|公 法 评 论
(reprinted from Philosophy & Public Affairs 25, no. 2 (Spring 1996)
Objectivity and Truth:
You'd Better Believe It
By Ronald Dworkin
New York University
Is there any objective truth? Or must we finally accept that at bottom, in the end, philosophically speaking, there is no "real" or "objective" or "absolute" or "foundational" or "fact of the matter" or "right answer" truth about anything, that even our most confident convictions about what happened in the past or what the universe is made of or who we are or what is beautiful or who is wicked are just our convictions, just conventions, just ideology, just badges of power, just the rules of the language games we choose to play, just the product of our irrepressible disposition to deceive ourselves that we have discovered out there in some external, objective, timeless, mind-independent, world what we have actually invented ourselves, out of instinct, imagination and culture?
The latter view, wearing names like "post-modernism" and "anti-foundationalism" and "neo-pragmatism," now dominates fashionable intellectual style. It is all but inescapable in the unconfident departments of American universities: in faculties of art history, English literature, and anthropology, for example, and in law schools as well. More sophisticated forms of the same deep skepticism have been influential within academic philosophy for many centuries.1 They come in two versions: a general, all-encompassing wholesale version, which attacks the very idea of objective truth about anything, and a limited, selective version that concedes objective truth to "descriptive" claims, including mathematical ones, but denies it to "evaluative"--moral or ethical or interpretive or aesthetic--ones.
In both the wholesale and selected versions, these influential theories are "archimedean," as I shall call them. They purport to stand outside a whole body of belief, and to judge it as a whole from premises or attitudes that owe nothing to it. Of course they cannot stand outside thought altogether, to deny real truth to every thought. For even archimedeans need some place to stand, as their progenitor conceded. They must assume that some of what they think (at an absolute minimum, their beliefs about good reasoning) are not just their own or their culture's invention but are true or valid--indeed "objectively" so. Otherwise they could only present their views as "subjective" displays in which we need take nothing but a biographical interest. Skepticism, in the sense of disbelief, must be built up from belief of some kind; it can't be skeptical, as we might put it, all the way down. The wholesale version of archimedeanism proposes, in extreme forms, to stand outside as much as possible. The selective version I shall mainly discuss proposes, more modestly, to stand outside all the evaluative domains.2 These selective archimedean skeptics offer to justify their skeptical claim--that these domains cannot provide objective truth--from premises that are not themselves evaluative. They argue, they say, not from moral or ethical or aesthetic assumptions, but from non-evaluative theories about what kind of properties exist in the universe, or how we can gain knowledge or reliable belief about anything.
In this essay I concentrate on this selective version of archimedeanism--about truth in the "soft" domains of morality and art rather than the "hard" ones of physics or mathematics. Selective skepticism about value, under the name of "subjectivism" or "emotivism," has for a long time been regarded as the most plausible form of archimedean skepticism. It is also the most dangerous. No one--not even the most committed post-modernist or anti-foundationalist--thinks his views should affect how physicists or mathematicians actually work. But it is now strenuously argued that since there is no objective truth about interpretation or art or morality there can be no standard of merit or success in artistic or moral or legal thought beyond the interest a theory arouses and the academic dominion it secures. This auto-da-fe of truth has compromised public and political as well as academic discussion.
I argue that even this selective form of archimedean skepticism is misconceived. Any successful--really, any intelligible--argument that evaluative propositions are neither true nor false must be internal to the evaluative domain rather than archimedean about it. So, for example, the thesis that there is no right answer to the question whether abortion is wicked is itself a substantive moral claim, which must be judged and evaluated in the same way as any other substantive moral claim; the thesis that there is no right answer to the question how a clause of the Constitution should be understood is a legal claim, which must, again, be judged or evaluated like other legal claims; the statement that it is indeterminate whether Picasso or Braque was the greater creative artist is an artistic or aesthetic claim; and so forth. So even this selective form of skepticism must be limited. We can't be skeptical, even about values, all the way down.
II. EXTERNAL AND INTERNAL SKEPTICISM
Archimedean or external skepticism is to be contrasted with internal skepticism.3 A skeptical thesis about value is internally skeptical if it presupposes the truth of some positive value judgment. I shall use moral skepticism as the leading example of internal skepticism, though it is easy enough to construct examples in other evaluative domains as well, as we shall see. I shall assume that all readers, including those drawn to archimedean skepticism, accept that our shared language and common experience include assessments on what we take to be a distinct moral dimension. I shall not attempt to define that dimension, or to separate the predicates we use to employ it. If I am right, no helpful definition of morality as a whole can be given. In any case, the existence of a moral dimension of assessment in our experience is not in question, though its status is.
The moral predicates--those we use in that dimension of assessment--include abstract ones like "morally good," "morally bad," "just," "unjust," "morally obligatory," and "morally prohibited," and also less abstract ones, including "thick" predicates like those used to identify moral virtues and vices. A positive moral judgment ascribes a moral predicate to an act or person or event; a negative one denies such an ascription. The class of positive moral judgments includes not just simple ascriptions, like the judgment that equality is good or abortion wicked. It also includes more complex forms of such ascription, including conditional ones like the claim, for example, that equality is morally desirable provided it does not lead to indolence and counterfactual ones like the judgment that abortion would still be wicked even if no one thought it was. It also includes counterfactual ascriptions that assume that acts or events or people would have moral properties were certain specified circumstances satisfied, though it declares that they are not satisfied; it includes, for instance, the familiar claim that since there is no God morality is bunk. It also includes claims about morality as a whole that embed or presuppose direct or conditional or counterfactual ascriptions of evaluative properties. The utilitarian claim that the most fundamental point of morality is to maximize overall human happiness, for example, assumes that human happiness is a good, and the rival claim that its most fundamental point is to recognize and honor the inherent worth of every human being assumes that human beings have inherent worth. It is an interpretive question whether a general statement about morality is a positive moral judgment. A sociological account of other peoples' moral convictions is not, because it does not itself endorse or presuppose any moral assessment.
An internally skeptical position, then, denies some group of familiar positive claims and justifies that denial by endorsing a different positive moral claim--perhaps a more general or counterfactual or theoretical one. Many people are internal skeptics about conventional sexual morality, for instance. They deny that sexual acts are inherently good or bad, right or wrong, because they believe that suffering is the only thing that is inherently bad, and they doubt that either heterosexual or homosexual acts, just in themselves, either prevent or promote such suffering. Other people purport to be skeptics about the place of morality in foreign policy: they say it makes no sense to suppose that a nation's trade policy can be either morally right or wrong. Though this opinion is indeed skeptical about moral judgments some people make--that American policy in Latin America was often unjust, for example--the opinion usually presupposes some positive moral judgment: for example, that a nation's officials should always act with only the interests of their own citizens in mind.
These are limited versions of internal skepticism, but there are also familiar versions that are broader or even global. The opinion that morality is empty because there is no God presupposes the substantive view that a supernatural will is a plausible and the only plausible basis for morality. The more modern opinion that morality is empty because all human behavior is causally determined is usually supported by a different but equally substantive moral conviction: that it is unfair to blame people or hold them responsible for what they could not help doing. The popular contemporary thesis that morality is relative to culture, which is skeptical about all universal moral claims, often relies on the conviction that universal claims are imperialistic and insufficiently respectful of other cultures than our own. People who might be tempted to some form of broad or global skepticism for one or another of these positive moral reasons might well be immune to others. Someone who believes that the only ground of morality is a divine judgment, and who is therefore vulnerable to the first argument if he loses his faith, may not be vulnerable to the others so long as he does not. He may think that God has solved the problem of free will, and that cultural diversity is amply explained by God's fastidiousness in sharing His insights or grace.
As these examples suggest, broad or global internal skepticism often presupposes a counterfactual positive moral judgment. It claims that certain conditions, which it supposes would support positive moral ascriptions if they did hold, in fact do not hold. Such skepticism cannot include skepticism about the counterfactual positive judgments on which it rests, and so it cannot be skeptical about value, as I put it earlier, all the way down, any more than the more limited forms of internal skepticism can. That feature is therefore common to all internal skepticism. So is a second feature. Since internal skepticism is a substantive position, it has direct implications for action. If we are skeptical about sexual morality, then we cannot consistently censure people for their sexual choices, or lobby for outlawing homosexuality on moral grounds.
I emphasize these two features of internal skepticism--that it rests on positive moral judgments and that it has direct implications for action--because the external, archimedean skepticism that I shall examine is supposedly different in both respects. It is supposedly austere, in the sense that it does not rely even on very general or counterfactual or theoretical positive moral judgments. And it is supposedly neutral in the sense that it takes no sides in substantive moral controversies. Both features come together in many archimedeans' description of their project. Their skepticism is directed, they say, not to substantive moral convictions but rather to second-order opinions about such convictions. An archimedean agrees with most people that genocide and slavery are wrong, for example. He only denies that these practices are really wrong, or that their wrongness is "out there" in reality. He insists, rather, that the wrongness is "in here," in our own breasts, that we have "projected" moral quality onto reality, that events are not, in themselves, right or wrong good or bad, apart from our emotions or projects or conventions, that our moral convictions are not, after all, true or false or part of what we do or do not know, but are only, in complex ways, products of our invention or manufacture. He is skeptical, in other words, not about convictions but about what we might call the "face value" view of these convictions.
That is the view you and I and most other people have. We think that genocide in Bosnia is wrong, immoral, wicked, odious. We also think that these opinions are true--we might be sufficiently confident, in this case at least, as to say that we know they are true--and that people who disagree are making a bad mistake. We think, moreover, that our opinions are not just subjective reactions to the idea of genocide, but opinions about its actual moral character. we think, in other words, that it is an objective matter--a matter of how things really are--that genocide is wrong.
Selective archimedeans distinguish between the first of these last three statements, which sets out a positive moral judgment, and the second two, which, they say, do not, but rather express metaphysical or philosophical about the nature of positive moral judgments. We might call statements like the first one I- (for internal) propositions, and statements like the second two E- (for external) propositions. The distinction explains why archimedeans often deny that they are skeptical about morality. They regard the face value view--that our belief about genocide is true and that it describes an objective matter--as a set of E-propositions, and therefore not part of substantive morality, and they think that when they contradict those E-propositions they leave all I-propositions, like the claim that genocide is wicked, untouched.
That view of the archimedean project explains its claims both to neutrality and austerity. It claims neutrality about the substance of ordinary positive moral convictions because it takes no sides between the opinion that terrorism is immoral and that it is not. It claims austerity because it purports to rely on non-moral arguments to defeat the face value view, which it holds to be bad philosophy, not mistaken morality. It relies, for example, on the argument that there is no objective moral reality "out there" for moral beliefs to match, and that the idea of objective truth about morality is an illusion for that reason. There seems no trace of any even very abstract moral conviction or claim in this argument. On the contrary, it seems to have been constructed on a special philosophical platform from which a philosopher might look down on morality as a whole.
It is the combination of these two dimensions of externality-neutrality and austerity-that gives archimedeanism its great appeal. Neutrality is particularly important. As I said, many people, particularly in Western secular society, feel that it is arrogant in the face of great cultural diversity to claim that everyone who disagrees with them is in error. But global internal skepticism about morality seems out of the question. We can't believe that there is nothing morally objectionable about terrorism or genocide or racial discrimination or clitoridectomy. Archimedean skepticism offers people torn in that way exactly what they want. It is agreeably ecumenical. It allows its partisans to be as culturally modest and relative as anyone might wish, to abandon all claims as to their own morality's ultimate truth or even superiority to other moralities. But it allows them to do this while still embracing their morality as enthusiastically as ever before, denouncing genocide or abortion or slavery or gender discrimination or welfare cheats with all their former vigor. They need only say that they have revised their view not about the substance but about the status of their convictions. They no longer claim objective truth for these convictions; they no longer think their thoughts "mirror" an external "reality." But they still hold them with the same intensity. They can be as willing to fight or even die for their beliefs as they ever were, but now with a difference. They can have their moral convictions and lose them too. Richard Rorty calls this state of mind "irony."4
Neutrality, the supposed first dimension of externality, gives archimedeanism that crucial popular advantage over internal skepticism. Austerity, the second dimension, gives it a more technical and defensive, but still crucial, philosophical advantage. The internal skeptic can't be skeptical all the way down, as I said, because he builds his skepticism on some positive moral position. If he claimed that no moral judgment or conviction or instinct of any kind could be true, he would condemn his own theory. But if the archimedean's argument can be constructed wholly independently of any positive moral claim or assumption, then he can be fierce and unrelenting in denying the objective truth of any positive moral judgment without contradicting his own enterprise, because he stands in firm philosophy rather than some soft or dissolving domain of value. Though these two dimensions of externality work together, they are logically independent of one another. An archimedean critique can purport to be austere without being neutral: it can offer a supposedly non-moral proof that all positive moral judgments are false, which would take sides with those who, as a substantive matter, reject those judgments. So I shall consider the two dimensions of external skepticism separately.
