Journal of Political Philosophy

One of the crucial questions that lawyers, philosophers, politicians, and journalists struggled with during the twentieth century was how to “guarantee that informative and accurate news would flow to the public through the press.”1 Traditional answers to this question assumed that the key to a well-informed citizenry lay within speech rights. The idea was that speech rights would create a rich flow of information from which diligent citizens could learn the important facts and form their own views about public issues. However, in digital democracies, speech rights are very well entrenched, yet many people are still largely uninformed. To be sure, ignorance is sometimes the result of negligence, but it is undeniable that citizens are often the victims of disinformation campaigns, fake news, and personalized online propaganda. These phenomena make it difficult to understand public issues even if one is disposed to do so. And, importantly, they seem to confirm what scholars like Lebovic himself—but also Lippmann or Habermas—have lamented: speech rights are not enough to guarantee that the public receives an adequate supply of news. The traditional answer to Lebovic's question is, then, at least partially incorrect. Drawing on these insights, this article shifts the focus from speech rights towards the rights of the public, and argues that, if we want to guarantee that an adequate supply of news reaches the public, we need to start taking more seriously the idea that citizens have a right to be well informed. As I will show, this idea repeatedly appears in journalism theory and practice, as well as in democratic and legal theory, but it remains a somewhat vague notion: it is not clear what it might mean, nor what its normative implications could be. This article aims to fill this conceptual gap by conceptualizing what I will call citizens' right to information (henceforth CRI). One motivation for theorizing CRI is to show that a positive right of the citizenry to be provided with quality news is—contrary to appearances—fully compatible with free speech and journalistic freedoms. The key here lies in conceiving this as a moral right which imposes moral—but not legal—obligations on journalists while, at the same time, compelling the state to foster quality journalism through adequate media policies. Seen in this light, CRI may become a useful landmark whose conceptual clarification could illuminate other discussions, such as journalism ethics. Indeed, it seems that journalists' special democratic rights and duties could hardly be explained without assuming something like CRI. Moreover, CRI might illuminate the discussion on what journalism can mean in our digital era—in which the line separating professional journalists from lay citizens is increasingly blurred—by linking the profession to an ethical commitment to provide citizens with the information they, as such, have a right to. Finally, the concept of CRI might also be useful to guide the discussion of alternative models for quality journalism, which has been recently sparked by the collapse of the traditional advertisement-based funding system. After all, such discussion—just as that of media policy, including Lebovic's question—could be seen as a debate on what is the best way to provide citizens the informational service they have a right to. Of course, all these discussions can be only superficially addressed here. But by connecting them to CRI, this article will offer a perspective to look at them afresh. The article is structured as follows. Section II lays the groundwork for a democratic right to be well informed, by reviewing previous appearances of this idea, and by contextualizing it within the map of communication rights. Section III offers a systematic framework for justifying and conceptualizing CRI. After showing four possible ways to justify this right, I proceed to spell out its four main characteristics. First, its function is to ensure that citizens are provided with a good informational service that enables them to become well informed and therefore competent political decision-makers. Second, and consequently, its right-holders are all those with participatory rights. Third, its content is not what people want to know, or any kind of information about politics, but specifically information that is useful for updating political knowledge (that is, democratically relevant information). Finally, its two main correlative duties apply, respectively, to journalists and to the state. These two last characteristics—the content and the correlative duties of CRI—are highly controversial: any complete answer to them requires committing to contested normative claims about democracy. My strategy for navigating these difficulties is to draw a distinction between the general concept of CRI and its different conceptions. The general concept, I contend, concerns the democratic, moral, and positive right of citizens to be offered democratically relevant information. As I will argue, accepting this general concept does not require endorsing any contested claims about democracy. It is only when we try to fine-tune the content or the correlative duties of CRI that we need to take sides on issues upon which disagreement abounds. Each one of these specific, but controversial, definitions of CRI constitutes one plausible conception of this right. This article, I have to say, will not identify—much less defend—any of these conceptions. Rather, its purpose is just to lay out an analytical framework for further discussion, about both the concept and the different conceptions of CRI. Despite being incomplete, this analysis will be enough to distinguish the right to be well informed from two other rights with which it is often confounded: freedom of information and the public's right to know. The idea that citizens have a right to receive democratically relevant information is well established in democratic societies. It repeatedly appears in journalism theory and practice, as well as in democratic and legal theory, albeit under different names. In this section, I first offer a brief overview of these appearances, which—despite using different nomenclature—share the intuition that citizens have a right to be well informed. I then locate this idea on the map of communication rights, defining it as a right of the public, rather than as a speech right. The field of journalism is where the idea of a citizens' right to information has had the most clear and persistent influence. Journalists often justify their more controversial activities, such as whistleblowing or intruding on individuals' privacy, by appealing to citizens' interest in being—or their right to be—well informed. Such journalistic practice is supported by most professional codes of ethics, which often present public service as the profession's main value. What is more, empirical research on journalists' self-perception shows that they usually think of themselves primarily as public servants who work for citizens, rather than for their actual employers, which are often private companies.2 Journalism theory echoes this idea. Scholars tend to agree that the profession's core element is its ethical commitment to informing citizens of what they need to know to become competent political decision-makers, rather than of what they are merely curious about, or have a private interest in knowing.3 As Elliot and Ozar point out, journalism is not usually conceived as serving mere audiences, but rather as providing a service to the entire political community: that is, to the collection of individuals who deal with public matters. Thus, unlike other professionals such as doctors or teachers, journalists are not committed to serving particular individuals or specific groups of people, but rather the citizenry at large.4 This explains why, according to most theorists, journalists ought to treat their audiences as citizens rather than as consumers. What is probably the most famous formulation of journalism's basic tenet was coined by Kovach and Rosenstiel, who claim: “The primary purpose of journalism is to provide citizens with the information they need to be free and self-governing.”5 Often, this predominant view of journalism explicitly invokes the idea of a right, as when Pamela Taylor Jackson claims that “[j]ournalism represents a right to public interest news and information.”6 Legal scholars and rights theorists have also entertained the idea of a citizen's right to information. Sam Lebovic, for example, refers to “the right to the news” as the right we need to protect if we are to solve the problem of how to adequately inform the public.7 More recently, Rowan Cruft has appealed to a human right “to education, specifically about current matters of public interest”.8 Although his approach to this right is incidental (as he only invokes it in order to ground journalism as a human right in itself), Cruft makes clear that, for him, citizens have a right to be well informed. In fact, he claims that journalists' rights and duties “correlate with human rights to the provision of (good) journalism.”9 Before Lebovic and Cruft appealed respectively to the right to the news and to the right to good journalism, free speech theorist Owen Fiss defended what he calls the “democratic theory of the First Amendment”, which similarly appeals to “the public's right to know and to be informed”.10 Legal scholar Stephen Sedley, for his part, has lamented that despite juridical rhetoric appeals to the public's interests, free speech and free press practice often focuses exclusively on protecting communicators' interests. Along similar lines, Sedley holds that whenever something is worth knowing, the interested audience may have a right to receive such information, even if they are unaware of that information's existence.11 One might argue, then, that just as patients have a right to be informed by their doctor, citizens have a right to be informed by journalists. Note that in both cases the right-holders are unaware of what information they have a right to be provided with. Legal scholar Mark Bovens also invoked citizens' right to be well informed when he argued that twenty-first-century citizens need new information rights. Among these, what he calls “tertiary information rights” would regulate horizontal relations between citizens and other private entities, which might be required to disclose information of crucial importance for citizens qua citizens. That information, which “should remain reasonably accessible to each and every citizen,” may include things such as art collections, national events, sports, or cultural manifestations, but—importantly—also “news items”.12 Democratic theory has also long held the idea that citizens should be properly informed.13 Of course, democratic theorists disagree on the quantity and quality of citizens' informational needs.14 However, they agree that for democracy to work properly, citizens need to be provided with quality information about certain relevant issues. Jürgen Habermas and Robert Dahl, two champions of democratic theory, might be taken as illustrative examples of this point. In Habermas's deliberative conception of democracy, decisions are only legitimate if they are the result of a rational discussion through which citizens transform their “mere opinions” into “reflected public opinions”, or—as he calls them elsewhere—“considered public opinions”.15 For citizens to acquire these public opinions, Habermas argues, professional journalists should offer them the news, reports, and commentaries they need as citizens.16 As he puts it, “audiences are not only consumers, that is, market participants, but also citizens who have a right to partake in culture, to follow political events, and to be involved in the formation of political opinions.”17 These considerations lead Habermas both to criticize news media for wasting people's attention on irrelevant issues, and to defend interventionist media policies.18 In a similar vein, Dahl finds it unquestionable that, to be competent, citizens need political and social institutions that help them become so. For Dahl, “[o]pportunities to gain an enlightened understanding of public matters is not just part of the definition of democracy. They are a requirement for democracy”.19 Indeed, according to Dahl, one of the fundamental principles of democracy is precisely “enlightened understanding”, which states that “each member must have equal and effective opportunities for learning about the relevant alternative policies and their likely consequences”.20 Enlightened understanding would be satisfied, in principle, by three basic institutions of democracy: freedom of expression, freedom of association, and the existence of alternative sources of information.21 Dahl, however, is aware that these institutions, created during the nineteenth and twentieth centuries, might no longer be an efficient way to promote enlightened understanding in the twenty-first century. “If this is so,” he adds, “then democratic countries will need to create new institutions to supplement the old ones.”22 Dahl here opens the door for deep structural reforms that would adapt our democracies to the specific challenges that our present century poses to relevant information-gathering processes. Remarkably, at the very beginning of the digital revolution, Dahl foresaw the need for new institutions to tackle what would eventually become known as “information disorder”: the huge amount of available information, much of which is indifferent or even detrimental to enlightened understanding.23 Given the increased availability of information, he contends, democratic countries must “improve citizens' capacities to engage intelligently in political life,” which would require enhancing old institutions with “new means for civic education, political participation, information, and deliberation.”24 The assumption that, in a democracy, citizens need to be well informed, leads both Dahl and Habermas to struggle with Lebovic's question about how to guarantee that informative and accurate news flows to the public. Neither of them gives a clear answer to this question, but both are keen to accept institutional reforms, precisely on the basis of what might be called citizens' right to be well informed. Indeed, regardless of whether they come from the field of journalism or from the theories of law, rights, or democracy, all previous reflections point towards a right of citizens to be offered the information they need to become well informed. This right is not always explicitly invoked, and when it is invoked, it receives different names—the “right to good journalism”, “right to the news”, or simply “right to information”. Behind this apparent diversity, however, there is a shared intuition: unless citizens are aided in their information-gathering processes, they will not be able to become well informed. And if that happens, the intuition goes, we cannot expect democracy to work well. As James Madison wrote, “[a] popular Government, without popular information, or the means of acquiring it, is but a Prologue to a Farce or a Tragedy; or, perhaps both.”25 Hence the belief that citizens must be offered the information they need to become well informed. It is not fully clear, though, what a democratic right to be well informed might mean or imply, for no sufficiently systematic analysis of this right has yet been carried out. However, previous appearances of this idea offer illuminating insights. For instance, by affirming that democracy cannot work well unless citizens are offered quality news, they point towards its grounds. When they suggest that the information that matters is not what one might want to know, but what one, as a citizen, needs to know, they offer a crucial hint about its content. Similarly, when authors like Cruft, Dahl, or Habermas identify plausible implications that might follow from this right, they offer clues about which duties might correlate to it. In the remainder of this article, I will draw from these insights, and I will lay out a systematic framework for both justifying and analyzing this right, which I will call citizens' right to information (CRI). My first step will be to locate CRI on the map of communication rights. To make sense of the right to be well informed, the first step is to understand its place in what Emerson calls the “whole system of freedom of expression”: that is, the set of rights that regulate the exchange of information, opinions, and ideas.26 According to Emerson, this system includes two sides: the right to communicate and the right to know, each of which encompasses several further rights.27 Inspired by Emerson, but departing from his terminology, here I propose to classify communication rights into the following two categories, each defined by the role the right-holder plays in the communication process. The first category of communication rights consists of speech rights. These are rights whose right-holders are conceived as the emitters of information, opinions, and ideas—that is, those who send or transmit most of the content in the communication regulated by the right. Among these rights, we may include rights such as freedom of expression, freedom of the press, and the right to communicate. The second category of communication rights consists of rights of the public. These are rights whose right-holders are conceived as the addressees of content—that is, the audience or the public who receives the information, opinions, or ideas. It is this latter category to which CRI belongs, and on which I will therefore focus.28 To better illustrate the distinctiveness of CRI, let me first briefly present the two best-known rights of the public, which will serve as points of comparison. The first of these rights is freedom of information (FOI). Strictly speaking, FOI is not a freedom, but rather a claim-right to access documents and information held by public institutions. Its name is owed to the influential US Freedom of Information Act, which since 1966 has inspired legislation on institutional transparency around the world.29 The other best-known right of the public is the public's right to know (PRK). This right is often conceived as a general and unqualified entitlement to information that is of interest to the right-holder. It includes, as Emerson puts it, “[f]irst, the right to read, to listen, to see, and to otherwise receive communications; and second, the right to obtain information as a basis for transmitting ideas or facts to others”.30 The fact that the public's right to know is of such a “seemingly unlimited nature”31 might explain the pervasive disagreements about what this right might actually mean; Maciejewski and Ozar, for instance, have identified twelve different interpretations.32 To simplify things, I will distinguish only two different, but interconnected, meanings usually attributed—often without sufficient explanation—to the phrase “the public's right to know”. In its wide sense, “the public's right to know” is used for the whole category of rights of the public, and might thus be taken as referring to a principle or a very general right encompassing all the other more specific rights of the public. When used in this wide sense, “the public's right to know” is often indistinctly employed to also name specific rights of the public within it, such as FOI or, as we saw in Section II.I CRI, which are conceived as specific instances of it.33 To avoid possible misunderstandings deriving from this usage, in the following I will refrain from using the term in this wide sense. Instead, I will use “rights of the public” to name the set of rights whose right-holders are conceived as an audience. I will speak of the public's right to know (PRK) only in the narrow sense of the term. In this sense, the PRK is the entitlement to search, access, and receive the information needed to develop and exercise one's conception of the good. In this narrow sense, the PRK is just one specific right of the public among others, such as FOI and CRI. Even so, the PRK remains—even in this narrow sense—rather a wide right, at least in comparison to other rights of the public. This will become clearer when I address the question of content. So far, I have reviewed previous appearances of the idea of a democratic right to be well informed. I have also identified the side of the system of communication rights where it belongs, alongside other rights of the public such as FOI and the PRK. However, this is still insufficient for clarifying what a democratic right to be well informed could actually mean. In the following section, I will take the first steps towards such a clarification. Against the backdrop of the previous section, in the remainder of this article I will put forward a systematic framework for both justifying and analyzing CRI. This framework comprises five features: grounds, function, right-holders, content, and correlative duties. In spelling out these traits, I will compare CRI to other communication rights, which will help not only to better clarify it, but also to shed light on its difference from these other rights—further emphasizing the existing conceptual gap within communication rights. In this subsection, I will present two ways in which CRI might be justified. In both cases I will first indicate a general way to justify rights. Then I will apply this strategy to the case of CRI. The first strategy for justifying a right is to show that there are interests important enough to hold someone to certain duties which, if observed, would protect those interests.34 An important caveat is in order here: rights do not only protect their right-holders' interests. Indeed, sometimes the most important interest served by a right is the interest of a third party, or the common interest of society as a whole, which would be indirectly protected by the direct protection of the right-holder's interest.35 On this view, for instance, the right to parental leave would be mostly grounded not in any interest of parents, but rather in newborns' interest in being cared for, or in the common interest of a society in its new generations being cared for. Similarly, journalists' right to withhold their sources of information would not rest so much on journalists' interests in keeping their sources secret, but rather on “the interest of all in the free circulation of information which is of public interest,” or on the “usefulness [of this right] to members of the public at large”.36 As these quotes show, communication rights are not alien to the complex relation linking rights and interests. Scanlon shed some light on this intricate relation when he famously argued that freedom of expression protects three kinds of interests. According to him, freedom of expression protects the interests of its right-holders—or, as he calls them, the “participants”—but it also protects the interests of those to whom content is addressed—that is, the “audience”—and of those who might be indirectly affected by the exchange of content—that is, “bystanders”.37 Keeping these complexities in mind, CRI can be justified by appealing to at least two important interests, one of the right-holders, and another of third parties. The first is the interest that every citizen has in becoming well informed, in order to be able to competently advance her own political views. In short, citizens need informative and truthful news in order to understand what political option is best for them. Most references to CRI—some of which I mentioned above—appeal to this individual interest. The second grounding interest of CRI is the interest that every citizen has in her fellows being well informed, so that they do not advance harmful or unjust policies. Note that this line of justification grounds CRI not in any interest of the right-holders themselves, but rather in one of third parties: each citizen's right is grounded in the interest that each of her fellow citizens has in her being well informed. This interest that we have vis-à-vis other citizens arises from the fact that in democracy we face the consequences of choices which, despite their ability to deeply affect our rights and well-being, are promoted by a body of citizens on whom our individual influence is fairly insignificant. Given our vulnerability before the body politic, every one of us has a strong interest in each of our fellows becoming well informed, so that they make sensible choices and avoid promoting harmful or unjust policies. The second general way of justifying rights does not appeal to interests, at least not directly. Instead, it appeals to other rights. This process—known as linkage argumentation—consists in showing that the proper enjoyment of an already justified right is impossible unless another right is granted too. This instrumental right is then justified as a necessary—or at least a very important—condition for the meaningful enjoyment of another right.38 CRI could be justified along these lines by at least two linkage arguments, as follows. The first starts from the assumption that we have a right to live in democracy, and then argues that having a right to be provided with information about public issues is necessary—or at least very important—for preserving democracy. It is hard to see how democracy could function properly, or even how it could last, if citizens were unable to access a supply of informative and truthful news. A largely uninformed citizenry, unprotected against misinformation and fake news, might easily fall prey to demagogues and end up favoring policies that undermine basic democratic values. What is more, manipulated citizens might easily be misled into choosing undemocratic leaders who attempt to overthrow democracy by subverting the institutional order. An unaware citizenry might not detect these subversive tactics until it is too late, and for that reason, if we have a right to live in a democracy, then we also need a democratic right to be well informed. The second linkage argument derives CRI from political rights, such as the rights to vote and to demonstrate.39 The core of this argument is that political decisions cannot be considered autonomous—and therefore the exercise of political rights cannot be meaningful—unless citizens are given a fair chance to inform themselves and reflect on what is best. The value of a choice resides in its expressing the autonomy of an agent who chooses freely. As Scanlon says, “lack of information (or false belief) can render a choice involuntary in the relevant sense, undermining its legitimating force”.40 For this reason, a meaningful exercise of political rights requires that citizens can easily access an adequate supply of informative news—just as CRI states—so that their decisions express their autonomous choices rather than manufactured or ill-formed preferences. As Levitsky and Ziblatt claim, “[c]itizens have a basic right to information in a democracy,” because “[w]ithout credible information about what our elected leaders do, we cannot effectively exercise our right to vote.”41 Of course, having a right to be well informed does not entail becoming well informed. It merely entails a fair chance to do so. But that fair chance makes a crucial difference, for an uninformed vote has a very different value depending on whether the voter could or could not have informed herself. In sum, CRI can be justified in at least four different ways, depending on whether one appeals to its (individual or collective) grounding interests, to the right to live in a democracy, or to political rights. A complete exploration of these lines of justification, their problems, and their compatibility, remains to be done. For now, let's focus on the main traits of CRI. As can be inferred from its previous appearances in different disciplines, the function of CRI is to promote the ideal of a well-informed citizenry as a means to further a well-functioning democracy. As a right of the public, its specific purpose is to guarantee that citizens are provided with the information that they need to become competent enough to carry out their democratic duties and make well-grounded political decisions. This suggests that CRI resembles political rights or, more specifically, democratic rights. Political rights are those which are constitutive of a specific form of government, in that they make it possible. Democratic rights are the subgroup of political rights that make democratic government possible. The right to vote, freedom of assembly, and freedom of the press are usually included among them.42 Given the democratic function served by CRI, we have good reasons to include it within the category of democratic rights. To further clarify the specificity of this function, let me contrast it with the functions served by the other two best-known rights of the public. The function of FOI is to promote transparency: that is, to enable citizens to learn about the administration so that they can understand it and hold it accountable. To be sure, transparency is a necessary condition for having a well-informed citizenry: if citizens could not learn about the administration, then they could hardly be competent political decision-makers. However, CRI does not aim to create a transparent administration. Instead, it aims to promote knowledge about the issues that are worth knowing for citizens qua citizens, regardless of whether those issues impinge upon the actions of the administration. This function presupposes that information about public institutions is generally available, but it is conceptually different from the function of promoting transparency. Indeed, it goes beyond that. Let me illustrate this difference with an example. During the Covid crisis, FOI gave citizens a right to access official do

In his article, “The Space between Justice and Legitimacy”, Kit Wellman offers a novel account of the relationship between political philosophy's two central concepts.1 He argues that states can be legitimate yet impose many unjust laws and policies. This is true, he suggests, because political legitimacy should be understood as a claim about wide proportionality.2 Just as a country's war can be widely proportionate overall, and yet contain many instances of injustice, states can be proportionate (that is, legitimate) overall, yet contain a fair amount of injustice. But although Wellman thinks legitimate injustice is, in this sense, a real and pervasive phenomenon, he goes on to argue that legitimacy doesn't have much practical relevance with regard to unjust laws and policies. It doesn't make it permissible to impose and enforce unjust laws, nor does it generate an obligation to comply with unjust laws. Like Wellman, I think legitimate injustice is real and pervasive. But I don't share his view of this phenomenon and, as a result, I also don't agree with him about the practical implications of legitimate injustice. Contra Wellman, I think state officials can act permissibly when they enforce unjust but legitimate law, and I think they have claim rights against being interfered with when they do so. Before we begin, it will help to clarify what's at stake in labeling a state legitimate. There are many competing conceptions of political legitimacy in the literature, but, as Wellman says, “virtually everyone agrees that legitimacy at least entitles a state to coerce its constituents”.3 Following Wellman, I will assume this is, at a minimum, what's at stake in determining whether a state is legitimate. Other things being equal, a legitimate political authority is presumptively permitted to coerce its constituents in at least some ways that illegitimate authorities are not permitted to do. Wellman believes legitimate states can act unjustly, and he defends this view by appeal to a particular notion of proportionality. Just as it can be morally permissible to launch a war even when we can foresee that some of our troops will commit unjust crimes during its course, a state can be legitimate even though we know it sometimes commits injustice. In both cases, we weigh the good things that will be achieved if we proceed against all the harms or costs that will be caused. Provided the goods are sufficiently great to outweigh the harms or costs, the proposed course of action or the institution can be defended on the basis that the benefits are proportionate relative to the costs.4 As Wellman puts it, “Because states perform such incredibly valuable functions … they are worth at least some moral costs”.5 This is, on Wellman's view, simply what it is for a state to be legitimate. When the benefits the state provides are proportionate relative to the costs (including the injustices the state commits), the state is morally legitimate. This is an innovative way of thinking about the relationship between justice and political legitimacy. But I do not think it can be correct. In the ethics of self-defense and war, the standard way to determine whether some act is proportionate is to weigh the costs and benefits that will be caused by our action against the alternative of doing nothing.6 Suppose the only way to defend one innocent person from having her legs unjustly broken is by throwing a grenade that will kill two innocent bystanders. The act of throwing the grenade is not proportionate, because the costs of doing so (two innocent people are killed) are much worse than the alternative of doing nothing (one innocent person's legs get broken). If, however, these costs were reversed—if doing nothing meant two innocent people get killed, whereas throwing the grenade meant only one innocent person's legs get broken—then we might conclude throwing the grenade is proportionate. If we are to use wide proportionality to assess a state's legitimacy, we need to engage in a similar comparison. The difficulty, however, is that the state is unlike a harmful act that we contemplate performing, and so it's less clear what it might mean to ask, “are the costs created by the state proportionate relative to the alternative of doing nothing?”. How should we understand the benchmark of “doing nothing”? To see why I remain convinced that the gap between justice and legitimacy exists, let us return to the contrast between a slave-owner who forces slaves to build a pyramid and the prime minister of Norway. If I were the slave-owner, there is no question that I would be morally required to immediately free all of my slaves, even if there is no other way for me to build the pyramid. If I were the prime minister of Norway with the extraordinary power to unilaterally dismantle this government, however, I would not feel obligated to do so.7 Dismantle. A state is proportionate, and thus legitimate, when we would not be morally required to dismantle the government in order to prevent or halt the injustices the state commits. But this doesn't look like a good test. First, it's not clear what we are supposed to imagine occurs if the government or the state is dismantled. Maybe we should consider what would realistically occur in this particular country if its government or state apparatus suddenly ceased to exist. In many cases, what would occur might be terrible. In some places, civil war or ethnic cleansing might be the consequence of suddenly dismantling a government. In these cases, we would not be morally required to disband the government even if it were very, very bad. Even a brutal authoritarian regime might be preferable to civil war or ethnic cleansing, but surely that does not make a brutal authoritarian regime morally legitimate; it does not give that regime the moral permission to coerce its citizens to do things that they are not independently required to do. A regime is not morally legitimate merely because it is preferable to some of the worst conditions on earth.8 One problem with trying to conceptualize political legitimacy as a species of proportionality is that we almost never face a binary choice between our current political institutions as they are or else the total absence of those political institutions. Consider the Southern United States during the Jim Crow era. If the only two options were (1) retain the existing racist laws and institutions, or (2) a descent into an extremely dangerous form of anarchy, it's at least possible that option (1) might be widely proportionate relative to the benchmark of (2). But, of course, those weren't the only two options. Radically reforming the existing laws and institutions to make them less racist and unjust was a feasible option, and the very fact that it was a feasible option is surely part of the explanation as to why the racist political institutions might not have been morally legitimate. the appropriate baseline for testing proportionality must be the feasible alternatives rather than simply doing nothing at all. Just as a military intervention would not be proportionate if there were an equally effective and less costly diplomatic option available, a brutal dictator cannot justify the injustices of her regime if a more just political arrangement is within reach.9 But Wellman's statement here is puzzling for two reasons. First, this isn't how proportionality has standardly been conceptualized in the ethics of self-defense and war. As I've already explained, proportionality judgments are typically made by comparing a given act of harm-imposition relative to the benchmark of doing nothing. The entire set of feasible alternatives, however, is relevant with regard to a different moral judgment, namely, necessity. Philosophers working on self-defense and war typically claim that for an instance of harm-imposition to be morally permissible it must meet the condition of being necessary. Although there is a great deal of disagreement about how exactly to conceptualize the necessity condition, there is widespread agreement that judgments of necessity are made by comparing the potential act of harm-imposition to all the feasible alternatives.10 Killing a wrongful aggressor, for example, is not necessary if one of the defender's feasible alternatives is painlessly rendering the aggressor unconscious. Thus, although it's a matter of some dispute, on the most widely accepted picture, wide proportionality and necessity are two separate constraints on the permissible use of force. To be morally permissible, it's generally believed a harmful act must, among other things, satisfy both these constraints. All Alternatives. A state is proportionate, and thus legitimate, only when the net balance of benefits and costs it currently provides is proportionate relative to every feasible alternative. This test isn't helpful if we seek to explain how any existing state could ever be legitimate. It's surely true that every existing state has at least some unjust laws or unjust institutions which could be reformed to eliminate the injustice without becoming more unjust in other ways. It's thus true that, for every existing state, the status quo is not proportionate relative to a feasible alternative, and thus every existing state is illegitimate. The problem is even more stark, since this view has the implication that even hypothetical states that are far more just than any existing state would still be illegitimate so long as they contain even a single unjust law or institution that could feasibly be reformed. I think it's clear that these are not the conclusions Wellman seeks to defend. Wellman might reply that it's not feasible for a state to be entirely devoid of unjust laws or institutions; realistically, our best efforts to design and sustain political institutions will always contain some unjust features. But this reply involves shifting from the standard of “feasible” to the standard of “probable” or “likely”. Although the term feasible is contested,11 I suspect it's most commonly used to refer to whether achieving a given outcome is possible given our resources and technological limitations. In this sense, changing the laws regulating gun ownership in the US is clearly feasible. Of course, it is extremely unlikely that robust gun control legislation will be enacted in the US for reasons of ideology and political self-interest. Suppose, then, we take All Alternatives to refer only to those alternatives that are sufficiently likely to obtain. This would then vastly lower the bar for legitimacy. Even highly unjust states can be legitimate so long as it is sufficiently unlikely that they will become more just. The problem with this interpretation, however, is fairly obvious. It's often the case that laws or institutions are unlikely to be reformed because those who hold power don't want to reform them. It might, for example, be very unlikely that we can reform the unjust institutions in a country because the country is ruled by a tyrant who refuses to relinquish power and rules by violently suppressing dissent. On this interpretation, the tyrant can render his rule legitimate by forming a settled disposition to hang on to power through unjust means. I take this result to be a reductio of the proposed interpretation of the proportionality test. In sum, I don't see how the idea of wide proportionality can be made to work as an account of political legitimacy. One test for proportionality that Wellman seems to suggest—Dismantle—is far too weak a standard. It's too weak, in part, because it ignores the fact that we never face a binary choice between the status quo and dismantling the state. We always have many other feasible alternatives, which include reforming the unjust laws and institutions of our state. But the other test that Wellman floats—All Alternatives—creates either an unacceptably high standard or an unacceptably low one. If we focus on all feasible alternatives, then the bar for legitimacy is far too high—no existing state meets it and no hypothetical state with even a single unjust law meets it. But if we focus on likely or probable alternatives, then the bar is too low: very bad actors can render themselves legitimate by refusing to meet minimal moral standards they could meet if they wanted to. In each case Wellman argues that the answer is “no”. So, although legitimate states act unjustly, the fact that these injustices are perpetrated by legitimate states doesn't have many interesting practical consequences for the unjust laws the state attempts to impose. The state's legitimacy doesn't give it a special permission to act unjustly, it doesn't ground a duty to obey unjust laws, and it doesn't give the state a protected right to act unjustly. I disagree with Wellman on all three questions, but here I will focus only on the first. I think we should expect there to be a great deal of reasonable disagreement about justice. By that I mean that smart, sensible people, thinking clearly, who aren't driven by self-interest or prejudice, and who are committed to trying to treat others fairly, can and will disagree about many matters of justice, including difficult questions such as “What rates of income tax are required as a matter of justice?” or “Should freedom of expression extend to protect those who deny the Holocaust in order to promote anti-Semitism?”. But the fact that such people will disagree about the requirements of justice doesn't imply that there are no correct answers to these questions or that the answers are indeterminate—there may sometimes be correct answers in cases of reasonable disagreement. I also think that in cases of reasonable disagreement about what justice requires, a legitimate state should have some sort of democratic process to resolve the reasonable disagreement. Indeed, I think people have a claim of justice to an equal (and positive) say in resolving reasonable disagreements about justice. To disregard democratic decisions whenever one thinks they are unjust is a refusal to participate on equal terms in the political process with other reasonable people. To be clear: I'm not suggesting that unreasonable views about justice—those that are, in some sense, beyond the pale, such as the violations of core liberal rights and freedoms—have a claim to be implemented if selected by a democratic process. I'm only suggesting that some disagreements about justice are reasonable—all the competing views represent plausible interpretations of what justice requires—and that in disagreements of this type, there is a claim of justice to have an egalitarian democratic process for resolving the disagreements. Obviously, what I've just said is only the briefest sketch, and many details need to be filled in. But here is one way in which the view that I have sketched is inconsistent with Wellman's answers to the three questions above, in particular the first question: do legitimate states act permissibly when they act unjustly? Suppose, for the sake of argument, that the correct theory of justice includes a very expansive principle of free speech, one that provides a protected claim right against interference to people who engage in hate speech—for example, Nazis who wish to march and chant anti-Semitic slurs in Jewish neighborhoods. But the issue is one over which reasonable people disagree, and after a fair democratic process, our political community enacts a law that prohibits certain forms of hate speech, including the Nazis' anti-Semitic march. I think justice requires giving people equal opportunity to exercise political power in cases of reasonable disagreement. If the government doesn't do what it has been directed to do by a fair democratic process (within the boundaries of the reasonable), it will be acting unjustly—failing to respect what justice requires regarding the distribution of political power. So if state officials decline to enforce the prohibition against the Nazis' march, they will be acting unjustly. However, if the state officials enforce the democratically enacted law prohibiting the march, they act contrary to the principle of free speech, another requirement of justice. So the state will act unjustly whatever it does. If we assume that at least one option from an agent's feasible set must be permissible, then it must be possible for the state to act permissibly and yet unjustly. P5. A fair democratic process sometimes results in the selection of a policy that is reasonable but substantively unjust. Therefore, I'm not sure which premises of this argument Wellman would reject. He might reject P1, but denying it is implausible. It's not credible to suppose that all our disagreements about justice are driven by self-interest, prejudice, or irrationality. I also do not think it is plausible to suppose, as some contemporary Kantians do, that almost all substantive issues of social or distributive justice are largely indeterminate.14 I am sure that many specific questions about the requirements of justice are indeterminate, but this fact is perfectly consistent with P1. P1 makes only the very modest claim that reasonable disagreement is possible even when there are determinate answers about the requirements of justice. I doubt that democracy can do the work we would ask of it here, though, because it is hard to see how voting (or even deliberating, discussing, and voting, if you prefer) on a course of action can magically change its deontic status. If it is impermissible for me to use slaves to build pyramids, for instance, how does it become any more permissible to do so after a democratic process?15 if the sincere but unreasonable beliefs of the majority do not change the moral status of a given course of action, why think that the beliefs of the majority have this moral power as long as the beliefs are not unreasonable? It strikes me as more natural to conclude that the majority has made a reasonable mistake than to suppose that the majority's not unreasonable judgment changes the moral status of a course of action.16 This statement could be taken as a rejection of P2, though Wellman is not explicitly considering P2, so we should tread carefully. In one sense, Wellman is surely correct. The mere fact that reasonable people disagree about whether Φ-ing is permissible surely does not, on its own, affect the deontic status of Φ-ing. But the argument that I've given doesn't move from the mere fact of disagreement to permissibility. Instead it claims that reasonable disagreements about justice require, as a matter of justice, some fair democratic mechanism of resolution. It then moves from this claim to a claim about the permissibility of injustice, via the assumption (P4) that at least one option from an agent's set must be permissible. My argument thus provides the answer to Wellman's rhetorical question: the beliefs of the majority might have this moral power because justice requires a democratic resolution of some of our reasonable disagreements. I worry, though, that it depends on a misunderstanding of the moral value of procedural justice. In my view, justice in both its procedural and substantive incarnations provides side-constraints to be respected, rather than competing values to be promoted … even if an egalitarian democrat's argument establishes that each of us has a right to democratic governance, it at most establishes that acting justly includes an additional side-constraint, but this is very different from showing that procedural justice is a value which sometimes outweighs, overrides, or undercuts the moral imperative to respect nonprocedural rights.17 In this passage Wellman points out something that egalitarian democrats sometimes ignore—the mere fact that there is something valuable about democratic procedures does not suffice, on its own, to show that the scope of the right to democratic governance extends to cover substantively unjust decisions. But the idea that claims of justice operate as side-constraints is notoriously tricky. Most obviously, what happens when two claims of justice conflict? Some proponents of the side-constraint view argue that this is conceptually impossible. Claims of justice, by their very nature, can never come into conflict.18 But this is very difficult to accept; it is sharply at odds with commonsense morality. This view requires either (a) that the content of justice be extremely sparse, so as to eliminate the possibility of conflicts arising, or else (b) a kind of specificationism about justice, such that whenever we have an apparent conflict of claims, we must conclude that in those specific circumstances, only one of the claims turns out to be a valid claim of justice. I think there are compelling reasons to reject both of these options, though I lack the space to lay those out here. It's also worth noting that Wellman doesn't give us any reasons to doubt that claims of justice might sometimes conflict—he just asserts the side-constraint view. But if we accept the commonsense view that claims of justice can sometimes conflict, then we can easily accept P2. And it's worth emphasizing how revisionary and counterintuitive it would be to deny P2. To deny P2 would mean that there need be nothing unjust about stuffing ballot boxes or suppressing other people's votes, provided doing so is a way of ensuring that only truly just policies are enacted. P3 seems uncontroversial if P2 has been accepted. P4 is not uncontroversial: a minority of philosophers claims that in some situations an agent will act wrongly regardless of what she does. Entering this debate is beyond the scope of this article, so I'll simply say that, along with many others, I think P4 is true: there must be a permissible option in an agent's set of choices.19 And P5 does not seem controversial if P1 is accepted. And so I disagree with Wellman: I think a legitimate state and its officials can sometimes act permissibly even though they act unjustly. I also disagree with Wellman about the second and third questions, but I won't pursue those disagreements here. The phenomenon of legitimate injustice is deeply puzzling, and I share the skepticism that Wellman expresses about the capacity of some recent work in political philosophy to provide an adequate account of this phenomenon. But we can't, I have argued, understand legitimate injustice as an instance of harms that are proportionate in light of all the good consequences that a state delivers. We should instead understand legitimate injustice as something that arises as a result of the complex and multi-dimensional nature of justice itself. There are principles of justice that apply to the distribution of political power in cases of reasonable disagreement. When these principles conflict with other substantive principles of justice, the phenomenon of legitimate injustice arises. This is because, in such situations, the state and its officials cannot avoid injustice. This fact, contra Wellman, has significant practical implications for permissible conduct in politics. On my view, unlike Wellman's, the moral legitimacy of a law does real practical work—it can render permissible conduct that would otherwise be impermissible, and it can explain why public officials have claim rights against interference even in some cases where they threaten to infringe claims of justice. Nothing I've said here constitutes an argument for the view that a principle of justice concerning the distribution of political power should ever take precedence over more substantive claims of justice. I do think this view is correct, but this isn't the place to present that argument. For comments and discussion, I am very grateful to Peter de Marneffe, Rebecca Stone, Paul Weithman, Kit Wellman, Leif Wenar, and an anonymous referee. None relevant. There are no potential conflicts of interest relevant to this article. The author declares human ethics approval was not needed for this study. All relevant data are included in the article.

