Truthfulness, Lies, and Moral Philosophers: What Can We Learn from Mill and Kant? ALASDAIR MACINTYRE THE TANNER LECTURES ON HUMAN VALUES Delivered at Princeton University April 6 and 7, 1994 ALASDAIR MACINTYRE is Arts and Sciences Professor of Philosophy at Duke University. He was educated at Queen Mary College, University of London, and at the University of Manchester. He taught at various British universities, including Oxford and Essex, until 1970. Since then he has taught at a number of American universities, most recently at Vanderbilt University where he was the W. Alton Jones Professor of Philosophy from 1982 to 1988, and from 1989 to 1994 at the University of Notre Dome, where he was the McMahon/Hank Professor of Philosophy. He is past president of the eastern division of the American Philosophical Association. His numerous publications include Three Rival Versions of Moral Enquiry (1990), First Principles, Final Ends and Contemporary Philosophical Issues (1990), Whose Justice? Which Rationality? (1988), After Virtue (1981), Secularization and Moral Change (1967), and A Short History of Ethics (1966). When children are still quite young, they learn not one, but two rules concerning truth-telling and lying and these in very different ways. One of those rules they learn by explicit instruction, characteristically when they have first been discovered in a lie. What they are then taught is that ti is wrong to lie, but what the rule is that is invoked notoriously varies from culture to culture and sometimes within cultures. For some lying as such is prohibited. For others some types of lie are permitted or even enjoined, but about which types of lie are permitted or enjoined there are also significant differences. It is not difficult to understand why. Among those types of lie that are often permitted or enjoined in different social orders are certain types of protective lie, lies designed to defend oneself or one’s household or community from invasive hostility, perhaps from religious persecutors or witches or the tax-collectors of some alien power, or to shield the vulnerable, perhaps children or the dangerously ill, from knowledge thought to be harmful to them. since who is judged to need protection from what varies from one social and cultural order to another, which of these types of lie are permitted oreenjoined can be expected to vary accordingly. But these are not the only types of exception that are sometimes accorded social recognition and sanction. And, unsurprisingly, reflection upon how the rule that provides for such different types of exception should be formulated and justified commonly gives rise to controversy. Consider as on contributor to those controversies a moral tradition that belongs to the background history of our own moral culture. One of the earlier statements of that tradition, often appealed to later on, is in Book III of the Republic (382c-d), where Socrates is represented as describing some lies as useful against enemies or for the prevention of evils. Some Greek patristic theologians, among them St. Clement of Alexandria, held similarly that on occasion untruths might be told, for example, to protect the Christian community from the invasive enquiries of persecutors. About precisely what classes of untruths were permitted they and later writers sometimes differ from each other, and they also disagree among themselves in the precise [309] 310 The Tanner Lectures on Human Values [MACINTYRE] Truthfulness, Lies, and Moral Philosophers 311 statement of the view that they share, some saying that all lying is prohibited, but that an untruth told for a just reason is not a lie, others that some lying is not prohibited. Newman in summarizing their shared standpoint emphasized that all of them agree that the occurrence of such a just reason “is, in fact, extreme, rare, great, or at least special” (Apologia pro Vita Sua, note G). Modern exponents of this view, he adds, include John Milton, Jeremy Taylor, and St. Alfonso di Liguorio. None of these were, of course, consequentialists. Their position was expressed succinctly by Samuel Johnson: “The general rule is, that Truth should never be violated, because it is of the utmost importance to the comfort of life, that we should have a full security by mutual faith .... There must, however, be some exceptions. If, for instance, a murderer should ask you which way a man is gone, you may tell him what is not true, because you are under a previous obligation not to betray a man to a murderer .... But I deny the lawfulness of telling a lie to a sick man for fear of alarming him. You have no business with consequences; you are to tell the truth” (James Boswell, The Life of Samuel Johnson, June 13, 1784). John Milton, Jeremy Taylor, and Alfonso di Liguorio would all have agreed with Johnson that there is indeed an hierarchical ordering of duties and obligations and that any type of exception to an otherwise universal