III. NEUTRALITY: THE TWO WORLDS HYPOTHESIS
The Natural Reading
We must be clear, first, what kind of neutrality is in question. Suppose an argument breaks out about abortion in particular circumstances-say, when a serious fetal abnormality has been diagnosed. One side insists that abortion is forbidden in these circumstances, and the other that it is mandatory. Of course a third view is available: that it is permissible but not mandatory. This third view is neutral as between the original contestants, but it is not neutral in the argument they are having. It takes a position-it holds them both wrong-and it has a direct implication for action because it holds that it would be wrong to interfere with an abortion decision, either way, on moral grounds. Now suppose a fourth party, who insists that all three of these views are in error because moral judgments cannot be true or false, or because there are no moral properties, or on some other archimedean ground. This fourth view is not neutral either. It insists that it is a mistake both to demand and to oppose an abortion on moral grounds. It therefore has the same implications for action as the third view, and though it is couched in more ambitious language, it is no more neutral. If the archimedean is to show that his kind of skepticism really is neutral, he must do more than just point to the logical space between judgments of obligation and prohibition, or remind us that saying that a proposition is neither true nor false is different from saying that it is false. He needs more structure and a more ambitious argument than that.
He must show that the target of his criticism is something altogether distinct from the ordinary substantive moral claims we all make, so that we may accept his arguments while continuing to judge and act, in the moral dimension, as we did before. Richard Rorty, who is the most prominent American exponent of wholesale archimedean skepticism, aims to show this by distinguishing between levels of thought or discourse. Here is a recent and characteristic statement of his view:
Given that it pays to talk about mountains, as it certainly does, one of the obvious truths about mountains is that they were here before we talked about them. If you do not believe that, you probably do not know how to play the usual language-games which employ the word "mountain." But the utility of those language games has nothing to do with the question of whether Reality as It Is In Itself, apart from the way it is handy for human beings to describe it, has mountains in it.5
Rorty imagines two levels of discourse. The first is the ordinary level at which you and I live: at that level mountains exist, existed before there were people, will exist, presumably, after there are people, and would have existed, presumably, even if there had never been people. If you don't agree, you don't know how to play the "mountain" language game. In addition, however, there is a second, archimedean, philosophical level at which a different question can be raised: not whether mountains exist, but whether Reality as It Is In Itself contains mountains. At that second level, according to Rorty, a dispute has broken out between misguided metaphysicians, who say that It does, and pragmatists, like him, who say that It doesn't, that mountains exist only in virtue of a "language game" people at the first level play.
It is a familiar criticism of Rorty that none of this can make any sense unless the proposition that mountains exist can be given a different meaning from the proposition that mountains are part of Reality as It Is In Itself. Rorty plainly thinks that it can: he thinks the first is, in our vocabulary, an I-proposition internal to geology and the second an E-proposition external to it. But he doesn't explain what the difference in meaning is; he only claims there is one through pointless capitalization. If we press the question of what difference there can be between the two propositions, Rorty seems confronted with a dilemma. If he gives the sentence "Mountains are part of Reality as It Is In Itself" the meaning it would have within our "language game" if any of us actually said it, then it means nothing different from "Mountains exist, and would exist even if there were no people," and the contrast he needs disappears. If, on the other hand, he assigns some novel or special sense to that sentence-if he says, for example, that it means that mountains are a logically necessary feature of the universe-then his argument loses any critical force or philosophical bite. His position is interesting only if it contradicts something that someone believes, something that has influence, if not in ordinary life at least in academic philosophy.
Many philosophers who would endorse this familiar refutation of Rorty would resist a parallel claim about selective archimedean skepticism. They would insist that E-propositions like the proposition that morality is a matter of objective fact or that there can be right answers to moral questions really are different from I-propositions like the proposition that genocide is wicked or that abortion is immoral, and that it does make sense, in virtue of that difference, to deny the first pair of claims while affirming one or both of the second. I shall argue, however, that much the same argument as defeats Rorty's general archimedeanism also defeats archimedean neutrality even in its selective form.
Imagine I am speaking at length about abortion. I begin by saying "Abortion in wrong." This is, according to the archimedean skeptic, a positive, first order, I-proposition of morality. Now, drawing breath, I add a variety of other claims that archimedeans classify as E-propositions rather than I-ones. "What I said about abortion was not just venting my emotions or describing or expressing my own or anyone else's attitudes. My opinions are true. They describe what morality, quite apart from anyone's impulses and emotions, really demands. They are, moreover, really and objectively true. They would still be true, that is, even if no one but me thought them true-even, indeed, if I didn't think them true. They are universal and they are absolute. They are part of the fabric of the universe, resting, as they do, on timeless, universal truths about what is sacred and fundamentally right or wrong. They are reports, that is, of how things really are out there in an independent, subsisting, realm of moral facts." (If I have left out any of the archimedean's favorite targets, please put it in yourself.) Call all the statements I made after drawing breath, which the archimedean classifies as E-proposition, my "further claims."
Two questions arise about these further claims. First, can we find a plausible interpretation or translation of all of them that shows them to be positive moral judgments themselves-either restatements or clarifications of the original first-order I-proposition I announced, or further moral claims that elaborate or expand those I propositions? If we understand them that way, any skepticism about them would have to be morally engaged rather than neutral. It would constitute a negative moral judgment-either first-order skepticism about my I-claim that abortion is wrong, or first-order skepticism about some other I-claim I might be thought to have joined to it. Second, can we find an interpretation or translation of any of the further claims-a reading of it that captures what anyone who made it could plausibly be thought to believe-that shows it not to be a first-order I-proposition but to be a philosophically distinct E-proposition instead? If we answer the first question yes, and the second no, then archimedean neutrality is an illusion. If anyone is persuaded to give up the face-value view of morality, he must surrender morality along with it.
It is easy enough to answer the first question yes, because the most natural reading of all of the further claims shows them to be nothing but clarifying or emphatic or metaphorical restatements or elaborations of the I-proposition that abortion is wrong. If someone thinks abortion morally wrong, he might well say, for example, in a heated moment, "It is just true that abortion is wrong." But that would be only an impatient restatement of his substantive position. Some of the other further claims do seem to add something to the original claim, but only by substituting more precise I-propositions for it. People who actually use the adverbs "objectively" and "really" in a moral context do so to clarify the content of their opinions; they mean to distinguish the opinions so qualified from other opinions that they regard as "subjective"-just a matter of their tastes. If someone says that soccer is a "bad" or "worthless" game, for example, he may well concede, on reflection, that his distaste for soccer is entirely "subjective," that he doesn't regard that game as in any "objective" sense less worthwhile than games he prefers to watch. Though he has a reason for not watching soccer, he might say, no one whose tastes are different has the same reason. So when I say that the badness of abortion is objective, or that abortion is objectively or really bad, it would be natural to understand me as explaining that I do not regard my views about abortion that way. I would be emphasizing that, in my view, the deliberate destruction of human life at any stage is impermissible for reasons that in no way depend on my or anyone else's personal reactions or tastes. The claim that abortion is objectively wrong seems equivalent, that is, in ordinary discourse, to another of the further claims I made: that abortion would still be wrong even if no one thought it was. That, read most naturally, is just another way of emphasizing the content of the original moral claim, of emphasizing, once again, that I mean that abortion is just plain wrong, not wrong only because people think it is.
Consider, next, the further claim that I know that abortion is wrong. My audience might naturally understand me as claiming, in that way, that I have compelling reasons for believing that abortion is wrong, reasons that I have no doubt justify that conclusion. That further claim, so read, is evidently an I-proposition-it insists that the fact that abortion involves the deliberate killing of an innocent human being is an unchallengeable, obviously compelling, reason for condemning it. The further claim that abortion is universally wrong can also be understood as only a clarification of my original moral claim. It clarifies that its scope by making plain that in my view abortion is wrong for everyone, no matter in what circumstance or culture or of what disposition or from what ethical or religious background. That is different from saying simply that abortion is wrong, or simply that it is objectively wrong. I might conceivably think that the wrongness of abortion is objective, since it depends on features of abortion not the reactions of people to it, and yet that the wrongness of abortion is not universal, because it is not wrong in certain kinds of communities-those whose religious life supports an entirely different conception of the sacredness of human life, perhaps. When someone says that the wrongness of abortion is universal as well as objective, it is natural to understand him as ruling out exceptions of that sort. What about the claim that the wrongness of abortion is absolute? It is most naturally understood as meaning not just that abortion is always wrong in principle, but that its wrongness is never overridden by competing considerations: that it is never true, for example, that abortion is the lesser of two evils, even when a mother's life is threatened. What about the baroque claims I added at the end, about moral "facts" being "out there" in an "independent" realm? These are not things people actually say; they are invented by archimedean philosophers for reasons we shall consider later. But we can make sufficient sense of them, as things people might say, by understanding them as inflated, metaphorical ways of repeating what some of the earlier further claims say more directly: that the wrongness of abortion does not depend on anyone's thinking it wrong, for example.
So we have no difficulty reading the further claims as further I-propositions either repeating or clarifying or supplementing the original I-proposition that abortion is morally wrong. If we read them in that way, and if we take archimedean skepticism to reject them, then archimedean skepticism is not morally neutral. If the further claim that abortion is really or objectively wrong means that it is wrong even when and where people do not think it wrong, and the archimedean denies that further claim, he is endorsing the negative moral judgment that there is no moral objection to abortion in societies in which it is approved. So archimedean neutrality can only be sustained by meeting the challenge of the second question I posed. Can we find a plausible interpretation of any of the further claims that reads it not as a substantive I-proposition, but as a second-order, non-moral, E-proposition instead?
One strategy for meeting that challenge seems natural to many archimedeans. They say that the further claims can be read as meta-ethical, philosophical claims about value judgments rather than as value judgments themselves. On this view, the further claims are E-propositions because they take up positions on such metaphysical questions as whether there are moral properties in the universe, and, if so, what kind of properties these are. They assert that moral properties exist as genuine properties of acts or events, and that such properties cannot be eliminated from a full and accurate description of how the universe is. Archimedeans who take this view of the further claims hold that they are mistaken, but in a philosophical not a substantive moral way. They think that correcting these philosophical mistakes leaves substantive morality as it stands.
That strategy begs the question we are asking, however, because we are trying to decide, not whether the further claims can be translated to make them seem more philosophical or metaphysical, but whether we can understand those philosophical translations as themselves anything but first-order evaluative claims. The philosophical-sounding proposition that there are moral properties in the universe, for example, is or entails a broad denial of global internal skepticism: it claims that some acts really are unjust, or some people really are good, or something of the sort. So read, the proposition is a (very weak) I-proposition, and a skeptic who denied it would hardly be neutral toward substantive morality. Now consider a more robust piece of metaphysics: some naturalists insist that moral properties "consist in" or "are the same properties as" more ordinary, natural, properties. They insist that just as the physical property of being hot is the very same property as the property of having excited molecules so the moral property of rightness is the very same property as the property of maximizing happiness, for example. But these identity-of-property claims are not semantic truths.6 When they are true, they are substantive discoveries. The identity claim about heat rests, in the end, on an empirical, scientific discovery about the relationship between felt temperature and molecular kinetic energy.7
It is, moreover, exhausted by that discovery: there is nothing extra and "metaphysical" added to the scientific facts by expressing them in claims about the identity of properties. The identity claim about rightness rests, in a parallel way, on a substantive moral thesis-utilitarianism-and it is exhausted, in the same way, by that substantive thesis. There is no difference in what two people think if one thinks that the only thing that can make an act right is its maximizing power, so that it makes no sense to evaluate rightness in any other way, and the other thinks that the property of rightness and the property of maximizing power are the very same property. The second opinion uses the jargon of metaphysics, but it cannot add any genuine idea to the first, or subtract any from it. It sounds more philosophical but it is no less evaluative. So we can't convert the further claims from I-propositions into E-ones by reading them to make metaphysical claims of that sort.
