¹ The familiar idea of altruism is notoriously hard to pin down, and philosophical definitions can vary as widely as those of the contrasting term, egoism. I do not want to limit my general discussion by insisting on a specific definition, but roughly I mean by “altruistic acts” those done to benefit others and not motivated by self-interest. They are done “for the sake of others” from motives such as sympathy, respect, group loyalty, or moral duty. The idea of “self-interest,” unfortunately, is almost as slippery as the ideas of “egoism” and “altruism,” but when more than a common-sense understanding is needed, I favor the characterization suggested by Gregory S. Kavka in his account of “Narrow Egoism.” That is, acts motivated by self-interest have as their ultimate end “personal benefits,” as best identified by a list of examples. Kavka's list includes “pleasure, [avoidance of] pain, wealth, security, liberty, glory, possession of particular objects, fame, health, longevity, status, self-respect, self-development, self-assertion, reputation, honor, and affection.” Kavka, Gregory S., Hobbesian Moral and Political Theory (Princeton: Princeton University Press, 1986), p. 42.Google Scholar

Here I set aside, as nonaltruistic, acts done both for others and for oneself, even though for some purposes these might be called “altruistic” in another sense. Also, importantly, I do not assume, as some do, that acts done for the sake of others are done out of compassionate or sympathetic feelings towards others, for I want to count among altruistic acts those done to benefit others because one believes that helping others on such occasions is what one morally should do. In such cases, on the Kantian view, the aim or end one seeks is others' welfare and nothing further, but the motivating principle is to act as one believes morally right, and the accompanying feeling may be respect for moral principle rather than compassion. It should be noted further that altruistic acts, as understood here, need not be motivated by general benevolence (i.e., a concern for the welfare of human or sentient beings in general) rather than concern to help particular individuals.

² The view in question is a combination of psychological egoism and ethical egoism, as these terms are generally understood. Whether Thomas Hobbes actually was an unqualified egoist in these senses is a matter of controversy, which turns partly on the exact definitions of relevant senses of “egoism” and partly on the interpretation of familiar passages. It is clear that Hobbes acknowledged that we sometimes act from apparently altruistic motives, for example, “pity” and “compassion,” but his definitions can be read as efforts to reinterpret such terms in a way compatible with psychological egoism and its denial of genuine, i.e., actual, altruism. For finer distinctions and varying views on Hobbes, see, for example, the following: Broad, C. D., Five Types of Ethical Theory (London: Routledge & Kegan Paul, 1930), p. 54Google Scholar; Brandt, Richard B., Ethical Theory (Englewood Cliffs, N.J.: Prentice-Hall, 1959), ch. 14, especially pp. 370–71Google Scholar; Frankena, William K., Ethics, 2nd ed. (Englewood Cliffs, N.J.: Prentice-Hall, 1973), ch. 2, especially p. 15Google Scholar; Rachels, James, The Elements of Moral Philosophy (New York: Random House, 1986), chs. 5 and 6Google Scholar; Hampton, Jean, Hobbes and the Social Contract Tradition (Cambridge: Cambridge University Press, 1986), ch. 1, especially pp. 19–24Google Scholar; and Kavka, Gregory S., Hobbesian Moral and Political Theory, ch. 2, especially pp. 44–51.Google Scholar

³ As I intend these terms, we take a “practical, deliberative” point of view when we think seriously about reasons for and against policies or courses of action, and this perspective is also “conscientious” when our background assumption, and indeed our point, in deliberating is that we intend to do what is morally right. Conscientious people, of course, may have different views about what is morally right, but they have in common a commitment to act on their best judgment about this. Here I assume only that in taking up the conscientious perspective we seek what is right with the intent to do it, not that we are all moral saints, always perfectly free from weakness of will, negligence, and perversity.

⁴ The point of specifying that we are to address the questions in this essay as conscientious deliberating agents is methodological, not rhetorical. What one can take for granted, what is in doubt, and even what one is looking for can vary with the context of discussion of normative matters. Purposes and working assumptions tend to shift as one moves within contemporary philosophical literature from abstract philosophical arguments about “moral realism,” to practical debates on specific moral issues, to general explanatory accounts of moral belief and behavior from a third-person perspective. Ideally, we expect that reasonable conclusions in these different contexts will eventually cohere, but in the meantime it is only good procedure to keep in focus the background aims and assumptions of each particular discussion.