binding rule can be justified only as required by some other binding rule that is superior in that ordering. But Johnson’s statement suggests at the very least consequentialist questions. If there is indeed an ordering of duties and obligations, what is the principle by which they are ordered, if it is not a consequentialist principle? The consequentialism of J. S. Mill, for example, was intended to provide, by means of the principle enjoining the promotion of the greatest happiness of the greatest number, a standard for just such an ordering. What an evaluation of consequences by means of that principle is to tell us is which binding rules in practice at least have no exceptions (or almost so; see the penultimate paragraph of chapter 5 of Utilitarianism) -the rules prescribing justice, for example 310 The Tanner Lectures on Human Values [MACINTYRE] Truthfulness, Lies, and Moral Philosophers 311 — and which do have a few well-defined classes of exception, such as that otherwise prohibiting lying. And the onus seems to be on the adherents of Johnson’s Christian anticonsequentialism to offer us an alternative and rationally superior principle of ordering. Moreover, if the rule prescribing truthfulness is to be defended as Johnson defends it, further consequentialist questions are raised. Conformity to the rule prescribing truth-telling seems for Johnson to be a means to a further end, what Johnson calls “the comfort of life,” a necessary condition for which is “that we should have full security by mutual faith.” But insofar as this rule is treated only as a means to some such further end, no matter how important, the possibility of evaluating the consequences of making a few well-defined exceptions to it has been opened up. And once again we need to know why we should not move to some more general consequentialist position, such as Mill’s. One answer to this question may well be that I have only reached a point at which it seems difficult to reply to consequentialist claims, because I erred in my starting point. I began after all by considering the kinds of explicit rules that are taught to young children when they are first detected in a lie, perhaps at three or four years of age, and at once noted that often such rules allow for exceptions to the general prohibition of lying. But, it may be said, I ought to have begun with another, more fundamental exceptionless rule, one learned somewhat earlier and not by explicit instruction. This is the rule prescribing truth-telling that we all learned to follow by learning to speak our native language, whatever it is. That rule governs speech-acts of assertion. To assert is always and inescapably to assert as true, and learning that truth is required from us in assertions is therefore inseparable from learning what it is to assert. So two Danish philosophers of language, H. Johansen and Erik Stenius, suggested that “the utterance of a falsehood is really a breach of a semantic rule” (Erik Stenius, “Mood and Language Games,” Synthèse 17, no. 3 [1967), 269), although Stenius understood the relevant rule as one concerning what he called the language-game of reporting, while in fact it is assertion 312 The Tanner Lectures on Human Values [MACINTYRE] Truthfulness, Lies, and Moral Philosophers 313 in general — acts of reporting are only one species of acts of assertion — that is governed by the semantic rule “Assert p, only if p is true.” Mary Catherine Gormally has more recently characterized the relationship of lying to assertion by saying that “a lie (in language) is a cheating move in the language-game of truth telling” (“The Ethical Root of Language” in Logic and Ethics ed. P. Geach [Dordrecht: D. Reidel, 1991), p. 53) and by further arguing that “ ‘assertion’ . . . carries moral weight, like ‘property,’ ‘right’ and ‘obligation.’ It is a value-laden concept” (p. 65). It is, that is to say, among those “concepts which are used to describe human actions in a way which makes it appear why our actions or omissions are bad if we act in certain ways, or fail to do so” (p. 67). Note that the rule enjoining truth-telling in speech-acts of assertion is constitutive of language-use as such. It is a rule upon which therefore all interpreters of language-use by others cannot but rely. And it is not merely a rule of this or that particular natural language. Hence Gormally concluded that about it “one cannot be culturally relativistic” (p. 58), in this following Peter Winch, who had argued that it would be “nonsense to call the norm of truth telling a ‘social convention,’ if by that were meant that there might be a human society in which it were not generally adhered to” (“Nature and Convention,” in Ethics and Action [London: Routledge and Kegan Paul, 1972), pp. 62-63). And David Lewis (who has also argued that in part our commitment to truthfulness in speech is a matter of convention, since “a language £ is used by a population P if and only if there