© Koninklijke Brill NV, Leiden, 2013 DOI 10.1163/17455243-4681035 Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 brill.com/jmp JOURNAL OF MORAL PHILOSOPHY When Justice Demands Inequality John Thrasher* Department of Philosophy, University of Arizona jthrashe@email.arizona.edu Keith Hankins** Department of Philosophy, University of Arizona Department of Economics, Rutgers University Hankins@email.arizona.edu Abstract In Rescuing Justice and Equality G.A. Cohen argues that justice requires an uncompromising commitment to equality. Cohen also argues, however, that justice must be sensitive to other values, including a robust commitment to individual freedom and to the welfare of the community. We ask whether a commitment to these other values means that, despite Cohen’s commitment to equality, his view requires that we make room for inequality in the name of justice? We argue that even on Cohen’s version of egalitarianism equality, freedom, and welfare are not always compatible. Justice will require trade-offs between these values. Sometimes, equality will need to be sacrificed. This is a surprising result and to show it, we use two informal impossibility proofs drawn from examples in Rescuing Justice and Equality. Keywords G.A. Cohen; Egalitarianism; Impossibility Proof; John Rawls; Amartya Sen In Rescuing Justice and Equality, G.A. Cohen argues that justice requires an uncompromising commitment to equality.1 Cohen also argues, however, that justice must be sensitive to other values, including a genuine concern for individual freedom and the welfare of the community. Our question is * John Thrasher is currently a Ph.D. candidate in philosophy at the University of Arizona specializing in political philosophy and normative ethics. He has published papers on con- tractarianism, Epicurus, social evolution, and the virtue of justice. ** Keith Hankins is a Ph.D. candidate in philosophy at the University of Arizona and is currently pursuing graduate studies in economics at Rutgers University. He works in politi- cal philosophy, social choice theory, and game theory. 1 This work is the product of both authors equally. The ideas were conceived of and developed in conversations over a period of several months. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 2 DOI 10.1163/17455243-4681035 whether the sensitivity of justice to values other than equality entails that we must sometimes make room for inequality in the name of justice. One of Cohen’s central claims in Rescuing Justice and Equality is that there is no inherent tension between freedom, equality, and Pareto (wel- fare).2 All three values, Cohen argues, are co-achievable. The purported tension between them, what he calls the Trilemma Argument, relies on the premise that people are not driven by sufficiently egalitarian motives. Cohen is concerned to dissolve the trilemma because it threatens the tight connection between justice and egalitarianism. If equality can only be pur- chased at the price of either freedom or welfare, it seems less attractive. Cohen’s argument rests on the claim that our motivational sets are not fixed. If “people believe in equality,” there is no real trilemma and, accord- ingly, our reasons for licensing inequality in the name of justice lose their force. Cohen refers to this idea—that justice requires us to adopt an egali- tarian ethos—as the Ethical Solution, and he suggests that it is capable of dissolving any worries that the trilemma argument presents for his brand of egalitarianism. We argue that Cohen’s understanding of what justice requires leaves little room for the commitment to individual freedom that lies at the heart of liberalism which he claims to share (even if he disagrees with liberals about its implications). As such, refusing to temper the demands of equal- ity generates manifestly unjust consequences. Although Cohen’s claim that justice requires individuals to adopt an egalitarian ethos has received much attention,3 our argument departs from other discussions of Cohen insofar as our aim is to show that, even by his own lights, Cohen’s Ethical Solution is not successful. We discuss the Trilemma Argument in §1 and the Ethical Solution in §2. There we suggest that the Ethical Solution fails by developing an informal impossibility proof drawn from an example that Cohen uses to illustrate his argument. Cohen’s proposed resolution of the trilemma implies that for an individual to be free she must be able to make at least some choices in 2 Cohen’s use of “Pareto” sometimes obscures the fact that “Pareto” is not itself a value, but rather a principle that governs how we ought to evaluate different states that are meas- ured in terms of welfare, levels of preference satisfaction, or some other metric. The tension that Cohen is concerned with then is between freedom, equality, and ways of promoting welfare. 3 See especially Michael Titelbaum who agrees with Cohen that Rawls has good reasons to accept that justice requires individuals to adopt a certain ethos, but argues that such an ethos would not look like the one Cohen sketches. Michael G. Titelbaum, ‘What Would a Rawlsian Ethos of Justice Look Like?,’ Philosophy & Public Affairs, 36 (2008), pp. 289–322. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 3 accordance with her own beliefs and values. The idea that individuals must have space to live their own lives is further supported by his discussion of personal prerogatives and the role that appeals to these prerogatives play in justifying our practices. Each person, he argues, has “the right to be something other than an engine for the welfare of other people: we are not nothing but slaves to social justice.”4 In §3, we argue that Cohen’s discus- sion of prerogatives and his concern with dissolving the trilemma shows that costs to individual freedom are not easily borne—even for an egalitar- ian as uncompromising as Cohen. If the Ethical Solution fails then, as we argue it does, then Cohen cannot have all he wants and his criticism of lib- eralism loses much of its force. To show that the Ethical Solution does indeed fail, in §4 we develop a second impossibility proof which draws upon an example borrowed from Aristophanes. There we argue that because freedom, equality, and Pareto will not always be co-achievable, a genuine commitment to each necessarily requires making trade-offs. In §5, we conclude that a genuine commitment to respecting individual freedom means that inequalities will sometimes be justified in the name of justice. 1. The Trilemma Argument At the heart of Rescuing Justice and Equality is a claim that freedom, equal- ity, and welfare are co-achievable.5 Justice involves a commitment to all three values, but does not require trade-offs between them. Cohen uses a number of examples to support this claim. Each example compares various states of affairs, distinguished by how hard certain individuals decide to work or what occupations they decide to take up, and asks how they fare as measured against various metrics. Although Cohen does not himself utilize these tools, the structure of his examples makes them especially amenable to analysis using some of the tools of social choice theory.6 Utilizing these tools we can compare, as Cohen does, various states of affairs and ask whether they can be achieved consistently with a range of suitably 4 G.A. Cohen, Rescuing Justice and Equality (Harvard University Press, 2008), p. 10. 5 Cohen, Rescuing Justice and Equality, pp. 181–225. 6 We are largely following an approach derived from the application of social choice theory to ethics pioneered by Amartya Sen, but this basic approach does not require all of the assumptions of traditional social choice theory. For a similar axiomatic treatment of related issues, see: Michael Huemer, ‘Non-Egalitarianism,’ Philosophical Studies, 114 (2003), 147–171. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 4 DOI 10.1163/17455243-4681035 formalized ideals. This is helpful because it delivers a degree of precision that Cohen’s arguments (persuasive as they might be) otherwise lack. Ultimately, however, by developing an informal impossibility proof we will argue, contra Cohen, that freedom, equality, and Pareto, as metrics for eval- uating states of affairs cannot always be mutually satisfied. As a result, if justice is to incorporate each metric it must include a mechanism for bal- ancing them in cases where they conflict. Before we look at the proof itself, though, we first need to define the dif- ferent metrics in question and, because many critics are likely to object that the social choice approach we adopt is foreign to Cohen, say a few things to defend our approach. Although, as we suggested above, Cohen does not explicitly use the framework of social choice theory, his examples are set up in a way that mimics the structure of problems analyzed in the social choice literature and his analysis of these examples sometimes parallels the social choice approach we utilize. More importantly, Cohen discusses in detail Amartya Sen’s famous argument for the impossibility of Paretian liberalism.7 Sen, of course, explicitly utilizes the sort of social choice approach we adopt, and Cohen’s discussion of Sen clearly suggests that he takes Sen’s argument to bear on the thesis he wishes to defend.8 We take it, then, that Cohen has opened the door for the use of these tools in analyzing his own argument. Turning to the formalized metrics we will utilize, for welfare we use the notion of weak Pareto described by Sen and others. Following Sen, the weak Pareto condition states that “if every individual prefers an alternative x to another alternative y, then society must prefer x to y.”9 This is clearly the Pareto condition Cohen has in mind and it implies that for justice to be consistent with Pareto it cannot identify a state of affairs as just when there are alternative states available that are unanimously preferred to it.10 Moving on to freedom, in his discussion of the trilemma argument Cohen focuses on freedom of occupational choice. Freedom, he writes, requires “that people not be coerced into particular jobs, whether by direct state order or by something else that also deserves to be called coercive.”11 7 Cohen, Rescuing Justice and Equality, pp. 187–188. 8 Amartya Sen, ‘The Impossibility of a Paretian Liberal,’ The Journal of Political Economy, 78 (1970), 152–157. 9 Sen, ‘The Impossibility of a Paretian Liberal,’ p. 153. 10 Cohen writes: “The relevant Pareto requirement is the (weak) one that condemns pre- serving a state of affairs in which everyone can be made better off,” Cohen, Rescuing Justice and Equality, p. 184. 11 Cohen, Rescuing Justice and Equality, p. 184. Of course, Cohen is not interested only in freedom of occupational choice, he merely uses it to illustrate his argument, but if we can J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 5 Utilizing our approach, however, requires something more formal and so for our freedom metric we again borrow a metric from Sen, namely his notion of minimal liberalism.12 Following Sen, this condition states that for every individual A, there is at least one pair of alternatives {x, y} such that if A prefers x to y, society must prefer x to y, in the sense that it cannot coerce A into choosing y over x.13 While there is considerable debate about the definitions of freedom utilized by Sen, we think the notion captures what Cohen has in mind within the context of his example.14 For instance, although Cohen criticizes the conclusion that Sen draws from his proof, he does not argue that the minimal liberalism condition is an inadequate model of liberal rights.15 Further, a simple example illustrates that this condition is analogous to the basic conception of freedom of occupation Cohen sketches. If Roxanne prefers one job over another and both are offered to her, free- dom dictates that society cannot coerce her into choosing her less preferred job. It is in that sense that her decision about what job to take is decisive.16 The third value, equality, is the hardest to specify because Cohen’s dis- cussion of it is surprisingly unclear. Cohen writes that “egalitarians like [him] think that justice is served only if people’s access to desirable condi- tions of life is equal,” but he suggests that for the purposes of the trilemma Argument, “the relevant equality might be only that no one is substantially better off than others are with respect to both income and job satisfac- tion.”17 The challenge then is to specify what it would mean for access to the desirable conditions of life to be equal, or for some to be substantially better off than others. Fortunately, Cohen spends much of the book discuss- ing Rawls’ difference principle that lends itself well to formalization, and Cohen defends a particularly demanding interpretation of the difference show that there exists an unavoidable tension between this sort of freedom and equality, then our argument generalizes to freedom more broadly construed. 12 Sen, ‘The Impossibility of a Paretian Liberal,’ p. 154. 13 Note that the Freedom condition also requires that A be free to decide which pair of alternatives she is decisive over. This is necessary to ensure that the condition is substan- tive; otherwise, the condition could be trivially met by providing an individual with deci- siveness over two worthless alternatives. 14 For an overview of the debate on Sen’s argument of the impossibility of a Paretian liberalism and Sen’s responses to critics see: Amartya Sen, ‘Rights: Formulation and Consequences,’ in Rationality and Freedom (Cambridge: Harvard University Press, 2002), pp. 439–461. 15 See Cohen, Rescuing Justice and Equality, p. 187. 16 In the case where Roxanne is presented with multiple job offers, freedom requires that she be able to eliminate at least one from consideration, so that she cannot be forced to accept a job that she least prefers. 17 Cohen, Rescuing Justice and Equality, p. 184. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 6 DOI 10.1163/17455243-4681035 principle that is easier still to formalize.18 Following Cohen’s discussion of the difference principle, we assume that equality means maximizing, by their own lights, the condition of the least well off.19 On this definition, equality implies that given a pair of alternatives {x, y} if x maximizes the condition of the least well off, then x is more just than y. As a result, at least as a matter of justice, society ought to prefer x to y. While there is much more that can and should be said about equality, this definition provides us with the specificity required to analyze the trilemma argument and Cohen’s proposed solution to it. Before moving on it is also important to say something about the con- cept of preference we utilize. There is often some confusion about what it means to prefer one alternative to another, and indeed, Cohen is somewhat unclear about what he means by preference. In order to avoid confusion, then, we employ a technical sense of preference that is clearer than the ordinary notion. Although the ordinary notion of preference allows us to say that we can sometimes choose an act that we might not prefer to do (which perhaps because we are required to do by some type of duty), the technical sense of preference we employ subsumes all considerations that lead to action.20 This notion of preference is both essentially comparative and all things considered, so that to prefer something is to prefer it to some- thing else.21 18 The version of the difference principle that Cohen favors requires that we maximize the well-being (or primary goods available to) the worst off, that labor burden be included in the distributive metric, and that the demands of the difference principle extend to individu- als’ actions and decisions as opposed to being restricted to the structure of our major social institutions. See: Cohen, Rescuing Justice and Equality, Chapter 1, Part III and Chapter 2. 