¹ Ideas of Kantian constructivism can be found in the work of Immanuel Kant (on some interpretations), John Rawls, Onora O'Neill, and others who comment on their work. See, for example, Kant, Immanuel, Groundwork of the Metaphysics of Morals, ed. and trans. Gregor, Mary (Cambridge: Cambridge University Press, 1998)CrossRefGoogle Scholar; Rawls, John, “Kantian Constructivism in Moral Theory,” Journal of Philosophy 77, no. 9 (1980): 515–72Google Scholar; Rawls, John, A Theory of justice (Cambridge, MA: Harvard University Press, 1971)Google Scholar; Rawls, John, Political Liberalism (New York: Columbia University Press, 1993)Google Scholar; O'Neill, Onora, Constructions of Reason (Cambridge: Cambridge University Press, 1989), esp. chap. 11Google Scholar; O'Neill, Onora, Towards justice and Virtue: A Constructive Account of Practical Reasoning (New York: Cambridge University Press, 1996)CrossRefGoogle Scholar; Barry, Brian, Theories of Justice (Berkeley: University of California Press, 1989), vol. 1 of A Treatise on Social justiceGoogle Scholar; Brink, David O., Moral Realism and the Foundations of Ethics (Cambridge: Cambridge University Press, 1989)CrossRefGoogle Scholar; Hill, Thomas E. Jr., Dignity and Practical Reason in Kant's Moral Theory (Ithaca, NY: Cornell University Press, 1992), chap. 11Google Scholar; and Hill, Thomas E. Jr., Respect, Pluralism, and Justice: Kantian Perspectives (Oxford: Oxford University Press, 2000), chaps. 2, 4, 8.CrossRefGoogle Scholar

² Onora O'Neill's views might seem to be an exception here because she criticizes Rawls for arguing from a hypothetical idealized choice situation (see O'Neill, , Constructions of Reason, 207–13Google Scholar; and O'Neill, , Towards Justice and Virtue, 44–48Google Scholar), but her arguments from the thought that everyone “cannot share” certain principles presuppose at least some modest rationality conditions (as well as other background conditions) in her procedure of construction. Thus, my broad characterization of Kantian constructivism includes O'Neill's position, but nothing substantive in my discussion depends on fine points about how we use this broad term of classification.

³ In this essay, I will only describe aspects of Kant's moral theory when they are immediately relevant to my questions about the justificatory roles of actual, possible, and hypothetical consent. Fuller discussions of Kant's moral theory are contained in Hill, , Dignity and Practical Reason in Kant's Moral Theory, esp. chap. 11Google Scholar, and Hill, , Respect, Pluralism, and Justice, esp. chaps. 1, 2, 4, 8.Google Scholar

⁴ Although for some purposes it might be important to distinguish “consent”; “agreement,” and “will,” I use them more or less interchangeably here. It may be more natural to speak of the “consent” of actual persons to particular proposals and the “agreement” of ideal rational agents on general principles, but using the terms more flexibly helps to highlight the comparisons I want to make.

⁵ For Kant's references to an original contract, see Reiss, Hans, ed., Kant: Political Writings, 2d ed. (Cambridge: Cambridge University Press, 1991), 77, 79, 80–83, 85, 91Google Scholar (from Kant, Immanuel, “On the Common Saying: ‘This May Be True in Theory, but It Does Not Apply in Practice’” [hereinafter “Theory and Practice”]Google Scholar), 94, 99–100 (from Kant, Immanuel, “Perpetual Peace”), 140, 158, 162–64Google Scholar (from Kant, Immanuel, The Metaphysics of Morals‘).Google Scholar

⁶ The universal law formula is one of several ways that Kant expresses the Categorical Imperative. Kant, , Groundwork, 15 [4:402], 31 [4:421]Google Scholar. Numbers in brackets refer to volume and page numbers in the standard Prussian Academy edition of Kant's works. A fuller discussion of the universal law formula follows in Section III.

⁷ Kant sets forth the kingdom of ends formula at Kant, , Groundwork, 41–44 [4:433–37]Google Scholar. A fuller discussion of this formula follows in Section III.