⁵ Kant, Immanuel's The Metaphysics of MoralsGoogle Scholar is available in translation in two volumes: The Metaphysical Elements of Justice, trans. Ladd, John (Indianapolis: Bobbs-Merrill, 1965)Google Scholar and The Metaphysical Principles of Virtue, trans. Ellington, James (Indianapolis: Bobbs-Merrill, 1964).Google Scholar Kant's position on the possibility of altruism is inseparable from his ideas of autonomy and practical freedom, discussed in many of his major works. For commentary and detailed references, see Allison, Henry E., Kant's Theory of Freedom (Cambridge: Cambridge University Press, 1990).CrossRefGoogle Scholar

⁶ Here I paraphrase what Kant expresses in his more specialized terminology and what I explain more fully in the last section of this essay. What I call “conscientious agents” are roughly, in Kant's terms, (imperfectly) rational agents that acknowledge that they are subject to duties, conceived as categorical imperatives. This implies a “will,” though not an invariably effective disposition, to conform to what one judges (or “knows”) to be morally obligatory. Kant held that all (even imperfectly) rational human beings have such a will; and so, though he implied that there are rational requirements that are not duties, he believed that “rational agents” are also “conscientious” (as I use the term here). In saying that (for Kant) conscientious agents are “committed to evaluative consistency,” I allude (rather imprecisely) to the standards expressed in the “universal-law” formula of the Categorical Imperative, which is discussed briefly in the last section of this essay. The constraint of defensibility “before others in morally appropriate deliberations” is drawn from the “kingdom of ends” formula, not discussed in this essay. I present a more thorough view of some of these matters in my collection of essays Dignity and Practical Reason in Kant's Moral Tlieory (Ithaca: Cornell University Press, 1992), especially chs. 1, 3, 7, 10, and 11Google Scholar, and in “A Kantian Perspective on Moral Rules,” Philosophical Perspectives, vol. 6 (1992).Google Scholar See also Paton, H. J., The Categorical Imperative (London: Hutchison, 1947)Google Scholar; and O'Neill, Onora, Constructions of Reason (Cambridge: Cambridge University Press, 1989).Google Scholar

⁷ Hobbes, Thomas, Leviathan, ed. MacPherson, C. B. (Baltimore: Penguin Books, 1951), part I, ch. 6, p. 126.Google Scholar This passage, as Jean Hampton has reminded me, can be interpreted in several ways. A reading that assimilated Hobbes's point to Hume's account of sympathy would allow that acts moved by compassion (as Hobbes defines it) could count as altruistic. See the references in note 2 above.

⁸ Hobbes, , Leviathan, part I, chs. 14 and 15.Google Scholar

⁹ This view is at least suggested by various philosophers who interpret moral judgments as expressing sentiments of impartial spectators, but my unqualified summary of the position no doubt oversimplifies their views. See, for example, Hume, David, Moral and Political Philosophy, ed. Aiken, Henry D. (New York: Hafner, 1948), especially pp. 175–84 and 249–61.Google Scholar See also selections from Adam Smith and Butler, Joseph in British Moralists, ed. Selby-Bigge, (Indianapolis: Bobbs-Merrill, 1964), pp. 255–336 and 181–254Google Scholar; and Firth, Roderick, “Ethics and the Ideal Observer,” Philosophy and Phenomenological Research, vol. 12 (1952), pp. 317–45.CrossRefGoogle Scholar

¹⁰ Kant maintained that the idea of agent-initiated causation (e.g., I intentionally moved the lever with my hand) is importantly different from the idea of event causation (e.g., the explosion knocked my hand against the lever and/or my hand's movement caused the lever to move). All events, Kant argued, have prior causes sufficient to produce them with necessity, but from a practical point of view, we can and must think of agents as initiating causal sequences without being determined to do so by prior causes. See his Critique of Pure Reason, trans. Smith, Norman Kemp (London: Macmillan, 1956), especially pp. 464–79Google Scholar; Groundwork of the Metaphysic of Morals, trans. Paton, H. J. (New York: Harper & Row, 1964), ch. 3, pp. 114–31Google Scholar; and Critique of Practical Reason, trans. Beck, Lewis White (New York: The Liberal Arts Press, 1956), especially pp. 74–92.Google Scholar

¹¹ Both the agent's (normative) reasons for choosing and explanatory causes may be referred to as “the reasons why” the agent acted, and this ambiguity can cause confusion. See Darwall, Stephen, Impartial Reason (Ithaca: Cornell University Press, 1983), pp. 28–29.Google Scholar

¹² See Kant, , Critique of Practical Reason, pp. 20–21.Google Scholar

¹³ Kant, , Groundwork, p. 83.Google Scholar

¹⁴ Ibid., pp. 85–86.