Recent history has seen a rapid growth in the involvement of private parties in war conflicts.In 2020, there were almost twice as many private contractors as US soldiers in Afghanistan. 1 In the ongoing war in Ukraine, private actors are allegedly deployed by both parties in the conflict. 2Originally hired by states to provide support services from catering to logistics, private military firms (PMFs) have progressively taken on functions, including combat tasks, that were deemed, at least in the last century, inherently governmental. 3 The phenomenon amounts to an unprecedented form of corporatized mercenarism. 4 The condemnation of mercenarism has an illustrious history.While Machiavelli famously deprecated mercenaries' lack of loyalty and tendency to corrupt the state, 5 Rousseau worried that hiring mercenaries, rather than having citizens fight wars, would lead the latter to value comfort more than republican freedom.Recent critics argue, among other things, that fighting for profit is inherently wrong; 6 that the privatization of war leads to an unjust distribution of access to security; 7 that it allows both states and private parties to escape democratic accountability; 8 and that it provides incentives to escalate conflicts and to increase the use of violence in the battlefield. 9Some contemporary philosophers have, on the other side, shown a friendlier face towards mercenarism.Most prominently, Cécile Fabre argues that, at least under ideal circumstances, private parties have a right to sell their soldiering 1 Bilmes 2021. 2

It is good for people to flourish. But does the state have the authority to promote the flourishing of its citizens? Some political philosophers—perfectionists—hold that it does.1 For perfectionists, the state has the authority to pursue policies meant to promote the flourishing of its citizens, and it is appropriate for the state, or state officials, to take considerations about what will promote flourishing into account when exercising their authority. Traditionally for perfectionists, the fact that a policy will promote the flourishing of the citizens may legitimate the state's using its authority to pursue that policy, even if the policy does not promote any non-perfectionist aims— that is, aims other than flourishing. Other political philosophers—anti-perfectionists—hold that the state acts illegitimately when it tries to promote flourishing.2 These philosophers hold that the state does not have the authority to try to promote flourishing and that flourishing does not legitimate the extension of state authority to pursue a particular policy. One influential anti-perfectionist argument for this conclusion is that, however good it is for people to flourish, considerations about flourishing are not the appropriate grounds for political authority, and so it is an illegitimate extension of authority contrary to people's rights when the state uses its authority to promote flourishing. Let's call this argument the legitimacy objection to perfectionism.3 How should perfectionists answer this argument? After discussing the legitimacy objection in Section II, I explore one common perfectionist response in Section III—that there are natural or political duties to promote people's flourishing that can ground the state's authority to promote flourishing.4 I am skeptical that responses of this kind are dialectically effective. Rather, in Section IV, I suggest that the appropriate space for perfectionist state action lies in using considerations about flourishing to select among courses of state action that the state may legitimately pursue for non-perfectionist reasons. This consideration-based perfectionism allows officials to take considerations about flourishing seriously, without extending the legitimate scope of state authority, thereby avoiding the legitimacy objection. This shows that there are two distinct roles that flourishing could play within political life. On the one hand, flourishing could be something that the state is sensitive to in making its decisions. On the other hand, flourishing could be something that partially grounds the domain of legitimate state authority. The ability of consideration-based perfectionism to avoid the legitimacy objection shows how these roles can come apart—the state can be sensitive to flourishing even if flourishing does not partially ground the domain of legitimate state authority. Proponents of the legitimacy objection collapse together two distinct roles that flourishing can play in politics, and assume that responding to one role, the ability to legitimate state action, is sufficient to reject perfectionism in politics. In Section V, I explore how recognizing these two roles alters the effectiveness of different anti-perfectionist responses to perfectionism, by comparing political liberal and Kantian responses. This puts into relief questions about what anti-perfectionists find to be fundamentally problematic about perfectionism. Before discussing the legitimacy objection, it is important to clarify the targets of the objection and the goals of those proposing the objection—what it means to be a perfectionist and an anti-perfectionist. Here I understand the distinction between perfectionism and anti-perfectionism as those presenting the legitimacy objection understand it: whether the state acts legitimately when it undertakes laws and policies to promote flourishing.5 Legitimacy, in the most general sense, concerns the state's moral authority to issue rules that citizens have an enforceable obligation to comply with. Some particular law or policy, then, is legitimate for the state to adopt if the state, first, has the moral right to create enforceable obligations for the citizens to comply with the law or policy—it has the moral authority to create those rules—and, second, the state exercises this right in a permissible way—for example, by exercising the right using a permissible procedure. States act illegitimately when they exceed their authority or when they exercise their authority in an impermissible way. Traditionally, perfectionists hold that the fact that some law or policy contributes to human flourishing can partially ground the state's right to issue and enforce the law or policy—the state has the authority to promote flourishing—and that the state can permissibly exercise its authority to promote flourishing. Perfectionists hold, then, that a legitimate function of the state is to promote flourishing. This can be held more or less strongly. Some perfectionists claim that the state has a duty to promote flourishing (duty-based perfectionism), while other perfectionists merely claim that it is permissible for the state to promote flourishing (permission-based perfectionism).6 My discussion of the legitimacy objection will not presuppose either form of perfectionism.7 This leads to the question of what is meant by flourishing. As a general concept, flourishing consists in living a life that has intrinsic value. Different perfectionist theories provide different accounts of what it is for a life to have intrinsic value—whether it is in making autonomous choices that guide one's life, developing and exercising certain capacities, possessing various excellences, and so forth.8 One commonality between these views is that claims about intrinsically valuable lives have some objective component—some component that does not depend on an agent's attitude toward their own life and its value.9 Considerations about flourishing within politics, then, are considerations about what will contribute to the intrinsic value of lives and what laws and policies will facilitate citizens having intrinsically valuable lives. To add to the previous paragraph, perfectionists hold that the state may seek to promote its citizens leading intrinsically valuable lives. the power of the state should not be used to promote some ideals or discourage others, but rather should be used to provide a fair framework of rules and institutions within which each citizen can pursue his or her own view of the good life.10 While the state may enact policies that differentially impact people's ability to flourish—the state need not be neutral with respect to the effects of its policies on people's ability to flourish—the state only has the authority to pursue non-perfectionist ends. For example, the Rawlsian conception of justice denies that the state has the authority to promote any particular conception of flourishing, but does hold that the state should promote certain “primary goods,” which citizens can use to pursue their own conception of flourishing.11 These goods are determined based on the conception of the citizen within Rawls's theory of justice, and not in terms of their relation to flourishing.12 Anti-perfectionists may therefore allow that non-perfectionist considerations can identify certain goods to be promoted, but deny that what grounds the legitimacy of promoting these goods is their contribution to flourishing. The dispute between perfectionists and anti-perfectionists, then, concerns the role that flourishing may play in political life. Perfectionists hold that it is legitimate for the state to seek to promote flourishing, while anti-perfectionists hold that undertaking state action to promote flourishing is beyond the legitimate scope of state authority. The basic idea behind the legitimacy objection is that considerations about flourishing are not the right basis to ground the state's moral authority, and the state has the authority to promote only what grounds its moral authority. For the state only has that authority necessary to satisfy the justification for its rule—states do not have the authority to pursue their own independent projects. Since flourishing is not part of the basis on which the state has the right to rule, the state does not have the authority to promote flourishing itself. Correspondingly, the perfectionist claim that the state may legitimately implement policies for the sake of promoting flourishing requires an illegitimate extension of the state's authority.13 Why hold that considerations about flourishing are not the right basis for grounding the state's right to rule? The basic thought is that people—as free beings—have certain natural rights against being subjected to authority for certain kinds of reasons—reasons concerning what will lead to people's lives being better off overall. Anti-perfectionists, such as Jonathan Quong and Japa Pallikkathayil, typically generate this thought through various intuitive cases where it seems inappropriate to hold that one person is subject to the authority of another, even though it might make their life go better overall. For example, the fact that my brother is best suited to manage my finances, and so I have most reason to give him control of my finances, does not justify him taking control of my finances.14 It is a good idea for me to give him control, but the mere fact that it is a good idea does not ground a claim on his part, or anyone else's part, to wrest control from me and give it to him. I have a claim against being subject to another's authority on these kinds of grounds. This claim extends as well to being subject to another's authority to promote the flourishing of others. Just as I have my own life to live, my brother similarly has his own life to live, and so I, and others, lack the authority to make my brother manage my finances to ensure my financial well-being.15 The mere fact that I, and perhaps others, will flourish if my brother provides me services does not ground a claim to wrest control from his life to provide me with that service. My brother is not bound to be subject to anyone's authority for the sake of promoting the flourishing of others; this does not seem like the right kind of reason to subject him to anyone's authority. Let's call people's claim against others having the authority to interfere in their lives for certain reasons their right to live their own life consistent with an equal right of others to live theirs, or, more simply, people's right to live their own life.16 The general thought behind this right is that each person has their own life to live and so has authority in terms of setting their own purposes to pursue with their life, compatible with a similar authority given to others. This is a way of conceptualizing persons as free beings—they are free insofar as they are not subject to anyone's constraints in setting what ends to pursue.17 To subject my brother to an authority to promote my flourishing, then, is inconsistent with this conception of persons. My brother's life is not simply a means for promoting value within the world or ensuring that other people's lives are valuable. Rather, his life is something for him to live. To subject him to an authority on the grounds that it will promote my flourishing or his own flourishing, is to treat his life as subordinate to the promotion of this value. The right to live one's own life, then, is understood, among other things, as a constraint on the kinds of justifications that can be given for authority—justifications premised on benefit, including flourishing, cannot ground claims to authority. The legitimacy objection then proceeds as follows. Perfectionist justifications of state authority hold that the state has authority over people for the sake of promoting human flourishing. That is, the state has the authority to promote human flourishing as a distinctive goal over and above whatever non-perfectionist goals the state may have. Correspondingly, non-perfectionist justifications of state authority will not legitimate perfectionist exercises of state authority—such exercises extend the state's authority without furthering non-perfectionist goals. This extension of authority, then, needs to be grounded on the basis of promoting flourishing itself. But this is incompatible with people's right to live their own life—this right protects against being subject to an authority on the grounds of mere promotion of benefit, such as human flourishing. The fact that some course of action will promote people's flourishing cannot justify the claim that the state has legitimate authority to pursue that course of action, and so the state illegitimately exceeds its authority through implementing laws for purposes that are beyond the scope of state authority.18 Anti-perfectionists do not hold that the legitimacy objection undermines all political authority, only that it undermines perfectionist political authority. For anti-perfectionists still claim that states have the legitimate authority to pursue non-perfectionist ends. Before continuing on to how perfectionists might respond to this objection, it is important to see why anti-perfectionists think that states have the legitimate authority to pursue non-perfectionist ends. This provides a standard for what kinds of arguments anti-perfectionists find acceptable as responses to the legitimacy objection, and so a potential template for perfectionists to follow in responding to the objection. One common account presented by anti-perfectionists focuses on people's enforceable natural duties.19 By “enforceable natural duties” I mean the duties that people have prior to and independent from entering into a political society or other association with people that can ground the use of coercive force on the part of others to generate compliance. I do not assume, though, that all moral duties are enforceable natural duties—some moral duties may fail to ground coercion and some might only arise subsequent to an association with others. The basic argument, then, is that the state's legitimate authority is grounded on people's natural duties. There are two ways that a natural duty account of the state's legitimate authority might proceed. The first, adopted by some liberals such as Quong, is to hold that there are some natural duties that override or limit people's right to live their own life—such as the duty to rescue others in peril.20 The idea, then, is that insofar as I have an enforceable duty to another, I do not have a right to fulfill that duty in whatever way I wish. Rather, I must take the course of action that will do the most to fulfill my natural duties.21 In cases where following the authority of another will greatly increase the likelihood that I and others will fulfill our natural duties, we are obligated to comply with that authority or be subject to coercion to enforce compliance. In these cases, I cannot appeal to the right to live my own life against this authority, because that authority is based on a duty that already limits my right to live my own life. Insofar, then, as the state is able to greatly increase the ability of citizens to comply with their various natural duties, the state correspondingly has legitimate authority over them, at least for those purposes.22 Let's call this the “liberal response” to the legitimacy objection. The second way that a natural duty account of the state's legitimate authority might proceed is by holding that people's right to live their own life itself provides people with enforceable natural duties. This account is favored by Kantians such as Pallikkathayil.23 The basic idea here is that people's right to live their own life gives people an enforceable natural duty to create the conditions for people to live their own lives. But, the argument goes, this duty cannot be discharged absent the state—the state is necessary to secure the conditions for people to live their own lives through setting up systems of property and contract, resolving disputes between persons, and so forth. As with the liberal response, people have no claim on the grounds of their right to live their own lives against the state having this authority, because it is that very same right which licenses the state to have this authority over them. Let's call this the “Kantian response” to the legitimacy objection. This presentation of these anti-perfectionist responses to the legitimacy objection is not meant to show that anti-perfectionists successfully demonstrate that the legitimacy objection allows non-perfectionist state action. Rather, it is meant to show the standard that anti-perfectionists hold themselves to in determining whether the state's authority is legitimate. In this respect, the central claim in the liberal and Kantian responses is similar—the justification of state authority relies on some enforceable natural duty that already constrains a person's right to live their own life. For these views, such duties define the scope of legitimate state authority and ground a range of non-perfectionist ends that the state may pursue. One perfectionist response to the legitimacy objection is to argue for the state's authority to promote flourishing using the same method as anti-perfectionists. That is, some perfectionists argue that the authority to promote flourishing is grounded on a duty to promote the flourishing of others.24 If there are such duties, then it is possible to ground the state's authority to promote flourishing on the basis of fulfilling such a duty, just like the natural duties anti-perfectionists invoke. To this end, some perfectionists hold that there is a natural duty to promote flourishing—a duty that arises prior to any form of political organization—while other perfectionists argue that there is a political duty—a duty that arises within political society—to promote flourishing. In Section III.I, I will discuss the position that the duty to promote flourishing is a natural duty held by individuals, and I will address the second possibility in Section III.II. The idea that people have a natural duty to promote flourishing just holds that, among whatever other natural duties people may have, one duty is to promote the flourishing of both oneself and others. Perfectionists often ground duties to promote flourishing on people's interest in flourishing itself.25 For instance, Joseph Raz argues that “[s]ince autonomy is morally valuable there is reason for everyone to make himself and everyone else autonomous” and so there is a duty to promote the conditions to make people autonomous.26 The value of people's flourishing—living a self-directed, autonomous life for Raz—grounds people's duty to promote flourishing. The assumption here is that people's interest in flourishing is sufficiently weighty to ground an enforceable duty to promote the conditions for flourishing, which can then ground the state's authority to promote flourishing.27 There is something dialectically odd about this argument. Anti-perfectionists appeal to intuitions which deny that people's interest in flourishing can ground claims to authority over people. Perfectionists, then, hold that people's interest in flourishing grounds a natural duty to promote flourishing, which then grounds claims to authority. But the intuitions to which anti-perfectionists appeal should also support denying that people's interest in flourishing can ground the relevant kind of duty that can ground a claim to authority. For basing a natural duty to promote flourishing on people's interest in flourishing seems to just amount to restricting people's right to live their own life for mere benefit. The interest in flourishing plays the same role in both cases in grounding a restriction on people's authority over their own lives; the difference is just whether this restriction subjects people to a duty or an authority. Anti-perfectionists who endorse the legitimacy objection, then, are already committed to denying that there are natural duties to promote flourishing.28 It is therefore not clear why this alternative conceptualization of how flourishing might ground the state's authority should undermine the general intuition that anti-perfectionists have. Perfectionists might respond to this by giving intuitions that support the duty to promote the flourishing of both oneself and others. But rarely are such intuitions actually given. Take, for instance, Steven Wall's argument that the right to live one's own life may be constrained by perfectionist considerations. He argues that, intuitively, “we do not have a right, grounded in autonomy [that is, our right to live our own lives], to form and execute plans designed to harm others, even if we take care not to contravene their equal right to autonomy,” where this “harm” is understood to be a failure to satisfy our duties to promote the flourishing of others.29 While Wall finds this claim intuitive, he provides no concrete, intuitive cases to support it, no instance where a person has a natural duty to promote flourishing that might ground another's authority without also including the violation of some other natural duty.30 It is unclear, then, why anti-perfectionists should accept such arguments as dialectically effective—it seems to be just a bare rejection of the anti-perfectionists' claims, without the appeal to a common set of shared judgments that might support the controversial principle. This, of course, does not show that there are no such perfectionist natural duties. But it does make it unlikely that appeals to such duties can make effective headway against the legitimacy objection. Such a response seems doomed to a dialectical stalemate, with perfectionists and anti-perfectionists each appealing to a competing set of intuitions. To make progress against the legitimacy objection, some alternative argument is needed. One alternative is for perfectionists to argue that political societies—that is, the state—have a duty to promote flourishing. What might the argument for the state's duty to promote people's flourishing look like? Collis Tahzib has recently proposed such a duty within a broader Rawlsian political framework.31 He argues that people in an “original position” scenario, selecting principles of justice to govern the state's conduct, would include a right against the state to the fair conditions for living a flourishing life, which gives rise to a duty on the part of the state. This duty then competes with some of the state's other duties, such as distributive goals like satisfying the difference principle. The basic thought is that given people have an interest in flourishing, they would prefer to flourish more rather than less. While no one can demand the right to flourish itself—since that would not be sensitive to their own responsibility with respect to their flourishing—it seems appropriate that agents might demand the right to the fair conditions of flourishing. The state determines, in part, whether we have a fair shot at securing our flourishing through the institutions that govern the structure of society and the public culture that the state cultivates. Agents in the original position would thus find it “reasonable and prudent” to select a set of principles to govern society which include a duty on the part of the state to secure and maintain fair conditions for people's flourishing.32 This gives us grounds for thinking that the state has a duty to promote flourishing. Suppose that Tahzib's argument succeeds in showing that the state has a duty to promote flourishing, or the fair conditions for flourishing. Does that show that the state has the authority to promote flourishing? I am skeptical that the mere demonstration that the state has a duty to promote flourishing shows that the state has the authority to promote it. This is due to a general fact about the relationship between an agent's duties and their authority over others. The fact that an agent has a duty does not entail that the agent has any additional authority—any additional powers or claim rights over citizens—to carry out that duty. For example, suppose I make a promise to drive my friend to the airport, and so have a duty to drive them there. This gives me a duty to acquire the necessary means to drive them to the airport—such as a car—but it does not give me any additional authority over others such that they must provide me the means necessary to drive my friend. If I lack a car, the fact that I have a duty to drive my friend does not give me authority over anyone else to commandeer their car. I only have authority over whatever I had prior to the duty. The basic reason for this comes from people's right to live their own life. If agents could acquire authority over others simply by incurring duties, then people's right to live their own life becomes subject to the duties incurred through the decision to associate by others, undermining their authority over their own life. It must be shown that providing the agents with the authority to fulfill this duty is compatible with people's right to live their own life, and so an additional argument is needed. A worry here may be that there is a difference between me having authority over others through making a promise and the state having authority over its citizens to fulfill its duties. For citizens compose the state, and so it might seem that if the state has a duty, then the citizens must also have that duty by virtue of their composing the state. I doubt that this move can be so easily made—the relationship between the duties of a group agent and those who compose the group agent is complex. For instance, a corporation might be subject to an obligation—for example, to pay a fine—without agents comprising the corporation also being subject to that obligation—that is, the employees are not subject to an obligation to pay the fine. Duties that group agents have do not automatically transfer to all those who compose the group agent. The authority to take the necessary steps for the group agent to discharge its duty has to rely on the pre-existing authority of the group agent over the members of the group, rather than on the fact that the group agent has a duty. Correspondingly, there needs to be an independent argument to show that the state has the authority to fulfill its duty. Further, even if citizens have a duty to promote flourishing based on the state's duty, this does not show that this is the right kind of duty to extend the state's authority. For citizens can have political duties that are also unenforceable, and so do not extend the state's authority. One potential example is the duty of civility—the duty of citizens to provide public reasons that other reasonable citizens can accept in political deliberations—proposed by some anti-perfectionists such as Quong, and to which Tahzib also subscribes.33 This duty is “a moral, not a legal, duty,” it is not enforced by the state or its laws, but rather, perhaps, by the disapproval or condemnation of other citizens.34 Anti-perfectionists can thus hold that it is an illegitimate assumption to argue from the fact that citizens have a political duty to the conclusion that this duty extends the authority of the state. For there still remains the question of whether this duty is an enforceable duty, a duty that constrains people's right to live their own life, and which therefore can ground the authority of the state to promote flourishing. An additional argument is thus required to show that the political duty of the state translates to an extension of state authority. The discussion in Section II.III shows how anti-perfectionists like Quong and Pallikkathayil might argue for which political duties extend the state's authority consistent with people's right to live their own lives: the fact that certain political duties follow from people's natural duties.35 For Tahzib to make a similar claim for the duty to promote flourishing, he would need to argue that the authority to promote flourishing follows from people's natural duties—that is, that there is a natural duty to promote flourishing. Otherwise, it seems to just be an assumption that it is already within the state's authority to promote people's flourishing, and the duty just directs how the state may exercise its authority. Tahzib's argument at most shows that the state has the duty to promote perfectionist conditions, but not that this duty extends the legitimate authority of the state to promote perfectionist conditions. This reveals a general problem for perfectionist arguments for the state having the authority to pursue perfectionist policies based on the claim that the state has a duty to promote flourishing. Demonstrating such a duty is insufficient to demonstrate the claim to authority, unless perfectionists can also show that such a duty limits people's right to live their own life. The clearest way for perfectionists to do this, though, requires that perfectionists show that there is a natural duty to promote flourishing. As I argued in Section III.I, arguing against the legitimacy objection on the basis of a natural duty to promote flourishing is dialectically ineffective, as the same intuitions that support the legitimacy objection also support denying that there is a natural duty to promote flourishing. A dialectical stalemate remains. Both perfectionists and anti-perfectionists assume that the debate over whether the state has the authority to promote flourishing is about whether the state may legitimately exercise authority when that authority is exercised for the sake of promoting flourishing. But why not use considerations about flourishing to choose from among the range of legitimate policy options, rather than partially determine that range? That is, instead of considerations about flourishing extending

When discussing theories of justice, most philosophers take the moral equality of human beings as their starting point. As Will Kymlicka says, in all contemporary plausible theories of justice, moral equality constitutes an "egalitarian plateau".1 Arguably, the most prominent novel theory of justice in recent years is relational egalitarianism—a theory on which justice requires people to relate as equals. Relational egalitarians are no exception to Kymlicka's claim. They too start from the idea of moral equality. As one of us previously put it, "as a matter of fact, we are one another's moral equals and in relating as equals we honour that fact, and this is what grounds the ideal of relational egalitarianism".2 Either the proposed basis [the property, or properties, proposed to ground equal moral status] will turn out to vary by degree, and variations above the claimed threshold that establishes equality will give rise to inequality of moral considerability, or the proposed basis will turn out to be one that applies in an all-or-nothing fashion, and then it will turn out that the basis proposed as justifying equal moral considerability is too flimsy or insubstantial to do this justifying work.3 In this article, we will assume that not all human beings are moral equals. This is an assumption, not an assertion, on our part. It is motivated partly by the challenges mentioned in the previous paragraph, and partly by the nature of the present inquiry: to wit, examining what, if anything, relational egalitarianism implies when it comes to relationships between moral unequals. Must moral unequals relate as moral equals?4 Or as unequals? Or in some third way? We will show that relational egalitarianism has much to say about such relationships. And we will show that what it has to say is plausible. Before proceeding, we need to defend our line of inquiry in view of the following skeptical challenge. For its supporters, what grounds relational egalitarianism is the fact—so they claim—that people are moral equals. For instance, this seems to be what the following passage from Kolodny implies: "Insofar as we are to have ongoing social relations with other moral equals, we have reason to relate to them as social equals".5 Hence, to ask what relational egalitarians are committed to saying about social relations in a—in their view—hypothetical situation, where people are not moral equals, is to ask a moot question. It is like asking what a utilitarian is committed to, as regards the right thing to do, if welfare is not valuable. The question makes no sense, because the notion that welfare has value is built into, and therefore presupposed by, utilitarianism. Similarly, the notion that people are moral equals is presupposed by relational egalitarianism.6 While this challenge makes sense, we think that, ultimately, we are asking a perfectly justifiable question, and one we have the resources to answer. First, Kolodny's remark is most naturally taken to mean that moral equals must relate as social and political equals given that, more generally, the way people relate, socially and politically, should fit the way they relate in terms of moral status. If this is correct, Kolodny's view is underpinned by a general commitment to some kind of fittingness; and this general commitment, surely, has implications for the question of how moral unequals should relate, even if Kolodny thinks that people are not moral unequals, and thus even if the implications in question do not materialize in the real world.7 Admittedly, as a matter of logic, one can consistently hold both that if people are moral equals, that justifies certain claims about how they should relate socially and politically and that if they are moral unequals, that does not justify any claims about how they should relate socially and politically. However, such a view seems arbitrary. Why would moral status have implications for how we ought to relate only in cases where people's moral statuses are equal? So the view is not especially plausible. This implausibility is reflected in the complaint: "They treated us as if we were animals". The complaint assumes, precisely, that people and non-human animals do not enjoy the same high moral status, and that the bad treatment complained about would have been justified (or at least less unjustified) had there been a difference in the moral status of the wrongdoers and the victims of the sort that exists between people and non-human animals. Second, for reasons brought out in our first response, we believe the analogy with utilitarianism is misleading. Consider a utilitarian who believes that morality requires us to do what realizes the greatest amount of value, and who treats welfare, understood as preference satisfaction, as the only value. Surely, utilitarians of this kind are committed to a view about what morality requires agents to do if what is of value is not welfare as preference satisfaction, but, say, that people's level of welfare match their level of moral desert (as Kant thought would be ideal). That is, they are committed to the view that, morally, we should act so as to maximize the overall fit between levels of welfare and levels of moral desert. [E]ven those who do deny the existence of desert may find our investigation of some interest. For even if they are convinced that desert would only exist if this or that metaphysical condition is met, and even if they believe that the requisite conditions are not, in fact, met, presumably they are themselves open to the possibility that their favored metaphysical views on the relevant issues might be somehow mistaken. Thus it seems epistemically possible, if nothing more, that desert claims might turn out to be legitimate after all. Accordingly, there may be some value in getting clear about the nature of desert—what it would be like, if only it were real—even if one is skeptical about whether anyone really does deserve anything.8 Similarly, in the light of forceful challenges such as those put forward by Arneson, McMahan, and Singer, relational egalitarians should be open to the possibility that people are not moral equals. They should recognize the value of exploring what, if anything, their view commits them to, as regards how we should relate socially and politically, if they are mistaken and not all human beings are moral equals. Exploring the issue of how moral unequals are required to relate in the relational egalitarian framework is important for several reasons. First, it helps to develop relational egalitarianism as a theory of justice—a theory which still leaves many questions unanswered, because it is relatively young. As we will show, our discussion helps to shed light on the place of children and non-human animals—individuals that are arguably not moral equals of persons—in relational egalitarianism. Second, the exploration will show that relational egalitarian objections to paradigmatic relational inequalities—such as discrimination, racism, and sexism—can be constructed, even if not all human beings are moral equals. That means that the plausibility of relational egalitarianism does not stand or fall with the idea that all humans are moral equals.9 This result is important, given the dilemma Arneson describes above. Third, the exploration illuminates the important point that relational egalitarianism is a large family of theories that differ along many different dimensions, including their responses to the question of how moral unequals should relate. The article is structured as follows. In Section II, we introduce relational egalitarianism and distinguish two forms: deontic and telic relational egalitarianism. We further distinguish two accounts of deontic relational egalitarianism and explain that these provide different accounts of how moral unequals should relate. We show, moreover, that the least demanding conception of deontic relational egalitarianism provides plausible answers to the question of how moral unequals should relate.10 In Section III, we turn to telic relational egalitarianism. We show that the reasons proposed by relational egalitarians explaining why inegalitarian relationships are bad and egalitarian relationships are good apply to relationships between moral unequals as well: these reasons are not tied to people with equal moral status. We explore further implications of these arguments in Section IV in relation to the place of children and non-human animals in relational egalitarianism—something about which relational egalitarians have not said much. We provide a relational egalitarian explanation of why adult-adult paternalism may be regarded as more objectionable than parent–child paternalism. Additionally, we respond to a concern which may arise when the claims of people who are moral unequals are being discussed. Section V concludes and presents the main takeaway point of the article: that, fortunately, our commitment to relational, social, and political equality is not hostage to the philosophical discussion about whether all people are moral equals. The relational egalitarian theory of justice requires people to relate as equals.11 It has received much attention in recent years as a result of relational egalitarians' trenchant criticisms of distributive theories of justice. On these theories, justice is ultimately a matter of distributions.12 Distributive theorists, relational egalitarians argue, fail to see that, ultimately, justice is not about distributions. A given distribution in a society may accord with distributivist requirements of justice, but still not realize justice because, for example, racism and sexism are prevalent in the society. What ultimately matters with justice is, instead, whether relations are suitably egalitarian. Justice requires that people relate to each other as equals.13 For X and Y to relate as equals, relational egalitarians argue, they must (1) regard each other as equals; and (2) treat each other as equals.14 Derek Parfit famously distinguished between telic and deontic egalitarianism. According to the first, "it is in itself bad if some people are worse off than others".15 This is an axiological view. According to the second, "it is not in itself bad if some people are worse off than others … What is unjust, and therefore bad, is not strictly the state of affairs, but the way in which it was produced".16 Whereas in telic egalitarianism, inequality is bad, in deontic egalitarianism, inequality is unjust.17 These views are different. To see why, imagine a case in which an inequality is unavoidable. Thus, suppose the inequality was created by a natural disaster, such as an earthquake. Since, in this case, the inequality has not come about through anyone's wrongdoing, it is not unjust on the deontic view. According to the telic view, on the other hand, the inequality is in itself bad even though it is unavoidable.18 Telic Relational Egalitarianism. It is, in itself, good (bad) if egalitarian (inegalitarian) relationships between people exist. Deontic Relational Egalitarianism. It is morally required that people relate as equals, not unequals. So, telic relational egalitarianism says it is (dis)valuable that people relate as (un)equals, whereas deontic relational egalitarianism says that people ought to (not) relate as (un)equals. Lippert-Rasmussen refers to Christian Schemmel and Elizabeth Anderson as examples of deontic relational egalitarians.20 Anderson says that "[Relational] egalitarians base claims to social and political equality on the fact of universal moral equality".21 Schemmel says, "the objection to [inegalitarian] relationships is not merely that they are, in some sense, bad for people, but that they constitute unjust treatment".22 Telic relational egalitarians include Martin O'Neill, who says, "The existence of these kinds of social relations [egalitarian social relations] should itself be seen as intrinsically valuable, independent of the positive effects that such relations may have for individual welfare."23 Narrow Deontic Relational Egalitarianism. Moral equals must relate as moral equals. To relate as moral equals has to do with interests and agency. If we are to relate as moral equals, our interests must be given equal weight and our agency must be equally respected—in a fundamental sense for both.25 If X treats Y in a racist manner, X treats Y's interests as if they are less important, in a fundamental sense, than the interests of non-Y people. X and Y thereby fail to relate as moral equals, and that is unjust on narrow deontic relational egalitarianism. Broad Deontic Relational Egalitarianism. Moral equals must relate as moral equals and (many) moral unequals must relate as moral equals.26 Of course, it is not enough merely to claim that moral unequals, or some, or many of them, must relate as moral equals. We also need an explanation of why moral unequals, despite being moral unequals, must relate as moral equals. Perhaps the justification for narrow deontic relational egalitarianism can also be used to show why, as broad deontic relational egalitarianism prescribes, moral unequals must relate as moral equals. Lippert-Rasmussen offers an explanation of why moral equals must relate as moral equals with reference to fairness.27 On his interpretation of fairness, "it is unfair if people are differently situated if the fact that they are differently situated does not reflect their differential exercise of responsibility".28 This explanation of why moral equals must relate as moral equals looks promising. It certainly seems unfair, and therefore unjust, for a black person to be treated in a racist manner by a white person given that they are moral equals.29 But can fairness also explain why moral unequals must relate as moral equals? Can we say that it is unfair if moral unequals, or some, or many of them, relate as moral unequals? identify available courses of action she might take, discern reasons for and against the options, weigh and assess the reasons she discerns, deliberate and make choices, carry out the action chosen, and do all this not simply for a single decision problem at a time but with respect to long-term plans of action and projects she might undertake.32 These cognitive and volitional abilities come in degrees. To be a rational agent is to possess them at a certain given level. On this account of moral status, moral unequals possess rational agency to different degrees. Now, let us suppose that both X and Y have moral status, but that X has higher moral status than Y in virtue of possessing rational agency capacity to a higher degree. One fairness-based argument for the view that it is unjust if X and Y relate as moral unequals runs as follows. Suppose X's possession of a higher degree of the capacity for rational agency than Y is due to nature and nurture—for example, X is bright, and their parents have raised them in a way that has been conducive to their achieving a high level of rational agency. In this case, the differences in the capacity for rational agency between X and Y are not due to their differential exercise of responsibility. Thus, if Y relates as an inferior to X because X has the higher moral status, the fact that X is differently situated (for example, in that Y should treat X's interests as more important than their own in a fundamental sense) does not reflect X's and Y's differential exercise of responsibility. And since being differently situated when this does not reflect differential exercise of responsibility is unfair, it is unjust that X and Y relate as moral unequals despite their being moral unequals. A significant challenge to this argument is that it only goes through if X and Y's differing capacities for rational agency are not due to their differing exercises of responsibility.33 This is a challenge, because it seems highly likely that the abilities constituting the capacity for rational agency can be affected by the exercise of responsibility. For instance, one of the components of rational agency is the ability to "deliberate and make choices". It seems reasonable to assume that the more one deliberates, the better one becomes at deliberating. If X decides to devote a considerable amount of their time to deliberating and Y does not because they would rather do something else, we may expect that, over time, X will become better than Y at deliberating and making choices.34 And if that is the case, and all else is equal, X will at that point have a greater capacity for rational agency than Y, and thus, at least on some views, higher moral status than Y. In this case, their possession of unequal degrees of the capacity for rational agency is a result of X and Y's differential exercise of responsibility. So, it would be unfair if they were not differently situated: that is, if they did not relate as moral unequals. It follows that fairness can ground broad deontic relational egalitarianism only if we assume that human beings are not responsible for their degree of the capacity for rational agency (and, more generally, are not responsible for their full or partial possession of the property, or properties, grounding moral status). If they are so responsible—which they seem to be, at least to some degree—it would be unfair for moral unequals to relate as moral equals. Deontic Relational Justice. Moral equals must relate as moral equals, and moral unequals, assuming they have sufficient moral standing, must relate as moral sufficients. This account of deontic relational egalitarianism is not egalitarian "all the way down", in the sense that it denies that moral unequals must relate as moral equals. We therefore refer to it as deontic relational justice. But it is still egalitarian in the sense that moral equals must, as relational egalitarians prescribe, relate as moral equals. When it comes to moral unequals, the account only applies to entities with sufficient moral standing.35 It does not say anything about, say, rocks, assuming they have no moral standing at all. The account does not say, for instance, that a human being and a rock must relate as moral sufficients. But where entities with sufficient, but unequal, moral standing are concerned, the account says that they must relate as moral sufficients.36 The Egalitarian Deliberative Constraint (EDC). If you and I have an egalitarian relationship, then I have a standing disposition to treat your strong interests as playing just as significant a role as mine in constraining our decisions and influencing what we will do. And you have a reciprocal disposition with regard to my interests. In addition, both of us normally act on these dispositions. This means that each of our equally important interests constrains our joint decisions to the same extent.38 If, in a marriage, one party's interests always trump the other's, the parties fail to relate as moral equals. If they are moral equals, this is unjust. Moral unequals, on the other hand, do not have to grant each other's interests the same weight in their collective decisions. Arguably, to relate as moral sufficients, the marital parties do not have to satisfy the Egalitarian Deliberative Constraint in their dealings with each other. It suffices that they have a standing disposition to treat each other's interests as playing a role which is somehow fitting given their relative moral statuses.39 But it is a long way from acknowledging this to saying that "anything goes, morally speaking". Importantly, by adopting deontic relational justice, relational egalitarians will be able to object to most of the types of relational inequality to which they would want to object, even if some human beings are moral unequals. As Lippert-Rasmussen posits, the possession of "'sufficient moral standing' means that there is, in some sense, a sufficient number of sufficiently important things that one cannot do, morally speaking, to that individual and which this individual is permitted to do."40 Arguably, two of the sufficiently important things one cannot do to individuals with sufficient moral standing are discriminate against or dominate them. Suppose a two-year-old child and one of her parents are not moral equals, but they both have sufficient moral standing.41 We take it that we would find the relational inequality which would result from the parent treating the child in a sexist manner objectionable, because the parent thereby fails to treat the child as a being with sufficient moral standing (we would reach the same verdict, we take it, in cases of domination and racism as well). Similarly, if a normally functioning adult treats a person with Down syndrome in a sexist manner, they fail to treat them as a human being with sufficient moral standing. We can imagine cases where the gap in moral standing between two human beings is even smaller. Suppose an adult, Bert, has a very slightly higher moral standing than another adult, Carl. If Bert discriminates against Carl, it is hard to see how he could be treating Carl as a being with sufficient moral standing. Importantly, these ways of treating others—discrimination, domination, racist and sexist treatment—are paradigmatic relational inequalities.42 This shows that even if human beings are moral unequals,43 relational egalitarians, by adopting deontic relational justice, are not barred from objecting to paradigmatic relational inequalities.44 At this point, one might have a serious concern about our exploration of the idea that, even if people are moral sufficients, the standard relational egalitarian objection to paradigmatic relational inequalities stands. In explaining the importance of this article's main question—how should people relate, according to relational egalitarianism, if they are not moral equals?—we noted that, in the philosophical literature, there are forceful challenges to the notion of basic moral equality, citing among other things Arneson's continuity question: if moral status results from rational capacities, why is the moral status of people not a scalar matter in such a way that those with greater rational capacities have greater moral status? The concern is that at least some of these challenges also cast doubt on the threshold notion of moral sufficients, and motivate a continuous, graduated view of individual moral status instead.45 In response, we note, first, that this concern is not about the truth of our main claim (that even if we are not moral equals, we should, from the perspective of relational egalitarianism, relate as sufficients), but its importance. We are not arguing that people are moral sufficients (or moral incommensurables). We are arguing that if they are, it remains the case that they should not relate in the inegalitarian ways (for example, involving discrimination or domination) that relational egalitarians think moral equals should avoid. This can be true, even if the continuity challenge gives us good reason to think that people are not moral sufficients. Second, and much more importantly, we think the present challenge suggests we should extend the scope of our robustness claim beyond people being moral sufficients or moral incommensurables. Suppose moral status is very much a scalar thing, and that people therefore vary in their moral status beyond the variation involved in the claim that they, or most of them, are moral sufficients. Even so, it will still be the case that discrimination and domination—the sorts of inegalitarian relations that relational egalitarians object to—are wrong when they occur in relations between moral unequals—at least, it will be, if those with inferior moral status have sufficiently high status (which might simply be some moral status). Even if animals, say, have lower moral status than human beings, it is still wrong to dominate them; and even if children have lower moral standing than adults, it is wrong to treat a child in a sexist manner. In claiming this, we do not have to assume that the wrongness of domination and discrimination is a binary, as opposed to a scalar, matter—that there are no degrees of their wrongness. It may be that discrimination is worse when the person subjected to it has higher moral standing than when the sufferer has lower standing, other things being equal. But this is compatible with discrimination still being wrong when directed at an individual with a lower moral standing. If that is correct, the present challenge points in the direction of an even more ambitious, and for that reason even more interesting, claim than the main claim regarding robustness that we defend in this article. In this section, we have explored what deontic relational egalitarianism implies in situations involving moral unequals. Broad deontic relational egalitarianism prescribes that moral unequals must relate as moral equals. We saw that fairness can explain why that is so if, and to the extent that, people are not responsible for the property, or properties, grounding moral status. Deontic relational justice avoids relying on this assumption by prescribing that moral unequals, assuming they have sufficient moral standing, must relate as moral sufficients. If X and Y relate as moral sufficients, there are ways in which they must not treat each other: for example, neither can treat the other in a racist or sexist way.46 This shows that relational egalitarians, by adopting deontic relational justice, put themselves in a position to object to discrimination, domination, racism, and sexism even if human beings are not moral equals (at least, as long as the involved parties have a sufficient moral standing). Telic Relational Egalitarianism. It is, in itself, good (bad) if egalitarian (inegalitarian) relationships between people exist.47 We will explore two questions: (1) is it bad if moral unequals relate as moral unequals? (2) is it good if moral unequals relate as moral equals? In answering these questions, we propose to look at the reasons relational egalitarians have provided for holding that inegalitarian relationships are bad and egalitarian relationships are good. We will explore whether these reasons read across successfully to relationships between moral unequals. Two caveats: first, we devote more space to the reasons why inegalitarian relationships are bad than we do to the reasons why egalitarian ones are good. This is not because we believe the latter are unimportant—they clearly are important. It is because we wish to investigate whether the reasons that relational egalitarians have themselves given to explain the value implications of (in)egalitarian relationships tell us anything about relationships between moral unequals. It so happens that these reasons have tended to be constructed around the idea that inegalitarian relationships are bad, rather than the idea that egalitarian relationships are good. Second, relational egalitarians have argued that relational inequality is intrinsically bad in various ways. It might be suggested that, at most, they have shown relational inequality to be reliably instrumentally bad: that is, bad not in itself, but in virtue of its effects, and that, accordingly, the arguments do not amount to sound arguments in favor of telic relational egalitarianism.48 In response, we note first that (as we occasionally indicate below) we share the suspicion manifested in this response. However, we shall refrain from going deeper into this matter, since ultimately it pertains not to whether inegalitarian relations are bad, but rather to the way in which, according to relational egalitarians, they are so. Our main concern is to ask whether the reasons proposed by relational egalitarians themselves for thinking that inegalitarian relationships between moral equals are intrinsically bad apply to inegalitarian relationships between moral unequals as well. It is not to examine whether relational egalitarians are correct in seeing (at least many of) their arguments as arguments for telic egalitarianism (though we suspect they are not), as opposed to arguments for unequal relations being bad in virtue of their consequences. With these caveats entered, we turn to examine the specific reasons relational egalitarians have given for holding that unequal relationships are (as they believe, intrinsically) bad. Do these reasons read across to relationships between moral unequals? T. M. Scanlon has claimed that inegalitarian relationships are bad because "it is an evil for people to be treated as inferior, or made to feel inferior".49 He explains that this is an evil because it leads to stigmatizing differences in status, with "damage to individuals' sense of self-worth".50 In a racist society, the inequality between a black person and a white person may lead the former to devalue their self-worth.51 Is it only an evil for a person to be treated as inferior, or made to feel inferior, in a way which leads to damage to that person's self-worth if the people involved are moral equals? It seems not. Suppose that Higher and Lower are moral unequals: Higher has higher moral standing than Lower, although Lower still has sufficient moral standing. Suppose now that they relate as moral unequals—for example, they both regard Higher's agency as more important than Lower's agency, and they both treat it as such. Even if that were so, the inegalitarian relationship could still lead Lower to devalue their self-worth more than they should given their differences in moral status.52 Suppose Lower undervalues their self-worth relative to their moral standing. Suppose, further, that they do so to the extent that they believe their own interests have no importance at all in decisions about their and Higher's collective affairs. Lower comes to feel excessively inferior, morally speaking. In this kind of case, inegalitarian relationships between moral unequals are bad, because they lead the morally inferior individuals to de

