¹ Mill, John Stuart, Utilitarianism, ed. Crisp, Roger (New York: Oxford University Press, 1998), IV, 3Google Scholar. Throughout this essay, passages from Utilitarianism will be identified using the chapter, paragraph, and line numbers from this edition.

² Kant, Immanuel, Grounding for the Metaphysics of Morals, trans. Ellington, James W. (Indianapolis, IN: Hackett, 1981), 36 (Ak. p. 429).Google Scholar

³ Mill, , Utilitarianism, IV, 10.5.Google Scholar

⁴ Mill distinguishes between the standard of conduct and “the criterion of morality,” treating the latter as determined by the former (ibid., IV, 9.5). In addition, Mill suggests, at least sometimes, that the morality of an action turns not directly on its effects, but on whether it accords with “the rules and precepts for human conduct, by the observance of which” the best results are secured (see, e.g., ibid., II, 10.10). Presumably, particular actions that are among the agent's best available options might not be in accord with the relevant rules and precepts, and particular actions that do satisfy the rules and precepts may be among the options that are less than the best.

⁵ More precisely, Mill sees happiness as a measure of the balance of pleasure and pain, and holds “that pleasure, and freedom from pain, are the only things desirable as ends” (because pleasure is good in itself and pain is bad in itself) and “that all desirable things (which are as numerous in the utilitarian as in any other scheme) are desirable either for the pleasure inherent in themselves, or as means to the promotion of pleasure and the prevention of pain.” Ibid., II, 2.10.

⁶ Among other things, it would leave completely unaddressed the suggestion that distributive considerations are (nonderivatively) relevant to what ought to be done.

⁷ Mill, John Stuart, BenthamGoogle Scholar, in Mill, , Collected Works of John Stuart Mill, ed. Robson, J. M. (Toronto: University of Toronto Press, 1969), 10:111.Google Scholar

⁸ Mill, , Utilitarianism, IV, 2.Google Scholar

⁹ Nonetheless, some have criticized Mill for implicitly assuming that people, either individually or collectively, do desire the general happiness. For instance, F. H. Bradley suggests that “[e]ither Mill meant to argue, ‘Because everybody desires his own pleasure, therefore everybody desires his own pleasure”; or “Because everybody desires his own pleasure, therefore everybody desires the pleasure of everybody else.’ Disciples may take their choice.” Of course, the first claim is trivial and of no help and the second is obviously fallacious. Fortunately, it is pretty clear that Mill, whatever he meant, did not mean to offer either of these arguments. See Bradley, F. H., Ethical Studies, 2d ed. (Oxford: Clarendon Press, 1927), 113–14 n. 3Google Scholar. Along the same lines, C. D. Broad sees Mill as committed to thinking that something being desired is one and the same with it being desirable, so that when Mill holds that the general happiness is a good that is desirable to the aggregate of all people, he is, according to Broad, committed to holding (falsely) that the aggregate desires something—that is, the general happiness. Fortunately, again, it is pretty clear that Mill neither equated being desired and being desirable nor meant to hold that the aggregate desires anything at all. See Broad, C. D., Five Types of Ethical Theory (Paterson, NJ: Littlefield, Adams, 1959).Google Scholar

¹⁰ Mill, , Utilitarianism, IV, 3.10.Google Scholar

¹¹ Ibid., IV, 4.4.

¹² Ibid., IV, 11.5.

¹³ Ibid., IV, 8.2. In this, Mill is largely following Aristotle, who identifies valuing something for its own sake with taking pleasure in it directly, rather than thanks to what it produces.

¹⁴ T. H. Green seems to be picking up on this point when he writes, “It is the realization of those objects in which we are mainly interested, not the succession of enjoyments which we shall experience in realizing them, that forms the definite content of our idea of true happiness, so far as it has such content at all” (emphasis added). I think, though, that Green goes astray in equating happiness with the realization of what is valued. See Green, T. H., Prolegomena to Ethics (Oxford: Clarendon Press, 1890), bk. III, chap, iv, sec. 228, p. 244.Google Scholar

¹⁵ As I have said, the source of my less than full-hearted support of Mill's proof is largely found in my unwillingness to embrace the psychological claims upon which the proof turns.

¹⁶ Mill, , Utilitarianism, IV, 10.Google Scholar

¹⁷ Moore, G. E., Principia Ethica (Cambridge: Cambridge University Press, 1903), 67.Google Scholar

¹⁸ Ibid., 66. Broad similarly claims that Mill “starts by assuming that ‘desirable’ means ‘desired by someone’,” in Broad, , Five Types of Ethical Theory, 183.Google Scholar

¹⁹ Specifically, on Mill's account, our concept of value and our evidence that the concept is satisfied are traceable to our capacity to desire, just as our concept of color and our evidence that the concept is satisfied are traceable to our capacity to see. Yet the concepts of value and of color are not concepts of the experiences that the capacities make possible.