prevails in P a convention of truthfulness and trust in j, sustained by an interest in communication,” “Languages and Language,” in Philosophical Papers [Oxford: Oxford University Press, vol. 1, 1983), p. 169) says about what he calls the “regularity of truthfulness and trust simpliciter” and characterizes as “the regularity of being truthful and trusting in whichever language is used by one’s fellows” that it “neither is a convention nor depends on convention” (p. 184). We stand, so all these writers agree, and surely 312 The Tanner Lectures on Human Values [MACINTYRE] Truthfulness, Lies, and Moral Philosophers 313 rightly, in the same relationship to speakers of other languages in respect of the semantic requirement of truthfulness in assertion as that in which we stand to other speakers of our own language, a relationship defined by the rules governing the use and interpretation of asserted sentences as such. What then are these rules, if they are not not conventional? Winch’s answer was framed in terms of the distinction that Aristotle drew between natural and conventional justice, by saying of the precepts of natural justice that they “have the same power everywhere and do not depend for it on being accepted or rejected” (Nicomachean Ethics V, 1134b19-20). This characterization of the natural holds equally of the semantic rule requiring truthfulness in assertion, which, like the precepts of natural justice, cannot but be accorded universal recognition, and in the vast majority of cases obedience, by the users of all natural languages. In Aristotle’s terms the generally tacit semantic rule enjoining truth-telling is to be accounted natural because recognition of it belongs to the essential nature of human beings as language users. We notice at once that liars cannot withhold recognition from it any more than the truthful can, and this not only because even habitual liars cannot but tell the truth far more often than they lie, sustained in their truth-telling by the interest in communication that, as Lewis emphasized, they share with everyone else. But liars have in addition their own distinctive interest in general conformity to that rule. For they can only hope to lie successfully insofar as it is taken for granted by others that the rule requiring truthfulness in assertion is respected, more particularly by the liar herself or himself. The liar, as Kant put it, cannot consistently will that the maxim upon which she or he acts in lying should be, and should be understood to be, the universal rule governing truth-telling and lying. What successful lies achieve for those who utter them is an advantage with respect to information over those who are deceived. And successful liars necessarily deceive us not only about the subject matter about which they lie, but also about their own beliefs and about their intention in asserting 314 The Tanner Lectures on Human Values [MACINTYRE] Truthfulness, Lies, and Moral Philosophers 315 what they assert falsely, and indeed about their further intention to conceal this intention from us. So that even in the simplest cases of lying there is a complexity in the liar that is absent from the truthful person. Truthful persons may have much to conceal, including their own intentions not to disclose what they are concealing. But they do not misrepresent themselves to others as liars do, with regard to the relationship of their beliefs and their intentions to their assertions. The kinds of advantage to be gained by lying are of course various and so therefore are the motives for lying. Many lies, as I noticed earlier, are protective, motivated by a fear of harm at the hands of others. Some are acts of aggression, motivated by a wish to damage others. Some are intended to maximize advantage in competitive situations. Some lies are acts of flattery and some are intended to make the speaker appear more interesting than he or she in fact is. Some lies are told by office-holders from devotion to what is taken to be the public interest and some are told both to and by office-holders to subvert that interest. But in each of these different types of case, if a lie has been successful, it may well be that the liar will have altered the relationships of power in her or his own favor, or, perhaps, on occasion in favor of someone else. Yet in so doing, whether the lie is successful or not, the liar will also have altered her or his relationship to others in general, by deliberately violating the norm presupposed in all human relationships involving assertive speech-acts. She or he will have relied upon the general human regard for truth, while failing to have regard for it. “Without truth,” Kant wrote, “social intercourse and conversation become valueless” (Eine Vorlesung Kant’s fiber Ethik, ed. P. Menzer, p. 285, trans. L. Infield, Lectures on Ethics [Indianapolis: Hackett, 1980), p. 224). And the offense of the liar, thus understood, is not a matter of the harmful