On 31 January 2020 the United Kingdom left the European Union. The legal terms of this “Brexit”, and the new relationship between the UK and the EU, were set out in two international agreements, the Withdrawal Agreement, setting out the terms of the UK's departure from the Union, and the UK–EU Trade and Cooperation Agreement (TCA), governing trading relations and cooperation post-departure. Three and a half years of often acrimonious planning, posturing and negotiations produced texts that were agreed between representatives of the UK and the EU, and ratified by their respective parliaments. Following a transition period (during which the TCA was negotiated), on 31 December 2020, the UK ceased to be a member of the EU customs union and common market. Yet within months of their entering into force, the UK government (the same one that had negotiated them) began to publicly criticize these agreements, and specifically the difficult compromise agreed in respect of Northern Ireland.1 This may have reflected genuine misunderstandings within government about the effects of the agreement, or bad faith in the negotiations themselves, or the changing domestic political situation. Regardless, within weeks of celebrating their “great”2 and “cracking”3 deal, UK ministers were directly contradicting its content4 and, as time went on, denying that they ever supported it, and branding its content “problematic”,5 economically harmful,6 and built on a premise that “does not exist in real life”.7 Public responses from the EU and its member states (and indeed many UK politicians)8 expressed predictable frustration: whatever the merits of re-examining the practical operation of the agreement, the UK could not simply attack, denigrate and potentially disregard an agreement it had itself agreed only a few months earlier.9 There might be problems with that agreement, and scope for improvement. There might be other agreements that could have been made, that might have governed matters differently. However, it mattered that the UK had in fact agreed to this agreement, and any criticism must be judged accordingly. Brexit provides a particularly stark example of a state criticizing its own agreement, and the frustrated responses this evokes. However, it is by no means the only such example. It frequently seems important, in evaluating criticism of international treaties or institutions, to observe that the critics themselves agreed to that treaty or institution. As US officials have grown increasingly critical of the WTO in general, and WTO dispute settlement in particular, responses have frequently highlighted that these rules and institutions reflect agreements the US itself accepted (and indeed led).10 Conversely, development-oriented critiques of the WTO (and especially the TRIPS agreement) meet invocations of developing countries' voluntary acceptance of the Marrakesh Agreement.11 Criticisms of investment treaties and arbitration are answered by reference to respondent states' consent.12 Similarly consent, whether to IMF membership or to specific loans, seeks to blunt criticism of loan conditionality.13 While there are many differences between these examples, they share a fundamental assumption: that the procedural fact of past consent is relevant to substantive critique of an agreement's content. It is often hard to say how much weight is placed on voluntarism, compared to other arguments. It is rarely the only response offered: defenders of agreements will typically combine invocations of voluntarism with substantive defences of the terms of those agreements. However, it does seem to carry at least some weight in these kinds of debates. My goal in this article is to examine the moral significance of voluntary consent to treaties by interrogating this style of argument, which I label the “Voluntarist Reply”. I ask whether and to what extent the kind of consent invoked by the Voluntarist Reply changes the moral circumstances of the parties, in ways that might give that reply force. I am somewhat sceptical of the Voluntarist Reply, but I do not set out to challenge it directly. Rather, I show that, even if we grant some important premises on which it depends, there remain significant limits on that reply, some going to whether agreement is in fact voluntary in the required sense, and others reflecting deeper concerns not answerable by any appeal to voluntariness. Putting the argument in its simplest terms: for the Voluntarist Reply to have any application, an agreement must meet a very demanding standard of voluntariness; and even when it meets that standard, there remain many complaints that invocations of voluntary consent simply cannot answer. My principal interest is economic agreements, and at various points I focus on trade agreements in particular. However, the arguments described have much broader relevance, applying to almost any international agreement, and in particular those with distributive effects, whether within or between states.14 I approach my task in three parts. Section I seeks to clarify the nature of the Voluntarist Reply, as a procedural answer to substantive concerns, contrasting the role of consent here with its contribution to establishing either the validity or the legitimacy of international law. Section II asks how the Voluntarist Reply might be motivated, and what this tells us about the sense of voluntariness it requires. While the absence of voluntariness is not the only way that the Voluntarist Reply can fail, it constitutes one important limit on its scope. I argue, in this section, that invocations of the voluntarist reply are undermined where an agreement is made in the context of prior injustices for which a counterparty has some remedial responsibility. Section III considers other ways that the Voluntarist Reply might fail. It highlights the extent to which the applicability of the Voluntarist Reply, and hence the moral effect of state consent, will vary given different answers to background questions about global economic justice, mapping some of the limits implied by four views on these questions. Before proceeding, it is worth pausing to consider the kind of claim that the Voluntarist Reply makes and to distinguish it from a number of other roles that voluntarism and consent play in international legal discourse. The most obvious role of consent in international law is as a condition for the validity of a legal norm. Validity here denotes the extent to which the relevant norm forms part of the international legal system.15 For many positivist international lawyers, state consent is a necessary condition for the validity of international legal norms, whether custom or treaty.16 Anti-positivists have challenged that view: Dworkin's recent interpretivist account denies consent any fundamental role.17 Voluntariness may also be relevant to the (related but distinct) question of international law's legitimate authority.18 It may be relevant to determining not simply “what is the law?” (the question of validity), but also “why should a given agent obey the law?” (the question of legitimate authority). For the analytical positivist, validity alone is morally inert.19 Knowing that the law requires x is not the same as having a good reason to x: something more is required.20 Valid law may or may not have legitimate authority, and whether a given agent has consented to that law may play a role in connecting the two.21 Fundamental aspects of international legal doctrine, including pacta sunt servanda and the duty of good faith, are premised on the moral significance of state consent. While positivist lawyers might hope to work with these legal principles without interrogating their normative basis, it is hard to explain why they are so fundamental unless we think the fact of agreement matters, morally as well as legally.22 When a state stands on its treaty rights, and refuses to countenance renegotiation, it implicitly invokes the political and moral significance of its counterparty's prior consent. However the Voluntarist Reply has a broader focus than either validity or authority. In the examples noted above, critics are not denying that agreements constitute valid international law. Nor are they necessarily denying that those agreements impose legitimate obligations (although this may also be part of their complaint). Rather, they are complaining about the substantive content of those agreements: that they are unjust, or unfair, or inefficient, or otherwise fail to live up to some relevant standards. They are saying that these are bad agreements; they should not be law; and, to the extent that they are, there are good reasons to change them. The Voluntarist Reply seeks to answer such criticisms, not by engaging with the substantive criticism, but by emphasizing the relationship between the agreement and the agent by, or on whose behalf, the criticism is advanced. It says that, while the agreement might be criticized, their standing to criticize it, or the force of their criticisms vis-à-vis a specific counterparty, is undermined by the fact of their having agreed to it.23 It is a procedural answer to a substantive complaint. If the Voluntarist Reply works, then it is as a reply to complaints of this kind, about substantive content (whether articulated in terms of justice, fairness, efficiency or otherwise) rather than simply validity or legitimate authority. The Voluntarist Reply, as I understand it, is a response to criticism, understood as a negative evaluation by reference to some normative standard, expressed with a specific (negative) illocutionary force. There is an important difference between, for example, “I do not like this agreement” and “this agreement is unfair”. Either might readily explain a proposal for termination or renegotiation, but only the latter is a criticism of that agreement.24 The Voluntarist Reply seeks to answer criticism, but it says nothing about proposals for revision or reform that do not also criticize. We might wonder if the distinction here is simply a matter of tone? However, criticism imports significant practical implications that other calls for revision may not: that there are reasons (independent of the complainant's preference) to repeal or renegotiate the agreement; that those who benefit from the agreement are profiting from a moral wrong, while those who are burdened are also wronged; that other parties to those agreements should not strictly enforce their terms. Criticism has a normative force for the addressee that more neutral proposals for revision lack. “I don't like this agreement!” can be answered with “Well, I do!”; but “This agreement is unfair!” or “This agreement is impracticable!” demand answers that go beyond the interests of the counterparty and their power to maintain the status quo. Criticism and complaint call for answers (and actions) in ways mere preferences do not. So understood, the Voluntarist Reply operates at the level of political morality rather than legal doctrine or practice. It serves to answer political criticisms of, rather than legal challenges to, the content of international agreements. How far we endorse the Voluntarist Reply thus determines how we should feel about and act towards international agreements that are subject to these kinds of criticisms. The Voluntarist Reply invokes the fact of agreement to answer criticism of an agreement's terms. How, we might wonder, can it do this? That consent might ground validity or legitimacy is a much weaker claim that that it can answer a substantive grievance. What is the moral mechanism that might bring this about? We might wonder about the moral ontology required to translate either argument to the international case. Is it really “the state” as opposed to “this official”/“this group” doing these things?27 That ontology is fundamental to how we think and talk about international affairs, but I return to consider at points below the importance of the fact that we are here dealing with collective agents rather than human persons.28 The other motivating ideas—that consent can cure a moral wrong and that consistency limits our standing to complain about acts we have ourselves authorized—are widely shared. We can interpret them in various ways but they have sufficient plausibility across a broad range of contexts and views to grant, if only in order to examine the kinds of claims they support. Assuming the Voluntarist Reply implicitly invokes one or other of these argument structures, this in turn imports a very demanding standard of voluntariness, which many international agreements will not meet. “Voluntary consent” may mean very different things across different views. For Hobbes, the fact that consent is given at the point of a sword in no way undermines its force,29 a view rejected by many subsequent critics, from Hume's direct critique of Hobbes's social contract view30 to Rawls's denial of the moral significance of “threat advantage”.31 To matter, morally, consent typically requires an appropriate context, which might non-exhaustively include reasonable alternatives, adequate information, and respect for rights. Absent such circumstances, consent is voluntary in only a formal, and morally ineffective, sense. Beyond this, the criteria for voluntariness will vary, depending both on the moral work it is called on to do, and on the broader moral view against which the appeal to voluntariness is made. The threshold for voluntariness in the Voluntarist Objection may be significantly higher than we would apply to determine whether an agent was morally responsible for an action, or whether a state's consent was legally effective.32 I may be morally responsible for a choice, while still feeling legitimately aggrieved at both the option I have chosen and the fact that I had to choose at all; and we might have good reasons to treat an agreement as legally effective notwithstanding significant defects in the manner of its formation. If the Voluntarist Reply seeks to answer substantive criticism, then we must ask what circumstances must obtain for a party's having consented to an agreement to undermine such criticism? The two interpretations of the underlying argument sketched above help us to do this. Consider, first, the no-injustice interpretation. Robert Nozick's discussion of equality and choice is instructive.33 While arguing against “patterned principles”, and especially economic equality, Nozick describes a scenario whereby agents move, through a series of free choices, from an initial, ex hypothesi fully just, equal distribution, to a subsequent unequal distribution. In Nozick's telling, the lesson is that the second distribution, reached by a series of just steps from an initial position of justice, must itself be just. Choice, he argues, is justice-preserving.34 Whether we accept that claim or not, it is probably the strongest claim that can be made for the power of choice to morally validate an outcome. Yet [(I) Justice + Choice = Justice] does not imply [(II) Injustice + Choice = Justice]. The fact that I chose some outcome, over one where I was morally wronged, gives no reason to think this chosen outcome respects my rights. At most, it shows I thought it was better than the status quo. If the status quo was morally defective, then the same may be true following my choice, however free. In Rawls's language, it is only in circumstances of “background justice” that the results of agreements freely entered can, in virtue of that fact, claim to themselves be just.35 On the no-injustice interpretation, the Voluntarist Reply requires a no-injustice background to generate its no-injustice conclusion. We might expect the no-standing interpretation to be less demanding. After all, this interpretation does not claim that consent remedies the ills of an agreement, focusing only on who can complain, and to whom, about that agreement. However, on reflection the same demanding conditions seem to apply. G. A. Cohen has written helpfully on standing and the ways perspective might preclude specific agents from making specific complaints or offering specific justifications to specific others.36 Moral argument is interpersonal. Claims, criticisms, defences and justifications do not exist in the abstract: it often matters both by whom, and to whom, they are made. I may, on Cohen's account, be precluded from complaining to or criticizing some particular others, not simply because of the objective situation, but because of the ways that I or they may be implicated in that situation. The no-standing interpretation exemplifies this kind of perspectivity. Sometimes, analysing the substance of an agreement may convincingly identify defects, by whatever standard. Nonetheless, it may be difficult for the author of that agreement, or one of its co-authors, to condemn it on that basis. This is an instance of the challenge Cohen labels “You're involved in it yourself!”, which says: “How can you condemn me when you are yourself responsible, or at least co-responsible, for the very thing you are condemning?”.37 Many agents might criticize an agreement, but surely not the states party to it, who are after all its authors? Insofar as they have previously endorsed it, they are limited in what they can now say against it without opening themselves to charges of inconsistency, bad faith, or hypocrisy. However, it is also important to consider the perspective of the counterparty, to whom the substantive criticism is offered, and who invokes the Voluntarist Reply in response. When I invoke the Voluntarist Reply, I assert that my (real or hypothetical) interlocutor has chosen this agreement, instead of an alternative that they could have chosen, and presumably should have chosen, had they wished not to undertake these obligations. This puts in play not only their relationship to the agreement, but also my own, and my relationship to their consent, and how far I am entitled to rely on it. If the reason they entered into an agreement is that I myself violated or threatened to violate their moral rights, leaving them no choice but to agree, then I can hardly invoke that consent to answer their subsequent criticisms. This is another route to Cohen's “you're involved in it yourself” response: “For you are yourself more or less implicated in the act you seek to condemn if you caused the legitimate grievance to which the act is a response.”38 We need not think the fact of prior injustice licenses the making of false promises to recognize that the perpetrator of that prior injustice lacks standing to complain.39 Prior injustice can thus undermine the Voluntarist Reply, whether understood in its no-injustice or no-standing interpretations. Yet it seems plain that international economic relations constitute circumstances of injustice, whether because of contemporary inequalities or because of the historical injustices of colonialism that have so thoroughly shaped contemporary distributions. Does this preclude invoking the Voluntarist Reply to defend any economic agreements? The answer turns on the relation between injustice and agreement, and on the extent to which the existing injustice is one for which relevant counterparties have correlative duties. I address below some specific ways that existing international economic circumstances might impact the Voluntarist Reply. For now, let me sketch a more general distinction, between injustice that motivates or shapes an agreement, and injustice that is merely contemporaneous with it. The difference lies in how the agreement, or its context, links the making of the agreement with the continuation of that injustice. This is clearest in the case of coercion by threats: in the event of non-agreement, the addressee suffers some unjust harm, whereas if they agree then they do not. It may be harder to discern and adjudicate where one party is, ex hypothesi unjustly, in adverse circumstances, and the agreement represents one way they might better their condition, and thereby ameliorate the existing injustice.40 However, it is also important how far the injustice is one that the particular counterparty has duties to remedy. If we regard existing economic injustices as ones that all agents have duties to remedy, then many more agreements will be tainted by those injustices, and the relevance of the Voluntarist Reply reduces to the vanishing point. (This is the implication of many critical, including Third World, approaches to international law.) If, on the other hand, we understand responsibility as more narrowly circumscribed, then the Voluntarist Reply has greater potential scope. Perspective matters, both in identifying injustice, and understanding its implications for the criticism and defence of prevailing arrangements.41 I return to this point below. This no-prior-injustice condition thus imposes significant limits on the scope of the Voluntarist Reply. It is presented here as a minimum necessary condition for that reply to obtain. It seems likely that other conditions will also apply. In particular, there may be full information conditions: insofar as the Voluntarist Reply understands consent as waiving complaints that a party might otherwise make, we would usually expect a waiver to be limited to facts that a party was or should have been aware of. Certainly, if one party has intentionally concealed information this will undermine the reply's force.42 I examine a number of other potential constraints on the Voluntarist Reply in Section III. For now, let us examine a little further how the no-prior-injustice condition might affect specific agreements. The no-prior-injustice condition implies that determining the scope of the Voluntarist Reply requires first identifying the moral rights and duties that apply to and among states in the relevant domain. To judge invocations of the Voluntarist Reply in respect of international economic agreements, we must first know what justice requires in international economic relations, in the absence of those agreements (or, at least, in the absence of the particular agreement under scrutiny). This is plainly a very large question that I cannot hope to answer here.43 Better, therefore, to leave it open, assuming only that there are some duties of international economic justice, and allow readers to fill in for themselves their specific content. This leaves the discussion somewhat schematic, and the conclusions necessarily contingent. However, I hope it also means the analysis may be of use to those with widely varying substantive commitments. To get things going, it suffices to assume that at least some states have some moral rights and duties in the international economy. Given the centrality of market access to much international economic law, I will focus on moral duties to liberalize trade, mapping the implications of such duties for invocations of the Voluntarist Reply in this context. In what remains of this section, I sketch a number of plausible grounds for such duties, showing how they might affect the availability of the Voluntarist Reply in different contexts. The first, and most prominent, view regards duties to liberalize trade as responding to defects in the prevailing international economic order. Many scholars identify moral duties on rich states to liberalize their own trade, and to support and contribute to an open, stable, international trading system, whether on egalitarian or sufficientarian grounds.44 These duties may be subject to a various limitations: they may be duties to do one's fair share,45 or duties of maximum effort;46 they may be limited by the needs of, and effects on, one's own citizens;47 and there are obviously many open empirical questions about precisely which trade policies are most likely to achieve which morally significant results.48 The key point, however, is that states may not be morally free simply to refuse to enter into a trade agreement, or to dictate terms without reference to these moral duties. (Of course, they remain legally free to do both these things, at least on the dominant view.)49 Where a state pursues a hard-bargaining strategy, contrary to this duty, then voluntarism provides no reply to criticism of the terms offered or agreed. We might judge the Uruguay Round agreements in these terms. Even if we deny that concerns for poverty, equality, or economic development directly support duties of justice, historic injustices may give rise to remedial or compensatory duties, and trade agreements may be an appropriate mechanism for vindicating those duties.50 The historic wrongs of European colonialism constitute the unavoidable backdrop to the contemporary international economy. To the extent we think that today's rich states have remedial or compensatory duties arising from past colonial activities, we may regard the provision of beneficial trading terms as one way to address this. EU–ACP trade relations have sometimes been conceived in these terms (particularly from an EU perspective). They reflect specific historical connections between these countries, and it is in terms of that specific history that they are typically discussed.51 To the extent we think that history is duty-generating, we might expect trade arrangements to reflect and in part address this. If we think that responsibilities for historic injustices give rise to remedial duties, and that those duties might be discharged in part through the provision of beneficial trading terms, then responsibility-bearing states cannot invoke the Voluntarist Reply to answer criticisms of the agreements that they ask beneficiary states to enter into. Duties might also arise where past practices have given rise to legitimate expectations and consequent reliance on the part of counterparties, and the failure to grant market access, whether unilaterally or by agreement, would unjustly upset those expectations.52 As well as responding to historic injustice, we might understand the EU–ACP case in these terms. Under the Yaoundé, Lomé and Cotonou agreements, ACP states enjoyed preferential access to the European market for a number of decades. Economies adapt to the existence of trade preferences, whether through increasing investment in, or non-diversification from, industries where preferences are offered, so a subsequent loss of market access may leave them not only worse off than they were while preferences were in place, but potentially worse off than they would have been had they never had that preferential access. Having encouraged the ACP's reliance on the EU market, this argument suggests, the EU was not free to unilaterally withdraw that access, and hence could not invoke the Voluntarist Reply to defend the agreements it asked ACP states to accept in order to continue it. (The UK's reliance on EU market access post-Brexit might also be understood in these terms, although the fact that Brexit was a unilateral choice of the UK obviously muddies the waters somewhat.) The no-prior-injustice condition establishes one necessary limit on the Voluntarist Reply: it is only available where the party invoking it has themselves acted justly towards the other party. However, even where an agreement meets this quite stringent condition, there remain significant limits on the work that consent can do in answering criticisms of its terms. The limits canvassed below do not concern whether an agreement is voluntary: rather, they show how the moral force of wholly voluntary agreements runs out along a range of dimensions.53 How we understand these limits depends in part on our other moral commitments and, in the context of international economic relations, those commitments vary widely. As with the previous section, rather than work within one such moral framework, and risk losing the attention of those with different views, the discussion below shows how the moral significance of voluntariness varies, given four quite different sets of such commitments. These move, roughly, from the morally most permissive to the most demanding. The international moral minimalist denies the existence of extensive duties of international economic justice, and in particular rejects any egalitarian duties across borders. Thomas Nagel, for example, defends a view of international justice that limits states' natural duties to non-aggression and non-interference, denying any natural duties of international economic justice (albeit acknowledging a humanitarian concern for absolute poverty).54 The agreements that states make are ‘bare contracts’, binding in their terms, but not subject to or importing any wider sense of socio-economic justice. To the extent we owe economic duties across borders, these are products of those positive agreements and contingent relations. Duties apply to us only where these have been positively assumed. How might this view understand the Voluntarist Reply? If (almost) all duties arise from agreement, then we might expect the fact of agreement to in turn provide an ironclad justification for the content of those agreements. What other criterion could we have for evaluating them? Further, on such a sparse account of our antecedent moral rights, we will presumably regard as voluntary almost any agreement that meets the minimal legal standards in the VCLT.55 The mere fact that a counterparty is economically disadvantaged, even in extremis, gives them no basis to demand better economic terms than others are willing to offer. Yet there remain at least two routes by which even moral minimalism might limit the Voluntarist Reply. These derive, respectively, from the function of agreement in the moral minimalist account, and from the nature of the agent making that agreement. Taking the first: the moral minimalist characterizes agreement as the basis of international duties. However, the agreement norm, pacta sunt servanda, in turn requires some justification.56 Its

Suppose that in a jurisdiction there are 2,000,001 white voters and 1,000,000 black voters, all of whom enjoy equally weighted votes. The question of white supremacy is routinely put to a majority-rule plebiscite. In each such plebiscite, all white voters vote yes for white supremacy and all black voters vote no. This has been going on as long as anyone can remember, and it will continue for as long as anyone can foresee. This is a paradigm of a persistent minority, to which, intuitively, each black voter has an objection. What is their objection? One answer is that black voters don't get their preferences satisfied. Another answer is that black voters are oppressed by the eventuating policies of white supremacy. Yet another answer considers the white majority as a group. As a group, they have greater power to determine the outcome than have the blacks as a group. Indeed, the white majority as a group is always decisive. In the last plebiscite, all the whites voted for white supremacy and it passed; and if all the whites had voted against, it would have failed. By contrast, the black minority as a group is never decisive. They voted against and it passed; and if any assemblage of them had voted yes, it would still have passed. This is the agential social power enjoyed by one who finds herself among others disposed to act in concert with her to effect the realization of her intentions, preferences, or objectives, even if she has no power over them. It is the power enjoyed by members of a persistent majority under majoritarian decision-making: if over a broad range of issues over time, conditional on a voter supporting an alternative, a majority would also support it, then he has greater decision-making power than a voter for whom there is no such correlation of voting intentions—even if he exercises no power over the persistent majority.3 In this article, we raise a number of doubts about Abizadeh's suggestion that the power-of-numbers thesis can vindicate the thought that members of the minority as individuals have less voting power and thereby account for their objection to belonging to a persistent minority. Perhaps the most serious doubt is that while Abizadeh correctly holds that voting power must be assessed in part by counterfactualizing on votes—by asking what would have happened if a voter had voted otherwise than he in fact did—he does not counterfactualize in the right way. We cannot vindicate the thought that members of the minority have less voting power if we assume an a priori measure of voting power that abstracts from information about the distribution of political preferences and its causes. An example is the Banzhaf (or Penrose–Banzhaf) measure of voting power, according to which a voter's power is her probability of casting a decisive vote if all other voters vote independently and with equal probability for either alternative.4 (A voter's vote in favor of (or against) a measure is decisive if the measure passes (fails), but would have failed (passed) if the voter had instead voted against (in favor).) An a priori measure of voting power will not register any differences between members of the persistent minority and anyone else, because it ignores, by construction, the facts about the distribution of preferences and its causes, in virtue of which some voters qualify as members of a persistent minority. Abizadeh therefore seeks to vindicate the power-of-numbers thesis with an a posteriori measure: power is calculated on the basis of, rather than in abstraction from, information about how social structure influences the distribution of political preferences. This measure of power also differs from the Penrose–Banzhaf measure in making voting power a function of the degree to which the voter can expect her actions to be “efficacious,” where even non-decisive, redundant votes for the winning alternative count as partially efficacious. So let us consider voting power in the a posteriori context, in which we take into account information about how voters are likely to vote. The information on the basis of which power is measured may be more or less predictive of voting behavior. At the limit, knowledge of a person's structural position, together with knowledge of how they will vote, fully predicts how everyone else will vote. For the sake of simplicity, we start by assuming that such fully predictive information is available and is the appropriate basis for measuring voting power. This assumption is obviously unrealistic, but it simplifies the discussion by relieving us of the need to calculate probabilities, and it is an innocent simplification because it does not prejudice the assessment of the power-of-numbers thesis. It would be bizarre to argue that a member of a persistent minority has an objection only when there is genuine uncertainty about the future distribution of votes, but not when the future distribution, and her status as a political minority, is certain. We will, in any case, relax this assumption in due course. We assume, in particular, that the relevant information perfectly predicts the scenario described above: in every plebiscite, the 2,000,001 white voters are certain to vote in favor of white supremacy and the 1,000,000 black voters are certain to vote against. Our question is whether a given white voter can be said to have significantly more power than a given black voter when voting power is measured on the basis of this information. In this situation, everyone's probability of casting a decisive vote is zero. Thus, if a voter's power corresponds to the probability of casting a decisive vote (an assumption behind the Penrose–Banzhaf measure), then every voter, white or black, has the same amount of power. What if, following Abizadeh, we grant that a member of the winning side can enjoy partial efficacy, even if they are not decisive? When everyone casts equally weighted votes, as will be true in all our examples, we can assume the degree of partial efficacy is a function of the size of the winning coalition (relative to the size of the electorate). At one end of the spectrum is the case of someone who votes for the winning side along with every other member of the electorate; their vote has some partial efficacy, but not as much as it would have if the relative size of the winning coalition were smaller. At the other end is the case of the fully decisive voter who votes for the winning side and is part of a minimal decisive coalition (for example, a bare majority under simple majority rule). In our example, when a white voter votes with the winning coalition, in favor of white supremacy, then his vote has some partial efficacy, namely the degree of partial efficacy that corresponds to voting with 2,000,000 other voters, in an electorate of 3,000,001 voters, for the winning alternative. If the white voter were to vote against white supremacy, the efficacy of his vote would be zero. If a black voter were to vote with the winning coalition, in favor of white supremacy, then her vote would have some partial efficacy, namely the degree of partial efficacy that corresponds to voting with 2,000,001 other voters, in an electorate of 3,000,001 voters, for the winning alternative. When the black voter votes against white supremacy, the efficacy of her vote would be zero. Their partial efficacy scores are virtually identical. These claims rest on the tacit assumption that if a voter were to vote differently from how they will actually vote, everyone else would still vote the same as they are going to vote in actual fact. One must reject this assumption if one wishes to argue for the power-of-numbers thesis. Some might object to the tacit assumption on the following grounds. Given our stipulations about the case, all white voters are certain to vote in the same way. So, it follows that if a given white voter were to vote against white supremacy rather than for, then all other white voters would also vote against, too. But that does not follow. Suppose newspapers A and B always report the same events; when the scandal breaks, they are each certain to report it. It does not follow that if A were to refrain from reporting the scandal, then B would refrain as well. Probabilistic dependence does not imply counterfactual dependence.5 One would reject the tacit assumption with good reason if one voter's decision about how to vote causally influenced other voters' decisions. But let us assume that no one's vote has any causal influence over anyone else's vote. If you like, assume everyone votes secretly and simultaneously. The power-of-numbers thesis is not supposed to rest on voters' abilities to influence each other. A different reason one might reject the assumption is if one thought that measurements of voting power ought to reflect counterfactualizing not only the voter's action, but also the past events that causally influence the voter's action as well as other voters' behavior. One might reason as follows. The example assumes that all white voters always vote as a block and all black voters always vote as a block. This pattern of correlation can only hold in virtue of underlying social-structural causes that have one (fully determining) impact on white voters' preferences and an opposite (fully determining) effect on black voters' preferences. If a given white voter were to vote against the proposed ballot measure, it would have to be because the underlying social-structural influences caused him and all other white voters to be opposed. Thus if a given white voter were to vote against, they would find themselves on the winning side and would enjoy some partial efficacy, just as they enjoy some partial efficacy when they vote in favor; by contrast, if a given black voter were to vote in favor, she would still be on the losing side and her vote would still be inefficacious, just as her actual vote against is inefficacious. Thus the white voter enjoys greater power than the black voter. This reasoning involves “backtracking” counterfactuals: the assumption is that when one counterfactualizes over a white voter's action and looks for the closest possible world in which he voted no, one should include worlds with different histories up until the time of that vote.6 This allows one to say that the white voter would be in the majority even if he were to vote against the proposal, provided that the closest possible world in which he votes no is a world with a different history, one in which some common cause led all the other whites to prefer no and so to vote no as well. In the next section, we will argue that the counterfactuals that enter into the measurement of power should not be interpreted in a way that permits backtracking.7 But even if we allow backtracking counterfactuals, it may not help the power-of-numbers thesis. Suppose that the closest world in which a given white voter votes no is a world in which the normal link between social-structural causes and political preferences is broken for him alone, but not for other white voters (he regularly converses about racial justice with a colleague, and the closest possible world in which he votes against white supremacy is one in which these conversations induce a moral epiphany, severing the causal link that continues to make other white voters' political preferences a deterministic function of their position in the racial hierarchy). Then the crucial counterfactual would be false: if he were to vote against white supremacy, the other whites would still vote in favor, and his situation is not appreciably different from the situation of any black voter. Nor would it help much if the closest world is one in which he and, say, a thousand others had racial justice epiphanies. One worry about the power-of-numbers thesis is that, offhand, it's not clear why the closest possible world in which a given white voter votes no (even among those worlds with different histories) is one in which the social-structural causes that influence all other white voters' preferences are different. So even if we permit backtracking, it is not clear why, in our example, it would be true that if a given white voter were to vote against white supremacy, then all other white voters (or at least a subset that constitutes a majority) would also vote no. Another worry about the power-of-numbers thesis, as a diagnosis of our intuitive objection to a paradigm case of a persistent minority, is that the intuitive objection does not seem to depend on which of these backtracking counterfactuals is correct, but the power-of-numbers thesis does. In any event, this backtracking counterfactual isn't the relevant conditional for assessing agential power generally or voting power in particular. If we are measuring the power of a given white voter on the day of the election, we should ask what would happen if he were to vote no, holding fixed the actual history of the world up until the time he votes. The following cases illustrate the importance of excluding backtracking from the analysis of agential power. Jones sometimes prays for rain after eating breakfast. To be precise, he prays on all and only those mornings when, while eating breakfast and reading the newspaper, he notices that the weather forecast predicts rain. Thus, conditional on praying for rain in the morning, the probability that it will rain in the afternoon is high. And conditional on not praying, the probability that it will rain is low (because, conditional on not praying, there was no rain in the forecast, and forecasting the weather in his area of southern California is not hard). Jones does not have the power to make it rain, of course. We can explain this fact with normal, non-backtracking counterfactuals, but not with backtracking counterfactuals. On any given morning when Jones prays, the probability of rain is high, but would remain high even if he were not to pray for rain; and on any given morning when he does not pray, the probability of rain is low, but would remain low even if he were to pray. If agential power is measured with backtracking counterfactuals, however, Jones does have the power to make it rain. With backtracking, we have: on any given morning, if Jones were to pray, then (it would be because rain was in the forecast and so) it would probably rain, and if he were not to pray, then (it would be because rain was not in the forecast and so) it would probably not rain. Backtracking counterfactuals are no more plausible a basis for ascriptions of power when we move from the single-agent example to a voting situation. A boss threatens to fire anyone who votes for a proposal to unionize. As a result, each worker knows that if she votes for the proposal to unionize and it fails, she will lose her job; otherwise she will keep it. Needless to say, no one sticks her neck out for the union. It seems false to say, and a cruel joke in the mouth of the boss, that each worker has the power to effect, with the others, the passage of the unionization proposal.8 And yet it is plausibly true if we assess a worker's power using backtracking counterfactuals. The closest world in which she votes for the union might be one in which the boss never made the threat and no one faces the prospect of losing their job for voting for it, in which case everyone else would also vote for the union and the unionization proposal would pass. Thus if she were to vote in favor of the unionization proposal, then—backtracking and counterfactualizing the boss's past behavior—it would pass, in which case she would have effected, with others, the passage of the proposal. Or suppose an American football coach is deciding whether to go for a touchdown on fourth down or instead have his kicker kick a field goal. One consideration when he makes such decisions is the wind: if the wind is too strong, the kicker will not have the power to make the field goal, and so in such circumstances the coach decides to go for the touchdown. Suppose in actual fact the winds are strong, so he decides to go for the touchdown. After they fail, his critics say: “Coach made the wrong call. The kicker did have the power to make a field goal. In the closest possible world in which he kicked, he kicked because coach told him to, and coach told him to because the wind conditions were favorable, and since they were favorable, kicker made the field goal.” We have taken Monday-morning quarterbacking to a whole new level. We have so far considered an example in which the measurement of voting power is made on the basis of information that is fully predictive of everyone's voting behavior. Under this assumption, the differences in voting power between an individual member of the persistent majority and an individual member of the persistent minority were negligible. If one relaxes this assumption and measures power instead on the basis of information that is only partially predictive of the distribution of votes, there are cases in which members of a structural minority have more power in expectation than members of a structural majority—provided, again, that one does not measure power with backtracking counterfactuals. Voter 1 belongs to a structural minority, voters 2, 3, 4, and 5 to a structural majority. For each plebiscite t, there is a social-structural variable S t $$ {S}_t $$ that influences the voting behavior of all five voters. Either S t = 0 $$ {S}_t=0 $$ or S t = 1 $$ {S}_t=1 $$ . If S t = 1 $$ {S}_t=1 $$ , then the probability that voters 1, 2, and 3 will support the measure and voters 4 and 5 will oppose it is p $$ p $$ ; and the probability that voter 1 supports it, but voters 2, 3, 4, and 5 oppose it is 1 − p $$ 1-p $$ . If S t = 0 $$ {S}_t=0 $$ , then the probability that voters 1, 2, and 3 will oppose the measure and voters 4 and 5 will support it is p $$ p $$ ; and the probability that voter 1 opposes it, but voters 2, 3, 4, and 5 support it is 1 − p $$ 1-p $$ .9 Assume 0 < p < 1 / 2 $$ 0 1 / 2 $$ 1-p>1/2 $$ of the time), voter 1 finds herself in the minority and voters 2, 3, 4, and 5 find themselves in the majority. We now show that if p > 3 / 23 $$ p>3/23 $$ , then in any given plebiscite, voter 1 has more power than voter 5, despite the fact that voter 5 has numbers on his side. This conclusion holds for every plebiscite t, whatever the history preceding the plebiscite and, in particular, whatever the realized value of S t $$ {S}_t $$ . The argument for the conclusion abstracts from all facts about the voters' situation other than (1) the voting rule and (2) the social structure and probabilistic dependence that it induces in voting behavior. Our argument uses the measure of partial efficacy proposed by Abizadeh.10 According to this measure, in an electorate of five voters, where everyone casts equally weighted votes, a vote for the winning alternative has an efficacy score of 1 if it is one of three votes cast for the winning alternative, a score of 3/4 if it is one of four votes cast for the winning alternative, and an efficacy score of 3/5 if it is one of five votes cast for the winning alternative. A vote cast for the losing alternative has no efficacy. Perhaps a different measure of efficacy could be proposed. But so long as the measure holds that a successful voter's efficacy is a function f ( x ) $$ f(x) $$ of the fraction x $$ x $$ of the electorate voting for the winning alternative, and (in the 5-voter case) the function satisfies 0 < f ( 5 / 5 ) < f ( 4 / 5 ) < f ( 3 / 5 ) $$ 0 3 / 4 ( 1 - p ) $$ 3/5+2/5p>3/4\left(1\hbox{-} p\right) $$ , which is equivalent to p > 3 / 23 $$ p>3/23 $$ . Thus, provided p > 3 / 23 $$ p>3/23 $$ , each of voter 1's actions is more efficacious in expectation than the corresponding action of voter 5, so voter 1 must be deemed more powerful than voter 5. But voter 1 belongs to a structural minority and voter 5 to a structural majority. Voter 1 is in the minority, and voter 5 is in the majority, a fraction 1 - p > 1 / 2 $$ 1\hbox{-} p>1/2 $$ of the time, due to their positions in the social structure and its influence on voting behavior. We assumed without loss of generality that we were dealing with a plebiscite in which voter 1 supports the measure. But we would have reached the same conclusion if we had instead assumed that voter 1 opposed it. Thus, our conclusion does not depend on the actual history leading up to the plebiscite (in particular on the actual realization of the social-structural variable S t $$ {S}_t $$ ). It relied only on the assumption that plebiscites are decided by simple majority rule and that the social structural variable S t $$ {S}_t $$ induces probabilistic dependence in voting behavior in the manner described. The intuition for the result is simple. Voter 1 can have a higher probability of being in the minority than voter 5, even though she also has a higher probability of casting a decisive vote. If the latter probability is sufficiently high (in our example, if p > 3 / 23 $$ p>3/23 $$ ), the conclusion will be that voter 1 is more powerful. The critical assumption is that we assess counterfactual suppositions about voter 1's actions without counterfactualizing the features of her history (the realized value of S t $$ {S}_t $$ ) that fix the probability distribution over other voters' behavior and thereby fix the probability that her vote will be decisive. The problem is that appealing directly to the content of other voters' preferences, but not the power-assessee's, is insufficient for explaining the inequality faced by a persistent-minority member: if the persistent-majority and persistent-minority groups are defined in terms of members' substantive preferences, then counterfactualizing over the power-assessee's preferences implies abstracting from her membership of the persistent minority. Her power inequality would consequently fail to register. … A persistent minority must therefore be defined and picked out in terms of its members' position in other social structures—such as language, ethnicity, race, class, or geography—which position induces a correlation in preferences.12 This claim about how persistent minorities must be defined is important, because it is all that Abizadeh invokes to avoid the result that the power of numbers thesis “absurdly requires compensating political eccentrics on democratic-equality grounds.” It is his explanation for why “If libertarians or Bolsheviks are persistently outnumbered, democratic equality does not call for formal-procedural inequalities to compensate them.”13 What exactly is the problem with defining “persistent minority” in terms of actual preferences? The worry seems to be that, so defined, there is something incoherent about asking “what if the persistent minority had different preferences?”, just as there might seem to be something incoherent about asking, “what if every student in the class got a grade better than the median grade?”. But neither question is incoherent provided we take “the persistent minority” and “the median student” to be “rigid designators,” rather than terms whose referents vary across possible worlds. If, in the actual world, the median grade is a B, then the second question is asking what if every student in the class got better than a B. And if you are the one person who, in actual fact, belongs to the minority in every election, then the first question is asking what would happen if you had different preferences and voted differently. The justification for defining “persistent minority” in terms of structural position cannot be that this is the only definition that permits a coherent formulation of counterfactual statements about the preferences and actions of the persistent minority. We can intelligibly ask what would happen if a libertarian or Bolshevik, who in actual fact persistently finds himself in the minority, were counterfactually to vote differently from how he will actually vote, and compare that with what would happen if someone who in actual fact is in the mainstream were to vote differently from how she will actually vote. If the answer—in virtue of backtracking counterfactuals—is that the Bolshevik would still be in the minority, and the mainstream voter would still be in the majority, then the power-of-numbers thesis implies there is an inequality in power. If one rejects the use of backtracking counterfactuals to answer this question, then the libertarian and the Bolshevik will have no less power than anyone else, at least not in virtue of being persistently outnumbered. But the argument for the power-of-numbers thesis needs backtracking counterfactuals, so the question is whether there is any justification for excluding them when the question concerns libertarians and Bolsheviks, but including them when it concerns structural minorities. Backtracking counterfactuals need not be any less plausible in these other “non-structural” cases. Suppose there are 1,000,000 voters who consistently vote in whatever way the Bolshevik party leaders advocate, and 2,000,001 voters who consistently vote against whatever the Bolshevik party leaders advocate. Take any election, and assume (without loss of generality) that the pro-Bolshevik voters are all going to vote yes, the anti-Bolshevik 2,000,001 all no. If a pro-Bolshevik were to vote no, it could only be because the party vanguard had instructed them to vote no, in which case all the anti-Bolshevik voters would vote yes, and our voter would still be in the minority. That judgment is no less reasonable than the judgment that if a black voter were to vote no, in a society polarized along racial lines, it could only be because of social-structural conditions that induce all white voters to vote yes, such that the black voter would still be in the minority. We have argued that backtracking counterfactuals should be excluded from analyses of power in all cases, but if there were any justification for including them in the case of structural minorities, it is unclear why the justification would not carry over to non-structural minorities. And then one gets the absurd conclusion that the power-of-numbers thesis, if it justifies compensating the power deficits of racial or religious minorities, also justifies compensating libertarians and Bolsheviks. It is also perhaps worth noting here that once backtracking counterfactuals are admitted, the majority or minority need not be persistent in order for there to be power-of-numbers inequalities. Imagine a one-hit-wonder at the momentary apex of their popularity. Wonder's fans, for a day, form a majority. A vote is taken on that day whether to declare the day Wonder-appreciation-day. Wonder's fans will vote for whatever Wonder decides. If Wonder says yes, then the fans will vote yes; if Wonder, in modesty, says no, then the fans will vote no. Each of the fans in the majority enjoys power of numbers that each nonfan does not. But none of this will persist. Even if an agent's power should not in general be assessed on the basis of backtracking counterfactuals, could there nonetheless be a justification for incorporating them into a specific kind of voting power measure? We anticipate an erroneous argument for doing so that proceeds from the reasonable assumption that the desired measure of voting power ought to abstract from features of a voter's actual situation, such as the actual history preceding the vote, and draws the errant conclusion that voting power ought to be assessed by counterfactualizing these features of the voter's situation. The assumption that a measure of voting power ought to abstract from certain features of a voter's situation is reasonable. The Penrose–Banzhaf measure, for example, only purports to register the power a voter enjoys in virtue of the voting rule.14 That is why it abstracts from all other features of a voter's situation (such as how the voter can expect other voters to behave) that are plausibly relevant to a comprehensive assessment of agential power. Analogously, if a measure only aims to capture the power a voter enjoys in virtue of (1) the voting rule and (2) their position in the social structure, then it ought to abstract from the voter's actual circumstances. One cannot say what would happen if the voter were to vote yes, unless one makes some assumption or other about her circumstances, but one's conclusion should not be sensitive to what her circumstances actually are. So one must ask, “if she were to vote yes, what would her circumstances probably be, in virtue of the history that would have preceded her decision to vote yes?”. One then considers all these probable circumstances, her efficacy in each, and weights these scores by the probability with which such circumstances would have occurred in virtue of the history that would have preceded a decision to vote yes. That is, one uses backtracking counterfactuals to compute the probabilities of different configurations of votes for the other voters. The second proposition does not follow from the first, however. There is a logical gap between the reasonable supposition that a measure of voting power ought to abstract from a voter's actual circumstances and the conclusion that it ought to do so by asking backtracking counterfactual questions. To see that the conclusion does not follow, it suffices to establish the existence of a measure that abstracts from the voter's actual circumstances, but eschews backtracking counterfactuals. Consider the following approach. For each of the possible circumstances the voter might find herself in, compute the power she would have in those circumstances with normal, non-backtracking counterfactuals in just the way we did for the paradigm case and the example in Section IV. For example, there is the scenario in which a worker is voting on whether to unionize, and her boss has issued a threat to her and all the other workers; and the scenario in which they are voting, but the boss has not issued a threat. In the example from Section IV, there is the scenario in which S t = 1 $$ {S}_t=1 $$ and another scenario in which S t = 0 $$ {S}_t=0 $$ . Compute the voter's power in each of these scenarios (just as we have already done for the first of these two scenarios in Section IV).15 Now average over the measurements of the voter's power taken in all these possible circumstances, weighting according to the probability of each circumstance. The result does not depend on which of the possible scenarios is the actual one. The power-of-numbers thesis will still fail to hold if one considers it a thesis about “aggregate” voting power in this sense, rather than a thesis about voting power in a particular circumstance. Since no voter derives special power from having numbers on her side in any particular circumstance, no special aggregate power will emerge when one computes a weighted average of her power across the possible circumstances. Abstracting is not the same as counterfactualizing. One can measure power in a way that abstracts from a voter's actual circumstances and their history—in a way that makes measurements of power independent of a voter's actual circumstances and history—without counterfactualizing them. Because the use of backtracking counterfactuals in the measurement of voting power does not follow from the (reasonable) assumption that a voting power measure should abstract from the voter's circumstances, a proponent of the power-of-numbers thesis must find some other argument for it. We are not sure what it could be. Provided that a posteriori power is measured using standard, non-backtracking counterfactuals, facts about who has more or less of it will not closely track membership in structural minorities and majorities. In our paradigm case, members of structural minorities and majorities have all but equal a posteriori power. In the case from Section IV, members of structural minorities have greater a posteriori power than members of majorities. In so far as this article's argument underscores the vagaries of a posteriori power distributions, it may make one skeptical that democracy could require equality in this dimension, or even an approximation to equality. Once we move from a priori to a posteriori measures of voting power, there is no reason to expect an equal scheme of basic political rights to produce an equal distribution of (a posteriori) voting power. Just the opposite. Almost all political institutions, except those that have been calibrated to the actual social-structural conditions just so, will produce asymmetries in a posteriori voting power. And when the social-structural conditions underlying the distribution of preferences change, political institutions will also need to change if they are to preserve equality. Moreover, the kind of changes in social-structural conditions that could undermine equal a posteriori voting power are not normally considered threats to democratic equality. So-called swing voters will enjoy more a posteriori voting power than voters who are reliably on one side of a partisan divide. But even when this voting behavior reflects structural causes—the ways in which race, class, and other social-structural identity categories pull some citizens towards one party, some towards the other, and leave some conflicted and open to appeals from both parties—the fact that the rigidly partisan voter has less a posteriori voting power than the swing voter is not plausibly considered a breach of democratic principles. One weird implication of the view is that some citizens could violate political equality, as though they were denying someone an equally weighted vote, merely by “freeing” themselves of social-structural influences and thereby altering the probability distribution of preferences in a way that reduces their efficacy. As a simple illustration, suppose an electorate of three voters, 1, 2, and 3, faces a series of binary votes, and due to social-structural causes they always vote as a block and with equal probability for either alternative. Whether one uses backtracking counterfactuals or not, everyone has equal a posteriori power.16 But now imagine that voter 1 “frees” himself of the causal influence of the background social structure and comes to vote independently of voters 2 and 3, who still vote as a block, and everyone still votes with equal probability for either alternative. Now voter 1 has less power than either of the other two. Each of his actions has an expected efficacy of 1/6: if he were to vote yes, then with probability 1/2, the block would vote yes and his efficacy would be 1/3; and with probability 1/2, the block would vote no and his efficacy would be zero. The expected efficacy of voting no is calculated analogously and also comes out to 1/6.17 But each of voter 2's actions has an expected efficacy of 2/3, assuming we use backtracking counterfactuals in the calculation:18 if she were to vote yes, then with probability 1/2, voters 1 and 3 would each vote yes, in which case her efficacy is 1/3; and with probability 1/2, voter 1 would vote no, but 3 would vote yes, in which case her efficacy is 1. Thus her expected efficacy of voting yes is 1/6 + 1/2 = 2/3. The expected efficacy of voting no is calculated analogously and also comes out to 2/3. Some might find this result counterintuitive, because they do not think one can become less powerful in virtue of increasing one's ability to form political opinions independently of social-structural influences. But that is not our point. Our point is rather that if one has reasons to try to preserve political equality, but political equality is understood in terms of equal a posteriori voting power, then the value of political equality gives voter 1 reasons not to cultivate an ability to form political opinions independently of social-structural influences. That implication seems strange. These considerations make us doubt that a posteriori voting power is the currency of democratic equality. The power-of-numbers thesis is supposed to vindicate the intuition that a member of a persistent minority has a complaint against majoritarian democracy. If the thesis is true, then majoritarian democracy does not in general respect a principle of democratic equality, and the member of the persistent minority can ground their objection by appeal to this principle. We have noted several grounds for doubts about the power-of-numbers thesis, most importantly its reliance on backtracking counterfactuals. Supposing one accepts our critique of the power-of-numbers thesis, how else might one try to ground the objection? We noted several alternatives at the outset. One is that the real objection is not to an inequality in voting power, but instead to an inequality in the satisfaction of preferences: namely, that black voters' preferences go unsatisfied vote after vote, whereas white voters' preferences are reliably satisfied. Another alternative is that, again, the real objection is not to an inequality in voting power, but instead to the substantive inequality that results from the policies voted into effect: that black voters are subordinated by a regime of white supremacy. A final alternative recognizes an inequality in voting power, but at the level of groups rather than individuals. Even if no individual member of the structural minority has appreciably less power than any individual member of the structural majority, the latter group has more power than the former. Perhaps the distribution of power across social groups takes on normative significance—above and beyond the distribution of power across individuals—in societies with histories of group-based injustice.19 We do not have space in this article, however, to develop these suggestions and evaluate their merits. Those are tasks for other work. The authors would like to thank Zeynep Pamuk, two anonymous referees, and especially Arash Abizadeh for helpful comments on previous drafts of the article. None relevant. There are no potential conflicts of interest relevant to this article. The authors declare human ethics approval was not needed for this study. All relevant data are included in the article. Here we explain why the argument from Section IV generalizes to any reasonable measure of partial efficacy. We restrict attention to the five-voter example for the sake of simplicity. Our assumptions ensure that the right-hand side is a number strictly greater than 0 and strictly less than 1/2. So if p $$ p $$ is greater than this threshold, but less than 1/2, voter 1 is more powerful than voter 5—each of voter 1's actions has greater expected efficacy than the corresponding action for voter 5—even though voter 1 is more often than not in the minority and voter 5 is more often than not in the majority, due to the structural causal influences on their and other voters' preferences.

