or the death of your child, are rare. In ethics, on the other hand, precisely be cause its aim is essentially practical, that is, concerned with conduct, while the aims of mathematics and physics are only incidentally or occasionally practical in this sense, such cases are common. And this is why we make special epistemic demands on ethics. For it is not at all obvious why we generally, though not always, believe what we see to he true even outside ethics. Or, if you think that seeing something to be true entails believing it, I can put the point as follows: it is not obvious why we generally believe the conclusion of an argument the validity and premises of which we have accepted, or why we generally believe a proposition we find self-evident. There is no contradiction in saying that someone accepts a certain argument as valid and its premises as true but does not accept its conclusion, or finds a certain proposition self- evident in the sense that he finds himself unable to imagine or conceive of being mistaken about it, but does not believe it. Indeed, as Hume pointed out regarding our belief in an external world, sometimes the opposite occurs: belief is unshakeable even if we think reason is unequivocally opposed to it. That usually, outside ethics, we believe, to use Sidgwick’s terminology, what we see to be true is a deep fact about human nature, whatever its explanation, the fact that we respect truth. But it is also a deep fact about human nature that often we do not believe, not just do not do, what we see to be true in ethics. And, I suggest, we do have an explanation of this latter fact, namely, that in ethics we often do not want what we see to be true to he true. So we look for ways of denying that it really is true and make extravagant epistemological demands on ethics to show us that it really is true, demands we would not ordinarily make on mathematics or physics.
For example, we ask for a derivation of ought-statements from is-statements in ethics, but in mathematics and physics it would not even occur to us to ask for a derivation of is-statements from ought-statements. (An exception is John Leslie, in his important book Value and Existence.)xviii Or we think that the reality of ethical properties must first be demonstrated if ethics is to be a genuinely cognitive discipline, but we seldom think, with several philosophers now being exceptions, that the reality of numbers must first be demonstrated if arithmetic is to be a genuinely cognitive discipline. Or we think that we must have a clear account of the nature of our knowledge of an ethical proposition such as “Friendship is a good” if we are to accept it as genuine knowledge, but we do not think that we must have a clear account of the nature of our knowledge of the proposition “Three and two is five” if we are to accept it as genuine knowledge. Or we demand that ethical facts play an explanatory rote, though, again with some philosophers being exceptions, it would not even occur to us that any indirect explanatory role mathematics may be supposed to play is what provides it with its lofty cognitive status. My suggestion has been that there is nothing in ethics that explains such discriminatory treatment. What explains it is in us, the special stake we all have in ethics. We are truth-respecting creatures, in the sense that we want to believe what we see to be true, for its own sake, not just because believing it would be useful. But we also have other desires, which may conflict with our desire for truth. It’s just that outside ethics such a conflict is not pervasive or systematic. So, outside ethics, we do not usually ask why we should believe what we see to he true or demand proofs that it really is true.
But our explanation is not yet complete. We need also an explanation of why we take ethical truth as relevant. Why don’t we just set it aside when it conflicts with our inclinations, even though still respecting it, still acknowledging it, and not making unreasonable demands for justification of our acknowledgment of it? What needs to be added to our explanation is that in addition to being by nature creatures who respect truth, we are by nature also creatures who love the good (and the right, if this is not reducible to the good), in the sense that we want to do it, preserve it, pursue it, and promote it. Logically, the ultimate question that generates the exorbitant demand for justification in ethics is “Why should we do or promote the good of whose goodness we are aware?,” and so Sidgwick’s statement of it was on the right track. But its special importance becomes evident only if it is seen to presuppose the significance of the question “Why should we believe what we see to be true?” Indeed, we ask it because our doing or promoting the good is so very frequently in conflict with our inclinations. But what is interesting and important is that very often we try to resolve the conflict by questioning the epistemic credentials of our belief that what we see as good is really good. For our respect for truth does not allow us to simply ignore our awareness of the good, and our love of the good does not allow us to simply ignore the good. So, unwilling to believe and do what we see to be good, we face three options. One is to scorn truth, but this would violate the substance of our life as truth-respecting, as rational beings, it would he incompatible with our self- respect. Another is to scorn goodness, but this would violate the meaning of our life as lovers of the good, it would be incompatible with our integrity. The third option is the one I have discussed in this paper: to question, perhaps even deny, the truth about the good we see and thus to seem to ourselves to avoid scorning truth as well as to avoid scorning goodness, and thus to enjoy with what seems to us clear conscience the comfort of doing what we feel like doing.
Of course, many questions may, and ought to he, asked now. I have appealed to human nature. How do we know what human nature is and what it involves? Clearly, no lexical or stipulative definition, or biological criterion, or psychological statistical generalization is what we want. I shall confine myself here to making the somewhat cryptic, hut radical in its implications, suggestion that one’s knowledge of human nature is like one’s knowledge of the core of one’s native language. Stanley Cavell writes: “[when a philosopher appeals to ordinary language he] is not claiming something as true of the world, for which he is prepared to offer a basis--such statements are not synthetic; he is claiming something as true of himself (of his world, I keep wanting to say) for which he is offering himself, the details of his feeling and conduct, as authority.”xix (My emphasis).