Philosophers distinguish between primary qualities, which things have in themselves, like the chemical properties of metals, and secondary properties, which consist in their capacity to provoke defined sensations or reactions in sentient creatures. Most philosophers think that the disgustingness of rotten eggs, for example, is a secondary property. It is true that rotten eggs are disgusting, they say, but that truth consists only in the eggs' capacity to provoke a sensation of disgust in most or normal people. Many philosophers also think that moral properties are secondary properties, that the wickedness of genocide consists only in the fact that most or normal people who contemplate genocide react in a particular way. Suppose we understand the further claims, on the contrary, as denying this and asserting that moral properties are primary properties of things or events, not just dispositions to provoke reactions in people. Would these further claims then be external claims, taking up a position in a philosophical controversy about the nature of moral properties, but not taking sides in any actual or possible substantive moral dispute? After all, philosophers who think that the wickedness of genocide consists in its impact on most or normal people still think, as a matter of their first-order conviction, that genocide is wicked.
But someone who holds that moral properties are secondary properties does take sides in actual or potential substantive disputes. Suppose we discovered that, contrary to our expectations, contemplating genocide does not in fact outrage even most normal people. Genocide would not then be morally wrong on that dispositional account, though, of course, many people would think it was. Even when there is no substantive disagreement as things stand, moreover, the dispositional account is not neutral. It claims not just that most or normal people do react to genocide in a particular way, but that the wickedness of genocide consists in that reaction, and that claim yields conditional or counterfactual statements that are both substantive and controversial. Which conditional or counterfactual claims follow from the dispositional thesis depends on the precise form the thesis takes; it depends, in particular, on how far and the way in which the extension of moral properties is taken to be fixed by our own natural history.
The most natural form, I believe, is this: what makes an act morally wrong is that contemplating that act in fact produces a particular kind of reaction in most people, or in most members of a particular community. It follows from that formulation that if one day people in general, or in the stipulated community, ceased to react in that way to genocide, genocide would cease to be wicked, just as rotten eggs would cease to be disgusting if they no longer disgusted anyone. But the thesis that genocide would cease to be wicked if it were no longer so regarded is plainly both substantive and controversial. The dispositional account might, it is true, take a different form. It might hold, for example, that what makes genocide wrong is the reaction, not of whichever kind of people happen to exist from time to time, but of us, that is, of people with the physiological structure, basic interests, and general mental dispositions that people actually have now.8 In that case, it would no longer follow that genocide would cease being wicked if human beings developed very different general interests or different neural wiring. But some plainly substantive and controversial claims still would follow: for instance, that genocide would not have been wicked if economic or other circumstances had been different as human reactions evolved, so that creatures with our general interests and attitudes had not been revolted by genocide. The dispositional account might take other forms than these two; it might attempt to fix the extension of moral properties in other ways. But just as any philosophically illuminating account of what the disgustingness of rotten eggs consists in yields counterfactual claims about the circumstances in which rotten eggs would not be or have been disgusting, so any illuminating account of moral properties as secondary entails counterfactuals that state substantive moral positions. That does not mean that moral properties are primary. But it does mean that the argument over whether they are primary is a substantive moral dispute. We may say, if we wish, that it is an argument about the kinds of properties there are in the world. That formulation does no harm, so long as we remember that it is also an argument about the circumstances in which institutions are just or unjust or people good or bad and why.
Correspondence with Reality?
Archimedeans often say that the further claims are misguided because they assert or presuppose that moral convictions "correspond to" or "represent" or "match" some moral state of affairs. But by itself the claim that moral convictions correspond to reality is just redundant. "The proposition that abortion is wrong corresponds to a fact" can be understood as just a wordy way of saying that abortion is wrong. It is true that this reading makes the idea of propositions corresponding to reality a simple tautology: it uses, as philosophers now like to say, a "minimal" sense of "correspondence." But can we find a stronger, more external sense? In some contexts, "corresponds to" or "represents" can be understood as claiming a causal relation, as when we say, for example, that a photocopy represents what it is a copy of. We may mean that the original plays the kind of causal role in producing the copy that, when the process works as it should, guarantees that the copy matches the original.9 Can we read the further claims to say something parallel about moral beliefs and moral facts-to say, that is, that the latter cause the former? If so, then perhaps archimedean skepticism can show its neutrality by insisting that it is only that causal claim it is rejecting when it declares that moral beliefs do not correspond to any reality, not the moral beliefs themselves.
But that strategy works only if the causal thesis the further claims are said to carry asserts a direct and independent impact of moral properties on human receptors. It would not work, that is, if the supposed thesis consisted only in the conjunction of a substantive moral proposition and an ordinary empirical one asserting an interaction between human beings and natural phenomena. Since I believe that pain is bad, the sight of people in pain causes me to think that their situation is a bad one. I might report that process by saying that the badness of their situation made me think their situation bad, but the archimedean cannot dispute that causal claim without denying that pain is bad. Some philosophers argue that moral properties are identical with natural properties-that an act's relative rightness, for example, just is its relative power to maximize happiness. On that view, when we say that the fact that an act promotes happiness causes people to think it is right, which is often plausible, we might as well say that the fact that it is right causes people to think it is. But once again this later claim offers the neutral archimedean no target, because he cannot reject it without rejecting the identity-of-properties claim, and that, as I emphasized, is an abstract moral conviction.
The claims just considered are not suitable targets for the neutral archimedean because they are combinations of innocuous causal reports, on the one hand, and substantive moral positions on the other, and they can be attacked only by contradicting the latter. We must try to imagine, then, a causal thesis in which the supposed causal mechanism is not the ordinary interaction between natural properties and human sense organs, but a direct and wholly independent action of moral properties themselves. If the further claims can be understood as supposing that kind of causal interaction, an archimedean could sensibly say that, though he is neutral about the moral beliefs people have, he rejects the further claims because he rejects that theory of their origin. But can we make sense of an independent causal impact thesis, and if so can we plausibly attribute it to any of the further claims?
The idea of a direct impact between moral properties and human beings supposes that the universe houses, among its numerous particles of energy and matter, some special particles-morons-whose energy and momentum establish fields that at once constitute the morality or immorality, or virtue or vice, of particular human acts and institutions and also interact in some way with human nervous systems so as to make people aware of the morality or immorality or of the virtue or vice. We might call this picture the "moral-field" thesis. If it is intelligible, it is also false. It is not even a remotely plausible thesis to attribute to anyone who might deploy any of the further claims, moreover, quite apart from its insanity as a piece of physics. If someone thought that his moral beliefs were directly caused by moral particles, he would regard that as a vindication of those beliefs, as we regard the discovery that our ordinary observations are caused by what we claim to observe. But no one who believes that abortion is wrong thinks that he gives an argument for his view, or even an explanation of how he came to accept it, by insisting that its wrongness is objective or a moral fact or (if he would say such a thing) part of the fabric of the moral universe. If he is asked for argument or explanation he will offer, if anything, very different propositions from any of the further claims. He might say that abortion is wrong because it is the deliberate destruction of an innocent human life, and that he realized this by reflecting on that fact, or because he was taught it in church.
It might be objected that these latter statements are inadequate as justifications or explanations-that something more or different is necessary. It might even be said that only a direct impact of morons on the brain could justify a moral conviction. It is important to distinguish these critical suggestions, however, which I shall consider in the next section, from the interpretive thesis we are now considering: that though substantive I-propositions-positive moral judgments-need not be understood as claiming any such impact, one or more of the further claims should be so understood. Only that interpretive thesis is pertinent now, because it (not the critical one) could sustain neutrality by separating the further claims that the archimedean rejects from the positive moral judgments he supposedly respects. But the interpretive claim fails, not only because the moral-field thesis is absurd, but because no one thinks that the further claims do the justifying work they would, so interpreted, be meant to do. We should also notice, finally, that even if were sensible to attribute the moral-field thesis to the further claims, that would not convert them into neutral E-propositions, since we should then have to understand them as endorsing the view that the moral wrongness of genocide consists in the field that surrounds its instances, and that view is itself a (preposterous) substantive moral claim.
We have so far considered two unsuccessful E-readings of the further claims: that they insist that moral properties are primary properties, and that they assert a direct causal correspondence between moral fact and moral belief. We must now consider a third suggestion: that the further claims are external because they contain hidden assertions about the best explanation of moral disagreement and error. Crispin Wright, in his illuminating attempt to find genuine issues in the disputes between so-called philosophical "realists" and "anti-realists," suggests that it is an important question about any body of thought whether it is a priori true that disagreements within it, when not traceable to vagueness, can in principle be explained by pointing to someone's cognitive incapacity-either a lack of pertinent information or a general defect in intellectual equipment manifested in ways beyond his inability to reach the right answer on the particular moral issue in question.10 Could Wright's suggestion supply E-readings of the further claims? Is it plausible to interpret the claim that abortion is objectively immoral, or that genocide is really wrong, or that the wickedness of terrorism is a moral fact, as including what we might call a general incapacity claim: that anyone who does not agree with this opinion is suffering from some cognitive impediment that explains his error?
Once again, it is important to distinguish two propositions. The first is that people with controversial moral convictions have a reason for trying to explain why others disagree with them. The second is that such people have a reason for asserting that there must be such an explanation of a particular kind even when they have no idea what it is. The first of these propositions is true. We have more confidence in our own opinion, and a livelier chance of persuading others, if we can show that those who hold the contrary opinion have made mistakes of fact, or that their arguments are demonstrably fallacious, or that they are incapable of sustained reasoning, or that they have reasons of self-interest for holding their view, or that they have shown themselves on other occasions unaware of or insensitive to the suffering of others, or that they lack some other virtue of character essential to sound moral judgment, or something of the sort. But the second proposition is not entailed by the first, and it is false. People have no reason to claim that those who disagree with them must lack some information they have, or suffer from some intellectual incapacity or character defect, when they have no evidence of any such ignorance or incapacity or defect. That claim, in those circumstances, would be empty rhetoric, adding nothing to the original substantive claim. No one could think that his argument or position was improved by it.
Wright says that a general incapacity claim follows platitudinously, about any intellectual domain, from the assumption that opinion in that domain can represent something external to and independent of human beings. He explains the platitude through an analogy. If two representational devices, like fax machines, produce different output, then this can be explained in only one of two ways: either the devices had different input-different information-or at least one device functions poorly as an instrument of representation and we can in principle identify the defect. So if human beings have the capacity to represent how things are in an independent moral realm, this must be true of them as well.
That might seem a powerful argument for reading one or another of the further claims as embodying a general incapacity thesis, because these further claims do seem to assert the independence of moral fact from human will, and to presuppose the ability of humans to form opinions that represent or correspond to moral fact. But Wright's platitudinous inference assumes a particular sense of correspondence or representation-the causal sense I identified earlier. Something represents something else, in that sense, only if that something else leaves an imprint on it through some physical process, the way light leaves an impression on a photographic plate, or, in Wright's example, print causes an electronic representation of itself in a fax machine. People's convictions could not represent or correspond to morality in that sense of representation unless the moral-field thesis were true. As I said, we have no license to attribute that thesis to anyone, and, if we did, we would not be understanding their further claims as morally neutral ones.
There are other senses of correspondence and representation in which human beings can form opinions that correspond to and represent what is independent of them. Wright might well suppose, for example, that he has offered a good representation of the debate between realism and anti-realism and that his interpretation corresponds to the best way to understand that debate. Others disagree, but it is hardly a priori true that either Wright or they will turn out to have different information at hand, or to suffer any independent cognitive incapacity. Of course this different sense of "correspond" and "represent" lacks the explanatory power of "represent" in the causal sense that produces Wright's platitude. We help to explain photography when we say that light creates a representation on film of the object that reflects it. We do not explain anything when we say that our opinions about abortion represent or correspond to moral truth. But that observation only brings us back to an earlier point: the further claims are not meant to explain our moral opinions, but only to emphasize or clarify or expand their content in the way the natural reading of them reports. So we cannot interpret someone who says his moral opinions represent or correspond to the truth as claiming anything stronger than Wright might claim about his own philosophical opinions.