Many of us will remember the sense of foreboding induced by the simple threat, usually uttered by a teacher: ‘It will go on your permanent record’. This administrative bogeyman exploits our early awareness of the importance of being able to leave some things in the past. While some of the things we do may go on public record at various points in our lives, it used to be possible to comfort ourselves with the thought that these would soon be buried deep in the archives, where most people would not care to look. In the digital age, however, search engines do the digging and can serve up previously long-forgotten results on a simple search of a person's name. This article takes up the question of what kinds of claims we have against information being dug up from our past. Specifically, it focuses on information that is legitimately a matter of public record. When some information has been stored in an archive and has, in principle, been accessible to anyone, what claims, if any, do we have against that information being brought back to light? Many people would find it intrusive for someone to dig through archives and publicize afresh the information they found out about you there, but it is not clear what the basis of such a complaint would be. Moreover, it is this kind of complaint that is provided protection through data-protection provisions colloquially referred to as the right to be forgotten. Such measures are typically framed as privacy protections. While some have argued that we can have privacy rights over information that has been made public,1 I argue that appeals to privacy fail in cases concerning information that is legitimately a matter of public record. Paying attention to the reasons we have to object to the dredging up of outdated information reveals a new category of claims that are distinct from claims to privacy, but serve the same general interest in self-presentation that privacy scholars have long been concerned with. I call these claims against distortion. We can understand such claims as falling under a general principle of reputational control embedded in the historic right of personality. That general principle provides the basis for claims to privacy, claims against defamation, and, as I will argue, claims against distortion. The purpose of invoking the right of personality is to identify a general principle embedded in that concept, and to then draw out a taxonomy of claims that serve that principle in distinct ways. One advantage of separating these three categories of claims is that it allows us to retain clear boundaries around the concept of privacy, which has often come under fire for suffering from a plethora of meanings.2 On my account, claims to privacy are claims to prevent certain information from becoming publicly available beyond one's audience of choice. Privacy provides one aspect of reputational control by allowing us to decide who we share various aspects of oneself with. Claims against defamation are claims against having false statements made about oneself in a way that is detrimental to one's reputation. Claims against distortion are claims against having one's public profile distorted in ways that involve the dissemination of true information. Such distortion can happen when true information from someone's past is presented in a way that suggests it would be appropriate to hold them accountable for it, when it is no longer appropriate to do so. Identifying the category of claims against distortion and analysing its place alongside claims to privacy and claims against defamation allows us to broaden the scope of the justificatory basis of data-protection provisions. In doing so, it widens the scope of debates about when and why we have reason to give individuals more control over their personal information. Data-protection provisions have tended to be framed and defended as measures which protect the privacy rights of individuals. There is an important place for data-protection regulations which serve that purpose. However, not all data-protection regulations can be understood or justified as privacy protections. Trying to analyse all data protections through the lens of privacy therefore unduly constrains our thinking about how such regulation may be justified on the basis of providing important protections to the individual. The main argument proceeds by examining the data-protection provisions known as the ‘right to be forgotten’. Specifically, it addresses the aspect of this regulation that gives individuals claims to the erasure of certain results that are displayed when their name is entered into a search engine. The aspect of the right to be forgotten that requires explaining is the rationale for erasing search-engine results that point to some content, while leaving the same content publicly available at the original source. Privacy-based approaches fail to adequately explain the purpose that is served by this aspect of the right to forgotten, as it has been implemented under the EU's General Data Protection Regulation (GDPR). This is because what is at stake in the case of the curation of a person's search profile is not always what information is publicly available, but how that information is framed or contextualized. When a search engine serves outdated or irrelevant search results for a person's name, this can amount to a distortion of their public profile. We have reason to care about these practices, and give individuals claims against them, but such claims are not based in a right to privacy. They are better understood as claims against distortion. While the right to be forgotten is a relatively narrow aspect of data-protection regulation, it shares a feature that is common to many concerns about our current situation with respect to the collection and use of personal data. It is a case in which the information in question has already been made public, and either cannot or should not be made entirely private again. And yet, concerns remain about the erosion of the individual's control over how their personal information is presented to others. Under these circumstances, we need to find a way to shut the stable door after the horse has already bolted. Claims against distortion provide a normative basis for regulations aimed at providing such control. What is colloquially referred to as the right to be forgotten is not a general right as such, but a set of provisions laid out in the GDPR.3 Article 17 of the GDPR sets out a ‘right to erasure’, which gives data subjects the right to obtain from data controllers the erasure of personal data concerning the data subject. When relevant criteria are met, such as when the data is no longer necessary to the purposes for which it was collected, the data controller must respond to erasure requests and delete the data in question without undue delay. One of the controversial aspects of this legislation is that search-engine providers such as Google are considered to be data controllers, and search-engine results are classed as personal data under certain circumstances.4 This means that an individual can request the removal of certain search results that are displayed when her name is typed into Google as a search term, and Google is obliged to remove them if the relevant conditions set out in Article 17 of the GDPR apply. This aspect of the right to be forgotten has come under intense criticism from search-engine operators, media commentators, and scholars alike.5 Critics have argued that such measures go too far in curtailing freedom of information, especially when the search results in question pertain to information that was lawfully made public. It has been suggested that allowing individuals to make private information that has already been in the public domain, and indeed remains publicly accessible through other means, is futile at best, and at worst based on confused or incoherent premises.6 Finally, the way in which the right has been implemented has been criticized almost to the point of ridicule for being counterproductive on its own terms. The case that formed the basis of the landmark ruling on the matter by the Court of Justice of the European Union (CJEU) is the prime target of those charges of inconsistency. The plaintiff in the case, Mario Costeja González, had requested the erasure of details of a foreclosure notice issued against him, which had been published on the web pages of the newspaper La Vanguardia. The court decided that the link to the original newspaper record that featured as a search result on Google when Costeja's name was entered as a search term should be erased. However, La Vanguardia was not required to remove the original article from its website. This inconsistency in the application of the rules of erasure between the original source and the Google search-results page is a principal bone of contention in analyses of the ruling. Moreover, in an ironic twist, the case became subject to a phenomenon known as the Streisand effect—when the attempt to censor or protect private information results in it becoming much more widely publicized.7 The landmark ruling was widely reported. As a result, now when you enter Costeja's name as a search term on Google, the first page of search results is entirely populated with links to reports about the ruling, which contain reference to the facts of the case, including mention of the original foreclosure notice Costeja was seeking to have ‘forgotten’. Because the facts of the legal case are a relevant and sufficiently weighty matter of public interest, it has not been deemed appropriate for these search results to be erased, but nor has the judgment about the erasure of the original link to the foreclosure notice been reversed. Defending these aspects of the ruling through the lens of privacy presents a challenge, given that the information in question remains in the public domain via the original source, and has subsequently been amplified through the Streisand effect. I will argue that the ruling can nevertheless be interpreted as serving an important interest in self-presentation if we understand it as based on a claim against distortion, rather than a claim to privacy. This interpretation allows us to explain away the seeming inconsistency in treating the original source and the Google link differently with respect to erasure. While the Streisand effect is likely to be a rare side-effect particular to the landmark case, analysing this aspect of the case helps to illustrate the rationale for treating search-result links differently from source material. It sheds light on the importance of how information is framed in a given context, over and above the question of whether some piece of information is in the public domain. This idea is central to the case for claims against distortion. Before making the positive case for a distortion-based justification for the right to be forgotten, further argument is required to motivate the move beyond privacy-based justifications. Several scholars have made the case in favour of the right to be forgotten by drawing on the idea that the forgetting of information over time has long been an important aspect of our cultural practices. Such gradual forgetting, it is argued, has been crucial to our ability as individuals to lead adequately autonomous lives, unencumbered by mistakes or embarrassing mishaps from our past. From this perspective, the widespread accessibility of old information made possible in the digital age represents a radical and destabilizing shift, which threatens our ability to live our lives on our own terms as the years go by.8 Proposed solutions to this problem have focused on the principles and pragmatics of how to strike the right balance between the protection of individual privacy on the one hand, and the public interest in information on the other, once we've recognized that the weighting of this balance might shift over time. Viktor Mayer-Schönberger, for example, argues that information should have an expiry date after which it automatically gets deleted.9 Less radically, Giovanni Sartor makes the case that regulators should direct data controllers as to when the balance has shifted such that information warrants deletion.10 On the basis of an argument from moral autonomy, Meg Leta Jones has argued that we should move towards a model of information stewardship for individual users, allowing them ultimate control over their online profiles.11 I share the general concern that the widespread accessibility of outdated personal information may lead to a stifling social environment in which people are constantly reminded of and held back by events from their past. I am also sympathetic to the idea that individuals should have more control over their personal information profiles. However, to the extent that the focus in this domain has remained within the framework of assessing whether information in various cases should be ‘public’ (accessible) or ‘private’ (erased), this framing of the debate still implies a univocal approach to erasure across different information sources. The univocal implications of the privacy framing are demonstrated by Robert Post in his critique of the Google Spain ruling.12 Post argues that the decision conflated two conflicting senses of the right to be forgotten, one based on a bureaucratic principle of data protection, and the other on the notion of dignitary privacy. According to Post, the bureaucratic principle assumes that data are always collected for instrumentally specified purposes, and warrant erasure when no longer required for those purposes. Dignitary privacy rights, by contrast, seek to ‘define and enforce social norms of respectful expression’.13 Protecting dignitary privacy requires a balance to be struck between protecting the dignity of the individual against the public interest in freedom of expression. Post points out that the public interest extends broadly to information that ‘can or will become part of the agenda of public action’, as well as to ‘maintaining the integrity of the structure of communication that makes public discourse possible’.14 Post argues that Google forms part of that structure of communication. The ruling therefore should have followed a principle of dignitary privacy. According to a principle of dignitary privacy, the standard by which to assess whether some information ought to be accessible to the public is the extent to which that information is deemed offensive or newsworthy. If deemed sufficiently offensive to cause harm to the dignity of the individual, the individual has a privacy claim against its publication. In the context of dignitary privacy, according to Post, ‘what matters is not the content of information, but the context and meaning of particular communicative acts’.15 A key tenet of Post's argument is that Google links carry no further meaning over and above the content of the underlying web pages to which they point. As a result, they cannot be deemed any more or less offensive or newsworthy than that original content. On Post's account, the court should therefore have applied the same outcome to Google search results as to the underlying websites to which they point. Post's point about the contextual meaning of communicative acts is an important one. He interprets the significance of context in terms of the appropriateness of the audience to which the information is exposed, as evidenced by the following example: ‘It may be humiliating to reveal the CT scan of a private person to the general public, but unobjectionable to provide it to a doctor for evaluation’.16 I will argue in Section III that search-result links do carry meaning over and above the content to which they point, and that this gives us reason in some cases to treat them differently from that underlying content. However, the way in which they convey such contextual meaning is not by virtue of reaching a different or wider audience than the original source material. The contextual meaning of such communicative acts can thus fall foul of standards of offensiveness and give rise to concerns about dignitary harms in a way that is not adequately captured by the concept of privacy. The reason for this is that the concept of privacy is constrained to concerns about the appropriateness of the audience to which some information is exposed. Instead of expanding the concept of privacy, we have good reason to keep it so bounded, and instead expand the category of claims that function in the service of protecting individuals' interests in self-presentation. In order to bring out the way in which the concept of privacy turns on the question of audience exposure, it will be instructive to consider key conceptions of privacy. Conceptions of privacy standardly invoke a distinction between public and private spheres.17 On this general view, it is assumed that a person's privacy interest is frustrated whenever some information that ought to be restricted to a specific audience crosses over into a more public domain. For example, in Judith Jarvis Thomson's influential view of privacy rights, once a person has made some information public, whether intentionally or unintentionally, they are considered to have thereby waived their privacy right over that information.18 It is this feature of approaches to privacy that makes it difficult to make sense of the outcome in Google Spain, if we interpret the data-protection measures in question as aimed at protecting a privacy right. This general structuring idea underpins a range of approaches to privacy: privacy as control, privacy as accessibility, and privacy as contextual integrity. The conception of privacy as control paints a picture of the individual as sovereign over her personal domain, with her consent required to let others in, at her own discretion.19 In doing so, it most directly invokes a sharp distinction between the private sphere of the individual as a domain that requires protection against unwanted encroachment from the public. The conception of privacy as accessibility departs from the notion that privacy is best served when the individual has ultimate control over the flow of information between the private and public spheres. Arguments in this camp point to the fact that the value of privacy can be undermined when individuals willingly open the floodgates, especially if the resulting dissolution of their private sphere is irreversible.20 This represents a more paternalistic stance towards protecting privacy, but one nevertheless based on the premise that some things should be kept within a private sphere protected from public scrutiny. The third approach, privacy as contextual integrity, more directly challenges the simple picture of private and public as two distinct spheres. Instead, it emphasizes privacy as a right to an appropriate flow of information, with different information being appropriately known by different parties in different social contexts.21 This approach rejects the idea that the dividing line of privacy is between activities that are carried out in private spheres such as the home or within the family, and those carried out in public. It does so to make space for the idea that certain ways of collecting personal information can constitute violations of privacy even where the individuals in question are acting in public. However, in positing the idea of an appropriate flow of information, it nevertheless draws on the same general structuring idea of different spheres of accessibility of information. Each distinct social context in which some information is appropriately known can thus be thought of as its own restricted sphere as against the rest of the public who are not included in that particular context. While each of these approaches to privacy offers its own take on how best to think about the divide between public and private in any given context, they all share the idea that some privacy interest is frustrated when information becomes publicized in a context in which it ought to be restricted to a smaller or different group of people. One response to this might be to seek to expand the concept of privacy beyond this framing, in order to capture concerns about the ways in which different sources present information to us, over and above the question of the audiences to which they make it available. Such concerns about the contextual presentation of information ultimately explain what is at stake in cases like Google Spain. However, expanding the concept of privacy to attempt to capture these concerns risks undermining its coherence, leaving it open to the critique that it is an ill-defined term used to cover a range of disparate interests.22 This gives us good reason to keep the concept of privacy bounded and specific, instead of expanding it to capture all the situations that give rise to concerns about reputational control. In keeping the concept of privacy bounded to the question of audience exposure, I follow Carissa Véliz's description of privacy as the quality of having one's personal information and ‘sensorial space’ unaccessed. You have privacy with respect to a certain person to the extent that that person does not know anything personal about you, and to the extent they cannot see, hear, or touch you in contexts in which people do not commonly want to be the object of others' attention.23 This general structuring framework of the concept of privacy underpins the contention that the ruling in Google Spain was inconsistent, if based on a privacy claim. If privacy is about keeping information restricted to appropriately restricted spheres, then it is hard to see how one could justify treating Google search results differently from the original source material they point to, when that original source material is in principle publicly accessible to the same audience. It would be an even harder task to justify the differential treatment of Google links pointing to different sources, when those sources contain the same informational content, as happened with respect to the links to newspaper articles reporting the details of the Google Spain case compared to the links to the original foreclosure notice in La Vanguardia. If we cannot justify these elements of the right to be forgotten by appeal to the importance of privacy, we need to find another way to justify the measures implemented in Google Spain or accept that the case was dealt with inconsistently. A separate question that arises in response to understanding privacy as functioning to keep information restricted to appropriate spheres is whether we can have claims to privacy over information that has already been made public. Benedict Rumbold and James Wilson have argued that people can retain a right to privacy over information which has been made public.25 They argue that one infringes another person's privacy right when one deduces some piece of information about that person that they did not intend to reveal, from information which they did intentionally make public. While it may be possible to apply privacy rights in this way to information that has been made public in the specific way identified by Rumbold and Wilson, the cases relevant to the right to be forgotten do not fit that mould. Instead, they are cases where some piece of information has intentionally been made public, either willingly by the data subject in the first instance, or by virtue of it being a legitimate matter of public record. The particular challenge this article takes up is whether there is any justification for treating different sources for reporting that information differently with respect to erasure. Rumbold and Wilson's argument thus does not apply to the features of the right to be forgotten that are of central concern to this article. An alternative path to justifying the right to be forgotten can be drawn from the principles underpinning the more general concept of the right of personality. To the extent that this alternative justification stands up to scrutiny, it extends the theoretical tools we can draw on to defend various aspects of data-protection provisions beyond the constraints of what privacy-based arguments alone can offer. The argument is not meant to assume that the measures applied in Google Spain were intuitively correct, and that there must be an explanation for their consistency. Rather, the aim is to show that there is a plausible path to justifying data-protection measures which need not draw on a right to privacy, and provides a more coherent way to make sense of the measures applied in Google Spain. The origins of a right to be forgotten pertaining to personal data can be traced back at least as far as a legislative project developed in France in 2010.26 The French legal concept of a droit à l'oubli, sometimes translated rather dramatically into English as a right to oblivion, has historical precedent in setting standards for the way in which spent criminal convictions are expunged from the public record after a certain amount of time has elapsed. Within the legal tradition of continental Europe, the right to be forgotten stems from the right of personality, which has historically been conceived of as protecting the dignity, honour, and right to private life of individuals.27 As Rolf Weber points out, the right of personality is a right that incorporates a number of different concepts and terminologies, and is mainly intended to protect the moral and legal integrity of the person. It delineates a sphere of privacy as part of that goal. It is in this context that the right to be forgotten and the way it has been enshrined in the provisions of the GDPR has been interpreted and implemented as a privacy right. However, as argued above, the framing of the right to be forgotten as a privacy right constrains the scope of data-protection provisions that can be included in such a right. In order to provide a better justificatory basis for those data-protection provisions, we need to shift focus away from a narrow framing around privacy, and instead lean on some of the other concepts invoked by the right of personality to motivate and justify the claims in question. In doing so, the aim is not to provide a historical account of the legal origins of these existing data-protection provisions. Rather, it is to provide a philosophical analysis of the principles already embedded in the right of personality which can provide foundations for the right to be forgotten, in a way that provides a path to justifying a broader range of data-protection provisions under that umbrella than a focus on privacy alone. More specifically, the task will be to outline three different categories of claims which serve to protect against different threats to reputational control. These are claims to privacy, claims against defamation, and claims against distortion. The concepts of dignity and honour embedded in the right of personality are broad ones, as is the goal of protecting the ‘moral and legal integrity of the person’. When it comes to regulating the sharing of information, what is at stake is the integrity of a person's public persona. The key battleground over the integrity of one's public persona is located in the gap between how other people see, understand, and talk about us, compared to how we would like to present ourselves to others. The question this raises is how much control over the shaping of one's reputation to leave in the hands of individuals, and how much to allow to be shaped by others. This idea of reputational control embedded in the more general right of personality has been invoked in support of privacy protections. Consider the following arguments that have been offered in defence of the importance of privacy. As we saw, Warren and Brandeis argued that the right to privacy was based on a principle of inviolate personality, and was required to secure peace of mind for the individual.28 Contemporary accounts have argued that privacy rights are required to enable valuable social relationships, allowing us to regulate the boundaries between intimate relationships and more distanced ones. Relatedly, appeals are made to our interest in having a reasonable degree of control over how one presents to others, or to make decisions about one's own life, protected from the interference and scrutiny of others, with the suggestion that these are key aspects of living an autonomous life.29 A common thread in the literature on privacy is thus the idea that privacy protects an interest in self-presentation. Marmor, for example, argues that the distinctive thing about privacy rights is that they safeguard our interest in ‘having a reasonable measure of control over ways in which we present ourselves to others and the ability to present different aspects of ourselves, and what is ours, to different people’.30 The second part of Marmor's formulation of the interest reveals again the general framing of privacy in terms of the appropriateness of the audience to which some Information is exposed. Marmor explains the distinctiveness of privacy by reference to the importance of being able to conduct intimate relationships, and the formulation of the second part of the interest is key to that case. However, if we focus on the more general interest in having control over how we present to others (as expressed in the first half of Marmor's formulation), we can notice that privacy plays a relatively narrow role in safeguarding that interest. The distinct way in which it does so is by giving us a reasonable measure of control over which audiences we share different aspects of our lives with. That more general interest in self-presentation, however, can be undermined by ways of sharing information which do not infringe any privacy claims. Some ways of presenting information can constitute a claim-infringing distortion of one's reputation