19 In utilizing this variant of the difference principle as our equality metric, however, we do not mean to be taking a stand on the question of whether the difference principle licenses inequalities to some that might be necessary as incentives to induce them to do work that redounds to the benefit of others. Further, we do not mean to be claiming that the difference principle is ultimately the best way of understanding what equality demands of us. Cohen clearly thought that equality was best understood as requiring that persons have “equal access to advantages.” For analyzing the trilemma argument, however, using the difference principle as a metric of equality is sufficient and does not contradict the spirit of Cohen’s argument. 20 One account, among many, of the technical sense of preference employed here is found in S.I. Benn and G.W. Mortimore, ‘Technical Models of Rational Choice,’ in Rationality and the Social Sciences, 1976, pp. 160–161. This is also the account of preference normally employed in the social choice literature. A good recent discussion can also be found in Daniel M. Hausman, Preferences, Value, Choice, and Welfare (Cambridge: Cambridge University Press, 2011), Chapters 1 & 2 . 21 There is considerable philosophical debate on the nature and grounding of prefer- ences as well as their relationship to psychological states such as beliefs and desires. On these questions, we remain silent. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 7 Having described the approach we will use, we can now turn our atten- tion to the primary example that Cohen uses to illustrate the trilemma. Cohen’s example involves a person that we call Epicurean Doctor, who enjoys gardening, but will work as a doctor if her wage is sufficiently high.22 Epicurean Doctor is willing to give up the pleasure and satisfaction she gets from gardening if she is paid $30,000 more dollars to be a doctor, but if offered the same wage for each job, say $20,000, she prefers to be a gardener. Whatever she chooses, Cohen stipulates, she is much better off than most other people are in her society, and further, doctoring imposes no hardships on her over and above the fact that she prefers gardening. Her society, however, needs doctors and so society would most prefer to pay Epicurean Doctor only $20,000 to work as a doctor, but would also prefer to have her work as a doctor at $50,000 than to have her spend her days gardening. Assuming nobody else cares whether Epicurean Doctor is able to pursue her passion for gardening, and (to simplify things) if society as a whole is responsible for directly paying doctors, then we have a divergence between Epicurean Doctor’s preferences and society’s preferences.23 As we suggested above, we can model justice as a social choice rule that ranks the three occupational choices that confront Epicurean Doctor given her preferences and everyone else’s which are represented in Table 1 below: 22 Cohen, Rescuing Justice and Equality, chap. 5 section 2. 23 In a forthcoming review of Rescuing Japa Pallikkathayil argues that it is unclear on what grounds Cohen is committed to caring about the absolute condition of the worst off (or anyone for that matter). As such it’s not clear how justice should constrain the choice of occupations with equal wages (or levels of welfare). For instance, Pallikkathayil argues that although Cohen sometimes suggests that a commitment to equality is enough to motivate the choice of, for example, more socially useful occupations, this seems like a mistake. See Japa Pallikkathayil, ‘Review of G.A. Cohen’s Rescuing Justice and Equality,’ Journal of Philosophy (Forthcoming). Table 1 Epicurean Doctor Everyone Else (a) Doctor at $50,000 (c) Doctor at $20,000 (b) Gardener at $20,000 (a) Doctor at $50,000 (c) Doctor at $20,000 (b) Gardener at $20,000 J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 8 DOI 10.1163/17455243-4681035 Epicurean Doctor has the preference ordering a > b > c, while Everyone Else has the preference ordering c > a > b. The question that the proponent of the trilemma argument forces us to ask is how justice responds to these divergent preferences, and in particular whether a social choice rule mod- eling justice can mutually satisfy the constraints presented by equality, freedom, and Pareto as we have defined them? Both Epicurean Doctor and Everyone Else prefer a to b, so according to Pareto b should not be chosen over a. However, option c is also on the Pareto frontier because there is no alternative that everyone prefers to c. We can say then that the Pareto principle is not decisive here because both a and c are compatible with it. Alternatively, if we appeal to freedom, per- haps claiming that Epicurean Doctor has a right to freedom of occupation, then we must make (some of) Epicurean Doctor’s preferences concerning her occupation decisive and a will be the likely winner.24 Appealing to equality, however, yields a different result. If we assume that alternative c (doctoring at $20,000) maximizes the condition of the least well off, then even though Epicurean Doctor ranks option c last, it is the option picked out by the equality metric.25 We summarize these outcomes in Table 2 below: 24 Strictly speaking freedom picks out either a or b depending upon which pair of alter- natives {a, c} or {b, c} Epicurean Doctor is decisive over. The important point is that freedom effectively rules out c and ultimately yields a as the social choice either directly (if Epicurean Doctor is decisive over a and c) or indirectly (if Epicurean Doctor is decisive over b and c, in which case freedom picks out b and Pareto leads us to a because a > b by everyone). Note that we can ignore the pair {a, b} because freedom allows Epicurean Doctor to decide which pair she is decisive over and because society prefers that she be a doctor at $50,000 (a) to gardening at $20,000 (b), so there is no need for her to exercise her decisiveness over that pair. Our earlier discussion of Sen’s minimal liberalism condition explains why it is neces- sary that Epicurean Doctor be allowed to decide which pair she is decisive over, namely that if our freedom condition is to be substantive, Epicurean Doctor must not only be decisive over at least one pair of alternatives, but she must be free (within some substantial range) to choose which pair of alternatives she should be decisive over. 25 We grant for the sake of argument that this is a reasonable assumption since society needs doctors and c involves the Epicurean Doctor, working for less money, thus leaving society with more money which it might then use to benefit the least well off in other ways. Table 2 Freedom a (or b) Equality c Pareto a or c J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 9 Referring to Table 2 we can see that freedom and Pareto are compatible since they each yield a as an eligible social choice. Pareto and equality are, similarly, compatible since they each leave c in the eligible set. Freedom and equality, however, do not yield compatible choices. If we appeal to equality, freedom is swamped. Everyone Else becomes decisive over the pairs {a, c} and {b, c} and is thus able to dictate both Epicurean Doctor’s job and salary. This result suggests that an uncompromising commitment to equality requires that we sacrifice freedom in cases where an individual’s preferences conflict with the interests of the least well off. Alternatively, if we appeal to freedom and allow (again, some of) Epicurean Doctor’s prefer- ences to be decisive, then the least well off will be worse off than they might otherwise be. This is the deep tension between the values that concern lib- erals versus the values that concern egalitarians. Because Cohen is con- cerned with showing that a commitment to equality does not require us to sacrifice freedom, we focus on the former result and in the next section we look at his proposed solution to this problem. 