I cannot imagine every E-reading of the further claims that an archimedean might suggest. But those I have discussed show, I think, how difficult it is to resist the natural reading, and therefore how difficult it is to sustain an archimedean neutrality that depends on rejecting that natural reading. Any attempt to construct a different, archimedean, reading means sooner or later attributing to reflective people something as absurd as the moral-field thesis. That helps explain why archimedeans always describe the windmills they make war on in bad metaphors they never cash, why they say that they do not believe that morality is part of the "fabric" of the universe or that it is "out there," a phrase that appears hundreds of times in their texts, always in scare-quotes used like tongs holding something very disagreeable. The seductive claim of neutrality, which depends on a firm distinction between what I-propositions and E-propositions mean, seems so far supported only by silly figures of speech.
We have been trying to distinguish the further claims from substantive I-propositions by concentrating on what the former could be thought to mean. Some archimedean skeptics begin at the other end, however: they try to distinguish the two kinds of propositions by revising our opinions about the latter. They say that the positive judgments that make up the I-propositions of morality (or ethics or aesthetics or interpretation or law) are not actually propositions at all, but belong to a different semantic category: they are rather I-expressions of approval or disapproval, or I-recommendations of rules or conduct. When people say that Goering was a moral monster or that abortion is immoral or that terrorism aimed at children is wicked, they are not properly understood as describing anything. They are only expressing their negative feelings toward Goering or abortion or sadism, or endorsing a standard of conduct that would condemn such people or acts.
Some of these "non-cognitivist" theories are much more detailed and elaborate than that curt summary suggests. But the summary is accurate enough to suggest how dramatically revisionist they all are. People who say that it is unjust to deny adequate medical care to the poor do not think that they are just expressing an attitude or accepting a rule or standard as a kind of personal commitment. They think they are calling attention to something that is already true independently of anyone's attitude, including theirs, or of whether anyone, including them, has ever accepted any particular rule. So the non-cognitivist can't simply remind people of what they really mean to say, and expect them to agree, as the grammarian does who points out that we sometimes use the indicative mood ("There's a draft from the window") when we mean to make a request. The non-cognitivist must motivate people to change their view of what their moral practice amounts to-his message must be, in effect, "This is what you had better be meaning or doing when you say that Goering or abortion or terrorism is wicked, because otherwise your claims would make no sense at all."
Allan Gibbard, one of the most prominent non-cognitivists, puts the matter bluntly. If we continue to treat our normative judgments as descriptive reports, he says, we shall have to embrace "Platonism," which he defines as the idea that truths about what is rational or just or good are "among the facts of the world."11 That idea, he warns, is "fantastic" to an ordinary sensibility. If anyone seriously believed it, then it would be necessary to "debunk" his belief. So if we are to rescue morality, we must show that it does not depend on Platonism, and revisionary non-cognitivism aims to do this by proposing that morality is not a descriptive project at all, but an entirely different, expressive, enterprise.
But this explanation of why we need to rescue morality from the face-value view begs exactly the question we have been discussing. It assumes that the various further claims that make up Platonism (so defined) are not themselves among the I-propositions that are about to be reinterpreted as non-cognitive. For if this distinction is illusory-if the view that the wrongness of genocide is "among the facts of the world" can only be understood as a baroque repetition of the simpler claim that genocide is wrong-then the two supposedly different kinds of speech act must stand or fall together. If Platonism is really to be "debunked" as false, then morality must be debunked along with it. So the non-cognitivist must find a plausible reading of "It is a moral fact that genocide is wicked" that does not simply repeat that genocide is wicked. His strategy is not different after all: he needs the same E-proposition reading of the further claims that we have so far been unable to construct.
There is, however, another possibility. If "Platonism" and substantive morality are in the same boat, because the former is only a restatement of the latter, then rather than substantive morality falling debunked with Platonism, Platonism might perhaps be saved, reinterpreted as itself non-cognitive, along with substantive morality. Several philosophers have in fact made a beginning on this alternative, more ecumenical project. They are sensibly reluctant to dismiss some parts of the face-value view of morality as mistaken. They are reluctant to say that opinions almost all of us hold-for example, that slavery would still be wrong even if evolution and history had proceeded in such a way that almost no one thought it was-are metaphysical nonsense. So they widen the category of the non-cognitive. They say that the claim that slavery would still be wrong even if no one thought it was is not a failed effort to describe some moral reality, but only the expression of a somewhat more refined attitude than the attitude expressed by the simpler claim that slavery is wrong. In our terminology, they accept that the more complicated claim is itself a (non-cognitive) I-proposition rather than a (Platonic and false) E-proposition.
But they cannot find a place to stop this process of enlarging the domain
of the non-cognitive before it embraces and therefore destroys their own non-cognitivism.
Consider these remarks of Simon Blackburn, who describes himself as a "projectivist"
and "quasi-realist" about ethics:
The projectivist can say this vital thing: that it is not because of our responses, scrutinized and collective or otherwise, that cruelty is wrong. The explanation flows from the way in which quasi-realism has us deal with oblique contexts. It issues an `internal' reading of the statement of dependence, according to which the statement amounts to an offensive ethical view, about (of course) what it is that makes cruelty wrong.... [T]here is only one proper way to take the question `On what does the wrongness of wanton cruelty depend?': as a moral question with an answer in which no mention of our actual responses properly figures.... As soon as one uses a sentence whose simple assertion expresses an attitude, one is in the business of discussing or voicing ethical opinion.12
Blackburn is able to resist the view that the wrongness of cruelty depends on attitudes because, he says, that statement is internal to the "business" of expressing attitudes, and so must be treated like other first-order moral claims. It expresses an attitude which he, as well as almost everyone else, in fact holds. But then what claims about morality are external to that business? He must find some external statements of the right kind through which to declare his own "projectivism" or non-cognitivism. For if the further claim that (for example) moral facts form part of the ultimate structure of reality is itself internal to the enterprise-only expresses an attitude-then "realists" have committed no error which Blackburn's "projectivism" or "quasi-realism" corrects.
Here is his response to that challenge. "If one attempts to discuss external questions," he says, "one must use a different approach-in my case, a naturalism that places the activities of ethics in the realm of adjusting, improving, weighing, and rejecting different sentiments or attitudes." That is not yet helpful, because it suggests nothing that could contradict any fancied "realist." Of course a "naturalistic" description of the "activities" of ethics would be a psychological or sociological description. Blackburn also says, however, that "The projectivist ... has a perfect right to confine external questions of dependency to domains where real states of affairs, with their causal relations, are in question. The only things in this world are the attitudes of people ... moral properties are not in this world at all, and it is only because of this that naturalism remains true."13
What can this last passage mean? Blackburn says that "this world" is the world of "causal relations," so he might mean only that the moral-field thesis is false, that is, that moral facts do not enter into causal relationships with human beings or anything else. But if that is all that "projectivism" or "quasi-realism" comes to, it is consistent with a "Platonism" that asserts, as vigorously as can be, that there are genuine moral properties "out there," that these are independent of human will or attitude, and that moral judgments are true insofar as they correspond to or correctly represent these properties. If Blackburn is to preserve projectivism as a distinct meta-ethical position, he must find something stronger to defend, and the statement that "there are no moral properties" but only "attitudes" among "real states of affairs" seems intended to provide this. It can do so, however, only if we can understand this to say something independent of and consistent with his earlier declaration that the wrongness of cruelty does not depend on attitudes. But how can the wrongness of cruelty not depend on attitudes if attitudes are all that it can depend on, because they are the only things there are among "real states of affairs"?
So Blackburn's ecumenicism is no passport through the difficulty we have been exploring. He wants to distinguish "internal" questions of dependency, which are first-order substantive moral questions, from Aexternal" questions of dependency, which are philosophical questions to be settled by consulting the "world." But he has no way of separating the supposedly external mistakes the projectivist corrects in the name of naturalism from the internal convictions he embraces as part of the "business" of morality. If the thought that the wrongness of cruelty in no way depends on attitudes is an internal insight, then how can the claim that the wrongness of cruelty exists as a real state of affairs, which only says the same thing in more stilted language, be an external blunder? But if the latter judgment is also an internal insight, then the projectivist has nothing left to dissent from, and his theory swallows itself, the Cheshire Cat of moral philosophy.
AUSTERITY: THE PRIMACY OF VALUE
I conclude that both the familiar and baroque further claims so maligned by archimedeans are only, at least so far as they have been able to show, redundant or more elaborate forms of the I-propositions they wish to leave untouched. If so, the first dimension of archimedean eternality-its supposed neutrality-collapses. But that is not yet comforting for the face-value view, because it may only make the archimedean critique of morality more devastating than most archimedeans intend. John Mackie drew that conclusion. He was an archimedean who rejected neutrality: he insisted, as I have, that the face-value view is part and parcel of ordinary morality. But he concluded that ordinary morality is therefore false. He was an archimedean, in spite of spurning neutrality, because he insisted on austerity. He was not an internal skeptic relying on some deep substantive moral assumptions, like the thesis that there is no morality because God is dead. He was an external skeptic purporting to rely only on independent, non-moral, philosophical arguments.
Mackie relied on two arguments that are now staples of austere skepticism. The first is the familiar argument from moral diversity, which insists that the fact that people disagree so much about morality, from time to time, and place to place, and even within particular cultures, shows that the face-value view must be wrong and that no moral claim could be true. Moral diversity is sometimes exaggerated: the degree of convergence over basic moral matters throughout history is both striking and predictable. But people do disagree about fundamental matters, like abortion and social justice, even within particular cultures, and this fact does give people reason to reexamine their own convictions. Why should I be so confident that I am right if others, who seem just as intelligent and sensitive, disagree with me so deeply?
It is one thing, however, to reexamine one's own views, and perhaps change them after further reflection, and another to decide, as Mackie and other non-neutral archimedeans insist we should, that no positive moral claim is true. After all, we would not count the popularity of our moral opinions as evidence for their truth. Why should we count their controversiality as evidence against it? In any case, however, the popular argument from moral diversity is radically incomplete. Whether diversity of opinion in some intellectual domain has skeptical implications depends on a further philosophical question: it has such implications only if the best account of the content of that domain explains why it should. The best account of scientific thought does explain when and why disagreement in scientific judgments is suspicious. Suppose millions of people claimed to have seen unicorns but disagreed wildly about their size and shape. We would discount their evidence: if there were unicorns, and people had seen them, the actual properties of the beast would have caused more uniform reports. But when we have no such domain-specific account of why diversity of opinion impeaches all opinion, we draw no skeptical conclusions from that diversity. Since we do not think that philosophical opinions are caused by philosophical facts, we do not conclude from the diversity of philosophical views (which is more pronounced than moral disagreement) that no positive philosophical thesis is sound. If the moral-field thesis were true, then moral controversy would be like controversy about the properties of unicorns, and would excite similar suspicion. But once we reject that thesis, we are left with no connection between diversity and skepticism. Perhaps we will discover such a connection when we turn more directly to the question of moral epistemology, as we shall in a moment. But unless we do, we can set the argument from diversity aside.
Mackie's second argument is also familiar in the archimedean canon: it insists on the "queerness" of an idea that it declares is essential to morality as it is commonly understood-the idea that that moral properties are inherently motivating. The idea of an "objective good," Mackie said, is queer because it supposes that "objective good would be sought by anyone who was acquainted with it, not because of any contingent fact that this person, or every person, is so constituted that he desires this end, but just because the end has to-be-pursuedness somehow built into it. Similarly, if there were objective principles of right and wrong, any wrong (possible) course of action would have not-to-be-doneness somehow built into it."14
More metaphors! What could be meant by saying that an end has "to-be-pursuedness" or an action "not-to-be-doneness" "built into" it? That is not clear, and the popularity of the supposedly skeptical argument about morality and motivation may depend on not separating different ways in which these metaphors might be unpacked. On one reading, the proposition that morality is inherently motivating means that anyone who contemplates an end that is in fact good, or an act that is in fact wrongful, feels an emotional tug toward that end or away from that act. That could not be true unless something like the moral-field thesis were true; if it were, then the morons surrounding a genuinely good end or genuinely wrong act might have the power to suck people into an attraction or repel them into an inhibition. But the suggestion that good ends or bad acts have a built-in magnetic attraction or repulsion is plainly not essential to ordinary moral opinion or practice. We know that many of our moral opinions are controversial, and that otherwise perfectly normal people who contemplate abortion or economic equality feel very different emotions and impulses from those we feel. No one takes that obvious fact to undermine anyone's position.
