Bernard Williams
Bernard Williams; drawing by David Levine


Bernard Williams has been a distinctive presence on the intellectual scene for more than three decades. He might be described as an analytical philosopher with the soul of a general humanist. Notably sharp-minded and dialectically skilled, he is also impressively learned in history, classical scholarship, literature, politics, and music. He opposes the widespread contemporary desire to model philosophy on science, preferring to locate philosophy within its historical and cultural context; there are other ways to be intellectually serious than by aping the physicists. His interests have focused on ethics, in which he has contested the rationalist and universalist claims of both Kantianism and utilitarianism—though he has also written a first-rate book on Descartes and pioneered in the discussion of personal identity. He possesses a mind that is both flexible and muscular: open and imaginative on the one hand, rigorous and no-nonsense (and occasionally stinging) on the other. No one else can lance an opponent with a comparable twinkle in his eye. The phrase “withering wit” is unavoidable when witnessing Williams in action.

And yet no theory bears his name, nor has he made any notable technical contribution to the subject. In contrast to his late contemporary R.M. Hare, whose name will forever be associated with “prescriptivism” in ethics—the idea that moral utterances do not describe states of affairs but prescribe courses of action—Williams has not produced anything deserving to be called an ethical theory. Indeed, he has consistently mocked such theories. He gives the impression of having seen through all that—that he is simply far too intelligent and sophisticated to do anything as naive and crude as propose an analysis of moral discourse.

His influence lies more in the style with which he has discussed moral issues: his insistence on not losing sight of actual moral life, his sensitivity to moral subtleties, his Nietzschean disdain for high-minded moral posturing. His writings do not offer the dubious exhilaration of grand philosophical theory, in which messy reality is tamed and caged, but the thrill of seeing pretension punctured by a kind of high-voltage common sense (backed up by impressive erudition). Williams is a principled anti-moralist, a brilliant anti-theorist, an original blend of logic and humanism, an insider’s outsider. There is no one in philosophy quite like him.

Williams’s new book, the defiantly titled Truth and Truthfulness, bears all the marks of its author. The argument is heavily armored, both in its range of reference and in the structure of its sentences, which almost always coil around some anticipated objection and skewer it; Williams is always one step ahead of his reader. Every sentence seems constructed in such a way as never to need withdrawing; it is fully shielded, immune from refutation. Williams is so well protected that it is sometimes hard to make out the shape of his position. The sentences seldom descend to elegance, and lucidity seems less highly prized than impregnability, though there are certainly flashes of humor and no lack of verbal resource. Williams is often praised for his prose style, but I find his sentences frequently lopsided and lumpy (he is excessively fond of the phrase “very significantly”), and his transitions are often opaque. Still, there is no doubting his verbal firepower, his subtlety, and his argumentative strength.

His book is about a very important subject, especially in the current academic climate: the centrality of the concept of truth in our intellectual lives, and the value of respecting it in our speech and the formation of our beliefs. Williams argues that truth is both indispensable to responsible discourse and unscathed by recent “postmodernist” critiques (such as Richard Rorty’s). Such critiques “depend on the remarkable assumption that the sociology of knowledge is in a better position to deliver truth about science than science is to deliver truth about the world.” But his main concern is with what he calls, felicitously, “the virtues of truth,” which he divides into two, labeled Accuracy and Sincerity. Accuracy is the disposition to take care that in acquiring our beliefs we do our utmost (within reasonable limits) to ensure that we have the best evidence possible, that we weigh it impartially and thoroughly, and that we remain always ready for counter- evidence—that is, that we do everything we can to make our beliefs sensitive to the truth. This means not indulging in wishful thinking and self- deception, as well as being alert for propaganda, bias, and sheer laziness.