2. The Ethical Solution As Cohen recognizes, his example involving the Epicurean Doctor is analo- gous to Sen’s Prude/Lewd example from his argument for the impossibility of Paretian liberalism. Contra Sen, however, Cohen argues that it is a mis- take to think that minimal rights must conflict with Pareto or, carrying the point over to his own example, with equality. Cohen’s primary contention is that “liberalism does not require that you exercise the sovereignty it grants in accordance with your other-things-equal first preferences.”26 He also argues however that the notion of liberalism embodied in Sen’s “minimal liberalism” and captured by the freedom principle we have used, is not at all similar to the liberalism that he claims to share with Dworkin and Rawls.27 With respect to his second point, Cohen is simply mistaken. To be valu- able, freedom or liberal rights must give significant weight to protecting individuals from having the preferences of others imposed on them. Of course, an individual may decide not to exercise a right in a given case— Cohen is right about this much—but to be significant, that decision must presuppose the individual’s ability to exercise the right should she desire.28 26 Cohen, Rescuing Justice and Equality, p. 187. 27 Cohen, Rescuing Justice and Equality, p. 188. 28 It is true that elsewhere Cohen famously characterizes freedom in terms of ability, and in his example the Epicurean doctor is surely able to choose the occupation of her choice. Our J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 10 DOI 10.1163/17455243-4681035 To say that a minimally liberal set of rights conflicts with Pareto and equal- ity then is to say that if some individuals exercised their rights either Pareto or equality must be sacrificed. Indeed, this was the point of Sen’s original argument, namely that rights and social welfare will not always coincide, and it is only once we understand this point on a formal level that we are led to more substantive philosophical questions concerning which values justice requires us to be sensitive to and in what kinds of situations.29 Cohen is making a deeper point as well though, viz. that the demands of justice extend beyond those placed on the basic structure of society, and it is on this deeper point that his main line of argument against the problem posed by the trilemma rests.30 Cohen’s argument is that if the decisions we made in our individual lives were guided by a belief in equality, the tri- lemma dissolves. Justice, he argues, requires us to internalize certain values and once we do so, our preference for exercising our rights in ways that generate the sort of conflict the proponent of the trilemma worries about will vanish. Cohen refers to this as the Ethical Solution and motivates it with the following imaginary exchange: The trilemmist says: “We shouldn’t be egalitarians, because equality requires a sacrifice of either Pareto or freedom.” To which I reply: “That isn’t so, because, if we were egalitarians, we should be sacrificing neither.” And that reply isn’t trivial.31 Cohen’s argument is that once the Epicurean Doctor realizes the impor- tance of equality she would (or should) change her preferences. If she really cared about equality, freedom would not be an issue. point is simply that if freedom is to be substantively valuable, justice must carve out a range of choices, from which an individual is free to choose, consistent with the demands of justice. 29 Indeed, this point, and the larger point that the point of theorizing about justice is to provide practical guidance in alleviating suffering and solving disputes is the primary thesis of Sen’s recent book and has also been powerfully articulated by David Schmidtz. See Amartya Sen, The Idea of Justice (Cambridge, MA: Harvard University Press, 2009) and David Schmidtz, ‘When Justice Matters,’ Ethics, 117 (2007), pp. 433–459. 30 This is one of the main thrusts of Cohen’s book and is an issue about which there has been much debate. See: Cohen, Rescuing Justice and Equality, chap. 4; Liam B. Murphy, ‘Institutions and the Demands of Justice,’ Philosophy & Public Affairs, 27 (1998), 251–291; Thomas W. Pogge, ‘On the Site of Distributive Justice: Reflections on Cohen and Murphy,’ Philosophy & Public Affairs, 29 (2000), 137–169; A.J. Julius, ‘Basic Structure and the Value of Equality,’ Philosophy & Public Affairs, 31 (2003), 321–355; Thomas Porter, ‘The Division of Moral Labour and the Basic Structure Restriction,’ Politics, Philosophy & Economics, 8 (2009), 173–199; Kok-Chor Tan, Justice, Institutions, and Luck: The Site, Ground, and Scope of Equality (Oxford University Press, USA, 2012). 31 Cohen, Rescuing Justice and Equality, p. 196. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 11 Of course, the Epicurean Doctor might simply respond that if Everyone Else really cared about freedom, they would leave her alone to pursue her preferred profession without interference. In either case, though, solving the problem this way merely sidesteps the serious philosophical questions concerning how social justice should accommodate competing values. Cohen’s argument is that in a truly just society, people’s preferences would align with equality and, hence, it would never conflict with the other two metrics. This amounts to saying that if the Epicurean Doctor cared about equality she would prefer to work as a doctor for $20,000 rather than gar- den or work as a doctor for a higher wage. Her preference ordering would change from a > b > c to c > a > b and since her preferences would now be the same as Everyone Else’s no trilemma would arise. We say this approach sidesteps the difficult philosophical issues at hand, because it is hard to see how it shows that the trilemma is not really a problem. In this case, justice free from conflict is merely the product of philosophical fiat. Cohen “solves” the trilemma argument by stipulating that it never arises.32 For his solution to be of general interest, Cohen needs to show that the preference profiles that give rise to the trilemma are inadmissible. Essentially, Cohen needs to show that we have good reason to reject the unrestricted domain condition that we implicitly assumed in the Epicurean Doctor example.33 Sen describes unrestricted domain as the condition that “every logically possible set of individual orderings is included in the domain of the collective choice rule.”34 In other words, no individuals or preference orderings should be excluded from the scope of the collective choice rule— in this case, the rule being used to model the requirements of justice. Of course, it is by no means obvious that justice must license the range of 32 Note that the problem is more dramatic still. For Cohen’s solution not only requires that individuals be motivated to pursue egalitarian policies, but the same egalitarian poli- cies; otherwise the problem of divergence of preferences will arise even assuming an egali- tarian ethos. Rawls in particular, whose animating concern was the problem posed by the fact of reasonable pluralism, would have been particularly attuned to this worry. 33 Cohen’s solution to the trilemma exploits the well-known fact in the social choice lit- erature that only certain preference profiles generate the paradoxes of social choice that the famous impossibility theorems threaten. Like Cohen here, many have attempted to defang the threat of various impossibility theorems by weakening the unrestricted domain assump- tion that Arrow’s impossibility theorem, among others, relies upon. Unfortunately, one of the most dispiriting lessons in social choice has been that many forms of impossibility are resil- ient in the face of a substantial degree of weakening of their axioms. Furthermore, the type of restriction that Cohen proposes is not trivial as he is essentially proposing restricting the domain of individual preferences so that they are identical to the set of social preferences. 34 Sen, ‘The Impossibility of the Paretian Liberal,’ p. 153. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 12 DOI 10.1163/17455243-4681035 preferences that the unrestricted domain condition allows, particularly if justice requires (as Cohen thinks it does) that individuals internalize a cer- tain ethos, and so Cohen’s strategy perhaps looks promising. In the next section, however, we will see that Cohen lacks the resources needed to restrict the domain of acceptable preferences to the extent required by his solution to the trilemma. 3. Vacations, Sick Days, and Prerogatives In his earlier work, Cohen argues in favor of a principle of symmetrical jus- tification for coercion.35 In Rescuing, however, his discussion of justifica- tion is framed first in terms of something he calls the interpersonal test,36 and later in terms of the idea of a personal prerogative first sketched by Samuel Scheffler and subsequently developed by David Estlund.37 Cohen begins the trilemma chapter trilemma with the observation that the demands of egalitarian justice must fall “within the constraint of a reason- able personal prerogative, deference to which informs the whole of the fol- lowing discussion”—only a moral rigorist, he argues, could deny a place for such a prerogative, and he affirms that he is not such a rigorist.38 Prerogatives are important because valuing the freedom of persons to pursue a conception of the good requires that they have considerable room within to make decisions about their life and prerogatives provide people with this space.39 Prerogatives protecting occupational choice are espe- cially important in this regard. No less important is being free to choose how to trade off time spent working with time spent with one’s friends and 35 G.A. Cohen, Self-ownership, Freedom, and Equality (Cambridge: Cambridge University Press, 1995), pp. 56–57. 36 See Cohen, Rescuing Justice and Equality, chap. 1, Section  5. According to Cohen, as free and equal members of a moral community our relationships ought to be structured by a norm of comprehensive justification and a necessary condition of comprehensive justifi- cation is that arguments be capable of passing the interpersonal test, where this requires that an argument must be capable of providing justification under any and all dialogical conditions. In other words, an argument’s ability to provide justification must be indepen- dent of both who is presenting it and to whom it is being presented. 37 David Estlund, ‘Liberalism, Equality, and Fraternity in Cohen’s Critique of Rawls,’ Journal of Political Philosophy, 6 (1998), p. 99 and Samuel Scheffler, The Rejection of Consequentialism (New York: Oxford University Press, 1982). 38 Cohen writes “I do not wish to reject … what Samuel Scheffler has called an “agent- centered prerogative,” Cohen, Rescuing Justice and Equality, p. 61. 39 This point is especially clear in the case of love and relationships, which we discuss in section 5. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 13 family, or pursuing other interests and passions. Treating people as free then requires that individuals have the prerogative to make their own choices about how to live their lives even if this means that there will be some who do worse than they otherwise might. While Cohen accepts that there is a place for a modest personal preroga- tive within his view of justice, there are two ways of understanding such prerogatives. On one hand, prerogatives can be conceived of as “vacations” from the demands of justice. Vacations are justified departures from the requirements of justice that allow individuals a certain amount of space to freely pursue their own projects, regardless of how these pursuits relate to distributive justice.40 Conceived of in this way, prerogatives prevent justice from being all-encompassing. Alternatively, prerogatives can be thought of as “sick days.” Thought of this way, prerogatives are excuses for failing to live up to the demands of justice. However, just as it would perhaps be bet- ter if individuals didn’t have to take sick days from work, it would also be more just if people minimized the number of prerogatives taken to depart from the demands of justice. It sounds odd though to say that we should treat prerogatives just as we do sick days. This requires us to think that individuals’ pursuing their own projects is something to regret—a necessary evil. Acting in accordance with justice, on this model, is a kind of Stakhanovism.41 If this is Cohen’s view, he is in danger of repeating the mistake that Bernard Williams and Rawls attri- bute to utilitarianism: that it does not take the separateness of persons seri- ously, viewing people instead as mere nodes for the fulfillment of justice. To make the distinction we have drawn even starker, note that individu- als are often expected and even encouraged to take vacations from work, lest they allow their work to cause them to lose sight of the other important things in life. Although one’s occupation is often bound up with one’s iden- tity, and in many cases helps to give life meaning, what makes the freedom to choose one’s occupation important are the myriad values, interests, and larger-scale projects that individuals have which give them reasons to shape their identities in various ways. While these are issues that Cohen is sensitive to, unfortunately he does not fully appreciate the constraints they impose on a view like his. Although he accepts that justice must make room for a reasonable personal prerogative, he tells us little about what 40 For a more complete discussion of a view like this, see: David Estlund, ‘Liberalism, Equality, and Fraternity in Cohen’s Critique of Rawls,’ p. 99. 41 ‘Stakhanovism’ refers to the ideology of the Stakhanovite movement of the 1930s in the Soviet Union under Stalin. The movement was inspired by the example of Aleksei Stakhanov a miner who, in 1935, mined 14 times his quota of coal in one day. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 14 DOI 10.1163/17455243-4681035 falls within the scope of such a prerogative nor, as we have seen, does he tell us how such prerogatives mediate the demands of justice.42 This second point is particularly important, for if prerogatives are best conceived of along the vacation model as we have argued they should be, then it is hard to see how Cohen can constrain the domain of admissible preferences to the degree required to get his ethical solution off the ground. This is because the only way the ethical solution succeeds is if individuals could come to have such a strong egalitarian ethos that their occupational preferences coincide with what would best promote the needs of the least well off. As we will see in the next two sections though, even if individuals could come to adopt such an equality driven ethos, there are reasons to doubt whether they should. Before exploring this argument, though, it is important to first clarify why the ethical solution succeeds only if the domain of free choice is severely curtailed. The problem is that prerogatives license inequalities and we have good reason to concede that individuals have prerogatives within which they are free to act. Cohen admits as much. His disagreement with those like Estlund and Scheffler who have drawn attention to the inequality generating feature of prerogatives is directed solely at the amount of inequality that reasonable prerogatives license and the way in which they do so.43 Cohen accepts that prerogatives allow individuals to justifiably make choices that generate inequality, but contra Estlund he argues that the amount of inequality that is thus justified is not extensive. This is a debate about the extent of inequality that is justified, though, and not a debate about whether it is justified. Similarly, Cohen emphasizes that the prerogative justification provides a different sort of justification for inequality than the incentive argument that grounds the difference principle, but this too does little to cut against the claim that some inequalities will be justified because it only 42 In the introduction to Rescuing, for instance, Cohen points out that there are many forms of motivation lying on a spectrum that stretches from unrestrained market-maximiz- ing at one end to fully self-sacrificial restraint favoring the worst off at the other. He then argues that the first extreme is permitted by Rawls, something that he finds absurd, but he also argues that the second extreme isn’t required because requiring such self-sacrifice is “excluded by a legitimate personal prerogative.” Because individuals have their own lives to lead, Cohen suggests “they are therefore permitted to strike a balance between the claims of the difference principle and their own legitimate concerns.” The difficulty, on Cohen’s view, is that what constitutes an appropriate balance is a “vague matter.” Cohen, Rescuing Justice and Equality, pp. 10–11. 