The second reading supposes a connection, not between bare contemplation and impulse, but between two kinds of belief: it holds that anyone who believes that an end is good or an act wrong must also believe, on pain of contradiction, that he himself ought to behave in some indicated way. That implication fails in Mackie's first case of supposed "built-in-ness," the case of objectively good ends. If I believe that the world would be morally better if there were less suffering, then perhaps I ought to do what I can to lessen suffering. But that is a matter of moral or ethical judgment, not conceptual connection. In Mackie's second case-the objective wrongness of acts-the connection does seem more conceptual. If I acknowledged that it is morally wrong to cheat on my taxes, but denied that I had any reason not to do so, you would understandably be bewildered. There is nothing bizarre in the idea that a moral duty necessarily supplies a moral reason for action, however. That can be true only in virtue of what "duty" and "reason" mean.
A third reading of the proposition about "built-in-ness," which combines elements of the first two readings, is a more plausible interpretation of what archimedeans have in mind. On this view, the claim that morality is inherently motivating means that no one really accepts a moral or valuational judgment unless he feels some actual motivational impulse to act in the direction that judgment points. Once again, the proposition, so understood, seems wrong in the case of beliefs about objectively good ends. Someone who thinks that the world would be better with less suffering may be defective in character, as I said, if he is not thereby moved to action. But if he is not, it does not follow that he does not have that thought. Once again, however, the proposition does seems plausible in the case of beliefs about objectively wrongful acts. I may claim to think that cheating is wrong, but if you see that I am in no way deterred from cheating whenever I have the opportunity, and that I show no regret or hesitation or discomfort when I do, you may well think that I am either insincere or out of touch with my own real convictions. There is nothing queer in the idea that whether we ascribe a certain mental state to someone depends on more than his own opinion about that state, however. Someone may honestly think he is jealous or in love or trusting when his impulses and behavior show that he is not.
It is true, as the archimedean will emphasize, that we do not make it a condition of attributing beliefs about ordinary matters of fact that the agent display any particular motivational impulse, prescinding from all other motives he has. My believing that arsenic is a fatal poison does not depend on my showing any aversion for it unless it is also true that I want to stay alive. In the case of moral judgments about action, on the other hand, the claimed connection is independent of other motives: if I really believe that cheating is wrong, according to the present reading of the "built-in-ness" claim, I will feel a tug against cheating no matter what other motives I have. That is what is meant by that understanding of the idea that morality in inherently motivating. But we can explain this distinction between moral and pharmacological beliefs simply on grounds of their differing content. We ascribe beliefs to people as part of a complex process in which we also ascribe motives and meanings so that all three fit together in a way that makes best sense of their behavior as a whole.15 Given what we take judgments about wrongful conduct to mean, we therefore withhold their attribution unless we find it plausible to suppose that the agent would be moved to some degree to avoid the act he deems wrongful quite independently of other motives he might have for avoiding it, at least absent circumstances that show weakness of the will or emotional disorder. There is nothing queer or bizarre about that attributional strategy.
But suppose it is now said that what is queer is the supposition underlying the strategy, which is that people can be moved by moral considerations alone, independently of self-interest or other desires. Psychological egoism is an extremely implausible empirical hypothesis, but it might conceivably turn out to be true, and later in this essay I shall consider what the consequences for morality would be if it were. It is enough to say now, however, that even if its truth entailed that no one ever held a moral belief, it would not entail that no moral proposition was true.
I have insisted, in the argument thus far, that the absurd moral-field thesis is not essential to morality. I must now confront a general archimedean argument: that that thesis, or something like it, indeed is essential, because it makes no sense to suppose that acts or events or institutions have moral properties unless we have some plausible account of how human beings could be "in touch with" or aware of such properties, and if we reject the explanation offered by the moral-field thesis we must appeal to some other account of a moral faculty that would be equally occult. I have already referred to Crispin Wright's account of this argument. He suggests that a full-blooded version of "moral realism" would have to describe the mechanism through which human beings come to have moral opinions, and do this in some way that shows how moral error could be explained, in other than a trivial way, as the malfunctioning of that mechanism. In the last section, I said that it would be a mistake to interpret any element of the face value view of morality as announcing or presupposing that morality meets this test. The issue now is not that interpretive question, however, but the different, more directly philosophical, question whether morality does fail to meet it, and, if so, what the consequences of that failure are.
Wright says that the disposition of these questions may turn on which side has the burden of proof, and that the "realist" side has that burden. Unlike an argument in a court of law, however, the course of a philosophical investigation is fixed not by any free-standing methodological postulate, like Occam's razor which Wright cites, but by how opinion stands when the investigation begins. No skeptical argument can succeed, for anyone, unless it brings him skeptical conviction, and that means that none of us can accept such an argument unless we find its premises convincing even when we grasp their skeptical import. We must find these premises more plausible than what they require us to abandon. Let us accept, for the sake of the argument, that we are forced to choose between the following two propositions. (1) Human beings have a special though sometimes fallible faculty of judgment that enables us to decide which moral claims to accept or reject, a capacity whose malfunctioning may sometimes result only in moral misjudgment with no spillover impairment of other cognitive activity. (2) There is no moral objection to exterminating an ethnic group or enslaving a race or torturing a young child, just for fun, in front of its captive mother. Which should we abandon?
That is not a question of where a burden of proof lies, but of what, considering each possibility as fully as we can, and noting its implications for the rest of what we think, we find that we believe. Of course I do not mean that our convictions are right just because we find them irresistible, or that our inability to think anything else is a reason or ground or argument supporting our judgment. On the contrary, these suggestions are forms of the skepticism I am opposing. I mean that any reason we think we have for abandoning a conviction is itself just another conviction, and that we can do no better for any claim, including the most sophisticated skeptical argument or thesis, than to see whether, after the best thought we find appropriate, we think it so. If you can't help believing something, steadily and wholeheartedly, you'd better believe it. Not, as I just said, because the fact of your belief argues for its own truth, but because you cannot think any argument a decisive refutation of a belief it does not even dent. In the beginning, and in the end, is the conviction.
If I were forced to choose between the two propositions just described, I would accept the first and reject the second. I have considerable evidence in my own experience-as I think you have in yours-of a capacity to make moral judgments that bring conviction, that are mainly durable, that agree with the judgments of a great many others, and that are amenable to the normal logical combinations and operations. I have an open mind about the character of the neural strategies and intellectual processes that are deployed in this capacity: I know next to nothing about these. But if the hypothesis of the first proposition-that the moral capacity is ad hoc and not systematically integrated into other intellectual powers-is the only alternative to denying any capacity to reach credible moral opinions at all, I would be content to accept it. It involves nothing mysterious or artificial or counterintuitive, as the moral-field thesis, for example, does. On the other hand, it is startlingly counterintuitive to think there is nothing wrong with genocide or slavery or torturing a baby for fun. I would need very powerful, indeed unanswerable, reasons for accepting this, and I think most other people would as well. Can such reasons be found?
They cannot be found in what we might call archimedean epistemology. This is a hierarchical epistemology that tries to establish standards for reliable belief a priori, ignoring the differences in content between different domains of belief, and taking no account of the range of beliefs we already hold to be reliable. Consider Gilbert Harman's suggestion that we cannot regard any belief as reliable unless we think that the best causal explanation of why we hold it refers to the state of affairs it describes.16 In some form, this test does seem appropriate to beliefs about the physical world. But that is in virtue of the content of those beliefs. Since they are beliefs about objects and events that can interact causally with the human nervous system, it is sensible to include some requirement of direct or remote or at least potential interaction among our tests of their reliability. But nothing in the content of moral (or aesthetic or mathematical or philosophical) opinions invites or justifies such a test. On the contrary, the content of these domains excludes it, because an adequate causal explanation of a belief includes showing that the belief would not have occurred if the alleged cause had not been present, and we cannot understand or test that counterfactual claim with respect to moral or aesthetic beliefs because we cannot imagine a world that is exactly like this one except that in that world slavery is just or The Marriage of Figaro is trash. If the "best explanation" causal test is universally sound, therefore, no moral (or aesthetic or mathematical or philosophical) belief is reliable. But we can reverse that judgment: if any moral belief is reliable, the "best explanation" test is not universally sound. Either direction of argument-taking either of the two hypotheses as axiomatic and using it to deny the other-begs the question in the same way.
We must rather find our epistemology as part of an overall search for broad harmony-what John Rawls called, in a different context, reflective equilibrium-among our opinions as a whole, and none of these can be given an automatic or antecedent veto over the rest. It is true that, in a different and less troubling sense, this equilibrium epistemology also begs the question. We assume along the way whatever standards for reliable belief we take the process ultimately to justify. That is as much true for physical or scientific epistemology as for any other-the "best explanation" test assumes that the various psychological hypotheses about perception and belief that give us reason to accept the test themselves meet it. In the end the whole intellectual structure stands or falls together like the struts of a geodesic dome.
So the epistemology of any domain must be sufficiently internal to its content to provide reasons, viewed from the perspective of those who begin holding convictions within it, for testing, modifying or abandoning those convictions. Of course, we cannot simply stipulate that some set of opinions-astrological propositions, for example, or religious doctrines about a God with causal powers-are true and then declare that whatever methods of investigation would confirm those opinions, no matter how apparently scandalous, are for that reason reliable. That would make the opposite mistake of giving any conviction we happen to have immunity from any critical review. Since astrology and orthodox religion, at least as commonly understood, purport to offer causal explanations they fall within the large intellectual domain of science, and so are subject to causal tests of reliability. Since morality and the other evaluative domains make no causal claims, however, such tests can play no role in any plausible test for them. We do need tests for reliability of our moral opinions, but these must be appropriate to the content of these opinions. That is why an epistemological challenge that comes to nothing more than insisting that moral properties are not physical properties must fail. Morons play no part in moral reflection or commitment, and debunking the moral-field thesis therefore leaves morality untouched.
Wright asks whether a "moral realist" would have any reason for concern if morality could not meet the general incapacity test-if there was no good reason to think, a priori, that moral disagreement was always the result of factual error or an impaired general cognitive mechanism. He suggests one ground of concern. Once it is conceded that morality fails this test, then moral progress can no longer be explained as the result of the gradual elimination of ignorance or other impediments to the functioning of human intellect. That is true, but we must take some care in deciding how serious that would be. Anyone who is convinced that slavery is wrong, and knows that his view is now shared by almost everyone else, will think that general moral sensibility has improved, at least in that respect, since slavery was widely practiced and defended. Perhaps sufficient other examples can be found to allow us the much more ambitious claim that moral opinion has improved broadly on all fronts. How much progress we think we can claim, in that simple comparative sense, depends only on our own moral convictions and our sociological and historical beliefs about the distribution of parallel convictions now and in the past.
It would indeed be desirable, however, to be able to explain as well as report this putative progress, for at least two reasons. First, an explanation might give us reason to think that the progress would continue into the future. If we thought that progress had for centuries resulted from the cumulative impact of anthropological discoveries that made prejudice of different kinds more difficult to sustain, for example, we would have that reason for supposing that future such discoveries would continue the trend. Second, an explanation would give us added confidence that the changes were indeed progress-that we are right about slavery, for example, and the Greeks wrong-because we can have more confidence in any of our opinions if we can explain why those who disagree with it came to hold a mistaken view.
In fact we might well be able to explain much of what we regard as moral progress in this way. We might be able to show, for example, that people who defended slavery held false empirical beliefs about the biological humanity of races they enslaved, or that slave owners were subject to special economic stringency that blinded them to slavery's immorality, or that they lacked pertinent information of some other kind or were subject to other influences known to distort judgment. Some people, in virtue of their own moral convictions, may have other kinds of explanation available. People who think that God is the source of moral knowledge, for example, may also believe that He has gradually unfolded His moral plan to more and more of His children. Utilitarians-to take a very different example-may explain progress in economic terms. Moral error gradually disappears, they might claim, because people afflicted with unnecessary suffering have a greater incentive to press for principles that call for relief than other people have for resisting them.