⁸ The kingdom of ends formula invites us to ask what laws rational agents would will. As I reconstruct the idea, it is assumed that the hypothetical agents are fully rational, and the question is what they would necessarily will qua rational (and properly informed) agents regarding all the possible general permissions and prohibitions that we might want to assess. Thus, they would (necessarily) will a prohibitive law if and only if it is rationally necessary for them to will it in the context in question. See Hill, , Dignity and Practical Reason, chaps. 3, 11.Google Scholar

⁹ For an example of this sort of reconstruction, see O'Neill, , Constructions of Reason, 206–18.Google Scholar

¹⁰ Kant, , Groundwork, 32 [4:423].Google Scholar

¹¹ My remarks here are not meant as a literal interpretation of Kant's arguments. To provide such an interpretation, much more preparatory work would be needed, including identifying the relevant maxim. (A maxim is a subjective principle, or personal policy statement, that summarizes, in a way relevant for moral assessment, one's understanding of what one intends to do, one's purpose, and one's underlying reasons.) My suggestion is that, Kant aside, practices of various kinds typically determine when consent is (and is not) needed, and so justifications of the form “He consented and so it is permissible” require moral evaluation of the practice to which one is implicitly appealing. When the issue is what practices are justifiable in the way they demand consent (or not), Kant's formulas of the Categorical Imperative move us to another level. The universal law formula, for example, asks whether we can will our maxim as universal law, but “the maxim” needs to be described in a way that appropriately reflects our practice-laden understandings.

¹² Kant, , Groundwork, 32 [4:422], 38 [4:429–30].Google Scholar

¹³ Kant expresses the humanity formula of the Categorical Imperative at ibid., 38 [4:429]: “So act that you use humanity, whether in your own person or in the person of any other, always at the same time as an end [in itself], never merely as a means.” He then, almost immediately, applies this principle to the example of someone who tries to borrow money without intending to repay. Arguably there is more to the argument from the humanity formula, at least implicitly, than is captured in the idea that we must treat persons as those who can share (or “contain in themselves”) the end of the action, but the further ideas are not important for present purposes.

¹⁴ A complication I ignore here: Suppose the would-be borrower pretends that he intends to repay and the would-be lender does not believe him but pretends that he does. Has a loan been made? I suppose so, for the lender could complain afterward when the borrower does not repay, and the borrower could not defend his not paying, when he learns that the lender did not believe him, by saying, “I do not owe you anything because you never believed that I would give the money back.” Therefore, the “understanding” the lender needs, perhaps, for a loan to exist is not strictly that the borrower intends to return the money, but just that the borrower intends (in borrowing) to cause the lender to believe that he intends to return the money. The lender in Kant's example, we may suppose, has this understanding. In fact, it is natural to suppose that the lender will transfer the funds only because he thinks that the borrower intends to give them back, and with this understanding the lender, in making what he understands is a loan, cannot share the borrower's end, that is, that the borrower get money cost-free (i.e., without repaying it).

¹⁵ There are other ways to interpret Kant's humanity formula, but under any plausible interpretation, it places limits on what actual consent can justify and helps to explain why actual consent, though important, is not always decisive.

¹⁶ The universal law of nature formula is expressed at Kant, , Groundwork, 31 [4:421]Google Scholar. Scholars differ on whether the difference between the universal law formula and the universal law of nature formula is significant, but for my present purposes it is not. Because of their structural similarity, the two formulas are often referred to collectively as “the universal law formulas.”

¹⁷ That is, the universal law formulas do not themselves specify the underlying standards that determine what we can will reasonably and what we cannot. There are many different ideas about what these specific standards count as irrational willing: for example, willing what proves to be logically impossible, having an incoherent set of intentions, willing that everyone adopt one's maxim even though this would defeat one's initial purpose in adopting it, willing contrary to the rationally necessary value of humanity in a person, and so on.