The problem with police is not that they are fascist pigs but that our country is ruled by majoritarian pigs. Ta-Nehisi Coates To change the police without changing the police role in society is as futile as the labors of Sisyphus. John van Maanen This question is, I will argue, of considerable importance to political philosophy, because it exposes a tension in the liberal project as a whole. It has, unfortunately, been subject to relative neglect; there has until recently been very little philosophical or normative literature on policing, much less the policing of minorities.3 There are two reasons for this, both of them based ultimately on the perception that the problem is theoretically straightforward, and so does not raise any philosophically difficult questions.4 With respect to policing in general, many political philosophers appear to believe that the police are uninteresting, because they serve the purely instrumental role of enforcing statutory law. And with respect to minorities, widespread acceptance of the fascist pig theory leads to the perception that the problem is also normatively straightforward. If tensions between police and minorities are caused by the police being racist, then the solution is obviously for the police to become less racist. These two views, when combined, generate a highly simplistic account of the problem of policing minorities.5 It suggests that, since the police are there to enforce the law, all they need to do is enforce the law in a non-discriminatory way, treating all citizens equally, and the problem should go away. The argument, however, rests on a false premise, since the police do a great deal more than simply enforce the law, and it is primarily these other activities that give rise to problems. My preliminary contention in this article will be that both of the underlying theories about policing are false or misleading. Obviously, the police do enforce the law, but as several decades of work in the sociology of policing have shown, this is not all that they do, and, indeed, this does not even add up to a substantial fraction of what they do. And equally obviously, there are some police who are racist, both consciously and subconsciously, but this goes only part of the way to explaining the seriousness of the difficulties that practically every major police force experiences in its relations with particular minority groups.6 Indeed, a great deal of the motivation for the shift in emphasis toward “systemic racism” in expert discourse on the subject lies in the recognition that the emphasis on eliminating attitudinal racism or increasing officer diversity has reached the point of sharply diminishing returns in many jurisdictions.7 In order to understand the broader source of the problem, I will argue, it is necessary to see that effective policing is inherently communitarian. The result is that minority groups who are unable to implement a self-policing regime find themselves subject to coercive enforcement of social norms and standards of respectability that reflect parochial aspects of the majority culture, which they often resist. I refer to this, referencing Ta-Nehisi Coates, as the “majoritarian pigs” theory of oppressive policing.8 Liberalism, to the extent that it imposes a neutrality constraint only on statutory law, is powerless to remedy this problem directly, which is why the policing of minorities poses a fundamental challenge in liberal societies. To put the claim in highly condensed form: even when the state is liberal, society is not, and the police, in order to be effective, are obliged to meet society half-way.9 The article proceeds as follows. I begin by distinguishing two very different normative models of policing, which follow from different views about the basis of police authority. The first view, which I refer to as the legalist model of policing, rests on the claim that police power should be limited to law enforcement, narrowly construed (which is to say, preventing crimes in progress and apprehending those who have broken the law). In contrast to this, I will describe what I refer to as the communitarian model of policing, which regards the organized enforcement of statutory law as deeply intertwined with enforcement of the informal social order. The authority of the police, on this view, extends beyond the enforcement of law to include a wide range of interventions aimed at complementing the efforts of communities to maintain order. The communitarian model, as a reconstruction of policing best practice, is what underlies the widespread call for more “community policing” as a response to various ailments.10 It is not difficult to see, however, that this practice creates a number of normative difficulties in a pluralistic society. The central tension arises from the desire on the part of the liberal state to exercise a monopoly on the legitimate use of force while simultaneously enacting only laws that meet a fairly restrictive standard of justification. The consequence is that the police are subject to constant demands from the public for interventions that extend far beyond the ambit of law. In principle, this tension could be dissolved by adopting a legalist approach and restricting the activities of the police to law enforcement narrowly construed. I will therefore round out the argument by outlining several reasons why I think this would be neither desirable nor realistic. I unfortunately do not have any simple solutions to propose to the underlying dilemma. I do, however, think that a deeper understanding of the difficulty of the problem is important. Policing has been the subject of a great deal of debate in recent years, which has unfortunately exposed significant misunderstanding of policing as an institution, not just in the public at large, but among many academic commentators as well.11 This has many effects, one of which is to reinforce the pre-existing tendency among the police to see themselves as an embattled and misunderstood minority, which in turn provides them with grounds to dismiss normative demands made upon them. The first step toward amelioration of the situation must involve a proper understanding of the difficulty of the problem and the extent to which police are being subjected to contradictory demands. When sociologists first began to take an active interest in studying the police, in the late 1960s, they quickly acquired the Socratic wisdom—they came to realize that they didn't really understand what the police do most of the time, or why it is the police that are doing it. In particular, they discovered that statutory law enforcement constitutes a vanishingly small fraction of what the average patrol officer spends most of the day doing.12 Yet despite this puzzlement over what the police do, there was never any doubt about the distinctive characteristic of their institutional role, which is that they are agents of the state who are authorized to use force.13 It is also generally accepted within modern liberal societies that this authorization to use force is grounded in, or legitimated by, law—primarily criminal statutes, but also traffic and vehicle law, certain municipal and regional ordinances, and in some countries immigration law. The most common mistake, however, is to think that because the use of force by police is authorized by law, it is used only to enforce the law (or that its domain of application is limited to that of legally regulated behavior). Empirically, this is simply not the case, as police use the power that is granted to them by statutory law to pursue a range of extra-statutory objectives.14 This includes maintaining order, emergency response, protection, and rescue, and, perhaps most expansively, “brokering and enforcing social cooperation,” especially regarding the use of public space.15 This aspect of policing practice is obviously in tension with the dominant normative understanding of police power. The latter is closely tied to theories of political authority that regard the enforcement of law as the only legitimate basis for the use of force by agents of the state.16 This view follows from a number of different philosophical positions, but one can find a particularly clear version in Kantian theories of law, which draw a sharp distinction between morality and law, conceiving of the former as a system of rules that binds individuals internally, the latter a system of rules that binds them externally. To say that individuals are bound externally is to say that they are given incentives to act in conformity with the law, first and foremost through the threat of punishment in the event of noncompliance.17 Thus the use of force against individuals by agents of the state is justifiable only when it is used to secure compliance with a justifiable system of publicly promulgated laws.18 Although the police are often described informally as being responsible for “enforcing” these laws, technically punishment can only be imposed by the judiciary, and is carried out by the correctional system. The primary task of the police is to identify and apprehend individuals who are suspected of having broken the law, delivering them to the judiciary for trial. Thus the police operate at the front end of the criminal justice system, the point at which the enforcement of law is initiated. Secondary tasks then involve deterrence, which involves persuading individuals to obey the law through the threat of apprehension, and, in the rarer case, direct interdiction, which involves stopping individuals who are in the midst of committing an infraction. This way of thinking about state coercion gives rise to what can be characterized as a legalist theory of policing, on the grounds that it restricts the use of force by police to direct enforcement of the law.19 This implies that the police should have no power over the ordinary citizen, so long as the individual is not currently acting in an unlawful manner and is not suspected of having committed some other crime. It suggests that, absent some law to the contrary, individuals are under no obligation to speak to the police, to answer any questions posed, to obey their orders, or even to identify themselves, unless there is some direct connection between these actions and the investigation of criminal conduct. Individuals on this view enjoy an extremely broad sphere of liberty—such as the freedom to move at will through public space, to use offensive language or be verbally abusive toward others, perhaps even to brandish lethal weapons—so long as they have not committed any crime and are above any reasonable suspicion of having committed a crime. Public commitment to a legalist theory of policing has been encouraged by certain popular interpretations of the US Constitution, which become apparent when citizens of that country refuse to obey a direct order from a police officer, refuse to identify themselves, refuse to exit their vehicle, or refuse to circulate or disperse in public space, on the grounds that their ongoing activities are not in direct violation of any legal restriction. The situation is significantly complicated, however, by the fact that legislators in jurisdictions in which this view of police authority is widespread have often sought to undermine it by enacting extremely broad laws, which have the effect of legally prohibiting all conduct that is contrary to the will of the police. In other cases, courts have achieved the same effect by recognizing a class of “common law misdemeanors,” thus tacitly extending the criminal law.20 As a result, in some jurisdictions it is an offense to insult or show disrespect to a police officer, to fail to identify oneself, or to obstruct an investigation. Laws prohibiting loitering, vagrancy, nuisance, disorderly conduct, public mischief, and public intoxication can also be used in ways that essentially override any citizen's right to occupy public space. And finally, the offense of failing to obey a police officer—a common feature of motor vehicle law—can be extended to ordinary contexts of interaction. The result is the phenomenon of “overcriminalization,” in which the sphere of individual freedom is essentially extinguished by the legal prohibition of practically all conduct.21 But, of course, in most jurisdictions in which citizens are subject to this sort of overcriminalization, the sphere of individual freedom is seldom extinguished in practice, because the purpose of these laws is not really to prohibit the activity in question, but rather to grant police the discretion to control members of the public in response to situational exigencies. The important point, for our purposes, is that the existence of these broad prohibitions can render the legalist theory of police power trivially true. By making it the case that everyone is always breaking some law, any exercise of police power can be characterized as mere enforcement of the law. And yet the legal form clearly serves to conceal the true character of the underlying practice, which is not to arrest everyone who commits an infraction. While the police do engage in some by-the-book enforcement of statutory law, especially with respect to indictable offenses, they also use the power that is conferred upon them by statutory law to pursue other, extra-statutory objectives. They do so with the full knowledge and support of legislators, which is one of the reasons that the latter often respond to court-imposed restrictions on police power by crafting laws that expand the scope of police discretion. As we shall see, there are rival accounts of this “other objective” that police pursue, but before getting to that, it is important to consider why the legalist model of policing is not more widely embraced. Although philosophers sometimes speak of law and morality as though they were quite distinct, when one turns to the social reality of crime and enforcement it is difficult not to be struck by the fluidity of these categories. Emile Durkheim, in his pioneering sociological analysis of crime, argued that the regulation of social action should be understood in terms of a single category of rule, that of a social norm.22 The most important feature of norms is that they are all enforced, although some more obviously than others. This is because the punitive sanctions that are evoked by nonconformity have a socializing effect, by virtue of which the norm becomes internalized by individuals over time. Through this process of socialization, individuals acquire both intrinsic and extrinsic motives for compliance. As a result, sanctions become increasingly symbolic, aimed more at triggering internal self-control (as with mockery, humiliation, and moralizing condemnation). Because of this, among well-socialized adults, explicit sanctions in the form of rewards and punishments are reserved for only the most extreme cases, in which more subtle instruments of social control have failed. Moral norms, on this view, do not form a natural kind, but represent merely one end of a continuum.23 This gives them the appearance of being purely inner constraints, despite the fact that they are enforced by social sanctions like all other norms. Similarly with law, conformity is typically elicited through a combination of intrinsic and extrinsic motives. What makes law distinctive is not the fact that it is coercively enforced, since all norms are enforced, but that it is enforced through the organized sanctioning capacity of the state. In the informal social sphere, punishment of norm-violators is decentralized, ad hoc, and often based on withdrawal of cooperation.24 The state, by contrast, regiments this process through a set of formal mechanisms, including written specification of the rules, centralized enforcement by specialized agents, and costly, individually targeted punishment. At the same time, a great deal of the motivation for legal compliance remains intrinsic, arising both from malum in se convictions regarding the specific actions prohibited, as well as generalized respect for the authority of law. There is a complex division of labor involved in maintaining social order. Without some form of constraint, human society would resemble a Hobbesian state of nature, marked by widespread failures of coordination, cooperation, and collective action. The orderliness of everyday life is achieved through conformity to shared social norms, which not only prohibit various harmful or destructive forms of conduct, but also diminish the anxiety and cognitive strain involved in routine interactions. When this functions effectively, we are able to ignore most of the work that is involved in achieving this orderliness. The operations of socialization, symbolic sanctioning, and decentralized enforcement remain largely invisible, like the bulk of an iceberg sitting below the waterline. The law, by contrast, is like the visible portion of the iceberg. And yet, precisely because it is so easily observed, one can overestimate its importance. This dramatically inflates public perception of the efficacy of the sanctions that are imposed by the legal system. Particularly with respect to the criminal law, the vast majority of the population is deterred through a combination of informal social pressure (for example, concerns related to reputation, community standing, stigma, social role, and identity), and moral commitment.25 Fear of punishment is always in the background, but it provides a decisive consideration only with a small segment of the population. Criminologists often illustrate this through an “enforcement pyramid,” which has at its base purely intrinsic motives, then ascends through various mechanisms of informal social pressure, capped off by coercive punishment at the peak.26 The pyramid shape is intended to represent the fraction of offenders who can be brought into compliance without escalation to the higher mechanism. From this perspective, the police can be seen as specialized agents of social control, who are called upon to act when the upper levels of the enforcement pyramid are reached. This naturally makes them the primary agents of law enforcement, but it also involves them in a wide range of activities variously described as “order-maintenance” or “peacekeeping.” This view, which is the one most widely held by those who study the practice of policing, emphasizes the continuity and relations of mutual support between different systems of social control, and, in part for that reason, tolerates a great deal of ambiguity about the proper scope of police power. I refer to this model as communitarian, not only because it underlies most versions of “community policing,” but also because it is in tension with some of the commitments of political liberalism. While the state has long been characterized by its ability to employ coercive force in an organized fashion, the modern period has been marked by a much more ambitious attempt on the part of the state to assert a monopoly on the use of force in society. This ambition was clearly stated by John Locke, who thought that the primary feature of the social contract, through which the sovereign power of the state was constituted, was that individuals would surrender the natural right to the use of force in a punitive manner against one another.27 And yet, even though this ideal was clearly articulated in the 17th century, it did not become an institutional reality in Western societies until near the end of the 19th.28 Prior to that, the violence meted out by the state typically served as a supplement to the relatively high levels of retributive violence occurring in the informal social sphere (with the practice of dueling serving as the most high-profile example).29 The state did not attempt to prohibit this sort of violence until well into the 19th century, a process that went hand in hand with the development of policing.30 That this program of domestic pacification should coincide with the rise of modern policing is hardly an accident, since the state can only aspire to eliminate interpersonal violence once it has the institutional capacity not just to control it, but to offer at least a semi-credible replacement. Retributiveness is a powerful human impulse, which has so far proven resistant to all attempts at eradication. This means that when the forces of “law and order” appear unequal to the task of ensuring that the guilty are punished, members of the public will be inclined to take matters into their own hands. A significant fraction of crime is not first-order social deviance, but rather private vengeance in response to such deviance.31 As a result, the police are always subject to pressure from two directions. From “above” there are legislators and other state officials, who demand that the police enforce enacted laws, while from “below” there are members of civil society, who expect certain performances as a condition of their own willingness to refrain from engaging in private uses of force.32 This demand from below is what accounts for the widely noted phenomenon that the police, apart from the enforcement of statutory law, also act to maintain social order.33 Historically, this police power had an explicit constitutional justification in the relationship between the executive and other branches of government.34 European monarchs in the early modern period were taken to have two domains in which they could act without legislative authorization or oversight. Externally, there was the realm of foreign affairs, including military strategy and the prosecution of war, and internally, there was the “police”—in the original, archaic sense of the term—which involved promoting the well-being of the people and the good order of society.35 Democratic and republican revolutions led to the erosion of autonomous executive power and an expansion of legislative claims, in part because the legislature was seen as the conduit for popular sovereignty, which came to be regarded as the only legitimate basis of state power. This led to a dramatic expansion of the importance of statutory law, although the prerogative of the executive to conduct foreign affairs outside the scope of legislative control has largely been retained. The internal police power, however, fell into constitutional disrepute, leading to its preservation largely as an implicit police practice, given a veneer of legislative legitimacy through overcriminalization. The idea that the police enforce not just laws, but also the social order, sometimes evokes negative associations, and so it is important to be clear about its meaning. Social order, in the generic sociological sense, simply refers to the stability of the institutional systems governing everyday life.36 When private citizens reach the limit of what can be achieved through the informal enforcement mechanisms that are available to them, and, in particular, when they reach the limit of what can be achieved without recourse to force, they call the police. For the average patrol officer, this sort of call—to mediate a conflict between a landlord and tenant that is beginning to get out of hand, to intervene in a domestic dispute that has woken up the neighbors, to disperse a group of young people drinking in the park, to evict a bar patron who is refusing to leave, to deal with a mentally ill person who is scaring passengers at a bus stop—constitutes the overwhelming bulk of police work. When the police are obliged to use force in these circumstances, they appeal to statutory law in order to justify that use of force, despite the fact that the objectives they are pursuing, or the reasons for which they are using force, are often external to anything that is explicitly specified in any statute. As a result, modern policing winds up encompassing an eclectic list of responsibilities that defies easy summarization. (This is one of the reasons that police departments, despite possessing all the appurtenances of modern bureaucracy, are only superficially rule-governed organizations—most police decision-making is in fact driven by situational exigencies.)37 Part of the complexity is due to the fact that prohibiting civilians from engaging in uses of force has downstream effects (many of which were clearly unanticipated by social contract theorists who called for the prohibition). The recourse to force, or the availability of such recourse, has a structuring effect on other social practices, much of which escapes notice simply because, when effective, it results in force almost never having to be employed. For example, as Egon Bittner observed almost a half-century ago, there is a “profound misconception” about policing involved in the claim that, because there is substantial overlap between the tasks undertaken by police and social workers, the former could be replaced by the latter.38 The problem is not just that routine mental health and domestic disturbance calls have the potential to erupt into violence. It is also that the availability of recourse to force provides a form of decisiveness in action that is otherwise difficult to obtain. The central limitation of dialog and argumentation is that it can go on forever. At some point, however, discussion must cease and action must be taken. As Bittner emphasizes, citizens often call police in order to deal with situations that call for such decisive action.39 Furthermore, police intervention is often valued precisely for that reason (which explains, inter alia, why the police are involved in all emergency situations, regardless of the underlying nature of the problem). Because of these peculiarities of the role, the police stand in complex relationships of dependence with the communities that they serve. On the one hand, the effectiveness of legal regulation, and the capacity of the police to enforce statutory law, clearly depends on the quality of the underlying social order. This is a major theme in the work of Jane Jacobs, who emphasized that “no amount of police can enforce civilization when the normal, casual enforcement of it has broken down.”40 This is not just because police rely on members of the community to provide information, but also because the most important forms of deterrence are achieved through the informal social order. The real panopticon, Jacobs observed, are the “eyes on the street” of neighborhood residents, shopkeepers, and pedestrians, who create security simply by observing the comings and goings of others.41 By comparison, the police are extraordinarily limited in their ability to detect and deter crime. Yet, just as the police rely upon the informal social order to achieve compliance with the law (the way the visible part of the iceberg is supported by the larger bulk that rests below the waterline), members of the community depend upon the police to contribute to the maintenance and reproduction of the informal order, by intervening in situations that are beginning to get out of hand, or that escape the regulatory capacities of civil society. Police in these cases use force to supplement the informal, non-violent sanctions that are available to private individuals. This is where the most discretionary aspects of policing arise, because the patrol officer is forced to determine whether a situation is one that can be handled informally by the parties involved (perhaps once everyone is calmed down), or whether it requires formal intervention (and if so, what degree of sanction). Consider the case of individuals drinking in the park. In most jurisdictions, consumption of alcohol outside of a private home or a licensed public establishment is prohibited. And yet these laws are seldom enforced universally, but are merely used by police to break up groups whose behavior poses some other threat to public order (perhaps, but not necessarily, due to the influence of the alcohol consumed). The fact that some people are breaking the law by drinking in a public park says practically nothing about what the appropriate police response should be. Is it a young couple having a picnic, enjoying a glass of wine? Is it retired men from the neighborhood gathering in the afternoon? Is it a group of teenagers after dark? These are very different cases when it comes to assessing the disruptive potential of their behavior. Are they drinking socially, or are they getting drunk? And how are they drinking? Are they doing it discreetly, for their private enjoyment, or brazenly, in a way that shows disdain for the law? Is their presence intimidating, or interfering with other users of the park? And if intervention is warranted, should the alcohol merely be confiscated and the group ordered to disperse? Or should some or all be arrested? And once brought to the station, should they be released or detained overnight? An affront, as it is used here, is a challenge to the policeman's authority, control, and definition of the immediate situation. As seen by the police, an affront is simply a response on the part of the other which indicates to them that their position and authority in the interaction are not being taken seriously.43 Equally important is the role that judgments of respectability play in determining police response.45 Generally speaking, police are more likely to intervene, and to intervene punitively, when the person causing trouble belongs to a less respectable social category, just as they are more likely to take complaints seriously when they come from respectable individuals. The respectability of the “ordinary, law-abiding citizen” involves a generalized willingness to conform to social expectations and to defer to legitimate authority, and so is impaired by indications of either a history of, or propensity toward, social deviance. Thus the police focus their attention on individuals belonging to classes who pose the greatest threat to the social order. The most obvious targ

Throughout the Covid-19 pandemic, protests against the public health measures instituted by governments have become a familiar sight on the streets of major cities across the world. The policies these protests challenge, and the kinds of claims made by protestors, have differed across jurisdiction and have evolved through different stages of the pandemic, with protests across Europe and North America focussing at various times on the supposed injustice of initial lockdown measures, the rollout of vaccines, and the implementation of vaccine mandates. A significant number of these protests have been, in various respects, legally prohibited; many of the earlier protests violated emergency legislation restricting the number of people or households permitted to gather in public, while later actions include occupations of the offices of media companies and refusal to pay taxes.1 At least some participants in these protests have been animated by false claims that the pandemic is a hoax, or that it is in some sense deliberate, or that it is being exploited by politicians and businesses to implement policies of radical social control. Many participants and commentators have described some of these actions as forms of civil disobedience; a placard displayed by one protestor in Bristol in November 2020 stated that ‘civil disobedience becomes a sacred duty when the state has become lawless or corrupt’,2 while a Canadian anti-lockdown campaigner compared his refusal to provide details of his vaccination status in a coffee shop to the actions of Rosa Parks.3 Similar claims have been made regarding the protests in Ottawa in late 2021/early 2022.4 These protests raise important questions regarding the conditions for permissible civil disobedience in a pandemic specifically,5 but also highlight some more general questions for theories of civil disobedience. Some of these relate to the use of civil disobedience in the pursuit of ‘anti-democratic and illiberal goals’, such as whether broadly liberal theories of civil disobedience effectively apply across these cases as well as the justice-promoting paradigm examples.6 In this article I focus on a slightly different issue: is the justification of civil disobedience conditional on dissenters satisfying some epistemic conditions? If so, what kinds of conditions? Perhaps surprisingly, the two most directly relevant philosophical literatures suggest quite different answers to these questions. Within broadly liberal theories of civil disobedience they have received little direct attention. Such accounts generally view the justification of civil disobedience as primarily relating to the conduct of the dissenters during and after disobedience. The justification of an act of civil disobedience just relates to the question of whether agents act in accordance with certain constraints, which may include non-violence, civility, and accepting punishment. However, there are some accounts within this literature, such as those developed by Ronald Dworkin and Kimberley Brownlee, which incorporate explicit epistemic constraints. These identify these constraints as tests by which we can determine whether an agent's claim qualifies as a conviction, and thus as a commitment which deserves a special kind of consideration in liberal democratic societies. As such, the constraints are viewed as relatively minimal tests of coherence rather than as tests of epistemic responsibility. The literature on political epistemology presents a very different picture. Political epistemologists disagree about the nature and scale of the epistemic obligations or responsibilities of citizens in a democratic polity. But they broadly agree that citizens do have some epistemic responsibilities when they act politically in various ways, including voting, engaging in public discussion, and even dissenting. If citizens in a democratic polity have responsibilities to act in epistemically responsible ways when voting in elections or discussing political issues in public, it might seem straightforward that these responsibilities will also constrain their civil disobedience. I argue that the political epistemologists are right that the justification of our political acts is sensitive to our epistemic conduct, and outline a particular way of incorporating epistemic constraints into theories of civil disobedience. This article has three aims. The first of these is to show that even those liberal theories of civil disobedience which incorporate explicit epistemic requirements generate an implausible picture of what is wrong with misguided disobedience. Incorporating epistemic conditions solely within the concept of conviction produces an implausible picture of conviction itself and relies on an untenable view of how claims of conviction relate to a political order. My second aim is to provide a more plausible alternative. I argue that agents have a defeasible epistemic obligation to their fellow citizens to take steps to minimise the risk that the claim under whose name they intend to disobey is mistaken. Like the obligation of dissenters to ensure that their actions inflict as little harm as is compatible with a reasonable chance of success, the obligation to take these steps bears on the justifiability of a civilly disobedient action. I show how this obligation is continuous with the epistemic obligations of ordinary political action, but made stronger by virtue of the different justificatory conditions of the burdens generated by disobedience. Finally, I aim to show that this account does not impose unreasonably stringent restrictions on dissenters. The stringency of this obligation is sensitive to competing considerations: appearing in a stronger form when agents have meaningful and broadly equal political influence under just institutions and in a reasonably healthy epistemic environment, and when the proposed disobedience imposes more severe costs; and in a weaker form, and in some cases no longer applying, in contrary circumstances. Exploring the factors which can mitigate this obligation in different circumstances has the additional benefit of illuminating the imperfect epistemic conditions in which most dissenters act and the relevance of such conditions to questions of the justification of civil disobedience. Before proceeding, a note on my use of the concept of ‘civil disobedience’. This concept, especially in its most influential articulations, has been subject to a range of important critiques from political theorists and philosophers in recent years, including that it functions as a ‘counter-resistance ideology’ in assuming the basic legitimacy of the constitutional order and repackaging the strategic commitments made by those engaged in paradigmatic disobedience campaigns (such as the Civil Rights struggle in the US) as necessary features of permissible disobedience.7 Although I find aspects of these critiques compelling, I am more optimistic that a conception of civil disobedience which retains some of the central dimensions of the liberal conception can, when revised to include a wider range of disobedient actions and a broader range of complaints, and conceived as just one, rather than the sole, form of permissible disobedient action, continue to be a useful category. In lieu of a full defence of the concept here, I note two reasons for sticking with a broadly conventional use of it in this article. The first is that my focus is on the kind of action which is the precise target domain of the concept–that is, ‘principled and deliberate breach[es] of law intended to protest unjust laws, policies, institutions, or practices, and undertaken by agents broadly committed to basic norms of civility’, with ‘civility’ in this context interpreted variously in terms of conditions such as nonviolence, publicity, and the willingness to accept punishment or other forms of public accountability.8 I think there is good reason to think that some of the arguments presented apply to some other ‘uncivil’ forms of principled disobedience, but I do not argue for that here. Second, the prominence of calls for civil disobedience and the use of the term itself by those engaged in the kinds of protests I outlined at the start of the article, as well as many others, indicates that scholarly disputes regarding the coherence and application of the concept have not undermined the normative standing it is granted in the public sphere. Many of these protestors take themselves to be engaged in permissible civil disobedience; whether this claim is true or false, and why, has important implications for democratic theory and practice.9 Few philosophical theories of civil disobedience incorporate explicit, substantive epistemic components. No meaningful epistemic requirements feature in the influential liberal accounts developed in the 1960s and early 1970s by Hugo Bedau and John Rawls. These theories claim that agents engaged in civil disobedience must act conscientiously (that is, with a demonstrated recognition of the authority and legitimacy of the legal and political institutions of the state) and in support of a claim which is their ‘sincere and considered opinion’, but clearly do not intend these requirements to have epistemic content. The former can be satisfied simply by acting in a way which indicates ‘fidelity to law’, such as accepting punishment for one's act of dissent, while the latter is most plausibly interpreted as a requirement of good faith which rules out merely opportunistic or strategic disobedience.10 Those working within (and, indeed, against) this tradition have by and large maintained this focus on the conduct of dissenting agents and the justice of their cause. Our acts of civil disobedience can go wrong in many ways, according to these accounts, but not by virtue of our epistemic conduct. Some liberal accounts, though, have incorporated resources which enable them to say otherwise. These build epistemic constraints into the familiar component of conviction. The most influential contemporary account of this kind is developed by Kimberley Brownlee, who constructs an unusually rich and expansive account of the epistemic conditions internal to the concept of moral conviction. On her account, the moral right to engage in civil disobedience is correlative to the duty that society holds to respect the facts that persons have capacities for reasoning and for developing deep moral convictions and that coercing agents to follow the law instead of expressing those convictions is deeply burdensome.11 Respecting our capacity for deep moral conviction generates ‘a limited moral right of conscientious action as our expression of our conscientious convictions’, and a further right protecting the cognitive and psychological pre-conditions for the development of these convictions.12 The ‘communicative principle of conscientiousness’ lies at the heart of Brownlee's account of deep moral conviction, and generates limits as to the kind of action which can properly be termed ‘civil disobedience’.13 There are four elements of this principle: (a) a consistency condition; (b) a universality condition; (c) a non-evasion condition; and (d) a dialogic condition. As Brownlee notes, both (a) and (d) have ‘cognitive elements’. The consistency condition requires a level of consistency between our judgements, motivations, and conduct. If agents are not sufficiently consistent in the application of their stated convictions–say, if an agent launches a campaign against blood sports, but regularly participates in fox hunts–their claim to hold that conviction will itself appear to be dubious. What is important to note for our purposes is the pre-conditions for such consistency noted by Brownlee: that convictions ‘must meet minimal standards of intelligibility, internal coherence, and evidential satisfactoriness’.14 Although she does not provide a comprehensive account of these standards or what it means to meet them, it is notable that in this discussion Brownlee remarks that ‘conscientiousness is not about the correctness of moral judgements, but about the consistency of moral judgements and the credibility of factual assumptions that have moral implications’.15 The dialogic condition also directly concerns the nature of the beliefs of the dissenter, and the processes by which those beliefs are tested. Conscientiousness not only requires us to hold our beliefs sincerely, but to engage in discussion of them with others and defend them against certain kinds of criticism.16 Civil disobedience cannot be communicative unless we are willing to engage in this kind of dialogue. Nor can others be reasonably expected to regard our views as sincerely and seriously held if we are not willing to defend them publicly and rationally.17 But the dialogic condition is not satisfied if we merely defend our claim automatically, without due regard for our interlocutors as reasoning moral agents or the willingness to take their claims and criticisms seriously; this would be a crude performance of dialogue rather the real thing. As Brownlee puts it, ‘having a sincere intention to engage others in a dialogue means that we do not wilfully immunize ourselves from their communicative efforts.18 Our readiness to heed and interpret them correctly is necessary for communication to occur’. In at least some cases, satisfying the communicative principle requires dissenters to both have a particular stance towards the object of their conviction, and to be willing to sincerely subject their beliefs to criticism and, if such criticisms are sufficiently powerful, revision. beliefs that fail to meet the logical and evidentiary standards for conviction cannot claim the degree of toleration or respect that genuine convictions might claim because they lack both the determinate content and reflection that confirm our psychological and emotional investment. Without determinate content and reflection, our declared convictions are flighty, capricious, incoherent things that warrant neither respect nor toleration from others. Brownlee's account is unusual in being explicit about the epistemic conditions it relies on, and in the resultant richness with which those conditions are drawn. Nevertheless, it fails to provide a plausible explanation of what kinds of epistemic considerations should influence our determinations of the permissibility of an act of civil disobedience, and why. It is implausible to claim that all the epistemic requirements bearing on the moral right to civil disobedience are internal conditions of serious moral conviction. Doing so mischaracterises agents who fail to meet these conditions of intelligibility, evidential satisfactoriness, and so on. On Brownlee's account, agents who so fail are simply doing something other than civil disobedience. But it seems very strange to view those who make claims which fail to meet these standards in their protests against COVID-19 public health measures – which they themselves regard as civil disobedience, and which appear to fit into most definitions of the category fairly easily – not as launching unjustified civil disobedience, or civil disobedience which they do not have a right to perform, but as simply doing something else. Brownlee's account then misdescribes the epistemic wrong involved in these cases. The flip side of this is that Brownlee's account also gives too much weight to conviction, granting a moral right to engage in civil disobedience on the basis of any genuine conviction without any reference to the broader political context or the competing values that might bear on the case. There must, though, be some considerations which bear on the question of whether an agent has this right which go beyond an analysis of the conviction of the dissenter. If not, this seems to leave the very establishment of political authority in a precarious position.20 Brownlee herself seems to regard this as a question of justification rather than right, and places faith in the constraints of conscientiousness to ensure that dissenters act responsibly.21 Locating the epistemic requirements of civil disobedience exclusively in the concept of conviction leads us to identify the epistemic considerations bearing on civil disobedience as solely introspective. But acts of civil disobedience do not only involve deep moral conviction. They also rely on moral and non-moral beliefs about the world–about the structure of social reality, the kinds of claims being made by other agents, the nature of the political and legal system–which, though always contested, cannot be regarded as a matter of deep and intractable disagreement in the way that religious or political conviction can. We should then reject the idea that limiting the moral right to civil disobedience according to the interests of other agents or society at large is in all cases insufficiently respectful of conscientious conviction. It is therefore implausible to identify the concept of conviction as the site of all of the epistemic constraints bearing on the justification of civil disobedience. Freeing the epistemic demands of a theory of civil disobedience from the inside of a conception of moral conviction enables a more plausible process of determining whether an act of civil disobedience is justified. Viewing these as one of a number of elements bearing on that question, alongside others such as the nature of the political regime, the character of the epistemic environment, the extent to which disobedient agents are able to make their claims through other means, and the weight of competing claims and interests enables us to recognise the epistemic constraints as an independent criterion which can generate different requirements in varying circumstances. What, though, might the epistemic conditions bearing on these determinations actually be? In the next section I advance an alternative to the conviction-based account, in which agents hold (defeasible) epistemic obligations to others which constrain their epistemic conduct prior to engaging in civil disobedience. The idea that agents have epistemic obligations is widely held. On Richard Feldman's definition, these are ‘duties that one must carry out in order to be successful from an intellectual (or epistemological) perspective’.22 My argument does not rely on accepting any general account of these duties, such as that these are duties to believe propositions supported by the evidence available to us, or to behave in ways which maximise true belief. I am concerned only with a relatively minimal subset of these obligations; those obligations agents have to take reasonable steps to avoid holding false beliefs.23 Note that these concern only the process of belief formation and verification rather than the content of those beliefs. It is important to distinguish between purely epistemic obligations and the kinds of obligations with which I am concerned. An example of the former would be a (prima facie) obligation on an agent to believe P if and only if P is supported by the available evidence at the time.24 This is not a moral obligation, but an obligation which is conditional on the pursuit of something like epistemic excellence. Often, though, we have moral or political obligations (as, say, members of a jury) which have epistemic elements, or which impose on us obligations to seek out certain kinds of information, to alter or reflect on our epistemic practices, or to discount certain kinds of information from our reasoning. Sometimes, we come to hold these obligations because we take on particular roles or offices or engage in particular actions or practices. My claim is that we have an obligation of this kind when we seek to engage in civil disobedience. Agents have an obligation to be epistemically responsible–that is, to seek to become informed about important topics, to take steps to avoid forming false beliefs, to be careful in their reasoning, to participate in processes which promote these goals–when voting,25 participating in democratic deliberation,26 or advocating a political view.27 As Candice Delmas has argued, ‘since citizens of a democracy make decisions on the basis of their beliefs, they ought to form their beliefs responsibly’.28 If we don't, we fail to give due weight to the interests of our fellow citizens. Both Delmas and Eric Beerbohm have articulated compelling accounts of this kind of epistemic care as a practice which citizens must engage in to identify and discharge their political obligations more effectively (including, for Delmas, their obligations to engage in various forms of disobedience against injustice).29 There are any number of ways in which we can fail to meet these standards in our regular interactions with our fellow citizens.30 But there are reasons to think that this obligation is especially weighty when citizens are considering engaging in acts of civil disobedience. Those who engage in civil disobedience impose burdens on other citizens by doing so, and imposing these burdens without institutional authorization may, without further justification, be interpreted as a rejection of mutual cooperation between agents as the basis of shared citizenship. I will defend this characterisation before outlining the content of the epistemic obligations I posit in more detail. There are burdens involved in making space for civil disobedience.31 Some of these burdens are financial, based on the initial costs of making space for civil disobedience and the economic consequences of protests.32 Others might involve the obstruction of an agent's capacity to advance important interests or exercise their rights. In either form, the lives of other citizens are often seriously disrupted by such action. Whereas fair-play accounts of political obligation lead us to think of these costs in terms of undeserved advantage or free-riding, no such step is needed to recognise the significance of this distribution of costs. Instead, I suggest simply that those who involuntarily assume these burdens are owed a justification. Burdens distributed through political institutions can be justified by reference to the value and legitimacy of those institutions (though such justifications will generally fall short when provided in support of unjust policies; in such cases, even the burdens levied by legitimate institutions may count as unjust). An answer of this kind cannot provide a justification for acts of civil disobedience, which are extra-institutional and thus attempt to override or challenge the way in which burdens would ordinarily be assigned. In such circumstances, a justification must be attached to the token of civil disobedience itself, and must be of such a kind that no agent could reasonably object to being forced to carry the burdens the disobedience loads on to them. In the absence of such a justification, a dissenter may reasonably be viewed as indifferent to the claims that others have on her as a member of the community. In many cases, such a justification will be provided through the kinds of claims made in the act of civil disobedience. However any given agent might respond, the burdens taken on by those who are stuck in traffic as a result of civil disobedience aimed at highlighting the urgent need to reduce emissions can be justified by reference to the severity of the injustice being protested and the insensitivity of the democratic process to such claims; similarly, those whose property was damaged in the protests against racial injustice across the US in the summer of 2020 following the murder of George Floyd did not have a reasonable complaint against the protestors.33 In speaking of a ‘justification’ offered to those who take on the burdens generated by civil disobedience, we are not then speaking of agreement between actual persons, but rather invoke the contractualist sense of the term whereby acts or principles may be justified on the basis of meeting certain standards of reasonableness.34 It is striking that discussions of the burdens imposed and taken on by those engaged in civil disobedience rarely engage with the epistemic dimensions of civil disobedience. We may reasonably displace burdens (especially, but not exclusively, when they have been unjustly distributed in the first place) on to others if we engage in civil disobedience which is in support of a just cause, in which we do not impose excessive costs on others, and following which we take responsibility.35 My suggestion is that the justification of these acts rests not only on the justness of the cause and the insensitivity of political institutions to the claims being made, but also on protestors' epistemic conduct. We can easily imagine civil disobedience which serves good causes or is undertaken in conditions which permit such action but which other agents may have legitimate complaints against because of the conduct of dissenters; if, without sufficient reason, considerably more harm than necessary is inflicted to achieve a good chance of success, or if those burdens fall disproportionately on the least advantaged.36 In the same way, legitimate complaints may be made against agents who do not take sufficient epistemic care in advance of participating in civil disobedience.37 We are now in a position to identify the nature of this obligation more precisely. My claim is that the legitimate interest of agents in not being subject to significant unjustified burdens gives rise to an obligation held by those seeking to engage in civil disobedience to take reasonable steps to minimise the risks that they are mistaken about the claims under whose name they disobey. It is prima facie wrong for an agent to foreseeably inflict significant burdens on others through disobedience if that agent has not taken reasonable steps to ensure that the cause which they aim to promote in their actions is in fact truthful and just.38 It would indicate disrespect for those agents who involuntarily take on the burdens generated by those actions. How might agents operating in a reasonably healthy epistemic context satisfy this obligation? They will seek to corroborate and check the factual claims on which their case relies through trustworthy media sources, eyewitness reports, or experts in the relevant field. They may engage in dialogue with others (ideally including a diverse range of perspectives) to come to a considered analysis of a policy or institutional arrangement. This might include public debate or questioning of public officials or others. On topics which require technical expertise that agents cannot be expected to acquire, they will try and identify experts who are respected within that field and trust in their guidance, though not to the exclusion of other dissenting views within the field (especially on highly contentious issues). They will recognise to some extent those inferences and claims which can and cannot be supported by the relevant forms of evidence, and seek a sound logical basis for their claims. They will respect and recognise others as knowers and treat their relevant personal testimony accordingly. As these examples show, this obligation is not especially onerous even for advantaged agents operating in contexts where information is widely accessible and there is a wealth of independent sources of expertise and experience one can draw on in verifying one's claims. Of course, dissenters generally operate in less healthy epistemic contexts. Even those relatively epistemically advantaged actors can be faced with circumstances in which some of the above ways of acting become unduly onerous, or even impossible. As Boyd Millar has put it, ‘for a belief-influencing action to constitute an epistemic obligation it must be the case that a given individual could reasonably have been expected to perform or omit that action under the relevant circumstances’.39 What epistemic obligations, if any, do agents bear in hostile epistemic circumstances? The obligation I have outlined is, as I show in the next section, sensitive to epistemic barriers that dissenters may face. It is a defeasible obligation which may be weakened or obviated when dissenters are subject to certain kinds of epistemic or pragmatic barriers. Agents who engage in any kind of political action, including civil disobedience, always face some epistemic constraints, such as limited time and resources and incomplete information. Not all such constraints alter the epistemic political obligations of agents. Some, however, do. In hostile epistemic environments, the ‘reasonable steps’ agents can be expected to take will be more modest than those that would apply in better circumstances. In some instances, agents may be released from any such obligation at all. There are (at least) two kinds of epistemic obstacle which can have this effect: oppression and urgency. Oppression can make epistemic checking too difficult or costly for agents to be expected to engage in it to the extent expected in better conditions. The claims that other agents have on dissenters to take epistemic care will in these cases be dissolved by the weightier claims of all individuals in being able to resist severe injustice. My fellow citizen can have no reasonable complaint if I launch an act of civil disobedience without engaging in epistemic checking or corroboration when the public distribution of information is tightly controlled, or I have been prevented from accessing what information is available, or public discussion of political topics is prohibited. Any attempt to act in accordance with the requirements that apply in better circumstances would be supererogatory in these conditions. There are a number of ways in which oppression can have this effect. Firstly, it can make epistemic checking overly costly, or restrict the means by which we might more or less reliably verify claims. This is most obviously the case when the state imposes