Sincerity is related to communication between people: speaking the truth, expressing what one really believes, avoiding deception. These are virtues, and not merely dispositions, because there are temptations and obstacles, both external and internal, that interfere with achieving Accuracy and Sincerity (Williams capitalizes these words to indicate their somewhat technical use). A cigarette company executive, for example, may know accurately the evidence of his product’s harmfulness, but may not be willing to assert what he really believes. Or he may sincerely believe that cigarettes are harmless, having wishfully ignored the best evidence. Or he may simply lie. Fantasy and wish are strong forces in the psyche, and there is often a lot to be gained by lying. The truthful person, as Williams rightly says, is truthful spontaneously and naturally; his beliefs are linked to his speech without hesitation or calculation. He just comes out with the truth.


On the face of it, the virtues of truth necessarily involve truth itself: “If you do not really believe in the existence of truth,” Williams asks rhetorically, “what is the passion for truthfulness a passion for?” So if we can show the importance of these virtues, we will have indirectly shown why truth is a concept we cannot do without (I shall come back to whether this is really so later). Williams identifies a group of thinkers, typified by Rorty, that he calls the “deniers”—those who deny that truth is a useful or coherent or politically defensible concept—and he argues that they are in imminent danger of losing these important virtues: “To the extent that we lose a sense of the value of truth, we shall certainly lose something and may well lose everything.” The academy, and education generally, must define itself as dedicated to the twin virtues of truth or else it will lose its authority and our intellectual activities “may tear themselves to pieces.”

The deniers, Williams insists, must at a minimum accept that there are “plain truths,” such as that it is Tuesday, that I just ate an egg, that Paris is in France; and there is no alternative to the idea that belief is a mental attitude that intrinsically aims at the truth. Something has gone wrong if I believe what is false; indeed, I cannot form beliefs at all just as a matter of my will. By their nature beliefs are dedicated to respecting reality, which is why I cannot decide to believe what I know very well to be false (by contrast, I can easily decide to assert what is false).

Thus, Williams argues, I have a great many beliefs in plain truths, and these are not the result of anyone’s arbitrary exercise of power. This idea of truth as independent of will, and the human obligation to pursue it, is something that Nietzsche insisted upon, contrary to the views of many deniers; this is why, as Williams points out, he emphasized how difficult it can be to face the truth, especially about oneself. Nor is the concept of truth something that varies from culture to culture; the concept of truth is the same always and everywhere (though what people take to be true obviously varies).

This last argument is familiar enough, though worth repeating. Where Williams strikes out in a new direction is with his account of the underpinnings of the virtues of truthfulness. His aim is to “explain the basis of truthfulness as a value,” and his method follows a strategy of Nietzsche’s, namely to sketch a “genealogy” of that value. Nietzsche’s aim was to debunk Christian ethics by claiming that they arose from resentment on the part of the powerless of the power of their superiors. Williams’s aim is to vindicate the virtues of truth by showing that they can have a useful function within a society existing in a state of nature. Thus we are invited to consider a fictional society of primitive speakers who develop the practice of pooling information so that they can take better advantage of their environment: some people have information that others lack, owing to their specific position in relation to the world, and a division of labor develops, whereby information is shared to the benefit of all. Williams offers this genealogical story as a kind of naturalistic explanation of the virtues of accuracy and sincerity, rooting them in pre-ethical ideas of cooperation and human needs. He hopes that by seeing how truthfulness could naturally arise in a primitive society we will come to appreciate its value in a way that avoids Platonic conceptions of the basis of virtue as a special kind of knowledge of a supersensible reality.

A genealogy of this kind can be illuminating if it is unclear how a particular human faculty could have come about, but it is hard to see how it can work to vindicate a value. We can, after all, come to see how various vices might come about (for instance, stealing), but obviously this does nothing to justify them. We have to suppose that what the disposition brings about is itself a good thing in order to vindicate it, so that we need to make a prior judgment of value—which will raise the question of how to vindicate that value. In the case of truthfulness, what the disposition brings about in Williams’s view are the benefits of cooperation—a better life for all. The same might be said for other virtues, such as courage or integrity or generosity: they can all lead to people being better off, so that we can see how people in a state of nature might favor these virtues. Williams does not raise the question of how generally he wishes to apply his genealogical method, but reflection on it reveals the limitations of the method. There are two points, one of which he recognizes, and the other not.