²⁰ Elijah Millgram argues that Mill is relying on the optimistic prospect that people in the future will desire the general happiness, and using this to support the claim that the majority of people (past, present, and future) give a decided preference to the general happiness even if people of his day do not. The success of the proof, on this view, turns on whether what is being “proven” ends up ultimately being accepted, and does not itself constitute any sort of argument for accepting it. See Millgram, Elijah, “Mill's Proof of the Principle of Utility,” Ethics 110, no. 2 (2000): 282–310.CrossRefGoogle Scholar

²¹ Sidgwick, Henry, The Methods of Ethics, 7th ed. (London: Macmillan, 1907), 388Google Scholar. Broad presses the same point at Broad, , Five Types of Ethical Theory, 184.Google Scholar

²² Thus, to say that the general happiness is “a good to the aggregate of all persons,” on the ground that each person's happiness is a good to that person, trades on ignoring the difference between “all” being used collectively (i.e., for all of us, taken together, all of it, taken together, is good) and its being used distributively (i.e., for each of us, considered individually, all of it, taken together, is good). See Ryan, Alan, John Stuart Mill (New York: Pantheon Books, 1970), 200–201.Google Scholar

²³ These problems circulate around Mill's claim that “to desire anything, except in proportion as the idea of it is pleasant, is a physical and metaphysical impossibility” (Mill, , Utilitarianism, IV, 10Google Scholar). I think it is pretty clear that Mill does not intend this claim to be true by definition, but it is hard to see why he thinks the empirical evidence would support it.

²⁴ Some, hoping to defend Mill, have jumped on this disclaimer to excuse Mill for arguments that appear invalid. If no proof is to be had, they note, then Mill must not be offering the considerations he does as deductively valid grounds for his conclusion. Still, the argument he offers had better be reasonable, and so far the argument under discussion seems to violate even that requirement by turning on ambiguities, implausible definitions, confusions between an individual and the group to which she may belong, and so on. Complicating matters, Mill (at least in A System of Logic) advances a distinctive view of what counts as a genuine proof, according to which deductively valid “proofs” are not proofs at all, since anything to be found in the conclusion of such a “proof” must already be present in its premises, and so is assumed rather than proven. Rather, he supposes, genuine proofs need to establish conclusions that go beyond their premises—a proof establishes something new. Mill, John Stuart, A System of LogicGoogle Scholar, in Mill, , Collected Works, bk. II, chap. 1, sec. 2, 7:158–62Google Scholar. Whether Mill thinks the “ordinary acceptation of the term” “proof” includes deductively valid arguments is unclear. See Millgram, , “Mill's Proof.”Google Scholar

²⁵ Mill, , Utilitarianism, IV, 1.Google Scholar

²⁶ The interpretation I am advocating thus differs from those that suggest that Mill is appealing to the fact that people desire happiness as an end as establishing that it is possible to desire happiness as an end. According to these accounts, the appeal to what people desire plays out against the assumption that what we ought to desire is constrained by what we can desire, and is meant to show that happiness can be desired and is the only thing that can be desired as an end. On this interpretation, happiness emerges as the only candidate for being what ought to be desired. Mill's own observation that “[i]f the end which the utilitarian doctrine proposes to itself were not, in theory and practice, acknowledged to be an end, nothing could ever convince any person that it was so” is often taken as supporting this interpretation. This interpretation, though, leaves Mill with no positive argument for thinking happiness is good—it would simply have survived as the only candidate. Moreover, the argument has not established even this much by the end of the third paragraph, where Mill infers that happiness has been shown to be at least among the things that are desirable as an end. Thus, this interpretation has trouble both with finding Mill a positive argument for thinking happiness is valuable and with making sense of the structure of the text. For discussions of this interpretation, see Seth, James, “The Alleged Fallacies in Mill's Utilitarianism,” Philosophical Review 17, no. 5 (1908): 469–88CrossRefGoogle Scholar; Atkinson, R. F., “J. S. Mill's ‘Proof” Philosophy 32, no. 121 (1957): 158–67CrossRefGoogle Scholar; Kretzman, Norman, “Desire as Proof of Desirability,” Philosophical Quarterly 8, no. 32 (1958): 246–58CrossRefGoogle Scholar; and Clark, George, “Mill's Notorious Analogy,” Journal of Philosophy 56, no. 15 (1959): 652–56CrossRefGoogle Scholar. In contrast, on my reading, Mill's point here is not guided by ‘ought’ implies ‘can’, but instead, as I will suggest, by an appreciation of the fact that evaluative conclusions require evaluative premises (i.e., no ‘ought’ from only ‘is’s).