¹⁸ Consider a policy under which individuals drink a certain kind of wine on their birthdays. One can imagine a case in which, because of scarcity, it would be impossible for everyone to follow this policy. If in fact only a few people want to act on the policy, however, the scarcity of the wine should not be taken as any reason to condemn the policy for those few who want to adopt and act on it. The general point here has been noted often. For example, Kurt Baier presents a “universalizability” requirement analogous to Kant's (though different), and he qualifies his principle “doing X should be forbidden by the morality of the group if it would be harmful for everyone to do X” by adding, along with other stipulations, “provided doing X is an indulgence and not a sacrifice.” Baier, Kurt, The Moral Point of View (Ithaca, NY: Cornell University Press, 1958), 211Google Scholar. If we interpret the universal law formula to say that maxims are wrong to act on unless we can will them as permissible for everyone to act on, then, in cases where only a few want to act on a maxim, the maxim could turn out to meet the formula's test even though it is not possible in fact for everyone on Earth to act on it. See Pogge, Thomas, “The Categorical Imperative,” in Guyer, Paul, ed., Kant's “Groundwork of the Metaphysics of Morals”: Critical Essays (Lanham, MD: Rowman and Littlefield, 1998), 189–213Google Scholar. Whether or not maxims can be as specific as the policy in my example above remains a controversy.

¹⁹ For more on the Hypothetical Imperative, see Hill, , Dignity and Practical Reason, chaps. 1, 7Google Scholar. Other principles of rational choice that might well be taken for granted are, for example, those that Rawls calls “counting principles”; see Rawls, , A Theory of Justice, 411–16.Google Scholar

²⁰ Some may argue that these later formulas cannot add significantly to the universal law formulas on the ground that the later formulas are derivative, but this is debatable. In any case, a reconstructed Kantian moral perspective can make use of ideas of autonomy and humanity as an end in itself in attempts to apply the universal law formulas. Doing so may help to deflect some familiar arguments against the universal law formulas, though problems with those formulas will remain.

²¹ If, as it seems, in using Kant's decision-guiding procedures we are supposed not to rely on further, substantial intuitive assumptions about what we have “reasons” to do and prefer, then this is a way in which Kant's procedures for justifying our decisions to others differs significantly from the procedures proposed by Scanlon, T. M. in his recent book, What We Owe to Each Other (Cambridge, MA: Harvard University Press, 1998).Google Scholar

²² In general, it does not follow from “We could not rationally will X” that “We would rationally will not-X,” for in some contexts it is possible that we neither rationally will X nor rationally will not-X. My point, however, does not depend on this false inference. Kant's universal law formula and kingdom of ends formula provide the context here. The kingdom of ends is a highly idealized model. We assume that its legislators are perfectly rational, appropriately informed, and have a will in favor of or against all the possible permissions and prohibitions that we might put to them. The analogy with a divine will is obvious. In calling the formula of universal law a Categorical Imperative, Kant claims that fully rational persons, as such, will that they not act according to a maxim if it is impossible to will that maxim as a universal law. Therefore, if acting as described in a given maxim cannot be (rationally) willed as universal law, then rational legislators in the kingdom of ends must will that they, and anyone relevantly like them, not act in that way. This is just what “legislating” in this context amounts to, for the ideal kingdom of ends is not a legal system in which public offenses are defined and sanctions are imposed. Without any sanctions, necessary rational willing of the Kantian “legislators” against an act is supposed to make not acting that way imperative for us as imperfectly rational beings.

²³ Such “objects of intuition” would be, for example, nonnarural intrinsic values as conceived of by G. E. Moore, or Platonic Forms as they are often interpreted. See Moore, G. E., Principia Ethica (Cambridge: Cambridge University Press, 1903), chap. 1Google Scholar. Julia Annas's paper in this volume criticizes the common view that Platonic Forms are intuited independent objects or properties like Moore's intrinsic values.

²⁴ Hill, , Dignity and Practical Reason, 58–66, 243–50Google Scholar; Hill, , Respect, Pluralism, and Justice, 33–56, 220–30.Google Scholar

²⁵ Hill, , Dignity and Practical Reason, 65–66Google Scholar; Hill, , Respect, Pluralism, and Justice, 36, 51–55.Google Scholar

²⁶ Kant, , Groundwork, 44 [4:437]Google Scholar. It is the humanity formula that Kant appeals to most frequently in later moral arguments, especially in the second part of The Metaphysics of Morals. See Kant, Immanuel, The Metaphysics of Morals, trans. Gregor, Mary (Cambridge: Cambridge University Press, 1996), 173–218 [6:418–74].CrossRefGoogle Scholar