The first is that showing the function that a virtue serves can only give it instrumental value, not intrinsic value: we might learn what the virtue produces in the way of benefits, but we don’t learn why it might be valued in itself. Since Williams insists, rightly, that truthfulness has an intrinsic value, in the sense that we value it for itself and not merely for the good results it might have, his functional story fails, by his own standards, to capture that intrinsic value; so it does nothing, really, to vindicate the intrinsic value of truthfulness.

Secondly, the functional account looks like a thinly disguised form of utilitarianism, an argument to the effect that truthfulness is good because it increases the general level of human well-being. The problem now is that the genealogical story is no longer helpful once this element of it is highlighted. Why not simply say that truthfulness is a virtue because it increases human well-being, and forget the story about states of nature and pre-ethical societies? What does the genealogical story add to the simple utilitarian account? The only reason it seems to justify the virtues of truth to begin with is that it connects the value of truthfulness to the value of well-being; by itself the story about pooling information in a state of nature makes no progress toward vindicating truthfulness as a moral virtue. The simple point is that the supposedly novel idea of vindication by means of genealogy only works because it assumed something like a utilitarian argument for speaking the truth; and once this is seen the genealogy itself becomes theoretically redundant. What if the information did nobody any good at all?


Leaving aside the aim of vindicating the virtue of truthfulness, what should be said of Williams’s central contention in this book, namely that truth matters because the virtues of truth matter, the virtues being accuracy and sincerity? It may seem like the simplest tautology that truthfulness requires a truth to be true to; but the deniers, or some of them, may reasonably protest that this formulation is tendentious at best. What Williams is really suggesting is that the virtues of accuracy and sincerity can only be explicated according to the notion of truth.

We can see how substantive this claim is by considering Williams’s own views about “ethical truth,” for in this matter he himself counts as a denier, believing that while scientists can hope to find out “how things are,” in matters of ethics there is no such hope of establishing the truth.1 He is concerned in this book with the morality of respecting a factual truth that is not a moral truth, and it is striking how absent from his discussion is the idea of speaking the moral truth (though the book by itself is largely an example of this). Yet even for those who don’t believe that the concept of truth is applicable to moral statements and attitudes, surely there are still virtues of accuracy and sincerity, i.e., the virtues of care in arriving at moral attitudes and honesty in expressing such attitudes. We should do our best to arrive at considered moral judgments, not succumbing to wishful thinking and propaganda, and we should express our moral convictions sincerely.

There is, however, a dominant tradition in moral philosophy, of which Williams is a distinguished representative, that contests the application of the concept of truth to moral views. The emotivists, for example, took moral utterances to be mere expressions of emotion (like saying “Boo” or “Hooray” on seeing a performance), devoid of truth or falsity—there are, for them, no “moral facts.” Therefore, according to this tradition, one cannot be literally truthful in ethics (though, as I’ve said, one can certainly be careful and sincere in arriving at and expressing moral convictions.) It follows, for Williams, that one must characterize these moral virtues in terms other than that of truthfulness. But if that is so, then the general deniers, like Rorty, will want to know why they cannot accept that there are virtues such as care in arriving at beliefs and sincerity in expressing them while declining to describe them as leading to the truth, since they reject the very idea of truth. (In my own view, truth applies equally and univocally to moral and scientific statements.2 )

Rorty and other deniers will accept that intensive indoctrination is different from rational investigation, but they hold that this difference does not require us to accept a notion of objective truth—just as emotivists can accept that genuine moral thought is different from indoctrination, even though truth does not enter the picture, as they see it. After all, there is such a thing as a sincere expression of emotion—such as sighing or laughing—that is not even a candidate for being true or false. Williams does not confront this problem, and he needs to.