²⁷ Whether the experience is of something as red, or of it as good, the evidence we have is constituted by the content of our experience, not the fact that we had the experience.

²⁸ Mill, , Utilitarianism, IV, 10.10Google Scholar. Here Mill first equates desiring something with thinking it pleasant, and then he equates thinking of it as pleasant with thinking of it as desirable. (He actually writes “finding it pleasant” rather than “thinking it pleasant,” but goes on immediately to equate aversion to a thing with “thinking of it as painful,” not “finding it painful,” so in this passage he seems clearly to be using “finding it” and “thinking of it” interchangeably.)

²⁹ Within this broadly empiricist approach to concepts and evidence, there is an important question of how to understand the original experiences that give rise to concepts and provide evidence for their application. These experiences are often characterized not only by using the concepts they are supposed to explain, but in a way that seems to suppose the concepts are already available to the person having the experience. This obviously needs to be avoided on pain of circularity. In the case of desire and value, then, Mill needs to suppose that while desiring x is something like seeing x as desirable or good, the experience of desiring x does not require already possessing the concepts of desirability or value.

³⁰ Views similar to Mill's in this respect (though not others) have recently been defended in Stampe, Dennis, “The Authority of Desire,” Philosophical Review 96, no. 3 (1987): 335–81CrossRefGoogle Scholar; Quinn, Warren, “Putting Rationality in Its Place,” in Quinn, Morality and Action (Cambridge: Cambridge University Press, 1993)Google Scholar; and Scanlon, T. M., What We Owe to Each Other (Cambridge, MA: Harvard University Press, 1998).Google Scholar

³¹ While he believes that “[i]n what we call Desire there is… always included a positive stimulation to action,” he also thinks there is always included (to use his father's phrase) “the idea of something good to have.” See Mill, James, Analysis of the Phenomena of the Human Mind, 2d ed., ed. Mill, J. S. (London: Longman, Green, Reader, and Dyer, 1869), chap. 19Google Scholar; and John Stuart Mill's notes to that edition, ibid., 194–95. The notes can also be found in Mill, , Collected Works, 31:215.Google Scholar

³² In some cases, of course, Mill thinks that the correct standard will recommend desiring something other than the general happiness. At the same time, there is, on Mill's view, an important difference between justifying a belief—giving evidence of its truth—and showing that it would be good to adopt it.

³³ As is made clear by his discussion of it in Mill, , A System of Logic, bk. VI, chap. 12, sec. 6, 8:949–52.Google Scholar

³⁴ Note that even in the presence of contrary evidence, the content of one's desires or visual experiences will be on board as potential counterevidence.

³⁵ When he writes this in the third paragraph, Mill is just summarizing the case for thinking that happiness is good, not yet that it is also the only thing good, as an end. However, the stronger conclusion is supposed to rest simply on a redeployment of the same argument form. See Mill, , Utilitarianism, IV, 9.5.Google Scholar

³⁶ Ibid., II, 18.2.

³⁷ Mill, John Stuart to Jones, Henry, 06 13, 1868Google Scholar, reprinted in Mill, , Collected Works, 16:1414.Google Scholar

³⁸ Roger Crisp appeals to the passage just quoted from Mill's letter to Henry Jones to argue that Mill does not mean to show that the general happiness is an appropriate end for each individual. According to the interpretation I am offering, however, Mill is trying to give each of us an argument for accepting the view that the general happiness is valuable as an end. That the letter does not deny this becomes clear once attention is paid to the difference between showing that something is a good and showing that it is a good to someone. To do the latter is, on Mill's view, to show that it will contribute to that person's happiness. In a better world, Mill notes, people's concerns would be structured so that they would take pleasure in the well-being of others. Until that happens, though, we need to acknowledge that while the well-being of others is in fact good, it does not always contribute to the well-being of an agent in the way that would underwrite saying that the well-being of another is a good to that agent. Mill wants to offer an argument that shows that when something makes someone happy, we are all constrained to admit that something good has been produced, whether or not we take it to be good and whether or not we ourselves are benefited by it (which we will be if we care for the other person's welfare, but may not be otherwise). See Crisp, Roger, Mill On Utilitarianism (London: Routledge, 1997), 77–78.Google Scholar

³⁹ Given Mill's views, everything that contributes to the happiness of the person in question will thereby count as good (either in itself or as a means), but not everything that is good will contribute to the happiness of that person. At the same time, what a person values (what is good according to that person) may or may not, on Mill's account, actually be good; whether it is depends upon whether it contributes to someone's happiness—either to the happiness of that person or someone else's. Also, what might otherwise not be good may be made so by the person valuing it (though only if, in fact, it brings her happiness).