Political liberals argue that democratic citizens have a duty of public reason to ensure that important laws are justified by reasons fellow citizens can accept given their own moral and philosophical beliefs.1 In any real-world democracy, many will fail to comply with this duty. Most people have never heard of public reason, and many who have heard of it reject it. This raises an important question about the demands of public reason: is there ever a duty to ensure that laws are justifiable to those who are not willing to reciprocate? Most political liberals answer “no”—the duty of public reason is owed only to those who are themselves willing to comply.2 This reflects a more general view that has wide currency in democratic thought: individuals ought to moderate their political activity to accommodate disagreement with others, but only when others are willing to reciprocate. Many think, for example, that individuals ought to seek middle-ground policy compromises, but only with those who are willing to compromise. The prevailing view among political liberals is that the duty of public reason is always conditional on reciprocal compliance in this way. As Andrew Lister asks, “How could it be reasonable to ask me not to count a reason I think true and relevant on the basis that you reject it, if you are not likewise willing to exercise restraint with respect to reasons that you think true but which I reject?”.3 My aim in this article is to answer this question and, in doing so, to rethink the demands of public reason in the face of noncompliance. I argue that there is a wide range of political contexts in which citizens have duties to comply with public reason for the sake of others regardless of whether they reciprocate. This helps lay foundations for a non-ideal theory of political liberalism that gives public reason a more inclusive and morally significant role in the practice of democratic politics. In Section I, I examine what I take to be the strongest argument for the prevailing view that public reason is always conditional on reciprocal compliance. The civic friendship argument grounds the duty of public reason in the valuable communal relation that citizens realize when they reciprocally comply for the sake of one another. Because unreciprocated compliance with public reason lacks mutuality, it undermines rather than promotes relational equality. When individuals bear the “moral cost” of excluding what they take to be true reasons without demanding reciprocation from others, they allow themselves to be treated as subordinates.4 In Section II, I argue that the civic friendship argument misses an important truth about relational inequality—that whether unreciprocated sacrifice generates subordination depends on background features of the relationship in question, including the motivations of the parties and the history of their relationship. I identify two general political contexts—one with background equality and the other with inequality—in which unreciprocated compliance with public reason promotes rather than undermines egalitarian relationships. When citizens are committed to maintaining a fair overall balance of cooperative sacrifice, unreciprocated compliance with public reason can display a valuable form of respect without subordinating compliers. And in contexts with background oppression, complying with public reason without demanding reciprocation can help repair political relationships to a footing of equality and restore bonds of trust. I illustrate the reparative value of public reason by examining the case of caste-based injustice in India. In ordinary political contexts, then, citizens can have duties to comply with public reason without demanding reciprocation. Acknowledging such duties, I argue in Sections III and IV, helps reconceptualize the role of public reason in non-ideal contexts. If the duty of public reason is always conditional on reciprocal compliance, it will have only a marginal political role in democracies characterized by substantial noncompliance and injustice. The view I defend gives public reason a broader and more inclusive role in promoting the ideal of an egalitarian community. This does not mean that political liberals must abandon the idea that civic friendship has special value. Rather, civic friendship is best understood as a relational ideal that citizens of pluralistic democracies should strive to realize in the long-run. Liberals have traditionally sought to justify liberal rights and institutions by appeal to a wider view of comprehensive moral or philosophical truth. Mill, for example, defends equal liberties by appeal to a theory of human flourishing that rests on an individualistic understanding of human nature.5 Political liberals believe that comprehensive views of human nature and the good of the sort to which Mill appeals cannot serve as a proper basis for the public justification of a liberal regime. This is because individuals reasoning freely and in good faith will inevitably disagree about such matters. This “fact of reasonable pluralism” ensures that no comprehensive doctrine can provide a mutually acceptable justification for liberal institutions.6 Why does it matter that citizens can accept the public justification of their political regime? Political liberals answer that the principle of public reason requires important laws7 to be justified by reasons that all reasonable citizens can accept given their own moral perspectives. Rawlsians identify the set of mutually acceptable public reasons in terms of a “family” of “political” conceptions of justice that affirm and give priority to basic liberal rights, opportunities, and distributive entitlements.8 Political conceptions are “freestanding” of comprehensive doctrines—they are built up from liberal values implicit in the political cultures of democracies without relying on comprehensive religious, ethical, or philosophical claims.9 The principle of public reason implies an individual moral duty of public reason, which requires citizens and officials to be willing to justify their political activity by appeal to a political conception of justice.10 The prevailing view among political liberals is that citizens must comply with public reason only for the sake of those who are willing to reciprocally comply.11 This implies that only citizens willing to comply properly belong to the justificatory constituency of “reasonable” citizens to whom the duty of public reason is owed.12 It is not immediately obvious why this view is correct. Not all moral duties are conditional on reciprocal compliance in this way. Consider duties of beneficence to help prevent harm to third parties in dire need. If I am walking by a pond with a drowning person whom I can save at reasonable cost, I plausibly have a duty to do so regardless of whether they would be willing to save me. Why not think that we can owe duties of public reason to those who aren't willing to reciprocate? Political liberals have done surprisingly little to answer this question. The view that public reason is always conditional on reciprocal compliance is often assumed without argument. However, advocates of the civic friendship account of public reason have developed what I think is the most promising argument for the view in the literature. The civic friendship argument grounds the duty of public reason in the communal relation of mutual respect that citizens realize when they reciprocally comply for the sake of one another.13 Andrew Lister understands the relevant form of respect in terms of “non-alienation”—unless laws are mutually justifiable, some will see their terms of political association “as being animated by purposes they find fundamentally alien.”14 R. J. Leland appeals to the value of mutual “non-imposition”. Imposition occurs when we seek to advance others' interests in ways that “substitut[e] [our] own judgments about what's good for recipients' judgments". This disrespects those we intend to help.15 Civic friends avoid the alienating imposition of cooperative terms by ensuring that each can accept the justification of democratic decisions from their own moral perspectives.16 Lister and Leland disagree about the level of societal compliance required for the duty of public reason to become binding. Lister defends a bilateral account—we always have a duty to justify decisions to every person who is willing to justify decisions to us.17 Leland defends a multilateral account—the duty of public reason is binding only if a significant proportion of citizens comply.18 This debate concerns the background conditions required for the duty of public reason to arise. I am addressing a different question: to whom is the duty of public reason owed when it does arise? In particular, why accept the prevailing view that the duty is owed only to those who are willing to comply with the duty? The civic friendship argument supports two answers. First, because unreciprocated compliance with public reason cannot realize civic friendship; it is morally pointless and cannot be justified. From the standpoint of one's comprehensive doctrine, Lister argues, compliance with public reason comes at a “moral cost”—it requires forgoing the pursuit of what one takes to be true justice for the sake of ensuring a mutually justifiable political order.19 This cost could be justified only if compliance produces a significant moral benefit. But unreciprocated compliance produces no such benefit, the argument goes, since only reciprocal compliance with public reason realizes the value of civic friendship. Lister suggests a different rationale for the view that public reason is conditional on reciprocal compliance when he argues that “[r]efusing to comply with a duty to someone who I know would not reciprocate involves affirming my own dignity”.20 The point can be understood in terms of the ideal of relational equality. Social inequality often arises in relationships when one party allows their willingness to sacrifice their interests to be exploited by others to advance theirs. It is sometimes argued that individuals have an interest in being able to appeal to the full range of their moral convictions when deliberating about matters of great political importance. Public reason requires citizens to bracket their fundamental convictions for the sake of ensuring a mutually justifiable political order.21 When X bears this cost for Y but Y is not willing to reciprocate, Y displays a “failure of recognition” towards X.22 And by allowing themselves to unilaterally bear this cost, X displays a failure of recognition towards themselves. In what follows, I respond to both of these arguments by identifying two ordinary political contexts—one with background equality and the other with background inequality—in which unreciprocated compliance with public reason serves important values while maintaining and promoting egalitarian democratic relationships. As an initial point, it is important to observe that particular instances of unreciprocated cooperative sacrifice do not always generate subordination. Whether they do so depends on contingent features of the relationship in question, including the attitudes of the parties and the history of their relationship. When parties commit to a fair overall balance of sacrifice, they need not engage in in-kind reciprocation of every sacrifice to maintain equality in their relationship. If two friends settle on a routine whereby one always shops for groceries, this need not create any relational subordination if the other friend makes sufficiently many other sorts of sacrifices. Moreover, when the balance of cooperative burdens in a relationship is unfair, beneficiaries can help restore relational equality by sacrificing without demanding reciprocation. If partner B consistently sacrifices their projects for the sake of partner A, A need not problematically subordinate themselves by now making sacrifices without demanding reciprocation. Rather, they can help restore relational equality by repaying their debt of sacrifice and recognizing the equal importance of B's interests. As we've seen, subordination can arise in relationships when individuals allow their willingness to sacrifice their interests to be exploited by others. But noncompliance with public reason is not always or even generally exploitative in this way. Some might fail to comply because they have never heard of political liberalism or the duty of public reason. Some might sincerely reject political liberalism in good faith. Perhaps they would willingly comply if they believed that the duty of public reason existed, but they sincerely believe it doesn't exist. Given the burdens of judgement, we should expect that many people will find the best arguments for political liberalism unpersuasive.23 Those who don't comply with public reason might still give equal weight to the interests of those who comply. They might affirm a comprehensive conception of justice that distributes the benefits and burdens of social cooperation in a fair and impartial way. And they may also be willing to comply with egalitarian laws that others advance through fair democratic procedures, and to cooperate on respectful terms in the democratic process. In such contexts, unreciprocated compliance with public reason need not subordinate compliers. Perhaps they make a particular sort of unreciprocated sacrifice, but noncompliers display a more general willingness to do their fair shares of the overall sacrifice required to relate as equals. To help illustrate the point, consider an idealized democratic society with a high degree of background equality and respectful cooperation: Jordania. Jordania is a democratic community with two large political parties, the Lakers and the Heat. The Lakers accept the duty of public reason. They believe that non-alienation is valuable in both its unreciprocated and reciprocated forms. They include all citizens of Jordania—including the Heat—in the justificatory constituency. The Heat do not comply with public reason. This is not because they want the benefits of being included in the justificatory constituency by the Lakers without bearing the costs of compliance. They simply do not find the best arguments for Rawlsian political liberalism persuasive after careful consideration. They instead appeal to their diverse comprehensive doctrines to justify a political platform that guarantees for all citizens a robust and fair set of primary goods. Jordania has fair and free elections. The Lakers more frequently hold a parliamentary majority because they have broader support among the citizenry. As a consequence, the laws more closely reflect the Lakers' platform than the Heat's platform. The Heat willingly do their fair shares of the sacrifices required by these laws and respectfully cooperate with the Lakers to solve collective problems. There is no pattern of oppression or injustice between the Lakers and Heat; one-off injustices are minor and swiftly remedied. An advocate of the civic friendship argument might concede that unreciprocated compliance with public reason does not create relational inequality in this case, but argue that it is nonetheless morally pointless because unreciprocated compliance cannot realize civic friendship. However, I think the best version of political liberalism acknowledges that individuals have an interest in non-alienation that stands alongside their interest in standing in relations of reciprocal non-alienation. Suppose we ask: why does mutual respect in politics require that we reciprocally justify political decisions to one another? There is generally no problematic lack of respect when individuals do not justify their choice of profession or life partner in terms of reasons others can accept. Why is democratic decision-making different? The most plausible sort of answer appeals to some general feature of democratic politics—for example, that it is coercive,24 that it involves public claims to authority,25 or that democratic power is the collective power of free and equal citizens26—in virtue of which mutual respect demands mutual non-alienation. And this feature of democratic politics will imply that citizens have an independent interest in non-alienation whether or not they comply with public reason. Individuals are subject to the state's use of coercive power and claims to authority even if they don't accept public reason. In this sense, I agree with David Enoch that considerations of reciprocity cannot lie at the “most fundamental moral level” of public reason.27 We can add some clarity by drawing two distinctions concerning cooperative reciprocity. The first distinction is between reciprocity of compliance and reciprocity of benefits. Reciprocity of compliance exists when individuals comply with norms out of a motivation to comply provided others do the same. Reciprocity of benefits exists when norms ensure, as Rawls puts it, that “all who are engaged in cooperation and who do their part as the rules and procedure require, are to benefit in an appropriate way as assessed by a suitable benchmark of comparison.”28 Reciprocity of compliance and reciprocity of benefits can come apart—there can be reciprocal compliance with norms that do not ensure fair mutual benefit and unreciprocated compliance with norms that do ensure fair mutual benefit. The second distinction is between local reciprocity and global reciprocity. Local reciprocity exists when there is reciprocity (of compliance or benefits) with respect to a particular norm. Global reciprocity exists when there is reciprocity (of compliance or benefits) with all norms of a given system. Reciprocity can be more or less global; individuals can reciprocally comply with greater or fewer norms and norms can do a better or worse job of ensuring fair mutual benefit. In Jordania, the Heat are not committed to local reciprocal compliance with the duty of public reason—the Lakers comply while the Heat do not. But the Heat display an effective commitment to global reciprocity of compliance and benefits. In such contexts, unreciprocated compliance with public reason need not subordinate compliers. Of course, real-world democracies have enduring legacies of injustice and oppression. Tommie Shelby argues that victims of systemic injustice are not bound by civic obligations, such as the duty to obey the law. Civic obligations are conditional on global reciprocity of benefits; when a cooperative scheme fails to distribute benefits and burdens in a reasonably just way, victims do not have a civic duty to obey the rules of the scheme. Drawing on Shelby, R. J. Leland argues that victims of injustice are not bound by the demands of civic friendship, including the duty of public reason.29 This is because they can reasonably conclude that fellow citizens do not take their civic interests seriously. And we cannot relate as civic friends with those who don't take our civic interests seriously. Notice what follows if we combine this conclusion with the view that public reason is always conditional on reciprocal compliance. Victims of injustice are not bound by the duty of public reason because fellow citizens don't treat them as equal partners in a fair cooperative scheme. So, victims must go above and beyond the call of duty to be included in the justificatory constituency and secure the good of non-alienation. By contrast, fellow citizens are properly included in the justificatory constituency simply by doing their duty. This seems unfair to victims of injustice. Leland suggests a response to this worry in passing, though he does not develop it at any length. Although unreciprocated compliance with public reason cannot constitute civic friendship, the response goes, it can promote civic friendship in the future. Extending fair cooperative terms that victims of injustice can accept from their own perspectives facilitate a kind of “reconciliation”. After reconciliation has been achieved through “an extended demonstration of concern over time”, victims of injustice will be bound by the full range of their civic obligations, including the duty of public reason.30 This is valuable because it is necessary for realizing civic friendship. Leland never clarifies in his brief remarks what he means by “reconciliation” or how public reason helps to promote it. One interpretation is that public reason can help repair political relationships damaged by injustice and oppression.31 I think that there is something right about this idea, and I develop a version of it below. However, any attempt to develop the idea within a civic friendship framework faces two problems. First, the civic friendship account seems unable to explain why unreciprocated compliance with public reason can help repair democratic relationships. The most natural explanation is that justifying laws in terms that victims of injustice can accept displays a valuable form of respect. However, the asymmetrical form of respect involved in complying with public reason without demanding reciprocation cannot be explained in terms of civic friendship, which is constituted by the symmetrical respect of reciprocal compliance. A second problem with Leland's account is that it grounds the reparative value of public reason in the value of realizing civic friendship in the future. But the eventual realization of civic friendship seems at most an ancillary reason to care about repairing relationships damaged by serious injustice. The more important reason is that we owe it as a matter of basic respect to affirm the equal standing of those who have been oppressed. Suppose that majorities can reliably predict that victims will not eventually accept the hand of civic friendship and comply with public reason in the future. They still have an urgent duty to repair their political relationship to a footing of equality and trust. If the use of public reason can help do so, this can ground a reparative duty of public reason. If public reason is grounded in its reparative value, its demands are not properly thought to be conditional on reciprocal compliance. Reparative duties are unidirectional and, plausibly, unilaterally binding. They are unidirectional in the sense that if X wrongs Y, X must perform reparative actions for Y (for example, apologize) that Y need not perform for X. Moreover, they are unilaterally binding in the sense that X must perform reparative actions regardless of whether they have assurance that Y would do the same if their positions were reversed. We must take responsibility for remedying the wrongs we have committed, even if we do not have assurance that those we have wronged would do the same for us in the (possibly hypothetical) world in which they wrong us. This can ground a duty to comply with public reason without demanding reciprocation. I have argued that the civic friendship account has difficulty answering two questions about the reparative value of public reason. First, why think that public reason can be used to repair relationships damaged by injustice? In the next subsection, I argue that the use of public reason facilitates a higher-order recognition of the equal standing of victims of injustice. Second, why is repairing relationships through public reason valuable? Leland answers that it helps promote civic friendship through reciprocal compliance in the future. I argue that moral repair through public reason is intrinsically valuable regardless of whether it eventually leads to reciprocal compliance or not. Democratic citizens stand in a distinctive relationship as co-sovereigns and co-subjects of their laws. They are co-sovereigns at least in the minimal sense that they have equal shares of political power—equal voting and other democratic rights—to determine which laws are made and implemented (or which representatives make and implement them). Like Rawls and many others, I assume that citizens have a duty to use their share of political power to promote just institutions and oppose unjust institutions.32 Citizens' duties of justice as co-sovereigns correspond to their vulnerability as co-subjects.33 In a democracy, we depend on others to ensure that laws and institutions treat us as free and equal members of a fair cooperative scheme. Unless enough fellow citizens vote and advocate for just laws and representatives, we may be subject to a basic structure that fails to give us the rights, entitlements, and opportunities we are owed as free and equal. This vulnerability is heightened by the fact that closed borders, limited economic opportunities, and language barriers make it excessively costly for most individuals to exit their states. Most people cannot easily escape the confines of an unjust basic structure. Political majorities commit a serious wrong when they use democratic power to exploit the vulnerability of minorities to advance their own interests. A paradigm example is persistent majority tyranny. This occurs when majorities use democratic institutions to persistently impose unjust cooperative terms on political minorities, often based on race, gender, class, sexual orientation, caste, or creed. I examine the case of caste-based tyranny below. Majority tyranny is wrong for many reasons. It is wrong in part because it results in substantively unjust institutions. But it is also an intrinsically problematic form of relational subordination—members of majorities and minorities do not relate as equal co-sovereigns and co-subjects, but as rulers and ruled-over. Consciousness of this subordination is likely to be experienced as oppressive by victims and to damage self-respect and bonds of trust. Individuals are active participants in majority tyranny when they vote and advocate for unjust laws, and they are mere beneficiaries when they do not vote for such laws but benefit from them. I will use the more general term “beneficiaries” to refer to both groups, and the term “victims” to refer to those who are subject to majority tyranny. Beneficiaries have a duty to oppose the effects of majority tyranny in a way that repairs their democratic relationships with victims. Importantly, the task of repairing a relationship introduces a set of demands oriented towards recognizing the equal standing of those who have suffered wrongdoing. Public reason can play an important role in this reparative task. It does not replace the need for corrective measures that guarantee victims a fair share of rights, opportunities, and resources, or for appropriate restorative procedures such as trials or commissions that hold wrongdoers to account. The use of public reason helps repair relationships through the public justification of such measures. It does so by facilitating a higher-order recognition of the equal standing of victims of injustice. Higher-order recognition is X's recognition that Y recognizes X's equal standing in their relationship.34 Establishing higher-order recognition is often an important element repairing relationships after wrongdoing. Persistent mistreatment generates a justified perception on the part of victims that they have, and are viewed as having, a subordinate moral status. After this damaging failure of recognition, establishing relational equality requires that victims can agree from their own perspectives that wrongdoers recognize their equal standing.35 This partly explains why reparative gestures such as apologies must be sincerely and freely accepted to have their full reparative effect.36 Higher-order recognition can be contrasted with the mere lower-order recognition that occurs when Y recognizes X's equal standing according to Y's moral standards but not X's standards. Wrongdoers might believe that they treat victims as equals, but if their moral beliefs are fundamentally alien or at odds with victims' beliefs about moral equality, victims might be unable to agree that they are respected as equals. Moral repair remains incomplete in an important sense. The use of public reason recognizes victims' equal standing, and promotes victims' recognition that they are so recognized, in two ways. First, it opposes the continued political alienation of victims. Part of what makes majority tyranny wrong is that it imposes institutions that political minorities cannot see as congruent with their own values and self-respect. Majorities have a strong reason to establish relational equality by combating the political alienation they have created and benefited from. The use of public reason does so by ensuring that important decisions are justified in terms of liberal values and principles victims can accept from their own moral perspectives. Second, the use of public reason promotes higher-order recognition of the equal weight of victims' interests in the distribution of primary goods. Majority tyranny entrenches an unfair and exploitative basic structure. Ensuring that the distribution of cooperative benefits and burdens is justifiable in terms of public reasons victims can accept promotes their recognition that majorities recognize their equality. If there is a mere modus vivendi on cooperative terms based on the balance of political power, there is no adequate assurance that majorities won't engage in further tyranny once their political position is more favourable. Moral repair is best promoted when victims can accept the entire public justification for democratic decisions, including the underlying values, principles, and intuitions that justify the decisions. This provides a reason to include victims of injustice in the justificatory constituency of public reason. Justice as fairness, understood as a political conception of justice, provides an argument from moral equality—the original position argument—that individuals with different comprehensive doctrines can accept. Given these considerations, I think the reparative value of public reason

How should the injustices associated with the enduring segregation of Black Americans be addressed? In contemporary debates about racial justice, there are two broad answers to this question. "New integrationists" argue that integration is necessary for remedying racial inequalities.1 "Egalitarian pluralists" instead argue that we should promote strategies that seek to improve the material condition of Black Americans without integration.2 Political philosophers have also recently begun to examine an apparently unrelated phenomenon: gentrification. According to its critics, gentrification may undermine social equality, violate residents' occupancy rights, subject residents to domination, impede valuable forms of democratic communication and/or impinge upon self-respect.3 But so far—with some important exceptions that I discuss below—political theorists have not examined the relationship between gentrification and integration in much depth. This article explores how racialised contexts of gentrification relate to and can shed light on the debate between new integrationists and egalitarian pluralists. I take as my focal point the debate between Elizabeth Anderson and Tommie Shelby.4 I examine three arguments for residential integration that Anderson makes: the opportunity argument, the epistemic-democratic argument, and the relational-democratic argument. I argue that racialised contexts of gentrification reveal some important limits to each argument for integration. But the upshot of my argument is not that we should abandon integration in favour of egalitarian pluralism. Rather, my suggestion is that examining racialised contexts of gentrification gives us a better understanding of the conditions under which residential integration can—and cannot—promote racial justice. My argument leaves open space for a modified defence of integration that takes these conditions into account. Ultimately, however, I suggest that such a modified defence of integration faces some important challenges of its own. This article's second aim is to contribute to the emerging literature on gentrification in political philosophy by examining its racial dynamics. By analysing the racial dynamics of gentrification, I bring to light some features of gentrification beyond residential displacement—in particular, the dynamics of social interaction in gentrifying neighbourhoods—which are relevant to its moral evaluation. My argument also responds to some recent arguments made by Andrew Pierce and Hwa Young Kim and Andrew Walton to the effect that gentrification's benefits can be harnessed, and its burdens can be limited, for the project of integration.5 My focus in this article is primarily on residential integration, though I do also discuss its relationship to other forms of integration at points. Residential integration is usually taken to be a central component—and often the central component—of the broader project of racial integration, because of the central role that residential segregation has played in reproducing patterns of racial inequality. And, as we will see, gentrification bears most directly on residential integration. First, I characterise the integration debate and explain gentrification's relevance to that debate. Then I examine the three arguments for integration in turn with and identify some limits to those arguments revealed by racialised contexts of gentrification. Finally, I reflect on the upshots of my argument for the integration debate, and for some recent arguments about gentrification's integrative potential. Residential segregation has long played an important role in upholding and maintaining what Desmond King and Rogers Smith call the "white supremacist racial order" in US politics.6 A variety of social technologies, including overt violence, restrictive covenants, redlining, real estate discrimination, predatory lending, and exclusionary zoning policies, have cemented patterns of residential segregation between Black and White Americans.7 Those patterns persist today, with roughly one in three Black metropolitan residents living in a "hyper-segregated" neighbourhood.8 And as Douglas S. Massey and Nancy Denton have demonstrated, residential segregation has played a central role in generating the persistent gap between Black and White Americans in wealth, opportunity, and well-being, effectively confining many Black Americans to so-called "ghettos" and constituting them as members of an underclass.9 The debate between new integrationists and egalitarian pluralists is a debate about how to respond to these injustices. Anderson and Shelby each provide sophisticated accounts of how advocates of racial justice should respond to segregation, which have set the terms of the contemporary debate over integration in political philosophy. This contemporary debate is an outgrowth of a broader set of debates within Black political thought about—put crudely—how Black Americans should respond to their oppression.10 These debates have a long history and have branched out in many directions, including Frederick Douglass's assimilationism, Martin Delaney's Black nationalism, Ida B. Wells-Barnett's view of Black self-help, W. E. B. Du Bois's cultural pluralism, Kwame Ture (Stokely Carmichael) and Charles Hamilton's conception of Black Power, Martin Luther King Jr.'s vision of interracial unity, Audre Lorde's reflections on difference, and far beyond. None of these complex visions of Black liberation can be reduced to a simple defence of integration or egalitarian pluralism, but each is an important part of the intellectual heritage of the contemporary debate between new integrationists and egalitarian pluralists.11 Both parties to the contemporary debate agree that desegregation is morally required.12 That is, they agree that the barriers that uphold and maintain patterns of residential segregation along racial lines should be dismantled, for example by aggressively enforcing anti-discrimination legislation and removing facially neutral zoning regulations that systematically exclude Black Americans from White neighbourhoods. This would be no small feat: although legal barriers to residential desegregation were removed with the passage of the Fair Housing Act, racial discrimination in housing markets continues to explain residential segregation to a significant degree.13 Anti-discrimination legislation in the real estate sector is underenforced, and more subtle forms of discrimination such as "racial steering" significantly impact patterns of residential segregation.14 And zoning regulations that limit the development of affordable housing in White neighbourhoods continue to play an important role in maintaining patterns of residential segregation.15 But few believe that effective enforcement of anti-discrimination legislation and inclusionary zoning policies would end the racial injustices associated with segregation. Both new integrationists and egalitarian pluralists see desegregation as a necessary but not a sufficient response to racial injustice.16 Beyond desegregation, new integrationists and egalitarian pluralists disagree about how segregation should be addressed. For new integrationists, policies that actively facilitate the residential integration of Black and White Americans—such as housing vouchers that promote the entry of Black Americans into predominantly White neighbourhoods—are a demand of justice.17 Anderson makes three main arguments for residential integration. First is the opportunity argument, which claims that residential integration has an indispensable causal role in remedying Black disadvantage by creating material opportunity. Second is the epistemic-democratic argument, which claims that residential integration is a central part of the process of democratic learning needed to address racial inequality. Third is the relational-democratic argument, which claims that only a racially integrated society will realise the conditions of social equality that are constitutive of the democratic ideal. Egalitarian pluralists reject the claim that residential integration is an imperative of justice. Shelby argues that "we should not regard residential integration as a legitimate mechanism for correcting the unjust disadvantages the ghetto poor face".18 On his view, racial justice neither requires nor prohibits residential integration. Shelby argues that relying on residential integration puts the costs of addressing racial injustice onto Black Americans.19 Black Americans moving into White neighbourhoods are exposed to racial hostility, prejudice and discrimination—or the burdens of being "Black in White space", as Elijah Anderson puts it—and on Shelby's view they cannot be required to bear such burdens as a matter of justice.20 He also argues that residential integration can prevent Black Americans from enjoying goods such as solidarity and community that can be obtained in neighbourhoods with a Black critical mass. For Shelby, solidarity amongst the oppressed is an important component of an ethic of resistance to racial injustice and is a legitimate basis for Black residential clustering.21 My suggestion is that we can make progress in this debate by examining racialised contexts of gentrification. The dynamics of gentrification illuminate some limits to arguments that have been made for residential integration. But first, it is important to clarify the relationship between the debates over integration and gentrification. What is gentrification, and what has it got to do with racial segregation and integration? Gentrification is a process of spatial and demographic change that occurs as urban spaces used by the less affluent are transformed into spaces used by the more affluent, typically involving an influx of wealthier residents into relatively deprived areas.22 There are some disagreements about the precise boundaries of the concept in the urban studies literature—for example, about whether gentrification necessarily involves the displacement of low-income residents, and about whether it only applies to historically disinvested neighbourhoods in the central city or whether it can apply to other spaces, such as public spaces.23 But for our purposes, we need not worry too much about these disagreements. This is both because the cases of gentrification that I am interested in here typically take place in historically disinvested residential neighbourhoods in the central city—and so are, in this respect, core cases—and because displacement need not be a necessary feature of gentrification for my analysis to get off the ground. Defined in this way, gentrification is not necessarily a racialised process. Here, however, I focus on cases of gentrification that involve White incomers moving into predominantly Black neighbourhoods, because such cases are particularly relevant for the integration debate. Gentrification often has some racialised aspect, and incomers in gentrifying neighbourhoods tend to be disproportionately White.24 But gentrification does not always display these racial dynamics. In fact, White gentrifiers often avoid majority Black neighbourhoods25 and prefer racially mixed neighbourhoods,26 which is consistent with findings on the role of White preferences in maintaining segregation.27 There is evidence to suggest, however, that gentrification involving White incomers in Black neighbourhoods is becoming increasingly significant.28 The racial dynamics of gentrification also differ significantly across contexts. Qualitative research on gentrification in Latino neighbourhoods and on gentrification involving the Black middle classes in predominantly Black neighbourhoods, for example, raises a number of distinct questions about racial politics that I do not address here.29 So it is important to keep in mind that the cases of gentrification that I focus on here represent only a subset—albeit an especially relevant and perhaps a growing one—of all cases of gentrification. In cases of gentrification that do display these dynamics, race is key to understanding how the process unfolds. Zawadi Rucks-Ahidiana argues that race plays an important role in the processes of valuation that are involved in gentrification.30 Gentrifying neighbourhoods are typically disinvested before redevelopment becomes profitable,31 and race often explains which neighbourhoods face disinvestment and how it occurs—for example, through systematic predatory lending to Black Americans that leads to mortgage foreclosures.32 Race also plays an important role in the revaluation of neighbourhoods, with the "diversity" of a neighbourhood being commodified to make a neighbourhood more attractive to early-stage gentrifiers and investors.33 Tyler Zimmer draws on Charles Mills's discussion of the racial coding of geographical space to diagnose the ideological underpinnings of racialised processes of gentrification.34 He argues that Black neighbourhoods are often depicted as "dens of iniquity, crime, vice and violence". Such depictions serve a dual function: to both rationalise the initial neglect of such neighbourhoods and to mark them out as "morally open for seizure, expropriation, settlement and development".35 Set against the background of White flight from city centres to the suburbs in the 1950s and 1960s, which left many inner-city neighbourhoods disinvested, gentrification is also sometimes viewed as an expression of a "revanchist" ideology that valorises the re-taking of the central city for the White middle classes.36 Race is not merely an incidental feature of cases in which White incomers move into predominantly Black neighbourhoods, but a central part of the way in which it unfolds and is legitimated. These racialised cases of gentrification are relevant to the integration debate because they are, in demographic terms, instances of residential integration. That is, racialised processes of gentrification create more racially mixed neighbourhoods, in which a larger number of Black and White Americans live side by side, at least in the short to medium term.37 In the long term, some neighbourhoods do experience "resegregation" as White incomers begin to dominate, but others appear to reach stable racially integrated equilibria.38 Advocates of residential integration typically envisage Black Americans moving out of Black neighbourhoods and into White neighbourhoods. In racialised cases of gentrification, the process operates in reverse: White Americans move into predominantly Black neighbourhoods. Racialised cases of gentrification provide an interesting test case for the integration debate. By examining the dynamics of gentrification in racialised contexts, we can bring to light some of the potential costs of residential integration, as well as discovering which of its benefits are robust across even apparently more inhospitable contexts for remedying racial injustice. Turning our attention to racialised cases of gentrification helps us to probe the limits of the case for residential integration. The aim of this exercise is not to undermine the case for integration in general. Rather, it is to identify some problems that can arise in the process of integration and to expose some ways in which integration may, in some circumstances, even deepen racial injustice, whilst at the same time deepening our understanding of the moral terrain of gentrification. In understanding racialised contexts of gentrification as instances of residential integration, I am adopting a fairly narrow understanding of residential integration as a process of demographic change that involves members of different social groups coming to live alongside each other in the same neighbourhoods. But of course, advocates of integration are likely to point out that they understand "integration" to mean something much more substantive than this. Faced with the problems that I identify in racialised contexts of gentrification, they may suggest that certain conditions must be met—so that integration proceeds "on terms of equality"39—in order for a process of demographic change in a neighbourhood to count as a genuine or meaningful instance of integration. This is certainly a defensible response, though it does face its own challenges, which I examine in more detail below. Indeed, part of what is at stake in my argument is a better understanding of how we should understand integration as a political ideal. The idea is not that gentrification is a particularly promising form of integration, such that problems that arise in gentrifying neighbourhoods cast doubt on the project of integration more generally. Rather, the idea is that by bringing to light the problems that arise in the context of gentrification, we can get a better understanding of the conditions under which residential integration can—and cannot—promote racial justice. A better understanding of these conditions is essential if we are to develop a more substantive conception of integration, to understand the place of residential integration within it, and ultimately to judge whether such a conception can function as an action-guiding political ideal in the struggle to achieve racial justice. For the moment, however, it is important to maintain a narrow understanding of residential integration, according to which it is simply a process of demographic change that involves members of different social groups coming to live alongside each other in the same neighbourhoods. This enables us to analyse the conditions under which residential integration can promote or hinder the achievement of racial justice, without prejudicing our analysis by making it the case that only cases that meet certain substantive conditions count as valid tests of residential integration. Of course, those conditions may ultimately figure in a broader account of the political morality of integration. It may be that residential integration is only, or especially, valuable for the pursuit of racial justice under certain conditions. But those conditions should be identified through an analysis that takes a narrow conception of residential integration as its starting point, rather than one that presupposes them as part of the definition of residential integration. Andrew J. Pierce and Hwa Young Kim and Andrew Walton have also recently examined the relationship between gentrification and integration, but their approaches differ from the approach that I take here.40 Both Pierce and Kim and Walton see in gentrification the potential to avoid some of the costs associated with traditional approaches to residential integration, since it does not require Black Americans to bear the burdens associated with moving into hostile White neighbourhoods. Pierce argues that if gentrification's negative impacts can be limited—through policies such as community land trusts—then it can provide the benefits of residential integration without some of the costs to which egalitarian pluralists object.41 Similarly, Kim and Walton argue that what they call "advantaged relocation" into disadvantaged neighbourhoods can—so long as it avoids displacement and cultural imposition—reduce prejudice and tackle social inequality without imposing unfair burdens on, or expressing stigmatising judgements about, the disadvantaged.42 I return to Pierce's and Kim and Walton's analyses of the relationship between gentrification and integration below. For the moment, it suffices to note that where Pierce's and Kim and Walton's aims are to show how an idealised form of gentrification might be harnessed to tackle racial injustice through integration, my focus is on what actually existing processes of gentrification tell us about the prospects and pitfalls of residential integration in the first place. Anderson's central argument for residential integration is that it plays an important role in achieving racial justice by creating opportunities for Black Americans. Anderson's argument includes both a diagnosis of the pathologies that reproduce Black disadvantage and a prediction about the role of residential integration in remedying that disadvantage. Anderson's diagnosis is that Black disadvantage is reproduced through various mechanisms of social closure such as opportunity hoarding and exploitation. She argues that residential segregation perpetuates Black disadvantage by preventing Black Americans from accessing good job opportunities, preventing the formation of social ties that enable upward mobility, and undermining the tax base needed for the provision of local public goods.43 She also argues that residential segregation causes stigmatisation by limiting inter-group contact between Black and White Americans. This exacerbates cognitive biases that give rise to racial stereotypes and discrimination—which intensify patterns of segregation, in a mutually reinforcing cycle.44 Anderson's prediction is that integration will remedy Black disadvantage in two main ways.45 First, she argues that residential integration plays an important role in increasing material opportunity for Black Americans by improving job prospects and facilitating the acquisition of cultural and social capital. Second, she argues that increasing contact between Black and White Americans leads to destigmatisation. Here, she draws on the "contact hypothesis", according to which inter-group interaction can—at least under certain conditions—reduce prejudice by broadening the boundaries of the perceived "in-group" and by reducing reliance on stereotypes.46 Anderson takes her prediction to follow from her diagnosis: "if racial segregation is the problem, it stands to reason that racial integration is the remedy". But the connection between the two is not a necessary one, and she accepts that we must "put integration to the test by investigating its consequences in practice".47 My suggestion is that racialised cases of gentrification show the limits of Anderson's prediction without giving us reason to doubt her diagnosis. Indeed, racialised cases of gentrification give us reason to support Anderson's diagnosis: neighbourhoods that become candidates for gentrification are often initially disinvested through precisely the mechanisms of social closure that Anderson takes to be central in reproducing Black disadvantage. But racialised cases of gentrification also show the limits of the two main ways in which Anderson predicts that integration will remedy Black disadvantage. Take first the argument that residential integration will increase material opportunity for Black Americans. Gentrification is sometimes defended on similar grounds, with its advocates arguing that it will increase employment opportunities for the urban poor and/or provide them with social contacts that enable upward mobility.48 But in practice, the relationship between gentrification and economic opportunity is much more complicated than this suggests, and this complexity sheds light on the limits of the opportunity argument for integration. One version of this argument—which Loretta Lees calls the "networks and contacts" argument—claims that a greater social mix in integrated neighbourhoods will create "bridging capital" between members of different social groups, leading to greater opportunities for the worst-off.49 Shelby objects to this argument on the grounds that it requires us to view social relationships as a "resource to be used for socioeconomic advancement" and "puts blacks in an untenable supplicant position" vis-à-vis Whites.50 But even if such concerns can be overcome, gentrification shows that its prospects are limited. As Lees points out, the empirical support for this claim is weak in the context of gentrification, with most studies finding little evidence that mixed-income neighbourhoods create social connections that lead to increased economic opportunities for the poor.51 The networks and contacts argument assumes that those living side by side will form economically beneficial social relationships with each other, but in gentrifying neighbourhoods incoming residents and existing residents tend not to form these kinds of relationships. Though spatially integrated, they continue to live parallel lives. So, gentrification illustrates that residential integration need not lead to greater material opportunity through building bridging capital. In fact, the reverse may even be true: because gentrification can undermine local social networks and community institutions upon which the poor rely, it can actually make them worse off.52 Another version of this argument appeals to increased employment opportunities and better public goods, such as schools and police protection, as ways of improving the material condition of the disadvantaged. Here, the evidence from the case of gentrification appears less clear overall. Gentrification clearly does bring some important material benefits to a neighbourhood. In some cases, gentrifiers can leverage their social position to secure improved public goods in their neighbourhoods.53 Employment opportunities may also be expanded overall, though it is not clear that low-income residents always benefit from those jobs.54 Existing residents sometimes view the increased opportunities, services, and public goods as serving the interests of White incomers, rather than Black residents. For example, increased police presence in gentrifying neighbourhoods may serve to enforce the values and expectations of White incomers and make the neighbourhood less hospitable for Black residents.55 And where there are material gains that accrue from gentrification, they are typically distributed unevenly. The evidence on mixed-income redevelopment suggests that "the majority of benefits have been realized by private sector developers, local government and other stakeholders…[l]ow income households, on the other hand, do not share in many of these benefits".56 This is not to say that the least advantaged do not gain at all, but benefits of gentrification are typically skewed towards the advantaged. Racialised contexts of gentrification also cast some doubt on the claim that integration will lead to destigmatisation. In such contexts, inter-group interactions often exacerbate, rather than attenuate, animosity and hostility between Black and White Americans. Qualitative research on gentrification in Black neighbourhoods suggests that contact between different social groups is often shallow at best, and that at worst it involves mutual suspicion, hostility, and conflicts over the use of shared space.57 Rather than creating genuinely shared spaces of community across social groups, gentrification leads to the creation of "White space"—spaces in which Black people are "typically unexpected, marginalized when present, and made to feel unwelcome"—in Black neighbourhoods.58 Inter-group interactions in these contexts appear unlikely to achieve the kind of destigmatisation that Anderson envisages, at least in the short to medium term. These findings illustrate some limits of the prospects of residential integration for remedying Black disadvantage, either through increasing material opportunities for the Black urban poor or through destigmatisation. When the paths to the acquisition of social and cultural capital are blocked, the benefits of public goods are skewed towards the advantaged, or the terms of social interaction are inhospitable to destigmatisation, we should not expect residential integration to yield these benefits. They also appear to provide supporting grounds for some egalitarian pluralist concerns about residential integration, such as Shelby's concerns that new integrationists overstate the likelihood of Black residents forming economically useful relationships with White residents and "give too little weight to the bonding capital that disadvantaged blacks already possess".59 But these findings do not undermine the causal argument for residential racial integration altogether. Instead, what they show is that if it is to be successful, integration must proceed in ways that avoid these problems. Anderson's second argument for residential integration is the epistemic-democratic argument, according to which integration plays an indispensable role in the process of democratic learning that is necessary for achieving racial justice. On Anderson's view, democracy is—amongst other things—a "mode of collective inquiry" that enables citizens to solve collective problems through the use of social intelligence.60 Democratic practices enable us to solve social problems because of the way that they incorporate diversity, communication, and feedback. Diversity here refers to the idea that those who occupy different social positions have access to knowledge arising from their experiences. Communication is a way of marshalling this asymmetrically distributed knowledge. And feedback mechanisms force decision-makers to consider the validity of political claims made on the basis of this knowledge. Anderson argues that integration promotes these epistemic functions of democracy, whilst segregation hinders them. Segregation encourages White Americans to complacently ignore the situation of Black Americans.61 It thus perpetuates the epistemic phenomenon that Mills has dubbed "White ignorance"—a systemic form of group-based miscognition that is sustained by, and itself perpetuates, White racial domination.62 Anderson argues that meaningfully addressing racial injustice requires negotiation and deliberation and that "negotiation and deliberation, in turn, require integration". In order to overcome White ignorance, negotiation and deliberation must incorporate the perspectives of Black Americans. Integration facilitates forms of "direct confrontation between claimants and the addressees of claims" that enable elites to be held to account.63 Anderson is right to suggest that democratic practices can function as tools for collective problem-solving. But we should a