²⁷ Hill, Thomas E. Jr., “The Kingdom of Ends,”Google Scholar in Hill, , Dignity and Practical Reason, 58–66.Google Scholar

²⁸ Kant, , Groundwork, 44 [4:436], 45–46 [4:438–39].Google Scholar

²⁹ See Kant, , The Metaphysics of Morals, 20–22 [4:218–21].Google Scholar

³⁰ Kant refers to a possible “sovereign” of a kingdom of ends that is also head of the realm of nature, implying a power to harmonize the two, presumably in a way that allows the natural end (happiness) of the virtuous to be realized. Kant, , Groundwork, 41 [4:433], 46 [4:439]Google Scholar. This power, so used, would give us additional motive to follow the laws of the kingdom, but is not necessary for its authority (or even for adequate motivation).

³¹ Ibid., 45 [4:438].

³² Ibid., 42 [4:434], 45 [4:434].

³³ Ibid., 43–44 [4:436].

³⁴ See Reiss, , ed., Kant: Political Writings, 79Google Scholar (from Kant, , “Theory and Practice”Google Scholar).

³⁵ See, for example, ibid., 77, 79, 80, 83, 91 (from Kant, , “Theory and Practice”), 99, 100Google Scholar (from Kant, , “Perpetual Peace”), 143, 158, 162, 163, 164Google Scholar (from Kant, , The Metaphysics of MoralsGoogle Scholar).

³⁶ A Hobbesian state would not have separation of powers of the sort that Kant's ideal constitution would have, and there would be no grounds of justice on which enlightened critics could criticize the legislation of the sovereign. States that meet Hobbes's stipulations could differ in various other ways (e.g., they need not be a hereditary monarchy), and so my reference to “a particular kind of constitution” is relative.

³⁷ See Rousseau, Jean-Jacques, Discourse on the Origin of InequalityGoogle Scholar, in Rousseau, , The First and Second Discourses together with the Replies to Critics, ed. and trans. Gourevitch, Victor (New York: Harper and Row, 1986).Google Scholar

³⁸ Note that I include here two aims that I distinguish later. I leave aside whether there are other uses, though it seems that there are. In particular, Kant seems to appeal to the idea of an original contract, along with argument that it would be a duty in a state of nature to enter a civil order, to support a claim that we are morally, as well as legally, bound to obey the ruler of our state. What I have in mind here is the claim that under any legal constitution, no matter how badly designed and executed, we must see the ruler as the representative of the united will of the people, for the alternatives are a state of nature or a lack of final legal authority. The language of “united will” in both Rousseau and Kant strongly suggests an attempt at moral justification. This raises many problems, and so for present purposes I am limiting my discussion to the use of the idea of a possible original contract for arguments that purport to say what cannot (for conceptual or moral reasons) be in a constitution.

³⁹ The evidence for this is complex and ambiguous, involving a significant controversy about the interpretation of the Rechtslehre that Pogge, Thomas describes in his paper, “Is Kant's Rechtslehre Comprehensive?” Southern Journal of Philosophy 36, supplement (1997): 161–87CrossRefGoogle Scholar. Here I will only attempt to characterize some possible alternative readings in a general way, to show their relevance to the discussion of types of consent, without undertaking the detailed review of passages needed for a definitive account. Among the problems here is that Kant uses the idea of an original contract in several different works, and it is doubtful that we should assume that remarks in one context automatically carry over to the other contexts.

⁴⁰ Note that even though this is not a moral question, it has a proper (though limited) place in Kant's overall project to lay out the basic moral limits and requirements on political and legal institutions and conduct. This is so at least insofar as Kant has moral arguments that legal systems ought to be maintained and respected. Thus, even if, as some suggest, the doctrine of law in The Metaphysics of Morals is a module setting out the necessary conditions of a juridical condition independent of the moral principles in Kant's moral theory, it still has an appropriate place in The Metaphysics of Morals when this is viewed as a work the primary purpose of which is to lay out moral conditions on law, political institutions, and personal choices.