⁴⁰ At the very least, they benefit. Of course, if unhappiness is bad, then the costs of one person being unhappy will be compounded by others caring about her.

⁴¹ Kant, , Grounding, 36 (Ak. p. 429).Google Scholar

⁴² Needless to say, one might doubt that all people see themselves as Kant supposes they do, or doubt that those who do are relying on the grounds he supposes. Still, if he is right about these commitments, his argument kicks in directly.

⁴³ Needless to say, one might doubt that all people see their own happiness as Mill supposes they do, or doubt that those who do are relying on the grounds he supposes. Still, if he is right about these commitments, the argument kicks in directly.

⁴⁴ In this second case there may well be two desires in play, one depending on the other, with the first being, in effect, a matter of seeing the cake as good, and the second being a matter of seeing getting the cake as good because it is the getting of something good. The second desire, of course, does not always follow upon the first, for a number of reasons. For instance, it may be that what one sees as good about the cake is something one cannot get, or it may be that one could get it, but only at the expense of someone else's not getting the cake, where that other person's getting the cake would be better. Note also that the second desire might be present without the first, since all the second requires is that one accept (perhaps as the content of a belief) the value of what it is that one desires for oneself.

⁴⁵ It is also important here to keep in mind the contrast between my desiring my own pleasure, which I might get from the cake, and my desiring the pleasure (conceived of in nonproprietary terms) for myself. The former puts us back in the first case, where the relevant objects of desire come in proprietary packets, whereas the latter simply shifts from a case where it is the cake that is seen as valuable to one in which the pleasure is.

⁴⁶ That is, as much reason as any other if the only reason anyone has is found in her own happiness appearing to her to be good.

⁴⁷ If the rational ground we each have for thinking of our own rational nature as an end has this standing turn, in each case, on the rational nature being one's own, then we would each only be committed to holding that for each person, his or her rational nature is an end for him or her.

⁴⁸ Sidgwick, , The Methods of Ethics, 420.Google Scholar

⁴⁹ Moore, G. E., Principia Ethica, 98–99Google Scholar. I am not here recommending this as a sound argument, only as one that gives voice, as if they are obvious, to the sort of assumptions I am supposing Mill made without comment.

⁵⁰ That is not to say that they would be right. It would be a tricky and interesting challenge to try to settle whether people might ever have the sort of proprietary desire that Moore explicitly rejects as incoherent and Mill implicitly rejects as something we do not have.

⁵¹ Bradley, , Ethical Studies, 113.Google Scholar

⁵² Just how contingent the commitment is on Mill's view is a little unclear. He does, after all, claim that “to desire anything, except in proportion as the idea of it is pleasant, is a physical and metaphysical impossibility” (Mill, , Utilitarianism, IV, 10Google Scholar). However, his grounds for this strong claim are less than clear, to say the least.

⁵³ In offering the argument I am focusing on, Kant takes the claim concerning people's initial (and supposedly necessary) commitment as a postulate. It is defended in Kant, , Grounding, sec. 3, 49–61 (Ak. pp. 446–63).Google Scholar

⁵⁴ Sidgwick, , The Methods of Ethics, 382Google Scholar. This is just the assumption Broad takes on, arguing that someone might acknowledge the equal value of others' happiness, but hold that each person “has an obligation to produce good experiences and dispositions in himself, and no such direct obligation to produce them in… anyone else.” Broad, C. D., “Moore's Ethical Doctrines,” in Schilpp, Paul, ed., The Philosophy of G. E. Moore (La Salle, IL: Open Court, 1968), 45.Google Scholar

⁵⁵ As I understand Mill's position here, he is defending happiness as being the “sole end of human action, and the promotion of it the test by which to judge of all human conduct,” but seeing morality and rationality as each “parts included in the whole,” with neither being identical to the fundamental test of conduct, and each being different from the other. On his view, rationality is concerned specifically with what is good to the agent in a way that morality is not. Mill, , Utilitarianism, IV, 9.Google Scholar

⁵⁶ In at least some cases, the criticism will also involve the claim that she is failing to care about what is of value. In such a case, the criticism will run, she has reason to value what she does not value, and in failing to value it finds herself without reason to do what she should.

⁵⁷ Such arguments thus leave open, in an interesting way, what the implications of the defended standard might be when it comes to what people should value. For all such an argument shows, it might be that the standard defended turns around and requires of people that they value something different than what has been defended as valuable. Indeed, it is a familiar fact about some versions of utilitarianism that they might require rejecting utilitarianism on utilitarian grounds.