I have also used the Sartrean term “bad faith,’ though not exactly in his sense but in the broader sense of “self-deception,” which can be taken to involve not just believing and not believing (or even disbelieving) one and the same proposition at the same time but also the manifestations of the gap between reason and belief with which I have been concerned. And it may now be asked, “How is self-deception possible?” There has been valuable work in recent years on this question,xx but its authors probably would (and many explicitly do) agree that any answer proposed must be defended by answering the obvious next two questions. The first is, “What is the nature of belief?” We would hardly understand what self-deception is if we don’t have an adequate understanding of belief. And we don’t seem to have such an understanding. For example, it is most doubtful that there are occurrent states, conscious or unconscious, that can reasonably be called beliefs (“believings”), and the notion of dispositional belief is even more obscure than the general notion of disposition, since to be belief a dispositional belief must have propositional content, and how can a disposition have that? (A dispositional belief can he manifested in the assertions of many very different propositions, just as bad temper can be manifested in many very different actions.) My suggestion, again radical but this time hardly cryptic, is that recent philosophy has for too long been leaning on the crutch of the concept of belief, and that “the way of beliefs” is due for retirement, just as “the new way of ideas” of the seventeenth and the eighteenth centuries was due for, and forced into, retirement.
The second question about self-deception is, of course, “What is the nature of that self-deceiving self?” Contemporary analytic philosophers have discussed at great length the question of personal identity, hut not enough the logically prior question of the nature of that which may or may not remain identical through time, of the entity Hume (and Wittgenstein in the Tractatus) could not “find.” Sartre did devote much attention to it, precisely in order to explain the possibility of bad faith. We may not accept his radical theory of selfhood, but I suggest that any adequate theory would have to be radical.
Needless to say, this is not the place to embark on a discussion of any of these questions. Yet it would be either false sophistication or intellectual cowardice to insist that until they are answered we should not attempt to understand why exorbitant demands for justification are made on ethics.xxi
The Methods of Ethics, 7th ed. (Chicago: The University of Chicago Press, 1962), pp. 4-6.
See his “Does Moral Philosophy Rest on a Mistake, included in Moral Obligation and Duty and Interest (Oxford, London, New York: Oxford University Press, 1968). pp. 1-2. 16-17.
iv If we think that mathematics is an exception, let us recall Spinoza’s view that, by using what (in On the Improvement of the Understanding) he calls “the second mode of perception,” ”tradesmen” acquire knowledge of universal mathematical propositions “from their experience with simple numbers.” The propositions are learned inductively and continue to he accepted partly because they explain some of the tradesman’s successes.
v Thomas Nagel, The View from Nowhere (New York, Oxford: Oxford University Press, 1986), p. 144.
vi For various formal characterizations of the alleged relation of supervenience, see Jaegwon Kim, ‘Concepts of Supervenience,” Philosophy and Phenonenological Research, 65 (1984), pp. 257-70.
vii Skepticism in Ethics (Bloomington and Indianapolis: Indiana University Press, 1989).
viii See ‘Duty and Interest,” in Moral Obligation and Duty and Interest, pp. 225-26.
ix Groundwork of the Metaphysics of Morals
(New York: Harper and Row,
1964), p.73, trans. H. J. Paton.
x Judith L. Rapoport, The Biology of Obsessions and Compulsion,’ Scientific American, vol. 260, n. 3 (March 1989), pp. 82-89. Evan Fales drew my attention to this article.
xi Judith L. Rapoport, The Boy Who Couldn’t Stop Washing: The Experience and
Treatment of Obsessive-conipuirit’e Disorder
(New York: Dutton, 1989), p. 18.
Ibid., p. 25.
xiii 1 should note that the principle of prudence, which is one of the three princi ples Sidgwick found self-evident, merely enjoins impartial concern for all temporal parts of ones conscious life (p. 381), and thus, contrary to what Sidgwiek says much later (p. is compatible with the principle of benevolence. Sidgwicks third self- evident principle is that of justice: “if a kind of conduct that is right (or wrong) for me is not right (or wrong) for some one else, it must be on the ground of some difference between the two cases, other than the fact that I and he are different persons” (p. 379). This principle does not exactly contradict egoism, since the egoist may be willing to universalize his view, but it is obviously incompatible with the spirit of egoism.
xiv Principia Ethica (Cambridge: Cambridge University Press, 1903), p. 99.
xv Whether the question really contains its own answer would depend, of course, on how we interpret the crucial words “should” and “right.” If the question were ”Why ought Ito do what I see I ought to do,” then of course it does contain its own answer. To be interesting, it must be understood as presupposing a gap between either (a) duty and self-interest, or (b) duty and action, or (e) reason and belief, and then as asking for a bridge between them. The word “should’ is vague enough to serve in all three cases, as well as in the paradigmatic case in which it is a synonym of “ought.”
xvi 16 This does not resolve the Greek puzzle over akrasia, usually translated as incontinence or weakness of the will, for the gap between belief and action may still be puzzling, but the introduction of the distinct notion of the gap between reason and belief seems to me to render the puzzle more manageable, at least by reducing it to two, though intimately related, puzzles.
xvii See above, note 13.
xviii Totowa, N.J.: Rowman and Littlefield, 1979.
xx For example (and these are only examples from a vast literature), Herbert
Fingarette, Self-Deception (Atlantic Highlands, N. J.: Humanities Press, 1969); Mike
\V. Martin, Self-Deception and Morality (Lawrence, Kansas: University Press of
Kansas, 1986); the essays by various authors in Mike W. Martin, Self-Deception and
(Lawrence, Kansas: University Press of Kansas, 1985).
xxi Versions of this paper were read at meetings of the Metaphysical Society of America, The Central States Philosophical Association, and the Department of Philosophy of The University of Iowa. I am indebted to many philosophers who commented on it at those meetings. I have also benefited from generous and very valuable comments by the Editor.