⁴¹ The qualification regarding appropriate means is needed because Kant believes that the means by which different people may work for constitutional reform are strictly limited. Philosophers can use public reason to criticize a constitution, but active resistance and revolution are forbidden. (We employ “public reason” when, as citizens, we participate in reasonable critical assessment of governmental laws and policies through newspapers, books, public speeches, etc., as opposed to what we may say privately or as representatives of special nonpublic institutions, such as a church or a club.) Rulers under nonrepublican constitutions may (and should) gradually work for reform, but even they are restricted in what they may do. The reason for the reference to “hoping” is that Kant's ideal constitution serves not only as a practical action guide, but also as a point of reference when we look hopefully (as we should) for progress in history.

⁴² Reiss, , ed., Kant: Political Writings, 94Google Scholar (from Kant, , “Perpetual Peace”).Google Scholar

⁴³ See ibid., 80–83 (from Kant, , “Theory and Practice”), 127Google Scholar (from Kant, , “Perpetual Peace”), 162Google Scholar (from Kant, , The Metaphysics of Morals).Google Scholar

⁴⁴ Ibid., 83 (from Kant, , “Theory and Practice”).Google Scholar

⁴⁵ Ibid., 79 (from Kant, , “Theory and Practice”), 99Google Scholar (from Kant, , “Perpetual Peace”), 153Google Scholar (from Kant, , The Metaphysics of Morals).Google Scholar

⁴⁶ Ibid., 58 (from Kant, Immanuel, “An Answer to the Question: ‘What is Enlightenment?‘”).Google Scholar

⁴⁷ Ibid., 163 (from Kant, , The Metaphysics of Morals).Google Scholar

⁴⁸ Kant, , The Metaphysics of Morals, 53.Google Scholar

⁴⁹ Reiss, , ed., Kant: Political Writings, 73, 79–81Google Scholar (from Kant, , “Theory and Practice”), 99Google Scholar (from Kant, , “Perpetual Peace”).Google Scholar

⁵⁰ Kant, , The Metaphysics of Morals, 108 [4:335]Google Scholar. Kant refers to Beccaria's influential work, Bonesana, Cesare, di Beccaria, Marchese, On Crimes and Punishments (New York: Bobbs-Merrill, 1963).Google Scholar

⁵¹ Kant, , The Metaphysics of Morals, 92–93.Google Scholar

⁵² Reiss, , ed., Kant: Political Writings, 164Google Scholar (from Kant, , The Metaphysics of Morals)Google Scholar; see also ibid., 79 (from Kant, , “Theory and Practice”).Google Scholar

⁵³ Ibid., 79 (from Kant, , “Theory and Practice”).Google Scholar

⁵⁴ It can be questioned whether Kant means to include women in this category, because he assumes women are merely “passive citizens” without sufficient independence to be allowed the vote. If, as it should be, women were meant to be included as parties to the hypothetical original contract, then women could be treated as “passive citizens” only if enlightened women and men alike would accept a constitution with this provision when they take up an appropriate genderless point of view. Assigning women second-class citizenship would surely fail this test, despite what Kant himself apparently thinks.

⁵⁵ Presumably, at least for some uses of the idea of an original contract, this does not mean that tests of the acceptability of provisions completely abstract from empirical facts about human nature and the human condition in general; rather, it suggests that such tests only abstract from the specific preferences and temperaments that vary from person to person, culture to culture. Some more strictly a priori arguments about the acceptability of provisions may proceed just through analysis of the idea of law, but there is no way that Kant could reasonably suppose that he could spin out all the conclusions listed above—that is, (1)–(7)—without presupposing general facts about human nature.

⁵⁶ Reiss, , ed., Kant: Political Writings, 80–81Google Scholar (from Kant, , “Theory and Practice”).Google Scholar

⁵⁷ Ibid., 81 (from Kant, , “Theory and Practice”).Google Scholar

⁵⁸ I add the qualification because our common practice is to use “just” as a more general term of moral assessment than Kant did. Many traditional moral philosophers did use our broader sense; Hobbes and others narrowed the term's use such that its application was limited to those realms of activity that can be enforced by secular authorities.