The notion of rights is ubiquitous in philosophical discourse. As Allen Buchanan put it over thirty years ago, ‘Future historians of moral and political philosophy may well label our period the Age of Rights’.1 This notion is not only popular, but also complex. As the legal scholar W. N. Hohfeld famously suggested, rights are susceptible to multiple interpretations: they can be claims, liberties, powers, or immunities.2 Despite this variation, the consensus view is that the core instance of a right is a claim right.3 In the moral domain, claim rights designate a binary relation between a right-holder and a duty-bearer, where the former stands in a distinctive moral position vis-à-vis the latter.4 While there is controversy as to what, precisely, this distinctive moral position amounts to, the idea that claim rights capture it is seldom put into question. In this article, I challenge this way of thinking. I argue that the language of claim rights is ill suited for the purpose of picking out a distinctive moral position.5 I show that the notion of a claim right is susceptible to several disambiguations, just as the notion of a right itself is. From this, I conclude that we should either no longer appeal to the concept of a claim right in moral theorizing or rethink its purpose. The article proceeds as follows. In Section II, I set out two desiderata that a plausible definition of moral claim rights should satisfy. The definition should: (a) capture a distinctive moral position and (b) account for paradigmatic instances of claim rights in our ordinary language. In Section III, I show that the two most prominent accounts of claim rights fail to meet desideratum (b). Of course, the fact that prominent accounts are unsatisfactory does not mean that no satisfactory account could be developed. To support this stronger claim, in Section IV, I offer a systematization of our language of claim rights. I suggest that the greatest common denominator of such language is the idea of empowerment, and show that paradigmatic statements about claim rights track either the justification for certain forms of empowerment (justification rights statements) or empowerment itself and the particular status it confers on individuals (status rights statements). As I explain in Section V, this twofold connection between claim rights and empowerment reveals that, for structural reasons, our desiderata cannot be jointly satisfied. No notion of claim rights can both capture justification as well as status rights statements and pick out a distinctive moral position. In Section VI, I consider three possible implications of this conclusion. One is that we should abandon the notion of claim rights in moral theorizing. Another, less drastic possibility is that, in light of its disjunctive structure, the notion of claim rights should be given a different purpose. A third possibility, for those not persuaded by my arguments, is to treat my discussion as setting out a challenge that any satisfactory account of moral claim rights should meet. Section VII concludes. Before I start, let me make some clarifications. My discussion focuses on claim rights in moral theorizing specifically, where the notion is assumed to capture a distinctive moral position. As I shall explain towards the end of the article, the ambiguities that make the notion of claim rights ill suited for this purpose in moral theorizing also render it a powerful tool of political advocacy. I remain silent about the suitability of claim rights for the technical purposes of legal reasoning. Since those are, plausibly, different from the purposes of moral philosophy, the concerns raised in this article need not extend to the legal domain. Finally, a note on my terminology. In what follows, I shall treat ‘claim rights’, ‘directed duties’ (the violation of which wrongs particular others), and ‘duties owed to others’, as equivalent. This terminological choice is in line with standard use in the literature.6 An evaluation of claim rights presupposes an account of the work we want this notion to do in moral theorizing. If no plausible definition that does that job can be found, we will have reason to doubt the usefulness of the concept, at least with respect to the originally intended purpose. For an analogy, consider a society with a conceptual scheme different from ours, in which there are only two types of cutlery: forks and spives. Forks are defined in the familiar way, as tools for picking up solid food. The concept of a spife is disjunctive: it points either to tools for scooping liquid food or to tools for cutting solid food. We want these concepts to help us effectively select different ‘tools for eating’. As it happens, two kinds of objects fit into the category ‘spives’: what we (in our world) call spoons and knives. In this parallel society, when asking for a spife, one may have in mind either what we call a spoon or what we call a knife. Without further specification, one may therefore end up getting something one did not want. The category ‘spife’, it seems, is a recipe for confusion: it does not help us select our tools efficiently. Instead of having one concept—‘tools to either scoop liquid food or cut solid food’—it would be much better to have two: ‘spoons’ for liquid food and ‘knives’ for solid food. This would be a reason to abandon the concept of a spife, at least for the purpose of selecting specific pieces of cutlery. If, mutatis mutandis, the concept of claim rights were similarly unable to do the work we want it to do in moral theorizing, we would have reason to abandon it, at least for that purpose. So, what work do we want claim rights to do in moral theorizing? Claim rights, as I have noted in the Introduction, are meant to designate a distinctive moral position that some (that is, right-holders) hold vis-à-vis others (that is, duty-bearers). Claim-rights relations are ‘bipolar’, in the sense of always involving two relata: a duty-bearer on the one hand, and a claim-right-holder on the other.7 Each pole within the relation is marked by distinctive attributes. One pole bears duties: she is bound by certain oughts. The other pole possesses correlative rights. This, in turn, puts her in a special moral position vis-à-vis the corresponding duties, which are owed to her and the violation of which wrongs her. To be a claim-right-holder, then, is ‘more’ than being the mere beneficiary of someone's duty. A good definition of claim rights must elucidate what this ‘more’ precisely amounts to: what it means to have a claim (or, equivalently, to be owed a duty and be wronged by its violation). In addition, a good definition of claim rights should account for paradigmatic rights statements. This is a familiar desideratum. As Leif Wenar puts it, rights theorists ‘take extensional fit with ordinary language as a primary standard of success’.8 And as Matthew Kramer confirms, ‘nobody has ever denied that ordinary understandings are an important anchor for one's theorizing about rights and rights-holding’.9 This is not to say that a successful definition of claim rights must fit every conceivable instance of ordinary-language rights statements. Ordinary language is messy, and a good account of claim rights for the purposes of moral theory may well require a more disciplined use of this term than we find in day-to-day discourse.10 But if our definition failed to encompass paradigmatic rights talk, we would have to ask ourselves whether what we are capturing really is the special moral position intuitively associated with the language of claim rights, and not something altogether different. Emphasis on ordinary language invites the objection that the idea of a claim right is a term of art, introduced by Hohfeld to disambiguate different meanings of ‘right’ in the law. In ordinary language, we talk about rights simpliciter, not about any Hohfeldian incident in particular. So, how are we to isolate ordinary-language instances of the notion of ‘a right’ that pick out claim rights specifically?11 This doesn't pose an insurmountable obstacle to our inquiry, for two reasons. First, while many of our ordinary-language rights statements implicate bundles of Hohfeldian incidents, claim rights often lie at their heart.12 For example, whenever we insist that individuals have rights to a variety of objects—shelter, life, free speech, property, and so on—we typically imply, among other things, that they are owed secure access to them.13 When such access is denied, we can infer that some agents are failing to discharge the corresponding duties, hence that some claim rights are being violated. Second, to test whether any rights assertion is meant to involve claim rights, we can ask whether violations of the asserted right would wrong the right-holder in particular, as opposed to being wrong simpliciter. If our intuitive answers are affirmative, then we'll know that ‘right’, in the statement under consideration, is intended to mean ‘claim right’. a) Distinctive moral position. The definition vindicates the view that to be a right-holder (i.e., the addressee of a directed duty who may be wronged by its violation) is to stand in a distinctive moral position vis-à-vis duty-bearers.14 b) Consistency with paradigmatic claim-rights statements. The definition fits our core ordinary-language statements about rights; it neither over-generates rights nor under-generates them relative to those statements. As will become apparent, meeting both desiderata proves challenging. In fact, if I am right, it is impossible. Philosophical discussion about claim rights is dominated by the dispute between will and interest theories.15 In what follows, I briefly examine how these theories fare in relation to our two desiderata. Doing so will require going over some familiar ground. Although this exercise may seem a little tedious, it provides necessary background for my argument in later sections. Let us start with the will theory, prominently defended by H. L. A. Hart and Hillel Steiner among others.16 For the will theorist, to have a moral claim right is to possess the moral power to control another's duty (that is, to demand/enforce its performance, waive it, and seek compensation for its violation). This definition captures an important aspect of the phenomenology of claim-rights possession. After all, if your duty is owed to me, its performance is in some sense ‘mine’, and when something is mine, I have normative control over it. The will theory not only has independent appeal, but also meets desideratum (a). It assigns claim-right-holders a distinctive moral position, namely the power to control others' duties. But what about desideratum (b)? Here, as critics have famously pointed out, the theory encounters some difficulties, in that it under-generates rights.17 On the will theory, a capacity for choice—hence for normative control over others' duties—is a necessary condition for rights possession. Consequently, the theory cannot account for the rights of entities—such as children, people with severe mental disabilities, and non-human animals—who lack such a capacity.18 What is more, the theory has no room for so-called inalienable rights. Since, for the will theorist, to have a right just is to have the power to control a duty, including the power to waive it, inalienable rights—that is, rights that cannot be waived—are a conceptual impossibility.19 In sum, the will theory satisfies desideratum (a), but struggles with desideratum (b). Let me now turn to its main competitor. According to the interest theory, prominently defended by Joseph Raz among others, to have a right is to possess an interest weighty enough to justify another's duty.20 The interest theory, too, has much intuitive appeal. It is natural to suggest that a duty is owed to me when my interest is key to explaining its existence. Many rights—to life, to bodily integrity, to food, and so on—appear to fit the interest-theory rationale. Furthermore, ‘being the holder of an interest that justifies a duty’ can be aptly described as a distinctive moral position. Whoever possesses it stands in a special normative relation with respect to the duty-bearer. The interest theory thus satisfies desideratum (a). But, like the will theory, it struggles to meet desideratum (b). First, since there are several duties justified by others' interests that do not appear to give rise to rights, the theory over-generates rights.21 Consider my duty to assist an elderly lady who is struggling to cross the street. If I am uniquely positioned to help her, and can do so at very little personal cost, I ought to do it. Yet, most people would deny that the lady has a claim right to be helped by me in the same way as, say, my baker has a claim right that I pay for the bread rolls I bought from him this morning. My duty looks like a matter of beneficence, and although the beneficiary in this case is clear, this doesn't seem sufficient to generate a right.22 Second, as others before me have also noted, since there are cases where someone's right to X cannot plausibly be justified by appeal to their interest in X, the theory under-generates rights. For illustration, consider my property right over a cheap ornament gifted to me by a cousin. I have very little interest in possessing the ornament. Finding it rather ugly, I put it at the bottom of a closet some years ago and have since forgotten about it. Under any plausible construal of the situation, my interest in possessing the ornament is not weighty enough to place others under strong obligations not to steal it.23 One may respond that the relevant interest is not in possessing the ornament, but in being the one who decides what to do with it: it is a broader interest in freedom. But, as Raz himself admits, there are conceivable cases in which ‘I do not mind and have no reason to mind your taking it without permission’.24 That is, no freedom-interest is involved. In such cases, it seems, an interest theorist must conclude that I have no right. Raz's way out of this difficulty involves suggesting that, while I may have little interest in owning this particular object, I, together with everyone else in society, have an interest in there being a well-functioning system of property. That system is justified not by appeal to the interests of particular individuals, but by everyone's interests: by the common good. Such common-good cases are ones where the right-holder's interest and the public's interests are ‘harmoniously interwoven’: that is, ‘benefiting him is a way of benefiting them, and … by benefiting them the right-holder's interest is served’.25 In such cases, the weight of my individual interest is ‘augmented’ by that of others, thereby justifying my (property) rights. While this response may allow the interest theory to avoid the under-generation charge, it does so at the cost of making it unable to satisfy desideratum (a).26 For, if what explains others' duties not to steal my property are not my interests, but my interests as well as everyone else's (namely, the common good), from the perspective of the interest theory, I can no longer be said to stand in a distinctive moral position in relation to those duties. When it comes to generating those duties, my interests are no more important than everyone else's. This strikes me as a significant price to pay, greater than just accepting the costs of under-generation.27 In sum, the will and interest theories—in their original formulations—both satisfy desideratum (a), each pointing to a distinctive moral position. They also each have difficulties satisfying desideratum (b). My aim in highlighting these difficulties has not been to induce a change of heart in these theories' advocates. Objections to the will and interest theories have been around for a while, and yet the literature continues to be broadly divided between these two camps, with theorists on each side either ‘biting bullets’ or providing ever more subtle responses to the objections raised.28 But I do hope that a fair-minded reader can grant me this: neither the will nor the interest theory provides a fully convincing account of claim rights, judged by our two desiderata. In fact, as others before me have pointed out, the debate appears to have reached a stand-off.29 Explaining why this is so, and why the difficulties plaguing the will and interest theories are bound to extend to alternative theories as well, will be my task in the next section. Diagnosing the difficulties with will and interest theories requires us to systematize the conceptual domain of claim-rights talk. A good systematization of this domain must be responsive to the following three considerations. First, there has to be something that unites claim-rights talk, otherwise we would simply be unable to discuss claim rights without always talking past each other. Second, will and interest theories must be ‘on the right track’, otherwise their enduring influence would be inexplicable. Third, the conceptual domain of claim-rights talk must be sufficiently heterogeneous to explain these theories' inability to capture some core rights statements. My suggestion—which, as I will show, accounts for all three considerations—is that moral claim-rights language, in all its core uses, is linked to the idea of empowerment relative to others' duties. This is the ‘greatest common denominator’ of claim-rights talk. The relevant empowerment may vary depending on the case at hand, and consists in different forms of control over others' duties—such as the power to demand and enforce them, to waive them, and to seek compensation for their violation. Crucially, our language of claim rights implicates empowerment in two distinct ways. On the one hand, we can talk about claim rights while focusing on the justification for empowerment: the moral reasons (typically, interests) there are to empower individuals. From this justificatory perspective, a statement such as ‘A has a moral claim right to X' is shorthand for ‘A's interests justify certain forms of empowerment—either for A herself or for other agents—relative to others’ duties to X'.30 On the other hand, we can talk about claim rights while focusing on the morally justified forms of empowerment that are present in any given circumstance: an empowered status individuals enjoy. From this status perspective, a statement such as ‘A has a moral claim right to X' is shorthand for ‘A is empowered relative to others' duties to X'. Justification statement. Women's interests in bodily integrity justify women's power to, for example, demand that their husbands not batter them, enforce their husbands' duties not to batter them, seek compensation in case of violations, and (possibly) waive those duties.31 Status statement. Women have the morally justified power to, for example, demand that their husbands not batter them, enforce their husbands' duties not to batter them, seek compensation in case of violations, and (possibly) waive those duties.32 Many statements about claim rights are equally susceptible to both interpretations. Which one is most appropriate will depend on the speaker and context at hand. For instance, rights advocates may often be understood as making justification statements, pointing to the reasons why certain forms of empowerment ought to be granted to a certain class of individuals: for example, LGBT groups, the differently able, ethnic minorities, and so forth. Others may instead point to existing forms of empowerment, reminding their interlocutors that certain agents already enjoy them. The fact that, for most statements about claim rights, status and justification interpretations are jointly satisfied allows us to talk meaningfully about claim rights, and gives us the sense that claim-rights language points to a unique phenomenon. This, however, is not so. First, and somewhat trivially, the forms of empowerment associated with claim-rights language need not always co-occur. For instance, there may be cases in which waiver powers are inappropriate (for example, the right not to be tortured) or where powers to request compensation would not be justified (for example, when rights violations are too trivial). While all claim rights involve empowerment, then, they do not all involve the same forms of empowerment. Second, and more importantly, there are some paradigmatic instances of claim rights which can only be made sense of from either a status or a justification perspective. These are statements where the ‘locus of justification’ and the ‘locus of status (empowerment)’ diverge. In such instances of divergence, ordinary language sometimes locates rights in the individuals whose interests justify empowerment, and other times in the agents who possess an empowered status. The former cases are captured by justification-rights statements, but not by status-rights statements. For the latter, the reverse is the case. Such statements reveal that, in fact, beneath our language of claim rights, lie two distinct—though related—phenomena: the justification of empowerment and empowerment itself. To see this, consider the assertion ‘children have moral claim rights to education’. This can be easily interpreted as a justification-right statement, meaning that children's interests in education justify adults' powers to see to it that those who bear duties to provide children with education fulfill them. But this statement lacks an equally natural interpretation from a status perspective. The view that children themselves should be empowered in relation to others' duties is hard to sustain. Children, as we have seen, cannot plausibly have the power to control others' duties. Their interests may justify that power, but the power lies elsewhere—with parents, guardians, or the state. From a status perspective, therefore, we cannot say that children have a right to education. For another example, return to the assertion ‘I have a moral claim right to the ornament gifted to me by my cousin’. From a status perspective, this can be easily interpreted as meaning that I am empowered to, for example, demand that others not take the ornament, seek compensation in case of theft, and so on. But when we try to read this assertion as a justification-right statement, we encounter difficulties. It is just not true that my interest in possessing and controlling that ornament explains why I am (or should be) empowered in relation to it. Justification-rights statements are ill suited to capturing rights whose existence is justified not by the interests of what ordinary language identifies as the right-holder, but by broader considerations, typically the interests of society as a whole in the existence of certain power-conferring practices: for example, a well-functioning property system.33 To sum up, claim-rights talk is underpinned by a common reference to empowerment relative to others' duties, and our language of rights invokes such empowerment in two ways, captured by the distinction between status- and justification-rights statements. Many statements about claim rights are susceptible to both status and justification interpretations, giving us the impression that a unique phenomenon lies beneath them. But this impression is, in fact, illusory. My systematization of the language of claim rights is, I believe, clarifying in several respects. Apart from highlighting what assertions of claim rights have in common, it (1) helps us diagnose the failures of the will and interest theories, (2) makes sense of these theories' enduring influence, and (3) explains why it is impossible for any definition of claim rights to satisfy both of our desiderata. In what follows, I elaborate on these points. While the interest theory accounts for the core concerns behind justification-rights statements, the will theory largely articulates those captured by status-rights statements. This, I believe, explains their lasting popularity: each theory responds to an important dimension of our rights language. At the same time, each theory is deficient in several respects, and my empowerment-based taxonomy helps us see why. Consider the interest theory first. Its focus on the justification of duties—as opposed to the justification of empowerment relative to those duties—is responsible for its over-generation of rights: for declaring that rights exist even in cases of mere beneficence. Moreover, insofar as it identifies rights with justifications for duties, rather than with empowerment relative to duties, the theory is systematically unable to account for rights statements we can only make sense of from a status perspective (for example, my property right over the ugly ornament). Turning to the will theory, while this is better placed to make sense of status statements, it runs into difficulties for two reasons. First, as we have seen, the forms of empowerment associated with claim rights are heterogeneous and need not always be co-instantiated (as, for example, in the case of inalienable rights). By assuming that such forms of empowerment always co-occur, the will theory under-generates rights. Second, by not referring to the interest-based justifications for empowerment, the theory is systematically unable to capture those claim-rights statements we can only make sense of from a justification perspective (for example, children's rights). As anticipated, my analysis not only helps us diagnose the difficulties with prominent theories of rights, but also reveals that such difficulties are bound to extend to any attempt to define claim rights in line with the two desiderata introduced at the start of this article. The point can be made most effectively by recalling that justification- and status-rights statements locate rights, as a moral category, in different places. Justification statements locate them in the individuals whose interests justify empowerment, and it is an open question as to which form of empowerment is most appropriate in any given case. Status statements, by contrast, locate claim rights in the individuals who possess certain justified forms of empowerment. The fact that the individuals whose interests justify empowerment and the individuals who are empowered need not always coincide (think of children's rights and property rights), coupled with the fact that empowerment can take several different forms that need not always be co-instantiated, makes it impossible to develop a definition of claim rights that both captures all core instances of ordinary-language use and singles out a distinctive moral position.34 If we want to capture all core ordinary-language statements, we need a disjunctive definition of claim rights. Such a disjunctive definition, in turn, does not point to a distinctive moral position, but rather to a family thereof, just as the broader notion of ‘a right’ is disjunctive, and points to four possible normative relations.35 I have argued that, for structural reasons, no extensionally adequate definition of a claim right can enable this notion to do the job we want it to do in moral theorizing. There is no distinct moral position that the notion of a claim right captures, but a family thereof. The idea of a claim right is not much more specific, then, than the broader notion of a right, of which it is meant to be a disambiguation. What follows from this? I consider three possibilities and offer some arguments in relation to each. My aim is not to defend one in particular, but to show that, regardless of one's instincts, the present discussion should prompt us to rethink the direction taken by much of the contemporary conceptual debate about claim rights. The first possibility is to abandon the language of claim rights in moral theorizing. The thought is simple: if that language is meant to pick out a distinctive moral position, but cannot do the job, then it would be better to cut it out of our moral vocabulary. Instead of using the category of claim rights as if it pointed to a unique phenomenon, we could simply refer to the specific phenomenon we have in mind: empowerment-justifying interests or particular forms of empowerment themselves. In other words, we should not ask questions such as ‘Does A have a moral claim right against B?’, but should instead directly rely on one of the following possible disambiguations: Does A have the moral power to demand and enforce the performance of B's duty? Does A have the moral power to waive B's duty? Do A's interests justify this or that form of empowerment relative to B's duty? and so forth. Incidentally, this solution parallels the conclusion reached in the imaginary society in which, instead of talking about spoons and knives, people use the concept of a spife. There, employing separate concepts for spoons and knives seemed like the right way forward. Why should things be any different in the case of claim rights? A natural reaction to these concerns is to suggest that confusion could be avoided by being explicit about the definition of a right one is relying on. The problem, it might be argued, is not with the notion of a claim right, but with users of the notion being unclear about what they mean by it. While I agree that this way of proceeding would mitigate the confusion, it is not obvious that it offers any advantages over the option of just avoiding claim-rights language. Holding on to the language of claim rights is still likely to mislead readers. Rights language is so charged that, despite definitions, readers will still project their own understanding of claim rights on to whatever matter is being discussed. Furthermore, the suggested strategy—to be more explicit about one's definition of a right—seems motivated by continued attachment to a word (‘claim right’) even when this is decoupled from what was supposed to motivate such attachment in the first place: the idea that a claim right designates a

Deciding whether a state has political authority is arguably the most fundamental judgment we can make about its moral standing. When a state possesses legitimate authority, it has the exclusive right to govern the occupants of its territory. Even when its decisions are inefficient, irrational, or morally mistaken, its subjects are expected to comply with the legal order it establishes, and outsiders are expected to refrain from attempting to interfere with its internal affairs. A state that lacks authority, on the other hand, enjoys no such protection: provided there is an adequate justification for doing so, its subjects are permitted to actively resist attempts to enforce the law, and outsiders are permitted to intervene in order to reform or replace state institutions.1 Given these practical stakes, it's crucial to settle on a plausible account of the requirements for legitimate political authority that can guide our assessments of both nascent and established states. Recently, Alexander Motchoulski has developed a novel “hybrid” account of political authority2 that he claims is up to this task. This hybrid view, which combines elements from existing functionalist and fair-play accounts3, establishes the authority of particular states in two stages.4 In the first stage, we determine whether the state satisfies various functionalist criteria such as the consistent protection of basic rights, the maintenance of democratic institutions, and the reliable provision of essential public goods. If the state passes this test, we then examine whether the principle of fair play gives the territorially demarcated populations the state claims jurisdiction over reason to comply with its laws and support its institutions. This involves evaluating the goods provided by the state to see if they are either acceptable upon reflection or morally required. A good is acceptable if, given their interests, the recipients have reason to prefer acquiring the good plus a corresponding duty to reciprocate over not receiving this good at all. On the other hand, a good is morally required just in case the recipients have a moral duty to contribute to its provision regardless of whether the good is acceptable to them.5 If the aforementioned political goods meet either of these conditions, individuals benefiting from them have a duty to reciprocate the political cooperation of their co-citizens by obeying the laws of the state and doing their fair share to maintain its cooperative structure over time. It is this duty, owed by the citizens of a state to one another, that generates particularized relationships of authority between states and populations.6 According to Motchoulski, the primary attraction of his view is that, unlike orthodox functionalism, it is able to account for our beliefs that: (i) states generally lack authority over populations they incorporate through acts of non-consensual annexation or colonization, and (ii) historical injustices committed by the state will sometimes have an impact on its authority, but this impact can diminish or “fade” over time. Thus, by endorsing his view, functionalists can retain their basic normative commitments without running afoul of our intuitions about the boundaries of political authority.7 However, I argue that Motchoulski's view fails to deliver on this promise. Even if we accept his two-stage framework for assessing the authority of states, we cannot rule out the possibility that states can acquire legitimate authority over populations they unilaterally annex, nor can we reach reasonable judgments about when and how historical injustices undermine the authority of states. Having established this, I suggest that instead of looking for other ways to rehabilitate functionalism, we should accommodate our intuitions about these issues by endorsing a theory of political authority that recognizes the moral importance of collective self-determination. Cases where one state unilaterally annexes another pose a problem for orthodox functionalism because of the view's singular focus on whether candidates for political authority are performing various morally important tasks. Since functionalists contend that the adequate provision of basic political goods is sufficient to ground a claim to legitimate political authority, endorsing their view means accepting that, so long as they govern in a reasonably just manner, the annexing power has a right to govern its newly acquired territory. Yet, this strikes most of us as wrong. Intuitively, states cannot legitimately extend the boundaries of their authority this way, even if the annexation is bloodless and the annexing power does not violate anyone's individual rights.8 One of Motchoulski's central claims is that, by taking considerations of fair play into account when evaluating states, his version of functionalism is able to avoid this counterintuitive result. To develop this claim, he begins by demonstrating that the presence of pre-existing reciprocal duties among co-citizens is sufficient to make unilateral annexation wrong. As he contends, certain kinds of positive first-order duties, including duties of reciprocity, give rise to negative, second-order duties of non-interference. For example, if I have a reciprocity-based obligation to help my friend move between cities, my roommate has a corresponding obligation not to force me to stay in town by hiding my car keys because they believe my time would be better spent volunteering with them. By the same token, because co-citizens of a legitimate state owe reciprocal duties of political cooperation and compliance to each other, and unilateral annexation ostensibly interferes with the fulfillment of these duties, acts of annexation wrong the annexed individuals by violating a second-order duty that is owed to them. Having established this, Motchoulski then argues that when alternatives are available, benefits conferred through acts of wrongdoing clearly do not count as acceptable or morally required. And if the benefits of political cooperation are neither acceptable nor morally required from the newly annexed population's point of view, the annexing state cannot demand compliance or support from this population by appealing to the principle of fair play.9 Although Motchoulski's version of functionalism represents a major improvement over traditional accounts, I believe his response to the annexation objection ultimately fails. To see why, it's helpful to divide acts of unilateral annexation into two categories: total annexations and partial annexations. Total annexation occurs when State A forcibly claims authority over all of State B's territory and subsequently replaces all of B's political institutions with their own, effectively dismantling B as a political entity. By contrast, in cases of partial annexation State A forcibly claims authority over some of State B's territory and the corresponding residents, but leaves B intact as an independent state. In the remainder of this section, I will show that the hybrid functionalist/fair play account cannot reliably rule out the possibility of states obtaining political authority over their newly incorporated populations in either kind of case because both total and partial annexations can occur without interfering with existing duties of reciprocity. Starting with cases of total annexation, suppose that the United States successfully carries out a bloodless annexation of all of Canada and subsequently declares that all former Canadian citizens are now citizens of the United States with all the attendant rights and obligations. Instead of being covered by the Charter of Rights and Freedoms, their rights are now protected by the Constitution. Instead of electing members of Parliament in Ottawa, they vote to send representatives to Congress in Washington DC, and so on. Could Canadians complain that this change interferes with their ability to discharge their ongoing reciprocal duties to each other? I think the answer is clearly "no". This change would, of course, alter the bundle of political and civil rights held by Canadians: the precise scope of their right to freedom of expression would change somewhat, as would the extent and content of their right to political participation, and so forth. However, for functionalists like Motchoulski, the political goods that citizens have a reciprocal obligation to provide—e.g. institutions that grant all individuals a degree of political power and legal protections that safeguard individuals against arbitrary interference in their private lives10—are not precisely specified rights or entitlements. Rather, they are generic goods that can be secured through a variety of distinct schemes of rights.11 So, as long as the scheme maintained by the United States does an adequate job of securing the relevant generic goods, following annexation individual Canadians would still be able to reciprocate the benefits provided to them by their co-citizens through obedience to the law and contributions to the institutions that provide them with essential services. The only difference is that this would now involve complying with US laws and supporting the US institutions they are collectively subject to. Here, Motchoulski might object that my argument misinterprets the content of the Canadians’ reciprocal political obligations. There are two versions of this objection worth considering, so I will address each in turn. The first version says that since individual Canadians had been benefiting from the protection of particular Canadian laws and the goods provided by particular Canadian institutions, the principle of fair play obligates them to comply with the same set of laws and support the same institutions going forward. Therefore, the kind of total annexation we are examining does in fact wrong the annexed population by interfering with their ability to discharge their duties of reciprocity. The problem with this version of the objection is that it seems implausibly restrictive to understand reciprocal political obligations in this way. Motchoulski's observation that the reciprocal benefit one returns must be “fitting and proportionate”12 only dictates the goods that one has a duty to contribute to and the level of contribution required; it does not directly determine the exact channels through which one must make this contribution. If I have a duty of reciprocity to contribute to the provision of good G for the benefit of population P, and I can make an equal and satisfactory contribution by donating to either organization X or organization Y, the people to whom I owe the duty have no grounds for complaint if I donate to X instead of Y, or vice versa; what matters is that a fair contribution to G is made. Likewise, the fact that my fellow citizens and I have reciprocity-based obligations to protect each other's rights and contribute to public goods does not entail that we must comply in perpetuity with the particular set of laws and institutions that previously structured our cooperation, provided that there are alternative arrangements that would serve this purpose just as well. If it did, any change to the legal and political institutions we are subject to would be considered pro tanto wrong on Motchoulski's view, even benign changes enacted through internal, democratic mechanisms with the full support of the community. This takes us to the second version of the objection. In response to the issue just raised, Motchoulski might point out that the goods commonly associated with political cooperation require stable, long-term coordination to achieve. Maintaining justice, for example, is only possible if my co-citizens and I reliably understand, and conform to, the legal norms and rules that govern our mutual interactions. If we frequently opt to ignore the law in favour of our own moral judgments, or if we are often unsure about what is legally permissible in our dealings with each other, it is practically inevitable that widespread rights violations will often occur. For this reason, although the principle of fair play does not obligate members of a political community to comply with the exact same set of laws and institutions over time, it does require them to settle on public rules for determining which laws and institutions they are collectively subject to and only cooperate through the laws and institutions selected via these rules. Otherwise, stable political coordination will be impossible to achieve, and the state will be unable to provide the goods that justify its rule in the first place. And if this is true, it seems to follow that the unilateral annexation of Canada by the United States does interfere with the ability of Canadians to fulfill their reciprocal political obligations insofar as it replaces the legal system chosen by them in accordance with their internal rules with a new, unfamiliar system that they had no role in choosing and may find more difficult to coordinate through. I find the main premise of this objection compelling. Political goods are clearly secured through stable coordination over time, and this fact no doubt plays some role in determining whether a particular set of institutions will facilitate the reciprocal provision of these goods. But it doesn't follow that a population can only fulfill their reciprocal political obligations through institutions selected via internal rules that they have agreed upon. Though admittedly the imposition of a new legal system from the outside can destabilize practices of political coordination by increasing the rate of intentional non-compliance with the law, or by triggering widespread uncertainly regarding what the law requires, there seem to be feasible measures that an annexing power could take to mitigate these risks. Concerns about large numbers of Canadians refusing to comply with US law could, for instance, be addressed by introducing additional incentives for compliance, while concerns regarding public knowledge of the law may be avoided by implementing a phased transition of the legal code and providing public announcements detailing the most important changes. Assuming they successfully implemented such measures, US institutions could in principle be just as effective at facilitating political cooperation as the Canadians’ former institutions were, and therefore the annexation of Canada by the United States would not necessarily interfere with the ability of Canadians to discharge their reciprocal political obligations. So, because the total annexation of a country can occur without violating the unique, second-order duty of non-interference that duties of reciprocity give rise to, Motchoulski's theory fails to explain why the goods provided by the annexing power are not a morally acceptable basis for duties of fair play in such cases. What about cases of partial annexation? At first glance, it might seem obvious that in these cases there is always interference with the annexed population's existing duties of reciprocity. For example, if the United States annexed the province of Alberta but left the rest of Canada intact, it is hard to see how Albertans could continue to provide their fellow (non-Albertan) Canadians with reciprocal benefits by maintaining a shared system of legal rights and contributing to public goods, since Albertans would no longer be subject to the same set of federal political institutions as the rest of Canada. However, given the distinctive structure of political reciprocity, I do not think that these changes necessarily interfere with individuals’ duties of fair play either. Most reciprocal exchanges have an asynchronous structure: A gives some good to B at T1, so B owes her an appropriate return that she provides later on at T2, then B provides A with some additional benefit at T3 that A reciprocates sometime later at T4, and so on for the duration of their relationship. But the kind of political reciprocity we are concerned with has a unique, synchronous structure. As Motchoulski himself acknowledges,13 in a well-ordered society a state of equilibrium is reached where every citizen's continual compliance with the law simultaneously fulfills the duties they owe to their co-citizens and generates reciprocal duties to obey the law, which their co-citizens then discharge more or less automatically in an ongoing cycle. This difference is morally relevant because it affects the ways in which outsiders to these relationships can violate their second-order duties of non-interference. When two or more people are engaging in an asynchronous form of reciprocity, separating them so that they can no longer interact will often leave at least one individual without a fair return for the benefits they provided others, and correspondingly, prevent at least one other participant from fulfilling their outstanding reciprocal obligations. This does not seem to be the case when it comes to relationships involving synchronous political reciprocity, however. Sticking with our previous example, in the period immediately before annexation occurs, it's of course true that Albertans have an obligation to their fellow Canadians to comply with Canadian federal law because they are benefitting from the legal compliance of their co-citizens, and their fellow Canadians in turn have an obligation to them to do the same. But at the moment that the annexation of Alberta is complete and Albertans are no longer subject to Canadian institutions, these bi-directional duties of reciprocity are simply dissolved. Albertans do not have any leftover duties of reciprocity to non-Albertan Canadians since, by assumption, they continuously returned an equitable benefit to their fellow Canadians by complying with Canadian law for as long as they were benefiting from the protection of their shared institutions. For the same reason, non-Albertan Canadians do not have any leftover obligations to Albertans either.14 The upshot of this is that while partial annexations do terminate previous relationships of reciprocity, they do not necessarily interfere with anyone's ability to fulfill their presently existing political obligations. Motchoulski's argument thus fails to establish that partial annexations are inherently wrong and, by extension, that annexing powers generally lack political authority over their newly acquired territories and populations. Before moving forward, however, it is worth flagging that Motchoulski could respond to this conclusion by pointing out that, even if his particular account of the wrong of unilateral annexation is not viable, his hybrid functionalist theory will still cohere with our intuition that states lack political authority over territory acquired through unilateral annexation as long as we can provide some account of why such annexations wrong individuals.15 However, I am skeptical of this move, since I believe that the most natural explanation of the wrong of unilateral annexation appeals to the value of collective self-determination, and once we acknowledge this, we have reason to treat respect for self-determination as an independent requirement for legitimate political authority. I will return to this point in Section III. Having addressed Motchoulski's treatment of the problem of unilateral annexation, I now wish to argue that, contrary to what he claims, his hybrid functionalist/fair-play view also fails to provide a plausible solution to the issue of historical injustice that has long plagued functionalist theories of political authority. Historical injustices pose a problem for functionalist theories because, due to their exclusive focus on whether the state is currently doing an adequate job of protecting its subjects’ basic rights, these theories cannot account for the intuition that a history of unjust treatment—especially involving wrongful assimilation into the state—can void a state's claim to authority over the victims of these injustices as well as their descendants, yet the state may eventually gain authority over these groups as the relevance of these injustices “fades” over time.16 Motchoulski attempts to reconcile his version of functionalism with this two-part intuition by again appealing to the idea that individuals only have a duty of reciprocity to comply with the state's directives if the benefits conferred by the state are acceptable to them. On his view, to the extent that the descendants of groups such as indigenous peoples and national minorities that were previously incorporated into the state through annexation or colonization maintain “residual” cooperative practices that generate duties of reciprocity that compete with the demands of the state, these descendants are not required to accept the benefits conferred by the state. And, as previously established, if the political goods provided by the state are not acceptable or morally required from the point of view of a particular group, the state will lack authority over this group. Motchoulski contends that this line of reasoning reliably tracks our judgments about when past injustices cease to impinge on the state's political authority: since the victims’ descendants only retain a claim to independence for as long as they continue to participate in competing structures of cooperation, it is not surprising that we believe that groups who were the victims of injustice hundreds of years ago, but have long since been fully integrated into mainstream society and enjoy the benefits provided to them by their co-citizens, are now legitimately subject to the state's authority.17 Unfortunately, while this argument from competing duties may align with our intuitions about historical injustice in some cases, it seems to yield very implausible results in others. Consider a version of Motchoulski's Conquestia and Annexia case where, following annexation, the Conquestian state goes to great lengths to constrain the Annexian's residual practices of political cooperation precisely so that they do not conflict with the demands of the Conquestian legal order. Initially this is done through force, but over time the state succeeds in fracturing the Annexian resistance movement enough that they can employ increasingly subtle, legal mechanisms in order to shape the relationship between the state and the Annexian minority. This eventually culminates in a situation where, though the Annexians continue to exercise some limited self-government in areas approved by Conquestia, they have effectively lost touch with their more comprehensive practices of political cooperation that would compete with the legal obligations that Conquestia imposes on them. However, despite Conquestia now treating them as full citizens with equal civil rights, the younger generation of Annexians feel extremely alienated from the Conquestian state and share the long-term goal of regaining their political independence. I take it that, in this case, the Annexians do not have robust, reciprocity-based political obligations to Conquestia. On the contrary, it seems clear to me that they are entitled to reject the ongoing imposition of Conquestia's legal order and re-engage with their previously abandoned structures of political cooperation. Note, however, that Motchoulski's theory cannot affirm this judgment. On his view, since the Annexians did at some point stop engaging with the political practices that competed with the authority of the state, their descendants cannot reasonably reject the benefits of political cooperation they receive through Conquestian institutions, and hence they are bound by the principle of fair play to comply with Conquestian law.18 I think this is a decisive strike against the hybrid functionalist position. Although I agree that there must be some conditions under which historical injustices no longer threaten the legitimacy of particular states, it seems morally absurd to claim that a state can have authority over the descendants of a people whose territory it forcibly annexed merely because the state was so effective at supressing this group's traditional political practices that, for the time being, their descendants have little choice but to assimilate into their oppressors’ political community. Let's review the discussion so far. The aim of Motchoulski's hybrid theory is to provide an account of political authority that can adequately address the boundary problem and the problem of historical injustice while “maintaining basic functionalist commitments”.19 In the preceding sections, I have argued that his view falls short in relation to both of these goals. Incorporating the principle of fair play into the functionalist framework does allow us to explain why individuals have obligations to particular states. However, it does not enable functionalism to avoid the annexation objection, nor does it provide us with a reasonable account of when past injustices are no longer relevant to questions of state authority. At this point, we could look for other ways to rescue functionalism from these problems. But I think it is more useful to first identify the moral considerations that motivate our intuitive reactions to cases involving unilateral annexation and historical injustice, and then examine how these considerations should be integrated into a theory of political authority. Like Margaret Moore and Anna Stilz, I believe our intuitions about these issues reflect an appreciation of the moral significance of shared political identities and the associated interest in collective self-determination.20 Contrary to what functionalist theories seem to suggest, the citizens and long-term residents of a state typically do not conceive of themselves as atomistic, self-interested actors for whom politics is merely a means of securing individual rights and avoiding interpersonal domination. Instead, they often identify with and value their status as members of a specific community defined by a distinct history of cooperation on their territory and maintained through processes of mutual recognition. In short, they see themselves as being part of a unique political "people".21 Generally speaking, individuals have a substantial interest in achieving a fit between this kind of shared political identity and the political institutions they are subject to. As members of a people, we lead more autonomous and fulfilling lives when we are able to endorse the laws that govern us because these laws have been shaped by us in accordance with our shared commitments and beliefs.22 But this interest in collective self-determination is inevitably frustrated when one country non-consensually annexes another. Since this act necessarily involves the imposition of a new political order by outsiders who do not share their history of political cooperation, following annexation the newly incorporated population will no longer be able to readily affirm their participation in the political institutions that structure much of their lives. Hence, there is a way in which unilateral annexation wrongs the subjected population even in cases where it has no impact on their individual rights or duties of reciprocity. Does the presence of this wrong suffice to justify our intuition that states lack political authority over populations they annex? I think it does. Barring some kind of moral emergency, there is simply no reason why a population that has previously enjoyed a significant political good like self-determination should be expected to suddenly settle for less by recognizing the authority of new institutions that deny them this good. The moral logic at play here seems equally applicable when we turn our attention to the relationship between historical injustice and political authority. If the descendants of an annexed population maintain their identity as a distinct political group and continue to perceive a mismatch between this identity and the annexing state's institutions, they are no less entitled than their forbears to complain that their current political status undermines their interest in collective self-rule. Yet, it is clearly possible for this mismatch to be resolved over time. In some cases, this will be due to these descendants coming to identify with the annexing nation as successive generations are brought up and form attachments within this new social context. In others, the descendants will retain their original group identity and the annexing state will ultimately make institutional concessions—typically by implementing a form of federalism—that enable them to have a degree of political autonomy they find acceptable. Either way, the upshot is that individuals no longer suffer the distinctive harm that makes political rule following unilateral annexation objectionable, and thus past acts of annexation no longer bear on the state's current political authority.23 This account of when and how historical injustices fade from relevancy easily explains the intuitive distinction Motchoulski draws between cases involving the descendants of a people whose territory was seized many centuries ago (for example, the Mercians), and cases involving the descendants of a people who were subject to more recent historical injustices (for example, Indigenous populations). In the former set of cases, we usually conclude that past injustices do not undermine the authority of the state because the descendants of those affected usually no longer identify as a distinct people. By contrast, in the latter set of cases it is often apparent that the descendants have maintained their status as a distinct political people and oppose their current political situation, and so we often believe that the state's claim to authority over them is illegitimate. Moreover, unlike Motchoulski's view, it also yields the intuitively correct verdict in cases like the modified version of Annexia and Conquestia discussed in Section II. This is because, while residual practices of cooperation that compete with the state can be evidence of a desire for greater self-determination, their presence is not necessary to ground a complaint against the state's authority on this account. All that is necessary is that the descendants do in fact suffer from a lack of fit between their group identity and their political institutions, which by assumption the Annexians do. Given the central role that the value of collective self-determination plays in our moral judgments about these issues, there is a strong case for adopting an account of political authority that treats respect for self-determination as an independent requirement for state legitimacy that exists alongside various other requirements.24 After all, if the denial of self-determination is a significant enough wrong that, in cases of annexation, the state responsible lacks political authority over the annexed population and potentially their descendants, it stands to reason that if a state consistently frustrates this interest in other ways its claim to authority should be rejected as well. Perhaps, however, Motchoulski could propose that we instead incorporate the value of self-determination into his hybrid fair-play/functionalist framework by stipulating that political goods are only acceptable in the duty-generating sense if they do not come at the cost of this value. He would then be able to account for our intuitions regarding unilateral annexation and historical injustice, and hence the critique developed in this article would no longer represent a fundamental challenge to his view. Ultimately though, I think this concession would leave Motchoulski with a pyrrhic victory. The appeal of functionalism is closely connected to the view's singular focus on the role that states ought to play in securing a narrow class of basic, uncontroversial political goods. By grounding claims to authority in this relatively modest goal, functionalist accounts provide an alternative for those who are skeptical of tying political legitimacy to the achievement of more demanding procedural or relational ideals, including the ideal of collective self-determination we have been discussing. Indeed, defenders of functionalism often make a point of highlighting potential issues with treating collective self-determination as a requirement for legitimacy. Thus, for example, Jonathan Quong voices the concern that treating the state as a vehicle for the self-determination of a nation or people is incompatible with a commitment to liberal neutrality,25 while Kjartan Koch Mikalsen contends that making political authority contingent on whether the state's jurisdictional boundaries respect the self-determination of groups would threaten the freedom of individuals by undermining the ability of many states to secure a “rightful condition”.26 While I believe these issues can be addressed, for those that do not it is naturally appealing to instead employ functionalist standards when making judgements about a state's legitimacy. But, for obvious reasons, this appeal is undercut if we reconfigure functionalism in a way that ends up forcing it to take the ideal of collective self-determination into account anyway. So, while the structure of Motchoulski's account does allow him to move in this direction, doing so would leave him with a view that is unsatisfying, if not unrecognizable, to its likely proponents. Having addressed this last potential response, I think it's safe to conclude that Motchoulski's fair-play/functionalist hybrid cannot deliver on its promises. Assuming, then, that we do not want to give up our intuitions about unilateral annexation and historical injustice, does this conclusion entail that we should abandon the functionalist framework altogether? Not exactly. One could argue that while we have reason to treat respect for self-determination as an independent condition for legitimate political authority, this requirement is only applicable in cases where the population does in fact identify as a political people with a desire to be self-governing. In other cases where the population within a territory is so fractured that even an inchoate people cannot be identified, it seems reasonable to instead adopt a functionalist point of view when assessing candidates for political authority, since our primary focus should be on ensuring that individual rights are protected.27 Still, it's important not to lose sight of the difference between adopting this kind of contextual view and endorsing functionalism as a stand-alone theory of political authority. Even in cases where it is initially appropriate to grant institutions authority on functionalist grounds, we should anticipate that, over time, the people subject to these institutions will come to think of themselves as a political group with their own perspective on who should govern them. And when this occurs, we should be prepared to revise our initial assessment and demand that, going forward, these institutions must reflect the shared political identity of their subjects in order to retain their right to rule. For helpful suggestions and feedback on earlier versions of this article, I would like to thank Owen Clifton, Jordan Desmond, Michael Luoma, Margaret Moore, and an anonymous referee. Work on this article was financially supported by a doctoral fellowship from the Social Sciences and Humanities Research Council of Canada. There are no potential conflicts of interest relevant to this article. The author declares human ethics approval was not needed for this study. All relevant data are included in the article.