Nevertheless Wright is surely correct that nothing guarantees, a priori, that there will always be an explanation for any change we regard as progress. We may be forced to concede, in some cases, that those who held different views lacked no information we have, and were subject to no different distorting influences. All that we can say, by way of explanation of the difference, is that they did not "see" or show sufficient "sensitivity" to what we "see" or "sense," and these metaphors may have nothing behind them but the bare and unsubstantiated conviction that our capacity for moral judgment functions better than theirs did. Of course that is a less satisfactory situation, and our behavior, when possible, should reflect the difference. We should be less judgmental, more modest, more aware of the possibility that in the future we will be thought as insensitive as we now think others were.
But there is nothing in any of this that warrants what is plainly a further conclusion: that our moral opinions and the opinions of those who disagree with us are all wrong because no moral opinions can be right, which is the conclusion the archimedean skeptic who has abandoned neutrality and now presses this point must be urging. There is a great difference, which he ignores, between the thesis that we have no explanation of why others disagree with us that reinforces our belief, which is regrettable, and the thesis that we have no reason for thinking we are right, which does not follow from it. We do have reasons for thinking that slavery is wrong and that the Greeks were therefore in error: we have all the moral reasons we would cite in a moral debate about the matter. These are not necessarily reasons that contribute to a causal explanation of anyone's error on these matters. But it is only dogmatism to insist that the only reasons that can support a moral conviction are reasons of that kind. It is worth remembering here, as I suggested earlier, that the reasons philosophers offer for their own conclusions do not necessarily explain other philosophers' errors either. It is hardly a priori given, for example, that every disputant on at least one side of the controversy we are now considering-about whether people can sensibly claim to have a special moral capacity-is suffering from some lack of pertinent information or some generalized intellectual disability or some special distorting influence that would explain his failure to grasp the superiority of the other side's view. The archimedean employs his own autonomous philosophical capacity to declare that no intellectual capacity can sensibly be treated as autonomous.
We are considering whether skepticism about morality can be austere-whether, that is, a general moral skepticism can follow from a set of premises or assumptions no one of which is itself a positive moral judgment. I have been arguing that the most prominent candidates for an austere skepticism, which begin in a priori epistemological premises about the conditions of any reliable belief, must all fail, because they beg the question at issue in too bald a way. We must now explore, however, an austere strategy that might seem initially more promising because it begins within morality rather than outside it, and exploits the fact that most people's moral convictions embed non-moral assumptions that are hostages to non-moral refutation.
We have already noticed examples of that embedding phenomenon. Many people flirt with global moral skepticism, as I said, because they accept the following argument. "Unless people have free will, nothing they do can be morally right or wrong. Determinism is true and excludes human free will. So no human acts are morally right or wrong." This is not, however, an austere argument, because though the claim that determinism is true is not a positive moral judgment, the claim that free will, understood in some specified way, is indispensable to moral assessment is. It presupposes a positive counterfactual judgment-that acts would have moral properties if there were free will-and a theoretical moral judgment that the basic subject of moral assessment is free choice. These positive judgments may seem self-evidently true to you. But they have been rejected by many cultures, including the Greek, and by many religious traditions including some within the Catholic and Protestant churches.
I gave another example: many people believe that the discovery that God is dead (or otherwise engaged) would be catastrophic for morality. Though, once again, atheism is not itself a positive moral judgment, this argument also requires a premise that is-the premise that God is the one source of moral value, that His will, and that alone, can generate obligation and virtue. Perhaps much of contemporary philosophical skepticism has its forgotten source in exactly this logic: it may all be a lingering viral residue of the defeat of crude anthropomorphic religion. How else can we explain the widespread but plainly mistaken assumption that a successful Darwinian explanation of moral concern-that human animals with such a concern were more likely to survive-would have skeptical implications?
We may do better for austere skepticism, however, if we concentrate not on metaphysical assumptions buried in certain moral views, like these two, but on psychological assumptions common to almost all views. Any of these might be false, and we can test this fresh opportunity for austere skepticism by imagining the most destructive such possibility. One basic assumption we almost all make is motivational: that we are capable of identifying and acting out of the interests of others beside ourselves. What if that assumption were false-what if we had to accept the doctrine of psychological egoism that we can never act except in accordance with some conscious or unconscious perception of self-interest, that so-called moral reflection is only rationalization, that what we take to be moral conviction is inevitably and inescapably self-serving, only ideology fixed by narrow interests of class or role or situation? I know of no reason to accept this bleak thesis. It is firmly contradicted by experience and also, among other things, by the now most prominent Darwinian accounts of the development of morality.17 But let us imagine, for the purpose just described, that we have psychological or anthropological or some other scientific proofs that the bleak thesis is true. That would seem devastating, because we cannot regard the idea that morality is more than self-interest as a contestable hypothesis that we might consider dropping. It is of the essence of morality. If we accepted the bleak hypothesis, we would have to accept that we have no moral capacity at all, and that morality is, after all, bunk.
Or would we? What new reason would you actually have for abandoning any of your moral beliefs if the bleak hypothesis were proved? Suppose you think that justice requires higher taxes for redistribution to the poor. You take some pride in holding that view, because your income is large and higher taxes would be expensive for you. But now you are (somehow) persuaded by the bleak thesis that you must have unconscious selfish motives for your opinion, though you don't know what they are, and that you will be unable, even after more reflection, to shake off the determining effect of such unconscious motives. You do not yet have any ground for changing your mind, because the other shoe has yet to drop. You have been given no reason, as yet, not to think that you have two reasons for supporting higher taxation-your own self-interest and moral truth. It is true, you now think, that only one of these reasons played any actual motivational role in forming your views. But it did form your views, which include that view that justice requires higher taxes, and as long as you still believe that, you have no option but to treat it as a good moral reason for voting for higher taxes. How could you-thinking what you do about what justice requires-either not vote or vote the other way?
It will be said that it is unreasonable for you still to think that justice requires anything, one way or the other. But why is that unreasonable? Your opinion is one about justice, not about your own psychological processes or even about your own character. You never thought that your capacity to act for selfless reasons, when you still thought you had that capacity, counted as an argument for the justice of higher taxation. Why should the lack of such a capacity now count as an argument against it? You are now convinced that you wouldn't have seen the justice of a tax increase unless it had been in your own interests to see this. But why shouldn't you count it as a piece of luck-a special example of what Bernard Williams has called moral luck-that your self-interest and justice here coincide? You realize that you would have had a very different view if your own self-interest had favored that different view instead (just as you realize now, independently of the bleak thesis, that you would have had different views if you had been born into a very different culture). But that fact alone cannot undercut your conviction about justice, and therefore about your moral luck. You have to think something-either that you have an obligation to vote for higher taxes or that you do not-and you have no reason, just in what the bleak thesis declares, to turn your back on all the reasons (which, as I said, included no psychological assumptions) that inclined you to think that you do. You lack, so far, in other words, something that neither the bleak thesis nor any other piece of psychology can provide on its own. You lack a normative connection between the bleak psychology and any conclusion about justice, or any other conclusion about how you should vote or act.
I do not mean that there is no such connection. You might very well think you have found one in any number of substantive moral traditions or attitudes or assumptions. Kant argued, for example, that morality's point and value for people lies in freedom-the kind of freedom we have only when we act under the direction of reason rather than self-interest or some other heteronymous motive. If you are drawn to that view, then it will seem crucial to you, as it did to him, that you act not just in a way that matches what some Godlike creature who acted out of respect for the equal importance of every person would do, but that you yourself act out of that motive, that it be objective reason rather than self-interest that determines what you do. If you are then persuaded by the bleak thesis that neither you nor any other human being ever can act out of such a motive, you might well think that the very possibility of moral worth and behavior has been undermined.
But this Kantian view of morality's point, however compelling, is plainly a positive moral judgment in the sense I defined earlier. It supposes that acting out of reason-and only acting out of reason-has moral worth, and, in the imagined circumstances, it deploys that assessment to skeptical effect. A different substantive morality would not have the same skeptical results. If your morality was a consequentialist one, for instance, based on the positive judgment that human happiness is a moral good, and that whatever acts enlarge that good are morally required, the bleak thesis would give you no reason not to continue to think that you had a moral obligation to vote for higher taxes. Nothing in that thesis argues that human happiness is not an intrinsic good, or that the arguments were fallacious that convinced you that voting for higher taxes would produce more human happiness than any other act open to you at the time. You would therefore think that, in your case, self-interest and morality had indeed coincided. You would be in much the same state of mind as the theologians of predestination, from Augustine through Calvin, who supposed that people saw the truth only because they had been elected by God. Those theologians continued to think that they were among the elect, though they knew that people could think that who were not. Of course you would be disappointed to learn, from God or later history, that you had been led astray by self-interest. But not that you had been led in the right direction by the same master.
Even if you were a consequentialist, the new bleak psychology would present new moral issues for you to consider. Granted that people never act against what they unconsciously take to be their own deepest interests, is morality in fact better served if people give up morality and consult only those interests consciously, so that an invisible hand can maximize collective benefit? Or is the collective welfare better served if people suppress the bleak hypothesis and continue to act on moral calculations about what is generally best, because a social practice of that sort is itself in people's best individual interests, as they may unconsciously realize? These are not very different from questions that consequentialists have actually debated. In any case they would represent, in the dark circumstances we are imagining, not the defeat of morality by psychology, but the adaptation of the first to grim news about the second.
The argument of the last several paragraphs, I must remind you, is premised on a silly assumption, a worst-case scenario for morality. Even in such a case, non-moral discoveries cannot undermine or structurally change morality without morality's help. They furnish information that engages with deep substantive moral judgments, and their impact depends on what those deep judgments themselves declare or assume. Morality might wither in the face of this interaction, for the various reasons just mentioned or for many others. But it would not have been defeated, on the plane of argument, by austerity alone. There should be nothing surprising about this conclusion. Even the deepest skepticism is an opinion about what morality demands, and no argument can be decisive of that question that is not or does not include premises or assumptions about what morality is or is for.
Archimedeans are bad metaphysicians who think that the old-fashioned, full-blooded, shameless morality of the face value view needs non-moral foundations. Neutral archimedeans think that a pale, modest form of morality can limp along without such foundations, surviving by shedding pretensions. But that is just another form of the bad metaphysics, because it supposes that there is something by way of foundation that, if only morality had it, would make it more secure, more deserving of the honorifics of the face value view. Some so-called moral "realists" add to the confusion by accepting the archimedean challenge as sensible and trying to meet it. They declare that there really are objective and normative properties or facts in the universe, which is true. But they declare this in language that strives for metaphysical resonance, as if its truth was to be discovered in some philosophical domain other than that of substantive evaluation. If I am right, they share the fallacy of the archimedeans, which is to suppose that some sense can be assigned to the supposedly metaphysical claims that is not itself a normative sense, or that there is some way to establish a normative proposition other than through substantive normative arguments.
I regard my view of morality as a "realist" one (though, given the notorious ambiguity of the term, not much turns on the attribution) because it embraces the face-value view. I would not volunteer the more baroque formulations of that view, about timeless truths among the furniture of the universe. But if pressed I would insist that, so far as they mean anything at all, they are true. My realism, that is, knows no bounds. Some critics might nevertheless find this version of realism too pale because, they might say, it employs only a minimalist shadow of the full, robust ideas of truth or knowledge or property or existence or objectivity which they think a genuine realism would field. But once we dismiss the barely intelligible moral-field thesis, there seems no more point in calling the view I have been defending minimalist than in calling it maximalist, because there is no more robust thesis for any realism to deploy or any anti-realism to refute, no more metaphysical a meta-ethics for the former to embrace or the latter to mock. That is not a weak claim; on the contrary it has a powerful consequence that explains why so few philosophers seem willing to accept it.
The powerful consequence is this. Morality is a distinct, independent dimension of our experience, and it exercises its own sovereignty. We cannot argue ourselves free of it except by its own leave, except, as it were, by making our peace with it. We may well discover that what we now think about virtue or vice or duty or right is inconsistent with other things we also think, about cosmology or psychology or history. If so, we must try to reestablish harmony, but that is a process whose results must make moral sense as well as every other kind of sense. Even in the most extreme case, when we are offered grounds for scorching doubt, we still need moral judgment at some deep level to decide whether that doubt is justified and what its consequences for virtue and vice, duty and right, really are. No matter what we learn about the physical or mental world, including ourselves, it must remain an open question, and one that calls for a moral rather than any other kind of judgment, how we ought to respond. If morality is to be destroyed, it must preside over its own destruction.