He does mention an allied problem: why the notions of evidence and justification by evidence are not enough to explain what the virtues of accuracy and sincerity aim at; in his view, it is essential that we also employ the concept of truth. But is this convincing? What we most directly have a duty to do is to respect the evidence: we must form our beliefs according to the available evidence, and we must assert only what we are justified in believing. Our only route to the truth is by means of the evidence and the assertion of what it shows, so there cannot be an additional duty to believe and assert the truth. The concept of truth applies to the results of inquiry and assertion, not to the procedures for arriving at beliefs and making assertions; and what we have a duty to do is observe the right procedures. Whether our justified beliefs are true is a matter of whether they happen to correspond to reality—we have done our part only so far as we try to follow the best procedures for acquiring justified beliefs.

There are two reasons for drawing this distinction between justification and truth, neither of which is mentioned by Williams. First, no one can be reasonably criticized for believing and asserting what is false so long as the best efforts have been made to ensure that all the evidence has been respected; and it is clearly possible for an apparently justified statement to be false, as when the man caught holding a smoking gun over a corpse didn’t in fact do it, even though he claims he did. All you can do by way of aiming at the truth is to weigh the available evidence, but it is possible to do this responsibly and still end up with a false belief. So the virtues of accuracy and sincerity are really exhausted by the duty to respect the evidence—even if the evidence turns out to favor a false belief.

Secondly, there is no virtue in accidentally believing and asserting what is true: you may be guilty of the grossest wishful thinking and inept calculation in supposing you will win the lottery, and then it turns out that you have. In general, it is possible to believe what is true through entirely irrational means, such as wishful thinking, in which case you do not know what you believe. These two familiar points of epistemology—the study of what is knowledge and how we acquire it—show that the epistemic virtues attach most directly to the weighing of evidence, not to believing what is true; so it is the notion of justification—i.e., that we must assert only what we are justified in believing—that is really essential to the virtues of “truthfulness.”

Williams has a reply to this kind of argument, though he doesn’t raise the issue in quite the terms I have. He argues that the notion of justification must be characterized with respect to the idea of truth, since evidence is what justifies the claim that a given proposition is true. That may well be right, but it doesn’t show that the epistemic virtues of accuracy and sincerity are appropriately characterized by saying that they necessarily lead to the truth. Instead of referring to the truth, it is sufficient to refer to responsiveness to evidence, since there can be no more to the injunction to seek the truth than to respect the evidence and be willing to assert it. It is not that there is no distinction between evidence and truth; it is rather that our position as cognitive agents is confined to that of evaluators of evidence.

This raises the big question of precisely what notion of truth Williams is himself using. He says very little about traditional theories of truth, such as the coherence, pragmatic, and correspondence theories, supposing that he is working, as it were, with an idea of truth that does not require him to argue for one of these or another; but I think he cannot avoid committing himself to a concept of truth if he is to have a convincing position. It is fairly clear, though not explicitly stated, that he is assuming some sort of correspondence theory, according to which our beliefs are true when they correspond to the world independent of our minds. But a coherence theorist would be within his rights to complain that, for him, truth and justification come down to the internal logical coherence of beliefs, with no reference made to any supposed reality independent of mind, and that this is quite consistent with accepting that there are norms of truthfulness. For the coherence theorist the norm of accuracy is just the requirement that we make our beliefs as coherent as we can, and the norm of sincerity is the injunction to assert only what is consistent with what we believe.

In other words, accepting that there are virtues of truth does not entail a correspondence theory of truth, as opposed to a coherence theory. Similarly, verificationist philosophers, who maintain that truth is no more than propositions that can be asserted with justification, perhaps in the long run, can happily agree that there are epistemic virtues such as accuracy and sincerity; they can still deny that there is any notion of truth that transcends human powers of verification. What Williams hasn’t given us is any reason to suppose that the epistemic virtues demand any specific conception of truth, and so the deniers are given ample room for maneuver. In other words, the traditional notion of objective truth, defined as correspondence to the external facts, is not the only notion that can embrace the ideas of accuracy and sincerity (though I would myself, on independent grounds, strongly defend the correspondence theory).