¹⁵ Kant, , Religion Within the Limits of Reason Alone, trans. Greene, Theodore M. and Hudson, Hoyt H. (New York: Harper & Row, 1960), pp. 40–49.Google Scholar

¹⁶ See Kant, , The Metaphysical Elements of Justice, pp. 99–106.Google Scholar

¹⁷ See Kant, , “Perpetual Peace,” trans. H. B. Nisbet, in Kant's Political Writings, ed. Reiss, Hans (Cambridge: Cambridge University Press, 1970), p. 112.Google Scholar

¹⁸ Kant, , Critique of Pure Reason, pp. 218–33, 409–15, and 443–49.Google Scholar

¹⁹ Kant, , Groundwork, pp. 118–29Google Scholar, and Critique of Practical Reason, pp. 92–110.Google Scholar

²⁰ This is a conceptual point. That is, if a person really knows and is immediately and fully aware that it is impossible for him to do something, then it does not make sense to describe the person as “deliberating.” I might stand before a thousand-pound weight mouthing the words “Shall I—or shall I not—lift it?”—but anyone who knows that I know that I cannot lift it will also know that I am joking.

²¹ In most actual cases, perhaps, those who face torture do not know for certain that they cannot withstand the degree of torture that they will receive. Even a glimmer of hope that one might succeed, despite weighty but not decisive evidence, might be enough to enable one to “decide” to succeed. One can also self-deceptively “half-believe,” contrary to solid evidence and one's own “best judgment,” and sometimes, no doubt, doing so has good results. But the main point remains that deliberating about what to do presupposes seeing oneself as having options, and empirical evidence can sometimes show quite decisively that ways we might wish to act are not in fact options.

²² The special circumstance that might make it reasonable to try to do what seems almost certainly impossible (thereby treating it as an option) might be, for example, the fact that nothing else can save one's life or one's friends.

²³ James, William, The Will to Believe and Other Essays (New York: Dover, 1956).Google Scholar

²⁴ See especially “The Third Antinomy,” in Kant, 's Critique of Pure Reason, pp. 409–17, 422–30, and 439–59.Google Scholar Regarding Kant's views on causation, see note 10 above.

²⁵ See Kant, , The Metaphysical Principles of Virtue, pp. 45–55, 60–61, and 112–27.Google Scholar Kant claims that we do have a duty of beneficence of the sort he describes, but at this point in the essay we should consider it merely a supposed duty, because we have not yet considered any reasons for it and are still considering the question whether it is possible for us to do what it prescribes. If we assume now that beneficence is really a duty, then we could simply infer that we can conform to it because, in the relevant senses, “ought” implies “can.” But obviously that would be too facile a reply to those egoists who (relying on the idea that “cannot” implies “not ought”) would argue the reverse, i.e., that we have no duty of beneficence because acting altruistically (they say) is impossible.

²⁶ These distinctions are discussed at length in Gregor, Mary, Laws of Freedom (Oxford: Blackwell, 1963)Google Scholar; and also in my Dignity and Practical Reason in Kant's Moral Theory, ch. 8.

²⁷ See Bentham, Jeremy, An Introduction to the Principles of Morals and Legislation, especially chs. 1 and 4.Google Scholar The work is available in many editions, including A Fragment on Government and An Introduction to the Principles of Morals and Legislation, ed. Harrison, Wilfrid (Oxford: Basil Blackwell, 1960).Google Scholar

²⁸ See Ross, W. D., The Right and the Good (Oxford: Clarendon Press, 1930), pp. 16–47.Google Scholar A prima facie duty, according to Ross, is a feature of an act that would make the act an actual duty (i.e., a duty, all things considered) if there were no conflicting moral considerations. Examples of such features include: that an act fulfills a promise, that it returns a favor, that it promotes someone's happiness, or that it makes reparation for past wrongs.

²⁹ Kant's view, as I understand it, is that everyone is strictly required to maintain as an effective guiding principle (or maxim): “Promote the happiness of others, counting their (permissible) ends as among your own ends.” like Ross's prima facie duty of beneficence, this principle does not specify exactly how, when, or how much one must do for others. Unlike Ross, however, Kant held that the duty of beneficence cannot be fully satisfied by helping others unless one makes it a principle to do so. (Kant also implies, even beyond this, that one must maintain the principle for moral reasons, not merely for self-interested reasons; but this introduces complications best left aside here.) On Ross's view, adopting the maxim as one's personal action-guiding standard is neither a prima facie duty nor an actual duty. For Ross the principle “It is a prima facie duty to promote the happiness of others” is supposed to be a self-evident truth that once recognized tends to motivate, but one can satisfy this principle by doing what makes others happy even if one never makes it a principle to do so.