⁵⁹ I add “unqualified” because Kant holds that there could be, and we must assume that there is, a united will in an original contract on any constitution that establishes the existing juridical order, no matter how far from ideal that constitution might be. What the higher standards prevent is rational agreement on such a constitution as a permanent arrangement. With no other viable options in certain periods of history, we can and should endorse whatever de facto government gives us rule of law, but we can form a united will on imperfect constitutions only with the understanding that these will be gradually, continu ally, and legally reformed to resemble a republican constitution. This, I take it, is the spirit of the original contract. See Reiss, , ed., Kant: Political Writings, 163Google Scholar (from Kant, , The Metaphysics of Morals).Google Scholar

⁶⁰ Ibid., 79 (from Kant, , “Theory and Practice”).Google Scholar

⁶¹ It must be admitted that any pattern of argument used to justify political institutions, whether it appeals to hypothetical agreement or not, can be abused; one must also concede that the conditions stipulated as appropriate idealizations that must be realized for a hypothetical agreement to count can always be challenged. Furthermore, I am not claiming that Kant's own use of possible and hypothetical agreement, for example, in appeals to the idea of an original contract, are altogether satisfactory. My more modest aim is to argue that the appeal to hypothetical agreement in itself is not to be dismissed on the several grounds reviewed in this section. The devil, and maybe the angels, are in the details.

⁶² See Brandt, Richard, A Theory of the Good and the Right (Oxford: Clarendon Press, 1979), 113–29Google Scholar. Cognitive psychotherapy is a “process for confronting desires with relevant information, by repeatedly representing it, in an ideally vivid way, and at an appropriate time.” Ibid., 113.

⁶³ G. E. Moore famously objected to identifying normative concepts with descriptive ones. Such identification is prominent among the errors he called “the naturalistic fallacy.” See Moore, , Principia Ethica, chap. 1Google Scholar. Moore held that “intrinsic value” is a nonnatural, nonrelational, unanalyzable property that we can know by intuition. This identifies intrinsic value as a real metaphysical property that certain states of affairs have, and so it might seem at first to reduce values to descriptive facts. However, unlike typical metaphysical accounts of goodness, Moore's theory leaves no way to discover or even make sense of the property in question except by using the terms “good,” “valuable for its own sake,” and so on. What he calls “intuition” is, for all practical purposes, just evaluation without appeal to argument. Hence, despite initial appearances, his seemingly “metaphysical” account of intrinsic value does not reduce it to descriptive (nonevaluative) facts. In this one respect, I suggest, Moore's view is like Kant's, though Kant and Moore differ very substantially in other ways.

⁶⁴ See, for example, O'Neill, , Towards Justice and Virtue, 44–48.Google Scholar

⁶⁵ Practically minded critics would naturally focus on the specific reasons offered to show that such ideal deliberators would adopt the programs in question, but philosophers would probably also question whether it serves any useful purpose to characterize ideal deliberators first rather than turning immediately to moral arguments for the programs. This latter concern seems a natural one to raise about the simple argument in my example because its account of ideal deliberators is so thin, but whether it is an important worry about more subtle and complex arguments of the same structure is less obvious. My discussion of the fourth objection to hypothetical-agreement theories returns to this issue briefly.

⁶⁶ See Hill, , Respect, Pluralism, and Justice, chap. 2.Google Scholar

⁶⁷ Kantian arguments of this sort are analogous to arguments at what Rawls calls “the constitutional stage” of applying the Rawlsian basic principles of justice. See Rawls, , A Theory of Justice, 95–101.Google Scholar

⁶⁸ Andrews Reath, Shelly Kagan, and Philip Pettit convinced me of the need to respond to this apparent dilemma.

⁶⁹ The dilemma posed here is analogous to the “Euthyphro problem,” which is a dilemma that Socrates poses in Plato's dialogue, Euthyphro. Either what is pious (or righteous) is so because it pleases the gods or else what is pious (or righteous) pleases the gods because it is pious (or righteous). If the former is true, piety (or righteousness) seems to be arbitrary; if the latter is true, then the gods do not determine what is pious (or righteous), but merely respond to an independent truth about piety (or righteousness).

⁷⁰ Onora O'Neill's constructivism is more ambitious in this regard than is the general account of constructivism that I am discussing here (which is closer to Rawls's position), for she aims to build her arguments from only thin, formal assumptions. See O'Neill, Onora, “Constructivism in Kant and Rawls,”Google Scholar in Freeman, Samuel, ed., The Cambridge Companion to Rawls (forthcoming).Google Scholar