Perhaps the most engaging parts of Truth and Truthfulness are the discussions of lying. Williams makes a valuable and subtle analysis of whether any moral distinction can be made between outright lying and other forms of verbal deception—as when Saint Athanasius, rowing down a river, is asked by his persecutors “Where is the traitor Athanasius?” and replies accurately, “Not far away.” Williams argues convincingly that no such distinction exists, at any rate of the magnitude alleged in many traditions. He is also convincing on the question of whether lying can be morally defensible, arguing that certain types of questioner forfeit their right to be told the truth—as with the homicidal maniac seeking information, or the prurient gossip columnist, or the torturer. The practice of truthful assertion, he argues convincingly, rests upon a foundation of mutual respect and trust, of honor and shame, but these associations may be abrogated in certain cases. In the typical case, however, lying is a culpable violation of trust, which is why we take truthfulness to have intrinsic value—over and above any benefits we might receive from being apprised of the truth. Speaking of a person who has been deceived, Williams writes:

In allowing himself to accept the other’s belief as his own, and taking it that he has been given the truth through the speaker’s assertion, he will feel that he has come as close to the real thing as anyone in his situation could do. When he realizes that he has been betrayed, there is a complete reversal: the speaker’s will was entirely out of the picture, but now the picture is nothing but a product of that will. The victim recognizes the bare-faced lie as a pure and di-rect exercise of power over him, with nothing at all to be said for it from his point of view, and this is an archetypal cause of resentment: not just disappointment and rage, but humiliation and the recognition that in the most literal sense he has been made a fool of.

To this powerful statement one can only give heartfelt assent.

His basic thought about deception is that the deceiver substitutes his will for the reality of the world: he purports to be innocently reflecting how things are out there, but in fact he is imposing his will on his interlocutor. Truthful assertion is virtuous because there is no exercise of power over the hearer; the speaker acts merely as a conduit of the truth. This sounds right, but I think Williams misses an important aspect of our conception of sincerity as a virtue. It is simply this: in having a disposition spontaneously to tell the truth I thereby risk saying things that may rebound to my disadvantage and to your advantage. In other words, honesty implies self-abnegation. “Have you seen the $100 bill I left lying about?” an acquaintance asks me. Spying it under a plate, I realize that a quick lie could make me $100 richer. But I blurt out, “Yes, it’s right here.” What is important about sincerity is that it is not calculated to benefit the speaker.

This explains why lying to protect an innocent can also be virtuous, especially if the lie puts the speaker at some risk: the lie is not in the speaker’s self-interest. Much the same can be said about the virtue of accuracy, since giving up wish-fulfillment might make me less happy than accepting comforting falsehoods. It is not merely that belief and assertion are made subordinate to reality that brings virtue; it is that this can result in self-denial. So these virtues are such for the same reason other virtues are, like courage—they involve a kind of altruism, or at least a lack of egoism. Oddly, Williams does not mention this obvious point, possibly finding it too plodding.

There is a great deal in his book I have not addressed, particularly in relation to politics and history. There are rich discussions of Thucydides on the idea of the past, of Rousseau and Diderot on the nature of personal authenticity, on the connection (if any) between freedom of speech and the dissemination of truth, on how liberal democracy is bound up with a commitment to truth-telling, on whether historical interpretations can achieve objectivity. My focus here has been on the philosophical core of the book. The more applied discussions are independent of the philosophical material I have criticized, and I found them generally stimulating and enlightening. For example, Williams argues plausibly that the kind of open market in information afforded by the Internet is more likely to foster error than truth because, unlike the case of the scientific community, there are no professional standards that need to be met by the purveyor of information or misinformation. These parts of his book show Bernard Williams at his most persuasive and penetrating—when he can throw off the shackles of philosophical analysis.

This Issue

April 10, 2003