³⁰ See Section III (A) above.

³¹ Unrecognized feelings, of course, cannot enter the deliberation. That is, though sometimes they may unfortunately distort our judgment in ways we are helpless to prevent, the fact that they may be present is irrelevant to the question under deliberation, “What ought I to do?” Learning empirically that unrecognized feelings can skew our judgments gives us reason to be wary and to seek greater self-awareness; and in extreme cases, e.g., where afterthought reveals a pattern, suspecting that my judgment is likely to be skewed may be a reason not to try to deliberate at all. But once an issue is up for deliberation, I must treat facts about how I feel, like all other facts, as data for reasonable decision making, not as controlling forces.

³² This Kantian (and, I believe, common-sense) view of deliberation contrasts significantly with that of Hobbes. See Leviathan, part I, ch. 6, pp. 127–28.Google Scholar

³³ This distinction, admittedly controversial, is discussed at length in “Weakness of Will and Character,” in my collection of essays Autonomy and Self-Respect (Cambridge: Cambridge University Press, 1991), p. 9.Google Scholar Roughly, the idea is that what we call “weakness of will” is often not a disability beyond the agent's control (lack of willpower) but rather the agent's pattern, for which he or she may be responsible, of making half-hearted efforts, breaking and fudging resolutions, and not following through on projects and commitments (willing weakly).

³⁴ The ideas expressed here, as well as some in the preceding three sections, are part of what Kant seems to have meant in claiming that it is necessary in conceiving ourselves as moral agents to take ourselves to have autonomy of the will. For further discussion, see my Dignity and Practical Reason in Kant's Moral Theory, ch. 5; my Autonomy and Self-Respect, ch. 12; and Henry Allison, Kant's Theory of Freedom, ch. 5.

³⁵ Kant, , Groundwork, pp. 90 and 98.Google Scholar

³⁶ Ibid.

³⁷ Kant presents this formula in his Groundwork, pp. 100–102 and 105–7Google Scholar, and my efforts to reconstruct it are cited in note 6 above.

³⁸ Some other versions are the following:

Every morally practical relation of men to one another is a relation of them in the representation of pure reason, i.e., of free actions according to maxims which are suitable for universal legislation and which thus cannot be selfish. I want every other person to have benevolence for me; I should therefore be benevolent to every other person. (Kant, , The Metaphysical Principles of Virtue, p. 115).Google Scholar

It is a duty of every man to be beneficent, i.e., to be helpful to men in need according to one's means, for the sake of their happiness and without hoping for anything thereby.

For every man who finds himself in need wishes that he might be helped by others. But if he should make known his maxim of not wanting to give assistance in turn to others in their need—if he should make such a maxim a universal permissive law—then everyone would likewise refuse him assistance when he was in need, or at least everyone would be entitled to refuse. Thus the selfish maxim conflicts with itself when it is made a universal law, i.e., it is contrary to duty.

Consequently, the altruistic maxim of beneficence towards those in need is a universal duty of men; this is so because they are to be regarded as fellow men, i.e., as needy rational beings, united by nature in one dwelling place for mutual aid. (Ibid., p. 117)

³⁹ Kant, , The Metaphysical Principles of Virtue, p. 52.Google Scholar

⁴⁰ For some of the many interpretations, see the following: Korsgaard, Christine, “Kant's Formula of Universal Law,” Pacific Philosophical Quarterly, vol. 66 (1985), pp. 24–47CrossRefGoogle Scholar; O'Neill, Onora, Constructions of Reason (Cambridge: Cambridge University Press, 1989), pp. 81–104Google Scholar; Nell, Onora (O'Neill), Acting on Principle (New York: Columbia University Press, 1975)Google Scholar; Potter, Nelson T. and Timmons, Mark, eds., Morality and Universality: Essays on Ethical Universalizability (Dordrecht: Reidel, 1985)CrossRefGoogle Scholar; Singer, M. G., Generalization in Ethics (New York: Alfred A. Knopf, 1961), pp. 217–99Google Scholar; and Herman, Barbara, “Mutual Aid and Respect for Persons,” Ethics, vol. 94 (1984), pp. 577–602.CrossRefGoogle Scholar

⁴¹ Kant, , Groundwork, ch. 2, pp. 90–91.Google Scholar

⁴² Dennett's phrase is “intuition pumps.” See Dennett, Daniel, Elbow Room (Cambridge: MIT Press, 1984), p. 12.Google Scholar