We cannot climb outside of morality to judge it from some external archimedean tribunal, any more than we can climb out of reason itself to test it from above. Those facts do, I concede, have important implications for moral philosophy as well as for philosophy generally, and I hope to explore these elsewhere. But it is plain that these cannot be skeptical implications. That is not a result to cheer, however, and it is no kind of victory, cheap or expensive, for or over anything. The only kind of skepticism that counts, anyway, is the really disturbing kind, the chilling internal skepticism that grips us in a dark night, when we suddenly cannot help thinking that human lives signify nothing, that nothing we do can matter when we and our whole world will in any case perish in a cosmic instant or two. That kind of skepticism cannot be owned or disowned by semantic reclassifications or meta-ethical refinement. It takes hold as a terrifying, overwhelming, substantive fact, and until its grip is loosened by competing conviction we cannot be sophisticated or ironic, or anything else but hollow or paralyzed or sad.
Indeterminacy and Default
We are all skeptics-internal skeptics-about some evaluative claims, and in this final section I want to consider what kinds of arguments or convictions might support skepticism in particular cases. Skepticism can be expressed in many ways. People sometimes say, for example, that some evaluative proposition is neither true nor false, or that it is "indeterminate," or that two subjects being compared evaluatively are really "incommensurate," or that there is "no right answer" to some evaluative question. These familiar claims are different from one another, but they all mean to deny an entire, contextually defined, set of positive evaluative claims. In some cases this contextually given set is a series of putatively exhaustive claims-that the best red wine is better than the best white, for example, or vice versa, or that the best reds and whites are exactly equal in quality. The skeptical claim is then that these supposedly exhaustive claims are all false, because some ignored alternative-that the best wines of each type are both, in Derek Parfit's phrase, "not worse than" those of the other.18 In other cases none of the contextually given set can be said to be false without accepting that another is true-the propositions that abortion is morally wrong and that it is morally permissible form such a set, for instance-and the skeptical claim is that none of these propositions is either true or false. I shall call any form of this general denial of all apparently eligible positive propositions a claim of "indeterminacy," because the issues I shall discuss will not require careful distinctions of the different forms such claims take, and it will be sufficiently clear, in the contexts I discuss, what the position I call indeterminacy means to reject.
What positive reason-what arguments or convictions-are needed to support a judgment of indeterminacy? I begin with a threshold question. Do we need any positive reason? Do we need a positive reason, for example, for thinking that it is neither true nor false that Antigone did the right thing in burying her brother, or that there is no right answer to the question whether abortion is wicked? Though this may surprise some readers, a great many philosophers think the answer is: no. They think that indeterminacy is the default position in morality, ethics, art, and law. There are, in fact, two versions of this default thesis: a first-personal version, which assumes that a judgment of indeterminacy is the right one for someone to reach in the course of his or her own, personal, first-level moral experience when he or she is deeply troubled by some moral issue, and a third-personal version, which assumes that a judgment of indeterminacy is the right one for external observers to reach when they find that other people disagree in the personal, first-level moral judgments they make. I shall discuss mainly the internal version of the thesis, because once we see why the default thesis fails in that version, we see why it fails in the other as well.
Suppose I am bewildered whether abortion is wicked. Certain arguments or analogies make it seem so to me sometimes, when I am in some moods. But other arguments or analogies make it seem not so at other times. I confess that I lack any secure or stable sense of which of these sets of arguments is better. Then, according to the first-personal version of the default thesis, I ought just to conclude that the question of abortion is indeterminate. There is no right answer to that question, I should say, but only different answers. This premise assumes that indeterminacy about abortion is a theoretically less ambitious claim than either the proposition that abortion is immoral or the claim that it is not, that something more must be known or shown or supposed by someone who asserts a moral judgment of any kind, or who denies one, than by someone else who says that the matter in hand is just a matter of opinion or that the answer is indeterminate or something of that sort. If so, then the failure to supply that "more" is enough to establish indeterminacy, which is therefore true by default.
It is certainly true that what we might call U-propositions, which are claims that we are or should be uncertain about some issue, are theoretically less ambitious than more positive or negative claims: U-propositions do hold by default in the way just described. If I see arguments on all sides of some issue, and do not find, even after reflection, one set of arguments stronger than the others, then I am entitled without more to declare that I am uncertain, that I have no view of the matter.19 I do not need a further, more substantive, reason, beyond my failure to be persuaded of any other view, for claiming uncertainty. But in all these respects indeterminacy differs from uncertainty. "I am uncertain whether the proposition in question is true or false" is plainly consistent with "It is one or the other," but "The proposition in question is neither true nor false" is not. Once uncertainty is taken into account, in some such way, then the default thesis collapses because if one of the alternatives-uncertainty-holds by default, indeterminacy, which is different, cannot.
A belief in indeterminacy is a positive claim, and it needs a positive reason or assumption to support it. There are three possible views I might take about the abortion issue, excluding uncertainty, but I am as much (or as little) in need of an argument for the third view-indeterminacy-as I am of an argument for either of the other two. So the default thesis fails in its first-personal version. It therefore fails in its third-personal version as well. If (as a matter of their ordinary, everyday moral convictions) one person holds that abortion is wicked, a second that it is not, and a third that the issue is indeterminate, there are three, not two, substantive positions in the field, and a fourth-party observer needs as much or as little argument for siding with any one of the three as with either of the others.
Reasons of a Third Kind
What kind of positive reason do we need for thinking that it is neither true nor false that abortion is wicked? It is not enough simply to have no reason for thinking that proposition true and none for thinking it false. That would justify only uncertainty. We need a reason for thinking that there are no reasons for thinking it true or for thinking it false. The difference is important. Though reticence is generally appropriate under uncertainty, it would be wholly out of place for anyone genuinely convinced that the issue is not uncertain but indeterminate. The Catholic Church has declared, for example, that even those who are uncertain whether a fetus is a person with a right to live should oppose abortion, because abortion would be so terrible if it turned out that a fetus was a person. No comparable argument could move someone who is convinced that there is no right answer to the question whether a fetus is a person. He might of course have others reasons for taking a position about abortion-he might say that abortion displeases (or, I suppose, pleases) him, or that since those who think the fetus is a person are very upset by abortion it should be banned for that reason, or that since it is unjust for the state to limit liberty by taking up one position on an indeterminate issue, abortion should be permitted for that reason. But he cannot have the reasons for reticence or agony or commitment to further thought that someone who thinks the issue uncertain has.
Someone who believes there is no right answer to a particular moral question thinks, then, that no one could have a reason of a certain kind for acting or praising or condemning in a certain way. What kind of reason she thinks no one could have depends, of course, on what mode of acting or praising or condemning is in play, because the reason must be internal to the domain. If the issue is an aesthetic one-if the question has arisen, for example, whether Picasso was greater than Braque and she thinks no one could have a reason for preferring either to the other-then she thinks there are no aesthetic grounds that could justify either preference. But since whether there is a justification for some aesthetic position is patently an aesthetic question, her position, that there is no justification for either position, is an aesthetic one as well, and it must be supported, if at all, on aesthetic grounds. If the issue is a moral one-if she thinks that it is neither true nor false that abortion is wicked-then she thinks there is no moral reason for a view either way and her grounds must therefore be moral as well.
Some people think that the idea of nobility, as an aesthetic concept, can sensibly be applied to wine: they say, for example, that P?trus is nobler than any Beaujolais. Critics of wine-snobbery believe, on the contrary, that the concept is simply out of place in this context. P?trus may taste better than Beaujolais, and its taste may linger longer, but it is neither true nor false to say that it is nobler. If that is our view, we must have some basis for it. We might not be able to deploy any careful elaboration of the concept of the noble. We might rely instead just on a felt sense of the limits to which a complex aesthetic characterization can be pushed-music may be noble, we might think, but not wine. However elaborate or crude our view is, however, it is an aesthetic rather than logical or philosophical or semantic one. We reject the nobility of wine, perhaps, as oenophiles would say, out of prejudice or ignorance or fastidiousness. But if so it is aesthetic prejudice or ignorance or fastidiousness, not some other kind.
Now consider the more interesting claim someone might make that, all things considered, Picasso was a greater genius than Beethoven. I assume that you, like me, are willing and think yourself able to make at least some comparisons of artistic merit: we think Picasso a greater painter than Balthus and also, though the case is closer, a greater painter than Braque, great though Braque was, and we think Beethoven a greater composer than Lloyd-Webber and Mozart a greater composer than Beethoven. So we believe that comparisons about the merits of particular artists are in principle sensible. I believe, as I just said, that though Braque was a very important artist, all things considered Picasso was a greater one. If you challenge me, I will try to sustain that opinion in various ways-by pointing to Picasso's greater originality, inventiveness, and range of qualities from playfulness to profundity, while nevertheless admitting certain advantages in Braque's work: a more lyrical approach to cubism, for example. Because artistic merit is a complex subject and my claim is an all-things-considered one, the issue can tolerate a complex discussion. The conversation would not soon turn silly, as it would if I were trying to defend a claim about the greater nobility of P?trus compared with Lafite. I might or might not convince you, after sustained discussion, that I am right about Picasso and Braque; you might or might not convince me that I am wrong. But if neither convinces the other, I will continue in my opinion, as no doubt you will in yours. It would tell against my view that I could not convince you of it, but I would not count this as a refutation.
But if I were asked whether Picasso was a greater genius than Beethoven, my answer would be very different. I would deny both that one was greater than the other or that they were exactly equal in merit. Picasso and Beethoven were both very great artists, I would say, and no exact comparison can be made between the two. Of course I must defend the distinction I have now drawn. Why can I rank Picasso and Braque, but not Picasso and Beethoven? The difference is not that people agree about standards for comparing artists in the same period or in the same genre. They don't, and it would not follow that the agreed standards were the right ones even if they did. The difference cannot be based on any cultural or social fact of that sort, but must be based, if it makes sense at all, on more general, perhaps even quite theoretical, assumptions about the character of artistic achievement or evaluation. I would try to defend my judgment about Picasso and Beethoven in that way. I believe that artistic achievement can only be measured as a response to artistic situation and tradition20 and that only order-of-magnitude discriminations can be made across such traditions and genres. So though I do think that Shakespeare was a greater creative artist than Jasper Johns, and Picasso a greater one than Vivaldi, I believe no precise ranking makes sense among evident geniuses at the very highest levels of different genres. This is not an evidently stable view, and I might well change my mind. But it is the view I now hold.
We may now summarize. Claims of indeterminacy are not true by default: they need, if not argument, which may not be available at any impressive length, at least a basis in more abstract instincts or convictions. These must be convictions or instincts within the domain about which the judgment is made: in the case of aesthetics they must be (as the "theory" I just reported illustrates) aesthetic instincts or convictions. Indeterminacy is a substantive view to be ranked alongside the other substantive views in the neighborhood-that Picasso was greater than Beethoven or vice versa, for example. We cannot demand any greater character or level of demonstration for either of the two latter, positive views than we demand for the former. It would obviously be unfair of me to complain, against anyone who said that Picasso was greater than Beethoven, that he could not prove this to everyone's satisfaction, or that he could not set out an impressive theoretical position from which that view flowed in some smooth way. For that is exactly my situation, too. I claim that the comparison is indeterminate, but the best I can do, in explaining why, is something that will not convince everyone else either.
It is common for philosophers to ridicule, as woolly or inconclusive or dogmatic, the arguments of people who believe that one position in some deep controversy has the better of the case. They say these partisans overlook the obvious truth that there is no "fact of the matter," no "single right answer" to the issue in play. They do not pause to consider whether they have any substantive arguments for that equally substantive position, and, if they do, whether these might not also be ridiculed as vague or unpersuasive or as resting on instincts or even bare assertions in the same way. Absolute clarity is the privilege of fools and fanatics. The rest of us must do the best we can: we must choose amongst the three substantive views on offer by asking which strikes us, after reflection and due thought, as more plausible than the others. And if none does, we must then settle for the true default view, which is not indeterminacy but uncertainty.