43 See Cohen, Rescuing Justice and Equality, pp. 387–389 Section  4 of the General Appendix, “Incentives and Prerogatives.” J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 15 speaks to the question of how those inequalities are justified.44 This is important because our argument has been directed at Cohen’s claim that equality, freedom, and Pareto are not in conflict. We have denied Cohen’s claim, and in doing so, what we have argued is that Cohen must license some inequality if he is to maintain his commitment to freedom. What we have not tried to say, however, is how much inequality he must make room for. 4. The Women of Athens In the “Ecclesiazusae,” Aristophanes depicts an Athens racked by war and deprivation, having been run into the ground by the men. In response, the women of the city take over the legislative assembly and, in an effort to save Athens from total ruin, their first act is to legislate absolute equality between all Athenians. Their leader Praxagora decrees that, “Mankind should possess in common the instruments of happiness. Henceforth pri- vate property comes to an end …”45 The plan, as Praxagora elaborates, is to “destroy [the old] morality”46 and supplant it with a new ethos, the hope being that by working together and owning all things in common Athenians could come to build a society on a foundation of equality. Praxagora is, like Cohen, ultimately concerned with equality of welfare, not equality of money and because one of the aspects of life that can con- tribute the most to happiness and welfare is love, she becomes concerned about the vast inequities of love that naturally exist. Recognizing that it is hard for the old, the ugly, and the awkward to compete with those younger, prettier, and smoother for the affections of potential lovers, the women outlaw marriage in the city and even go so far as to make it illegal for the young and pretty to have their turn at love without first allowing the older and uglier a chance.47 In one scene, a pretty young girl is waiting for her lover while an old woman stands close by. Eventually the young man arrives, seeking the 44 It is also worth noting that in his reply to critics Cohen eventually acknowledges, fol- lowing Estlund, that prerogatives could serve to legitimate certain deployments of the incentive argument, his argument with Estlund on that point being that there is simply no trace of such an argument to be found in Rawls’ work. 45 Aristophanes, ‘Ecclesiazusae,’ in The Complete Plays of Aristophanes, ed. by Moses Hadas, Bantam Classic (New York: Bantam Classics, 1981), p. 438. 46 Aristophanes, ‘Ecclesiazusae,’ p. 439. 47 Aristophanes, ‘Ecclesiazusae,’ p. 457. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 16 DOI 10.1163/17455243-4681035 young girl. The older woman, spying her chance, demands that the young man obey the new law. Horrified at what justice now requires of him, the young man tries to trick the old woman, but to no avail. She prevails and drags him off into her room. At just that moment, though, an older, uglier woman shows up to claim her right and the first woman is forced to give him over to her older and uglier competitor. Naturally, as the second old woman is taking the young man away a third appears, older and uglier still. As the third old woman says, “I’m older, and I’m uglier; consequently this boy belongs to me.”48 Given the multiplicity of women arranged in a hier- archy of unattractiveness, the young man will have to work his way down the line if he intends to ever be with his lover. While the young man has no interest in being with any of the old women, alas, his wishes are irrelevant. If he could do as he wished and be only with his beloved, the old women would be much worse off in relation to the young couple. From the point of view of justice, at least as instantiated by the women of Athens, this is unacceptable. If the young man wishes to pur- sue the girl, he does indeed “belong” to the oldest woman first, because her happiness and welfare overrule his freedom to choose his partners. There is a sense, then, in which the young man has become merely a means to the old women’s satisfaction. This case clearly illustrates what it is like for indi- viduals to be treated merely as material for the fulfillment of justice. What makes it particularly compelling though is that Cohen is clearly committed to thinking that it is inconsistent with justice to treat individuals in this way.49 As we did with the case of the Epicurean Doctor, we can formalize Aristophanes’ story. In the Aristophanes case, however, unlike the Epicu- rean Doctor case, we have many preference orderings rather than simply two. Furthermore, the preferences range over who may be with whom in a romantic situation and so the example is even starker than Cohen’s origi- nal case insofar as it involves action that is even more personal than occu- pational choice.50 48 Aristophanes, ‘Ecclesiazusae,’ p. 459. 49 Recall, for instance, Cohen’s claim from the introduction to Rescuing that individuals have the right to be something other than engines for the welfare of other people. 50 Aside from the already discussed prerogative, there is no reason in principle why Cohen should hold that the demands of justice should not apply to romantic choice, espe- cially if romantic partners are importantly related to welfare. We need not be committed to the claim that Cohen would have us all be egalitarians in romantic matters though. Rather, the point of the example is simply to illustrate the tension between freedom and equality and thus the need to invoke a prerogative or some other mechanism for resolving this tension. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 17 In order to simplify things, in what follows, we represent this situation as involving three persons, though the older and uglier women that enter later could easily be incorporated without changing the result. The Young Man prefers to be with the Young Woman and only the Young Woman. We can call this state of affairs, x. Nevertheless, if he cannot be with just the Young Woman, he is willing to be with the Old Woman first and then the Young Woman. He finds the prospect of being with the Old Woman objec- tionable, but he is willing to pay that price for love if he must. We call this state of affairs where the Young Man will be with the Old Woman first, and then the Young Woman, y. The worst state of affairs for the Young Man is to miss the opportunity to be with his beloved entirely. This state of affairs involves everyone leaving the scene without having any romantic involve- ment at all. We can call this state of affairs z. Thus, the Young Man’s prefer- ence ordering in this situation is: x > y > z. The Young Woman’s first preference is also to be with her lover, the Young Man, without any interference from the Old Woman, and so she pre- fers x to all other alternatives. The thought of her lover with the Old Woman angers her so much, however, that she would prefer z to y even though this means she would have to forgo the affection of her lover. The Young Woman’s preference ordering, then, is: x > z > y. Finally, the Old Woman prefers y to all other states of affairs, but would be so wrought with jealousy and envy if the Young Woman were