consequences of particular lies. To tell a lie is wrong as such, just because it is a flouting of truth, and it is an offense primarily not against those particular others to whom this particular lie has been told, but against human rationality, everyone’s rationality, including the liar’s own rationality. By lying she or he has 314 The Tanner Lectures on Human Values [MACINTYRE] Truthfulness, Lies, and Moral Philosophers 315 failed not only to acknowledge truth as a good that is indispensable in rational relationships with others, but also to recognize that a failure to respect truth is a failure in respecting oneself as a rational being. This conception of the wrongness of lying was elaborated within a moral tradition whose central theses were in crucial respects at odds with those of the tradition that I described earlier. For where the exponents of that tradition, from Clement to John Stuart Mill, had agreed on the need to exempt certain types of lie from the general prohibition of lying, the adherents of the tradition of which Kant was a late and distinguished member agreed in insisting that the rule prohibiting lying was exceptionless. Instead of looking back to Plato, its protagonists look back to Aristotle’s condemnation of all lying as disgraceful and to his praise of the lover of truth who is truthful whether something further is at stake or not (Nicomachean Ethics IV, 1127b4-8). There are trenchant restatements of this standpoint by St. Augustine, by St. Thomas Aquinas, by the Catechism of the Council of Trent, by Pascal, and by Protestant theologians both before and after Kant. Augustine declared in the Contra M endacium (31C) that “it is said to God ‘Your law is truth.’ And for this reason what is contrary to truth cannot be just. But who doubts that every lie is contrary to truth? Therefore no lie can be just.” Aquinas argued that truth itself is a virtue, since to say what is true makes a good act and a virtue is that which makes its possessor good and renders its possessor’s action good (Summa Theologiae IIa-IIae, 109, 1). Of the vices opposed to the virtue of truth lying is the first (110, prologue). Aquinas captured a thought central to this tradition when he distinguished between the wrong done by intentionally asserting what is false and the wrong done by intentionally deceiving someone by that false assertion. Even without an intention to deceive, the intentional assertion of what is false is wrong (110, 1 resp. and 3 ad. 6). The offense is against truth. Some adherents of these two contrasting and generally rival traditions may in fact disagree about very little of moral substance. 316 The Tanner Lectures on Human Values [MACINTYRE] Truthfulness, Lies, and Moral Philosophers 317 For among some of those for whom lying is altogether prohibited, the definition of a lie is such as to exclude just those cases that some adherents of the other tradition treat as permissible or required lies. But it would be a mistake to conclude from these cases that the differences between the two traditions are unimportant, as Newman seems to have done. Those differences extend to three kinds of issue. First there is the question of how a lie is to be defined. Those for whom some types of lie are permissible or even required characteristically define a lie so that an intention to deceive is an essential defining property of a lie, and the wrongness of lies is the same as that of other acts of deception, while those for whom no lies are permissible characteristically define a lie in terms of an intention to assert what is false, sometimes, like Aquinas, denying that an intention to deceive is necessary for an assertion to be a lie. A second difference concerns the nature of the offense committed by a liar. For those for whom some types of lie are permissible or even required the wrong done by a lie is understood in terms of the harm inflicted upon those social relationships that need to be sustained by mutual truth and credibility. Because of the constitutive part played by such trust in every important human relationship, that harm is never held to be entirely negligible. But evidently there are occasions on which the utterance of a particular lie will prevent some harm greater than that which its telling will cause to the social fabric. By contrast, for those for whom no lie is permissible the wrong committed by making a false assertion is understood as a type of wrong that inescapably puts in question one’s standing as a rational person in relationship to other rational persons. A third set of issues concerns the kind of justificatory argument advanced within each tradition. Those who hold that some types of lie are permissible advance justifications that cite the effects of different types of lie, even when those who advance them are not consequentialists in general. Those who hold that all lies are forbidden advance justifications citing the nature of the act of lying. And at