These various lessons are equally applicable to morals and ethics. Is there a "right answer" to the question of what Antigone should have done? That depends, plainly, on complex, highly theoretical issues of substantive morality. If we are drawn to a utilitarian morality, we will think that there is a right answer, even if, so many centuries after the play was written, we are deeply uncertain about what it is. We will think that the right answer depends on which of the acts open to Antigone (or, in a different version of utilitarianism, which of the rules that might have guided her choice) would have produced the greatest net balance of pleasure over pain in the long run. So if we think there is no right answer to Antigone's problem-that whatever she did was the wrong thing to do, for example-we must have rejected utilitarianism. And we must also have rejected a great many other moral theories or approaches or attitudes that would also have insisted on the existence of a right answer.
Our conclusion (or instinct) of indeterminacy reflects, then, a special conception of morality, a conception that emphasizes, perhaps, the brute, unremitting character of duties of different kinds. We may not be able to say very much to the utilitarian in defense of our moral attitude; perhaps we cannot say even as much to him as he can to us. That does not mean that we are wrong-or right. Here, too, conviction is inescapable. We remain as unconvinced by his appeals to the fundamental importance of human welfare as he does by our reminders about the deep and uncompromising character of duty; neither of us must bow to the other just in virtue of the raw, logical force of the arguments either can deploy. But our view has no epistemic or logical head start over his. It is just as much grounded in a moral attitude or conception as any other view in the field: it needs, to repeat, as much or as little positive argument in its defense as they do.
Philosophers have overlooked that point in ethics to an even greater degree than they have in morals and aesthetics. It is a popular view, for example, that at least within a certain range of options there is "no right answer" to what choices people should make about how to lead their lives. Suppose a young woman must choose whether to pursue a promising career as a public-interest lawyer in Los Angeles or to emigrate to a kibbutz in Israel. (Of course, she will have many other options as well. But suppose these are the only two now in question.) She might be puzzled about a great deal that is involved in that choice. Which life would she find more satisfying in retrospect? In which role would she be more successful? In which would she contribute more to the well-being of others? She may be uncertain about the right answers to any of these questions taken separately, and she will very likely be uncertain about the right answer to the further question how to weigh these answers against one another. Now suppose someone says that she is silly to worry about any of this, because, since both of these lives are valuable, there is no right answer to the question which is, all things considered, the best. That surprising view might be right. But it cannot be true by default. It needs as much positive argument or instinct or conviction as the contrary claim that the best life, all things considered, really lies in emigration, and no such argument is supplied simply by stating the obvious fact that there are many values, and that they cannot all be realized in a single life. For the question remains-for the philosopher as well as for people who agonize over fateful decisions-which choice is nevertheless best. That is an ethical question, and the third answer-that neither is-needs, not truisms about the pluralism of value, but an ethical defense of the kind it almost never receives from philosophers who embrace it.
I will add one more example, because I can make the point quickly.21 It is a popular thesis that in very hard cases at law, when the legal profession is split about the right answer, there actually is none, because the law is indeterminate on the issue. This "no right answer" thesis cannot be true by default in law any more than in ethics or aesthetics or morals. It does not follow from the fact that no knock-down argument demonstrates that the case for the plaintiff is, all things considered, better or worse than the case for the defendant that the plaintiff's case is not, in fact, actually better or worse. Since the no-right-answer claim about law is a legal claim-it insists that no legal reason makes the case for one side stronger than the case for the other-it must rest on some theory or conception of law. It is not hard to find theories of law that do claim to support that conclusion: the simpler versions of legal positivism do, because they hold that only past decisions of officials provide legal reasons, and in many cases there was no past official decision on either side of an issue. There are more complex and plausible legal theories that might also be thought to generate indeterminacy in certain cases: a theory that makes moral conclusions sometimes dispositive of legal argument might be combined, for example, with a view of morality that made some moral issues indeterminate, as in Antigone's case.22 Because such indeterminacy-generating legal theories actually exist, in articulated and elaborate form, law provides a good illustration of my claim that indeterminacy is a substantive position and therefore counts as a case of internal rather than external skepticism about the more positive views it challenges. Most contemporary legal scholars who find it self-evident that there is no right answer to controversial legal questions do not subscribe to legal positivism or to any other theory that offers positive legal arguments for indeterminacy.23 They fall into the fallacy of supposing that indeterminacy holds by default.
Intelligible indeterminacy claims, then, are special cases of I-propositions. We make sense of them, if there is any sense to make, by treating them as internal, substantive positions based, as firmly as any other, on positive theories or assumptions about the fundamental character of the domain to which they belong. In law, for example, the functional need for a decision is itself a factor, because any argument that the law is indeterminate about some issue must recognize the consequences of that being true, and take these into account. Consider the difference between the aesthetic question raised by the pure form of the question about Picasso and Beethoven, and the different question that would be raised if Congress had instructed that a great statue be raised depicting one or the other, whichever was the greater artist. This kind of "delegation" does not preserve the issue that is delegated unchanged; the delegation changes the question because it adds a context of purpose and a demand for action.
So the question becomes: given that Congress has instructed, at a particular time, that a statue be raised, and supplied that test for deciding whose statue it should be, which of the decisions the agency might make is the right one? A great deal beyond pure aesthetic judgment is then relevant. Which choice, if made by Congress itself, would have been the better, all things considered? Would one choice, for example, better match the public's opinion of the two artists? Would one choice be more likely to foster greater appreciation for the arts in general? Not responding at all-building no statue-is of course part of the set of options. But it would not follow that that was the best option just from the failure of the pure aesthetic judgment to deliver a verdict. For the agency must treat that fact as part of its challenge. Given the congressional instruction, and given the failure of the pure aesthetic judgment, what should be done? The range of factors pertinent to that new question is much larger than the range pertinent to the pure aesthetic judgment, and the case for indeterminacy is correspondingly weaker.
One final point. Archimedeans often declare themselves to be against "theory," with the exception of their own. They say "theory" is produced by philosophers who fail to understand that, from the philosophical or archimedean perspective, there are no right answers to the questions these theories purportedly address. We can now see (as a coda to the argument as a whole) that the supposed alliance between indeterminacy and an antitheory stance is as bogus as everything else in the archimedean architecture. U-propositions, which concede uncertainty, might be said to be (at least relatively) untheoretical. But indeterminacy claims, as we have just seen, are different from uncertainty claims; the former require positive support, and there is no reason to think that the positive support they need draws any less on theory than other positive claims do. On the contrary, in the global form in which they are characteristically defended, they are likely to be much more ambitious because much more general. Consider the legal claim just discussed: that when lawyers disagree, and there is no knock-down argument available to reconcile them, it follows that the case for neither side is better than the case for the other. There are an unlimited number of reasons why some but not all lawyers might think that one side had the better of a particular legal argument. Someone defending the view that no such reason can in fact tip the balance either way in any controversial case faces an enormously difficult task, much more difficult than that faced by someone who wants to argue for one decision rather than another in a particular case. How can he avoid appealing to some very general and abstract theory, like legal positivism? Someone defending a global claim of indeterminacy about morals or ethics or aesthetics-that there is never a right answer to any question about what we ought to do or how we ought to live or what is wonderful-has an even greater problem, and his need for a very abstract theory delivering these global conclusions seems even more evident. These are truly heroic claims, of vast theoretical pretension, and trying to dress them in the modest clothes of common sense or practicality is more comic than persuasive.
A Pious Hope
We want to live decent, worthwhile lives, lives we can look back on with pride not shame. We want our communities to be fair and good and our laws to be wise and just. These are enormously difficult goals, in part because the issues at stake are complex and puzzling. When we are told that whatever convictions we do struggle to reach cannot in any case be true or false, or objective, or part of what we know, or that they are just moves in a game of language, or just steam from the turbines of our emotions, or just experimental projects we should try for size, to see how we get on, or just invitations to thoughts that we might find diverting or amusing or less boring than the ways we used to think, we must reply that these denigrating suggestions are all false, just bad philosophy. But these are pointless, unprofitable, wearying interruptions, and we must hope that the leaden spirits of our age, which nurture them, soon lift.
Copyright 1996 by Ronald Dworkin.
I have been lecturing and teaching on the subject of this essay for many years, and more people have discussed and criticized my ideas than I can acknowledge or remember. Particular thanks are due to Paul Boghossian, Ruth Chang, G. A. Cohen, Donald Davidson, James Dreier, Stephen Guest, Derek Parfit, Thomas Scanlon, Nicos Stavropoulos, Sigrun Svavarsdottir, and David Wiggins, to the members of the New York University Law School Colloquium on Legal, Political, and Social Philosophy, and to my teaching colleagues Thomas Nagel and Bernard Williams. The essay will appear in a collection of essays on incommensurabilty, edited by Ruth Chang, to be published by Harvard University Press in the fall of 1997.
1. "Skepticism" is used in different ways. I use it in the sense not of agnosticism but of rejection. I emphasize that different skeptics, even about morality, have different targets. The skeptics I mainly discuss claim to reject not morality but only certain philosophical opinions about it.
2. Whether a form of skepticism is properly understood as internal or selectively external to the domain it criticizes is often a complex interpretive question. Science-based skepticism about religion is internal, for example, if religion is understood to be itself part of the domain of science, as it should be if it includes causal claims about the origin of the universe competitive with other cosmological claims.
3. I have discussed this distinction before. See my Law's Empire (Cambridge, Mass.: Harvard University Press, 1986), pp. 78-86.
4. Richard Rorty, Contingency, Irony and Solidarity (Cambridge: Cambridge University Press, 1989).
5. Richard Rorty, "Does Academic Freedom Have Philosophical Presuppositions?," in The Future of Academic Freedom, ed. Louis Menand (Chicago: University of Chicago Press, 1996), 29-30.
6. Not, that is, if they are sensible. I am not considering here the patently false view that it is part of the very meaning of "correct" that right actions are those that maximize happiness. If that were right, only utilitarians could avoid linguistic error.
7. See Hilary Putnam, "On Properties," in his Philosophical Papers, Volume I, Mathematics, Matter, and Method (Cambridge: Cambridge University Press, 1975), at p. 305.
8. Compare the discussion of "subjective naturalism" in M. Davies and L. Humberstone, "Two Notions of Necessity," Philosophical Studies 00 (1980): 22-25. The subjectivism discussed there is personal; something being wrong is taken to be a matter of the speaker disliking it. The argument in the text, which addresses the less implausible example of intersubjective subjectivism, would of course apply to the personal version as well.
9. That is the sense of "represents" that Crispin Wright uses in exploring the question whether moral beliefs can represent reality. See Wright, Truth and Objectivity (Cambridge, Mass.: Harvard University Press, 1992). I return to Wright's discussion later.
11. Allan Gibbard, Wise Choices, Apt Feelings: A Theory of Normative Judgment (Cambridge, Mass.: Harvard University Press, 1990), p. 154.
12. Simon Blackburn, Essays in Quasi-Realism (Oxford: Oxford University Press, 1992), pp. 172-73.
13. Ibid., p. 174.
14. Mackie, Ethics: Inventing Right and Wrong (New York: Penguin Books, 1977), p. 40.
15. That is a crude statement of a prominent and influential account of interpretation. See Donald Davidson, Inquiries into Truth and Interpretation (Oxford: Clarendon Press, 1984).
16. See Gilbert Harman, The Nature of Morality (New York: Oxford University Press, 1977).
17. See, for example, the literature described in Robert Wright, The Moral Animal (New York: Pantheon, 1994).
18. Derek Parfit, Reasons and Persons (Oxford: Oxford University Press, 1984), p. 431.
19. I say "all" sides to include the possibility that I see no good reason for thinking the issue indeterminate either.
20. I defend a different but in some ways similar view about ethics in "Foundations of Liberal Equality," in Tanner Lectures on Human Values, Volume XI, ed. Grethe Peterson (Salt Lake City: University of Utah, 1990).
21. For a fuller exposition of the argument of this paragraph, see "No Right Answer?" in my book, A Matter of Principle (Cambridge, Mass.: Harvard University Press, 1985).
22. John Mackie argued that a legal theory that made moral argument relevant would characteristically produce indeterminacy. See his "The Third Theory of Law," Philosophy & Public Affairs 7, no. 1 (Fall 1977): 3-16.
23. See the discussion of "Critical Legal Studies" in my Law's Empire, pp. 271-74.