able to have the Young Man to herself that she prefers z to x and so her preference ordering is: y > z > x.51 These orderings are represented in Table 3 below: 51 We assume for the sake of argument that the Old Woman’s preferences accurately track her welfare. As we will see in the next section, though, there are perhaps problems with making this assumption that in turn give us further reason to doubt the efficacy of Cohen’s proposed solution to the trilemma. Table 3 Young Man Young Woman Old Woman x x y y z z z y x As in the Epicurean Doctor case, we have a dilemma. We say dilemma, rather than trilemma here because in this situation, every alternative satis- fies the Pareto principle and so it cannot conflict with our other principles. There is no pair of alternatives such that one alternative is preferred to the other by everyone. Both the Young Man and the Young Woman prefer x to J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 18 DOI 10.1163/17455243-4681035 y while the Old Woman prefers y to x and similarly, the Young Man and Young Woman prefer x to z while the Old Woman prefers z to x.52 On the other hand, the Young Man and the Old Woman both prefer y to z, while the Young Woman prefers z to y. There is no way then to reconcile their respective preferences with both equality and the requirement that each individual be decisive over at least one pair.53 5. The Demands of Equality In the “Ecclesiazusae,” the women of Athens deem equality to be the unique collective choice rule consistent with justice. The result is comical. Their uncompromising commitment to equality threatens to undermine even their most basic freedoms. We should be worried then by the fact that Aristophanes’ dilemma parallels the trilemma argument, particularly since Cohen’s proposal that the trilemma can be dissolved if only we could embrace an ethos of equality is not offered in jest. In light of Aristophanes’ dilemma, we must ask whether Cohen’s ethical solution really allows us to express an uncompromising commitment to equality without becoming “slaves to social justice.” Let us assume, following Cohen’s suggestion, that the characters in our example borrowed from Aristophanes have become perfectly committed to equality so that the young couple’s preferences come to coincide with the Old Woman’s. Their respective preference orderings can then be repre- sented as follows: 52 To get this result we need not assume transitivity, only quasi-transitivity since we are only dealing with strict preference orderings without indifference. 53 Note that the primary problem presented by this particular set of preference profiles is that the Young Woman ranks y, the option required by equality, as her least preferred option. Table 4 Original Dilemma Cohen’s Revised Dilemma Young Man Young Woman Old Woman Young Man Young Woman Old Woman x x y y y y y z z z z z z y x x x x J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 19 Note that the proposed revision of the parties’ preferences does indeed solve the problem, but as with the Epicurean Doctor case, it is only able to do this by changing the circumstances. Earlier we charged Cohen with avoiding the problem rather than confronting it. The point we now wish to make, though, is that Cohen himself gives us reason to think that we should be hesitant to adopt such an ethos. First, however, we must consider what can be said in favor of adopting such an ethos. There is a similarity for instance between this example and Cohen’s occasional approval of “everyone doing his or her bit” in war- time England.54 On Cohen’s view, despite the hardships that accompanied “doing one’s bit,” a certain amount of fellow feeling and patriotism moti- vated people to do more than they might have otherwise done and this is to be celebrated. It does not follow, however, that since people were moti- vated by fellow feeling that their hardships were no longer hardships. Indeed it is precisely because everyone “did their bit” in the face of these hardships that that fact is to be celebrated. Similarly, the Young Man, despite his newly acquired egalitarian ethos, will still find the prospect of being with the older women objectionable, but he will see it as his duty as well and so will perform the task in spite of this. He may even get a warm feeling knowing that he has done his duty. The problem we face is that in this case equality is providing the wrong kind of reason for action. Cohen opens the door to this argument in the final section of “The Freedom Objection.” Here, Cohen makes an intriguing argument against prostitution. Prostitution is wrong, Cohen argues, for the same reason that rape is wrong, namely because “the wanted thing is yielded for the wrong reason.”55 Of course, some might argue with this characterization of rape. It may not be merely sex that the rapist is after. Rape is, at least some of the time, an exercise of power and domination and not just a sexual act. Cohen argues, however, that it would be wrong to view rape merely as a “species of assault.”56 To think of rape in that way would be to base one’s disapproval of it on a “breach of self-ownership” and would make one indifferent to prostitution to the extent that prostitution is (at least sometimes) an expression of self-ownership. Cohen’s argument here is far from uncontroversial, but rather than dis- pute Cohen’s assessment of rape and prostitution we want to consider how Cohen’s claim puts pressure on his proposed solution to the trilemma. 54 Cohen, Rescuing Justice and Equality, p. 219. 55 Cohen, Rescuing Justice and Equality, p. 224. 56 Cohen, Rescuing Justice and Equality, p. 224. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 20 DOI 10.1163/17455243-4681035 In our Aristophanes example, the Young Man has a duty to have sex with the Old Woman if he ever wants to be with his lover. Of course, he does not love or desire the Old Woman and so, although he may be motivated to engage with her, this can only be because he has embraced his egalitarian ethos so completely. The Young Man’s commitment to egalitarianism may be commendable, but, like the prostitute, he is yielding the “wanted thing” for the wrong reasons. While the Young Man is not being forced to yield his body to the Old Woman because he is motivated to do it, his commitment to equality or justice does not translate into love or desire for the Old Woman. It follows, then, that if we are to condemn prostitution and rape along the lines that Cohen suggests, we should also condemn this particu- lar act of “yielding” for the same reasons. This is bad news for Cohen’s ethical solution, for it shows that even if adopting an egalitarian ethos could dissolve the trilemma, by Cohen’s own lights we sometimes have overrid- ing reasons not to do so. Sometimes, it seems, there are inequalities that it would be unjust to change. Of course, the defender of Cohen could argue that what an appropriate egalitarian ethos really requires of the young couple is that they simply refrain from being with one another. Thus, the Young Man would not be compelled to yield himself to the Old Woman for the wrong reasons, but the Old Woman would now have no reason to envy the young couple’s abil- ity to find love. This, we think, goes too far. While it perhaps does not make the young couple mere engines for the welfare of the older and less attrac- tive, it does threaten to swallow up any space they might have in which to live their own lives. More reasonably, Cohen could argue that romantic choices should fall within the purview of a personal prerogative. We do not disagree. The problem is that if Cohen grants that choosing one’s romantic partner should fall within the scope of a reasonable prerogative, then that romantic choices are protected in this way simply serves as proof that the commitment to equality is not, in the end, uncompromising. At this point we must consider whether we have given due weight to Cohen’s sensitivity to the distinction between the relation of the personal and impersonal on one hand and the relationship between justice and the personal on the other.57 The worry as we understand it is that Cohen’s sen- sitivity to this distinction mitigates much of the criticism that we have lev- ied against him. We appreciate Cohen’s sensitivity to issues like this. We recognize, for instance, that Cohen would prefer us not to be egalitarians about everything, least of all sex and love. We struggle to see though how 57 We thank an anonymous referee for this journal for raising this point. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 21 this insulates him from the charges we have laid at his feet. One reason is that it is hard to see how Cohen can draw a non-arbitrary line between those choices that do and those that do not fall within the scope of justice. After all, one of the main theses Cohen defends is that “the personal is political” and thus that the demands of justice extend to the sorts of deci- sions individuals make in their everyday lives. Indeed his entire criticism of Rawls rests on this claim. More importantly, however, we believe that there is good reason to think that something like a prerogative does not in fact compete with justice. There is a deep tension running through Cohen’s thought, though, that makes it unclear whether he can hold this view. Cohen recognizes that because individuals have their own lives to lead, neither justice nor equality can be all that matters. On one way of understanding things, this is not problematic because justice is simply one value among many. Indeed, Cohen suggests that “there is some justice that cannot, and some that should not, be implemented institutionally, or indeed, as Shakespeare’s Portia knew, at all.”58 On this view, we must make room for inequality in light of individuals’ reasonable personal prerogatives, but this is not a requirement of justice. Rather, as we argued in §3, it is a (potentially regret- table) situation that arises because these prerogatives lie outside the scope of justice. As Cohen also argues, though, the label “justice” is significant and one reason it is significant is that “each person possesses an inviolability founded on justice.”59 As we see things, making room for a prerogative is one way of giving force to this powerful idea and the implication of this is not just that we must make room for inequality, but that these inequalities are sometimes licensed in the name of justice. This, however, is a conclu- sion that Cohen is hesitant to draw. There is one further problem for the prospects of Cohen’s ethical solu- tion that bears mention. Whatever one thinks of Cohen’s objection to pros- titution, it does suggest an important question, namely whether the merely physical “yielding” of the Young Man can truly satisfy the Old Woman in the right way. Although we earlier assumed for the sake of argument that the Old Woman’s preferences accurately tracked her welfare, there are good reasons for believing that the Young Man’s yielding to her would not in the end satisfy her. For an extremely lonely person, any human contact 58 Cohen, Rescuing Justice and Equality, p. 304. 59 Cohen, Rescuing Justice and Equality, p. 305. Here Cohen quotes Rawls approvingly, the idea being that appeals to justice give strength to individuals’ claims that they would otherwise lack. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) 22 DOI 10.1163/17455243-4681035 is often welcomed. Being with the Young Man might, very well then, be able to eliminate some of the negative effects of isolation and alleviate some of the Old Women’s loneliness. She would no doubt be aware, how- ever, that the Young Man did not love her or desire to be with her in any way except insofar as his sense of equality required it. Peter Railton, in a similar context, describes this type of situation as a kind of alienation.60 In his example, a husband performs all the duties that a loving husband would perform, but his reasons for acting the way that he does is his devotion to morality, not, strictly speaking, his dedication to his wife. As Railton suggests, “something seems wrong”—after all, one doesn’t only want to be treated as a means of discharging moral duties or as a means of increasing the amount of well-being in the world.61 Imagine a son telling his mother upon visiting her at her home that he is visiting her because she is old and lonely and morality requires that the lonely be vis- ited periodically. Or imagine a wife embracing her husband, not because she has any particular affection for him but rather because she knows that her husband’s well-being will improve if he is embraced. Surely, both of these situations, along with our example borrowed from Aristophanes seem perverse in the same way that Cohen thinks prostitution is perverse. Love and affection are given freely for reasons of love and affection, not for reasons of morality or equality. The problem, we have argued, is that however strong the Young Man’s commitment to equality is, that alone will not allow him to satisfy the demands of justice. In the trilemma case, Cohen argues that someone with an egalitarian ethos acts without coercion when they act in accordance with equality and so, freedom can be consistent with equality. The egalitar- ian ethos alone, however, will not achieve equality in the Aristophanes case because acting from an egalitarian ethos provides the wrong kind of reason to the Young Man. The Old Woman needs the Young Man to love or desire her for the right reasons, and an egalitarian ethos does not provide those reasons to the Young Man, no matter how committed he is to it. Reasons of equality are dictated by how closely they track distributive desiderata. Reasons of love, on the other hand, come from the deeply held desires that are directed at the object of love. To truly achieve equality in our example, we would need to change the desires of the Young Man, but 60 Peter Railton, ‘Alienation, Consequentialism, and the Demands of Morality,’ Philosophy and Public Affairs, 13 (1984), pp. 134–171, at pp. 135–137. 61 Railton, ‘Alienation, Consequentialism, and the Demands of Morality’, pp. 134–171, at p. 136. J. Thrasher and K. Hankins / Journal of Moral Philosophy (2013) DOI 10.1163/17455243-4681035 23 this might not be possible. Further, even if it were possible, as our discus- sion of personal prerogatives has suggested, it may not be desirable for, as Cohen himself affirms, justice to not be all encompassing in this way. We should not be slaves to social justice. Cohen may be right that justice does not license the degree of inequality that many others have supposed. He may also be right that justice requires us to adopt an egalitarian ethos. Our suggestion here is simply that even if we grant Cohen’s claim that adopting an egalitarian ethos may help to alle- viate the tension between co-realizing the values of freedom, equality, and Pareto, we need to do more to fully dissolve that tension. It is not only minds that need to be changed, but also hearts, and even that may not be enough. Finally, although we have levied a series of objections against Cohen’s proposed solution to the trilemma argument, it is worth noting that the failure of Cohen’s Ethical Solution does not present the sort of problem that Cohen supposes. While we might still lament inequality, if justice is about more than just equality, licensing inequalities in the name of justice need not be self-effacing. On the contrary, sometimes constraining the reach of justice serves as a way of respecting the separate worth and importance of individual lives.62 62 The authors would like to thank Jerry Gaus, Bill Glod, Guido Pincione, David Schmidtz, Kevin Vallier, Steven Wall, and an anonymous reviewer for helpful comments on earlier drafts. Special thanks to the participants in a seminar on egalitarianism at the University of Arizona, especially to Tom Christiano who organized and ran the seminar. His early interest in the project was extremely important. When Justice Demands Inequality