Philosophers engaged in projects of ‘ameliorative inquiry’ offer accounts of social categories, such as those of race and gender, that set aside the descriptive question of understanding those categories as they currently exist in favour of developing accounts of how we ought to think of those categories given our political goals. For feminists whose goal is to combat gender injustice, the dictionary definition of ‘woman’ as ‘adult human female’ has, on the face of it, little to offer. If we see gender injustice as arising primarily out of the system of patriarchal oppression, then understanding ‘women’ and ‘girls’ as the classes of people who are the primary targets of that oppression might seem appropriate, even if it turns out that these classes exclude some human females and include some human males. And if we see gender injustice as also involving an unjust imposition of gendered expectations and gender categories on people regardless of their own gendered understanding of their selves, then an account of ‘women’ as ‘adult human females’ might appear even to exacerbate this kind of gender injustice, by forcing people into gendered categories that are contrary to their identities. As a result, the consequence of ameliorative inquiry is often to recommend that we revise our accounts of existing concepts so as to better serve our political ends. But what should we do if, having engaged in an ameliorative inquiry, we come to the conclusion that our concepts need to be amended? Concepts and definitions have a life outside of philosophy, and presumably those convinced that revisions are needed should have something to say about what should change in our use of our concepts outside of discussions taking place in philosophy journals. In the case of gender concepts and terms such as ‘woman’ and ‘girl’, these terms have existing legal meanings and uses. The natural consequence of ameliorative inquiry should then presumably be proposals to amend our existing legal categories to better represent the targets of our inquiry. Indeed, in recent years, many jurisdictions have been grappling with the question of whether to amend the ways in which sex and/or gender are recognized in law; and so an opportunity presents itself for those involved in ameliorative inquiry into gender concepts to offer some practical proposals for legal changes. The Right to Legal Recognition. Everyone has the right to legal recognition without reference to, or requiring assignment or disclosure of, sex, gender, sexual orientation, gender identity, gender expression or sex characteristics. Everyone has the right to obtain identity documents, including birth certificates, regardless of sexual orientation, gender identity, gender expression or sex characteristics. Everyone has the right to change gendered information in such documents while gendered information is included in them. STATES SHALL: While no state has formally signed up to be bound by the Yogyakarta Principles, they are taken very seriously as a guide to best practice. It is against the backdrop of Principle 31 that we can understand recent proposals in the UK (both in England and Wales as consulted on in 2018 by the UK government, and in Scotland as passed by the Scottish government in December 2022) to amend the 2004 Gender Recognition Act (GRA) to move to a system of self-determination of gender (self-ID) in line with Part C of Principle 31. What, then, should those involved in offering ameliorative definitions of ‘woman’ say about legal proposals around the recognition of legal sex or gender? In part, of course, that will depend on one's ameliorative account. This article will take the ameliorative proposal of Katharine Jenkins2 as a starting point, and consider the question of what someone convinced by Jenkins's ameliorative approach should conclude about the legal recognition of sex or gender. Given that this is currently a live issue in UK political debate, the article takes the UK legislative framework3 as its central case study, and considers how existing legislation (including the 2004 GRA and the 2010 Equality Act (EA)) currently serves to protect the interests of the target groups of Jenkins's ameliorative definitions of ‘woman’, and how best to amend this legislation if we wish to promote gender justice in line with Jenkins's analysis. However, given the broader context of international moves to implement Yogyakarta 31, the lessons of this case study in the UK setting should have implications elsewhere too. Jenkins's proposal is of particular interest, because the result of her ameliorative inquiry is to deliver two separate target gender concepts, and hence two separate (albeit overlapping) concepts of ‘woman’, which she takes to be equally important if our aim is to oppose gender-based injustice, and thus deserving of ‘equal status within feminist theory’.4 Jenkins's target concepts are being classed as a woman and having a female gender identity. What makes Jenkins's proposal of particular interest in relation to recent political debates over gender recognition is that it has typically been the case that opposing sides have argued for the primacy of one notion of gender over another (that is, for a broadly sex-class-based notion of woman over a broadly gender-identity based notion, or vice versa). A starting point that sees both notions as of equal importance is of interest if it can deliver concrete proposals that preserve the interests of both groups (or, at the very least, provide a framework within which the interests of both groups can be recognized, and balanced where they conflict). Even if there are many on either side of the debate who would reject Jenkins's position that both notions of ‘woman’ are equally important, to the extent that their interest is in ensuring that ‘women’ in their preferred sense are adequately catered for in legislation, a legal proposal that shows how the interests of both categories can be protected and balanced in law might offer a compromise for single-account views of ‘woman’. Such a proposal would be grounded in principles of toleration of alternative accounts of gender even if these accounts are not accepted, insofar as the main interest on both sides is in a legal framework that protects women in whichever sense they take to be important. We will start then with a reminder (in Section I) of the two notions of ‘woman’ that Jenkins takes to have equal importance in the feminist fight against gender injustices, before (in Sections II and III) considering how each of them is represented in the current UK legislative framework (particularly via the 2004 GRA and the 2010 EA). I argue in Section II that, given that the EA outlaws discrimination on grounds of perceived, as well as actual, possession of a protected characteristic, the protected characteristic of sex in the EA adequately protects the interests of women in Jenkins's class-based sense (that is, those who are observed or imagined to be female), even though not everyone who counts as a ‘woman’ in this class-based sense is counted as female according to the EA's understanding of this category. Section III notes, by contrast, that current provision for gender recognition in UK law only recognizes the gender identities of a small subset of people who identify as women, and that it is also ill suited for recognizing non-binary gender identities. Section IV considers how a move to amend the 2004 GRA to introduce gender self-ID (as consulted on by the UK government and as recently enacted by the Scottish parliament) might, while improving the situation vis-à-vis legal recognition for those who identify as women, nevertheless weakens existing protections in the UK EA for those classed as women (in Jenkins's sense).5 Section V offers an alternative route to legal recognition of gender identity that is both in line with Jenkins's aim of treating both notions of ‘woman’ as equally important, and also preferable to the current ‘self-identification’ proposal in offering a straightforward route to the recognition of non-binary gender identities and to the protection of all transgender people against discrimination on grounds of a transgender identity. The central contention of this article is that moving to a system of self-identification of gender in line with Part C of Principle 31 of the Yogyakarta Principles collapses sex and gender identity in UK law in a way that undermines important sex-based protections provided by the 2010 EA which, I argue, primarily serve to protect women as a class, as well as making it difficult to fully protect against discrimination on grounds of a transgender identity. The article argues instead for the legal recognition of two separate protected characteristics: sex (understood biologically as per existing UK case law)6 and gender identity (determined on the basis of self-identification). Separating sex and gender identity in law allows for the recognition of non-binary gender identities, and for the proper legal recognition and protection of trans people understood as people whose gender identities do not match their sex. It also allows for clear discussions about whether a service should or could be provided on sex-based or gender-identity-based lines (as permitted by current UK Equality legislation), without requiring a one-size-fits-all approach to the provision of single-sex or single-gender services. By advocating a proliferation of legal sex/gender-based categories in law, however, this article goes against not just Part C of Principle 31, but also against the more radical proposal of Part A: that states should altogether end the registration of sex/gender as part of an individual's legal personality. Part A has thus far seen little uptake, with transgender advocates typically focusing instead on making the legal case for self-identification of sex/gender, in line with Part C, rather than pressing for deregistration of sex/gender. Thus systems of gender self-identification in line with Part C have to date been adopted by 19 countries, starting with Argentina in 2012 and joined most recently by Scotland in December 2022. By contrast, there has been much less enthusiasm for adopting the recommendation of Part A, although the Australian state of Tasmania has taken some steps in this direction by no longer recording sex on birth certificates (though sex is still registered at birth). A recent ESRC-funded project, ‘The Future of Legal Gender’ (FLaG), explored the pros and cons of deregistration in the UK context; in their final report the project team proposed deregistration of sex/gender along with merging the protected characteristics of ‘sex’ and ‘gender reassignment’ in the UK's EA into a single protected characteristic of ‘gender’.7 While the FLaG team's eliminativist proposal of removing sex/gender entirely as part of an individual's legal personality is an intriguing one, it is not a route I advocate here, in part because it seems to me to carry similar dangers of ‘merging’ two important and importantly distinct categories in need of protection to those carried by the more standard proposal of continuing to record sex, but doing so on the basis of self-identification. For those, though, tempted by eliminativism, the contention of this article can be read as conditional: if states are in the business of legally recognizing sex/gender, they should do so by recognizing these as two separate categories (one biological and one a matter of self-identification), rather than stopping at the inadequate half-way house we currently have (which stands between complete deregistration on the one hand and full and separate recognition of both sex and gender identity on the other) of recognizing gender identity only by conflating gender identity with sex. Katharine Jenkins identifies two target concepts of ‘woman’, a class-based concept and a gender-identity based concept,8 arguing that both are equally necessary for feminist aims. I will briefly review these two concepts here. the recognition that males and females do not only differ physically, but also systematically differ in their social positions. What is of concern, to put it simply, is that societies, on the whole, privilege individuals with male bodies.10 Recognizing this hierarchical structure of oppression of one class of people by another, Haslanger's ameliorative proposal is to define woman in terms of membership of the subordinated class in this structure. Importantly, for Haslanger, this allows some people who are not biologically female (on one or other way of drawing the male/female divide) still to be classed as ‘woman’ for politically important purposes: given that anyone who is regularly perceived11 as female in a patriarchal society will be placed in the subordinate role and suffer from patriarchal oppression, the politically relevant category is not identical with the biological one. A person with Complete Androgen Insensitivity Syndrome (CAIS), for example, will on most biologically based classifications be classed as male on account of having XY chromosomes and (internal, undescended) testes, but they will typically be entirely female in outward appearance and, as such, their experiences under conditions of patriarchy will be just like those biologically classed as female. The same goes for trans women to the extent that they are perceived as female. It also recognizes that trans men may successfully transition out of oppression and into a dominant social role to the extent that they are perceived as male. So by defining women as those on the receiving end of patriarchal oppression on the grounds of being perceived as female, Haslanger's account has it that being an adult human female is neither necessary nor sufficient for being classed as a woman. S is classed as a woman within a context C iff S is marked in C as a target for subordination on the basis of actual or imagined bodily features presumed to be evidence of a female's role in biological reproduction.12 Although Haslanger offers her account as an ameliorative definition of ‘woman’, she doesn't think a great deal hangs on whether to use the terms ‘woman’ and ‘man’ to pick out members of the subordinate and dominant classes in the system of patriarchy, so long as the needs of the subordinate class are recognized as politically important. For the sake of clarity, we will use the term ‘womenCL’ to pick out ‘women’ in Haslanger's sense—that is, the class of so-called ‘sexually-marked subordinates’—and reserve ‘womenID’ to pick out ‘women’ in Jenkins's second, identity-based sense—that is, the class of people with a female gender identity. Regardless of their choice of terms, both Haslanger and Jenkins agree that an important feminist aim is to improve the lot of womenCL. S has a gender identity of X iff S's internal ‘map’ is formed to guide someone classed as a member of X gender through the social or material realities that are, in that context, characteristic of Xs as a class.14 S has a female gender identity iff S's internal ‘map’ is formed to guide someone classed as a woman through the social or material realities that are, in that context, characteristic of women as a class.15 To avoid circularity, Jenkins's norm-based account of internal ‘maps’ appeals to the different gender norms that exist to enforce the class-based oppression of womenCL by coercively imposing different social roles on those perceived as male from those imposed on those perceived as female. The presence of these different social roles, Jenkins hypothesizes, gives rise to internal mental maps that help to guide individuals through a sexist society by reminding them of the sex-based expectations that apply to them in various contexts. Jenkins further hypothesizes that, in the case of transgender individuals, their internal maps have somehow developed differently from those expected of them, so that someone assigned male at birth and socialized as male, for example, may nevertheless develop a mental map that corresponds to the norms typically expected of females. In other (non-binary) cases, individuals may have no stable social map corresponding to either of the typical ‘male’ or ‘female’ maps. Importantly, for Jenkins, the internal map account explains how people may have a male or female gender identity despite having deep misgivings about the norms expected of them, and despite being extremely gender non-conforming in their behaviour as a result of these misgivings. Insofar as a female person recognizes the typical female norms as applying to her, even if she consciously and deliberately transgresses these norms in her behaviour and presentation, she can correctly be counted as having a female gender identity. Questions have been raised about whether Jenkins's account ultimately succeeds in its aims of providing a plausible and non-circular understanding of ‘gender identity’ that serves to meet the political needs of transgender people while including everyone in the appropriate gender identity.16 However, I will set these aside here, and assume that, if not Jenkins's account, then some such account will be available that plays the roles that Jenkins sees as being required of an adequate ameliorative definition of gender identity. That is, I will assume that some account is available that supports the central claims that gender identity is an important enough feature of individuals to be worthy of political recognition and protection (especially in cases where gender identity does not match sex), and that while having a given gender identity is not simply a matter of believing oneself to identify as a man, woman, or non-binary, nevertheless individuals are best placed to know their own gender identity, so that in practice individuals should be presumed to be authoritative with respect to questions of their own gender identity. If this is so, then there is a class of individuals with a female gender identity, womenID, who face certain forms of gender-based injustices, and whose interests are therefore, according to Jenkins, equally important as those of the (overlapping) class of womenCL, when it comes to the feminist fight against gender-based injustices. Furthermore, on this view, the best way of determining membership of this class is through sincere self-identification. If we follow Jenkins in thinking that ameliorative inquiry into the politically relevant concept of ‘woman’ delivers two separate (albeit overlapping) classes, womenCL and womenID, both of whom are on the receiving end of gender-based injustices, the next question to ask is how to protect the interests of these groups in legislation, and whether legal protections for ‘women’ need to be clarified or amended so that it is clear that they apply to one or other of these groups (or indeed both). I will first consider womenCL, and then womenID, as these groups are reflected in existing UK legislation, and ask the question of whether recognition of Jenkins's two senses of ‘woman’ would require changes to the use of the term ‘woman’ in legislation as it stands. I will then consider the question of how proposed changes to UK legislation around gender identity (particularly the 2004 GRA) may interact with protections for these two groups. What legal and political mechanisms are in place that help to improve the lot of womenCL? In the UK, a significant piece of legislation protecting historically marginalized groups is the 2010 Equality Act (EA), which brought together over 116 separate pieces of previous equalities legislation stretching back to the 1970 Equal Pay Act. The EA identifies nine protected characteristics: age; disability; gender reassignment; marriage and civil partnership; pregnancy and maternity; race; religion or belief; sex; sexual orientation. The Act places duties on public bodies to take these protected characteristics into account in policies and practices, including making reasonable adjustments for those in possession of a protected characteristic, and outlaws direct or indirect discrimination on the basis of possession of a protected characteristic except in circumstances where it can be argued that discrimination is ‘a proportionate means to a legitimate aim’ (for example, in the provision of separate services for the sexes (section 26) and single-sex services (section 27), or sex-segregated sporting competition (section 195)). The Act also includes requirements to monitor potential discrimination on grounds of protected characteristics. The Explanatory Notes to the EA clarify that ‘references in the Act to people having the protected characteristic of sex are to mean being a man or a woman, and that men share this characteristic with other men, and women with other women’, where ‘man’ is understood to mean ‘a male of any age’, and ‘woman’ is understood to mean ‘a female of any age’.17 The sex-based terminology (male/female) might suggest that a traditional understanding of woman as ‘(adult) human female’ is at work in the EA (with the ‘adult’ qualifier being removed in this case for simplicity to ensure that references to sex in the Act also cover children). This understanding would be in line with English case law, which established a legal understanding of ‘sex’ in biological terms to mean the state of being male or female.18 This is complicated somewhat, however, by the provisions of the 2004 GRA, which enable some people, in specifically defined circumstances, to obtain a ‘Gender Recognition Certificate’ (GRC) and change their legal sex so that it no longer reflects their biological sex. Until recently, it has been unclear whether sex in the EA is to be understood as biological sex or legal sex—that is, whether to include as male (female) someone whose acquired gender is male (female) as a result of obtaining a gender recognition certificate. However, the recent (13 December 2022) decision in the judicial review case brought by For Women Scotland clarifies that ‘sex’ in the EA is to be read as legal sex rather than biological sex: that is, to include those in possession of a GRC in line with their acquired sex.19 As of May 2021 (latest available figures), fewer than 6,000 people in the UK are in possession of a GRC, in line with estimates of the likely ‘small number of people’ wishing to take up the provisions of the Act as discussed in parliament.20 This means that, despite the recent clarification that ‘sex’ for EA purposes means ‘legal sex’, aside from this very small number of exceptional cases, ‘sex’ in the EA overwhelmingly tracks biological sex. This would change, of course, if changes to the GRA resulted in many more people obtaining a GRC. (Indeed, this is one of the reasons why proposed amendments to the 2004 GRA to introduce self-ID, and particularly the recently passed Gender Recognition (Scotland) Bill, have been so contentious. I discuss this further in Section IV.) For now, though, let us work on the assumption that ‘sex’ in the EA closely tracks biological sex, returning later to the question of the effects of opening up gender recognition on the basis of self-identification on this. If we follow Jenkins in adopting Haslanger's ameliorative account of ‘womanCL’ as ‘sexually marked subordinate’, we might think that the natural upshot of this should be to advocate a change in the protected characteristic of ‘woman’ in the EA, so that, rather than picking out those who are female, it instead picks out those who are regularly, or for the most part, perceived as female. This, however, would be impractical, cruel, wrongheaded, and, given the nature of the protections provided by the EA, which outlaws ‘discrimination by perception’, unnecessary, as I will explain below. Speaking practically, making a legally protected category dependent on whether someone ‘passes’ or not as a given sex raises questions about how this is to be determined. Should the protected characteristic apply to those who are treated as female often enough? By whom, and how often? These questions should make clear also the cruel nature of a proposal to place people in a legal category on the basis of whether they ‘pass’ as a particular sex. It was in fact envisaged that the provisions to change sex of the 2004 GRA would be taken up primarily by individuals (described in the 2004 discussions as ‘transexuals’) who do ‘pass’ in their acquired gender, with the privacy needs of individuals not to have to disclose their cross-gender history being a key pillar in the motivation offered when presenting the Gender Recognition Bill, which includes provision for changing the sex marker on birth certificates so that people will no longer be required to ‘out’ themselves as having transitioned when proving their identity for a variety of purposes.21 Nevertheless, the Act quite rightly places no requirement on ‘passing’ for gender recognition. Aside from the cruel consequences of making ‘passing’ a requirement for transgender people to be legally recognized in their acquired gender, amending the EA so that the protected category is those perceived as female would also be problematic for female people who identify as female, but who are not regularly perceived as such. Certainly no one thinks that gender-non-conforming females who are regularly mistaken for males should not be counted as ‘women’ for the purposes of the 2010 EA. Aside from these issues with delineating the category, a further important reason exists for thinking it wrongheaded to replace ‘sex’ with ‘perceived sex’ as a protected characteristic, even if we recognize that much of the sex-based discrimination people face is as a result of their perceived rather than their actual sex. This reason is that not all discrimination or harms resulting from a society's devaluing of a given characteristic are due to a person's being perceived as being a member of the devalued group. Elizabeth Barnes and Matthew Andler make the point well with reference to the example of disability. They note that there are people who successfully hide their disabilities to the extent that no one perceives them as disabled. Nevertheless, to the extent that societies continue to devalue disabled people by, for example, failing to make reasonable accommodations for disabilities, someone with a hidden disability will still be harmed by such things as a built environment that fails to take their disability into account. ‘Simply put’, they tell us, ‘the social constraints and enablements of disability go beyond how others treat you or how you are perceived’.22 The same considerations apply in the case of sex. While it is true that many harms suffered by female people under conditions of patriarchy come from direct forms of discrimination that occur when people perceive them as female, society's devaluing of female people has negative effects that go beyond how individuals are treated on the basis of being perceived as female. Caroline Criado Perez offers numerous examples of how the ‘default male’ assumption in science and engineering has led to a built environment, technologies, and medicines that ignore the distinct needs of those with female bodies.23 Given that the devaluing of female people under conditions of patriarchy has contributed to the so-called ‘gender data gap’ and failure to take female bodies into account in numerous contexts, it would seem reasonable for equalities legislation to protect the interests of all female people, not just those regularly perceived as female, to ensure that they are not disadvantaged as a class due to failings to take their bodily differences into account, for example in medicine. It would likewise seem reasonable to require that equality-impact assessments (for example) should consider effects of policy proposals on female people. By making ‘sex’ a protected characteristic, the EA does just this. It also has an elegant solution for the protection of those who face discrimination on grounds of being perceived as female. For all protected characteristics, the EA outlaws ‘discrimination by perception’ that an individual possesses that characteristic. Thus, while it does not create a new legal category of ‘sexually-marked subordinate’ as per Haslanger's ameliorative definition of ‘woman’, it nevertheless does protect members of this class, by protecting against discrimination on grounds of the perception that one is female. So if one agrees with Haslanger that the class of those perceived as female (that is, womenCL) is politically important and in need of protection, then the retention of ‘sex’ as a protected characteristic in UK equality legislation should be welcomed. The presence of ‘sex’ as a protected characteristic in UK equality legislation serves to protect the needs of womenCL in the UK, by protecting against discrimination on grounds of both actual and perceived sex, so that even those members of the class of womenCL who are not legally female can be protected in UK equality law via the protected characteristic of sex. A person has the protected characteristic of gender reassignment if the person is proposing to undergo, is undergoing or has undergone a process (or part of a process) for the purpose of reassigning the person's sex by changing physiological or other attributes of sex.24 As a protected characteristic, individuals are protected against discrimination on grounds of (actual or perceived) gender reassignment. The 2004 GRA makes provisions for some adults to be legally recognized as the gender they identify with provided that they demonstrate to a Gender Recognition Panel that they have a diagnosis of gender dysphoria; that they have lived in their acquired gender for two years prior to their application for a GRC; and that they intend to live in their acquired gender until death. Individuals in receipt of a GRC are issued with an amended birth certificate, reflecting their acquired gender, so that the sex marker on al

One might hope that philosophy could reconcile us to our social world and each other. To entertain this as plausible is to think there is some perspective one could reach via philosophical enquiry that shows our life and society to be as they are for good reason, allows us to see it all as in some sense rational. Hegel is no doubt the great exponent of this ideal, his system promising to trace history's patterns and conceptual development, while he is so optimistic as to believe that, at its end, we would achieve the perspective whereby every agent's own actions and situation can be made intelligible to themselves and others. This was meant to be true for us the readers, so we would be able to see for ourselves how what we do makes sense, given our circumstances, and is plausibly tending towards a good end.1 Of course, the problem is that there may not be such a perspective. Perhaps to see the world aright is to recognize it as a jumbled mess, with no progressive tendency towards greater coherence, and no satisfaction to be had in achieving superior insight. Perhaps there is no good end we are collaboratively working towards, no possible reconciliation with each other; maybe we are perpetually on the brink of descending once more into a Hobbesian nightmare. Hegel hoped to reassure us that the existence of that clarificatory perspective is guaranteed; as free agents, once we achieve self-awareness we necessarily mutually recognize one another as engaged in a fundamentally cooperative project tending towards justified ends.2 But, alas, not all of us have been convinced, and a kind of existential anomie can befall a thoughtful person who surveys our present socio-cultural situation.3 What if there really just is no excuse for how things are, and no good reason for me to carry on? We ought then to make the social world worthy of reconciliation. The guiding idea here is that the ideal of reconciliation underlying Hegelian social thought is desirable, and if it is not yet possible given present social arrangements, we are called upon to change those arrangements until the ideal can be attained. To be clear, this is not a disagreement with Hegel's system at its deepest level; he may have jumped the gun on what a rationally reconcilable social order looks like, but in some sense that is a mere detail compared to his deeper point that we proactively seek a coherence that we can be reconciled to. Social and political philosophy can then play a dual role of identifying points at which our social order will throw up obstacles to attaining a coherent and reconcilable view of one's life, and suggesting means by which these obstacles can be removed.4 I shall illustrate these rather abstract ideas by constructing and analysing a narrative of the historical situation leading up to the current culture war; especially as it plays out concerning race, and black–white relations even more especially, among the middle class of the USA. The US being a culturally dominant global hegemon, the terms and structure of its culture war tend to be exported, so they are worth understanding even for those of us who are not US citizens. The pastime of the chattering classes, the culture war can be understood as a set of symbolic and political conflicts over emotionally highly charged issues du jour. Even if the particular topic of discussion is fleeting, the ultimate resolution to these debates can have drastic effects on the lives of citizens. How we decide to understand and enforce norms around gender and sexuality, for instance, touch upon some of the most intimate and important aspects of our lives. Here I am interested in how our peculiar socio-economic conditions shape the contours and possible points of resolution in the cultural debate around issues of race. We shall see that characteristic responses to our social order, which I shall describe through stylized character archetypes, make it impossible for participants in the culture war to achieve any lasting reconciliation. Instead, our responses both generate and constitute a kind of racialized psychopathology that I describe as white psychodrama. Given this analysis of the social order and its sources of psychic incoherence, I will suggest a way forward. My hope is that this will at least help people of colour caught in the midst of this to work towards a world we can live in and, by seeing ourselves as so working, to reconcile ourselves to our actual present social activities. We cannot, and ought not to, reconcile ourselves to a society wrapped up in its own contradictions, any more than we should seek to integrate into a burning house. But we can come to see ourselves as knowingly and self-consciously working to resolve those contradictions, quenching that fire, and laying the foundations for a better structure wherein we may all live comfortably. Once upon a time the United States of America was a de jure racist society. As was much of the world beyond its borders, controlled as it was by racist European empires. There was a broadly understood and explicitly codified and enforced racial organization of who could live where, how individuals could interact, what sort of jobs were appropriate for whom, how law and order would operate—or wouldn't.5 The science,6 sport,7 and artistic culture8 of the day were largely carried out in conformity with, if not support of, racist norms. Gross or subtle as it may have been in any given instance, the colour line ran through everything and one crossed it only at great personal peril.9 But mountains crumble and rivers disappear, new roads replace the old, stones are buried and vanish in the earth. Time passes and the world changes. So it was that eventually this de jure racist system went the way of all things. The Civil War overthrew the slave regime. Racial immigration laws were repealed. The Civil Rights Act made various sorts of explicitly racist laws and practices impermissible. By the latter half of the twentieth century, it was clear an officially endorsed de jure racial caste system was no longer to hold sway in American life. Likewise abroad, the great European empires fell, and in their place sprang up a plethora of nations governed by formerly colonized peoples. All things considered, the twentieth century saw de jure racism suffer a world-historic defeat. Along with these legal and institutional changes went cultural changes. Casual use of the most highly charged racial slurs became limited to the worst bigots, and nowadays one can hear the sentiment expressed that overt expressions of bigotry ought to disqualify someone from public office.10 Over the twentieth century, Americans steadily reported much less opposition to interracial marriage between blacks and whites.11 Careers which previously operated an absolute colour bar opened up to non-white people.12 Various of black Americans' artistic contributions came to define not just American, but much of the globe's popular culture.13 Mainstream right politicians ceased to explicitly identify as defending white dominance or white interests,14 a norm change which even President Trump made some nominal effort to respect.15 American social attitudes thus seemed to adjust in line with the twentieth century's legal changes. But change was not total. The twentieth century kicked off with an economist lamenting the black middle class's merger access to capital.16 The twenty-first century began the same way, as the gap between total wealth and assets owned by black versus white Americans was once again increasing.17 Black assets were then hit especially hard by the 2008 crash.18 Even setting the crash aside, the fact that what black wealth exists is often tied up in housing property is its own source of racial vulnerability. Black property tends to be worth less19 amid continuing residential segregation.20 This segregation can concentrate social difficulties that further hinder black Americans' life chances.21 And the rarity of interracial contact induced is no doubt related to the persistently low rates of racial intermarriage.22 All of which compounds the fact that inheritance law allows for intergenerational wealth transfers that maintain economic segregation;23 thus intergenerational mobility from being propertyless into wealth is difficult and rare.24 What's more, finally, the backdrop for all this is a global economy where the ability to inherit wealth increasingly determines one's life chances.25 Whatever else changed, the people who have the stuff still tend to be white, and blacks must still sell our labour to them if we are to get by. In this way American domestic politics mirrored the broader global trends of a post-imperial world.26 The European empires despoiled and depopulated nations.27 What they left in their wake were often underdeveloped economies28 and institutional structures ripe to be taken over by local elites who could simply continue the pattern of authoritarian wealth extraction.29 But the end of formal colonization did not generally lead to reparations. For the most part, agents based in the former colonial metropoles retained ownership of key resources and even infrastructure,30 and, if anything, inefficiencies in the credit market have led to a net capital flow from the former colonies to the former colonizers.31 Neither domestically nor internationally did a change in cultural attitudes and legal permissions correlate with a change in racial patterns of ownership. As such, many of the material patterns of inequality from the bad old days of de jure racist regimes have survived the demise of their former ideological superstructure. Returning to the US, these persistent material inequalities have consequences for occupational inequality. The legacy of de jure segregation plausibly goes a long way to explaining wealth and income gaps between black and white Americans.32 And contemporary de facto segregation generates social networks that concentrate access to opportunities for work and education among the already prosperous, further disadvantaging blacks.33 What's more, fulfilling and legal employment for 'less skilled' workers was already drying up by the early twenty-first century.34 In so far as there has been a coherent social response to this collapse in opportunity, what has stepped in to the place of those jobs has been the prison system.35 The rapid rise in the population of incarcerated persons has, of course, disproportionately affected black people,36 and despite mass incarceration being a persistent public concern, the American political establishment has been unable to effectively react.37 Finally, even if explicit appeal to white interests withered away, race continues to be in fact a powerful predictor of how Americans vote.38 The liberalization of social attitudes has not led to the total disappearance of overtly racist stereotypes.39 And the political intelligentsia are still largely white, which has arguably affected the content and focus of their work.40 This, then, is where history has placed us. The maddening ambiguity of our position is what leads to the titular white psychodrama. One cannot reconcile oneself to this society because it constantly pulls in two directions—it presents one with an ideological narrative that speaks of equality, and a material structure that witnesses rank inequality. At some level, this society just does not make sense to itself, its own ideology out of whack with the plain facts of its own existence. There are those who are tempted to focus only on the positives, and see in this a story of triumphant progress towards racial justice or a post-racial future. And there are those who are inclined to see in it a story of eternal recurrence, racism ever reinventing itself. But both of these perspectives are too tidy to capture the phenomenon. For this story is of a world and a nation in contradiction with itself. After much struggle, this world has publicly declared, and in some sense sincerely come to believe, that racial hatred is a social failure and a horrid character defect. We now welcome forms of love, friendship, and cooperation that were once unthinkable. And yet we carefully divide up the pie to ensure former slaves are kept poor and ashamed. The inevitable social discord generated by this immiseration are dealt with by brutality and caging. And the lingering suspicion remains that all this is the former slaves' fault. How then do people respond to the facts relayed in this historical narrative, and what does it mean for their ability to reconcile themselves to their own social order? The character archetypes are stylized representations of typical responses to the status quo. Each of the types below is assumed to be driven by some fairly normal psychological motivations—they do not wish to feel guilty, they would prefer to have more stuff rather than less—and respond accordingly to the evidence, incentives, and institutional structures their society presents them with. To that extent one can think of these as something like publicly available social roles which facilitate intentional action,41 or as agents for whose behaviour I am giving a structural explanation.42 In either case, allowing for the overly neat appearance of any stylized picture, what follows is meant to be a descriptively plausible picture of reactions by many actual politically switched-on agents to life in a society shaped by the circumstances of the historical narrative just relayed. They are caricatures for sure, but ones which I expect many readers will see resemblances to within their own lives. The agents we discuss are highly polarized people fighting a culture war. There is empirical evidence available which lets us situate what sort of person this would be. Partisanship has largely been a phenomenon only among wealthier and more politically engaged voters.43 These are mutually reinforcing categories; home ownership, for instance, predicts being more politically involved.44 Polarization encompasses far more than just party or policy preference.45 It includes, significantly for us, what news media and commentariat figures people listen to and engage with.46 This association with media consumption and lifestyle differences makes the polarization highly affectively charged.47 White people are more likely to be politically engaged48—if anything, black voter turnout is systematically over-reported.49 And, as already discussed, whites are also more likely to be wealthy. So all this means that the culture war categories I focus on will primarily be elite agents, and will largely (though not entirely) focus on various white responses to the status quo. Despite culture war polarization largely being a fight between whites, issues of race still turn out to be very important to how it plays out. Measures of racist attitudes still do a good job of predicting Americans' attitudes to candidates and policies.50 Since the 2008 election of Barack Obama, measures of white Americans' level of racial antagonism have spiked.51 And voters report racial issues as some of those on which they are most divided.52 Plausibly, at least some of the cultural divides now wracking this section of elite white America arose from differing responses to the changing social meaning of whiteness following the end of de jure racist regimes.53 So I shall rationally construct character types which allow one to appreciate how the sort of elite white agent engaged in a culture war revolving around race might behave and ideologically understand themselves.54 These responses may not always strike you as plausible or fully coherent. But I think that is to be expected in a situation where people are having to make sense of a society shaped by contradictory forces.55 That, after all, is why this is a story of psychodrama. The conflict between an ideology of racial egalitarianism and a material reality of strict hierarchy generates such a tension, and it is the desire to keep one's status atop that hierarchy (or make one's way within it) while avoiding guilt which thereby drives the culture war archetypes. One common type among the white majority may be called the Repenter. This character straightforwardly responds to our story with an overwhelming sense of guilt. They see the group they identify with as having committed horrible crimes globally and domestically, and they are ever so aware of the ways in which present material conditions generate continued deprivation for black people alongside relative comfort for many white people. They believe in racially egalitarian ideologies which tell them this ought not to be, and hence feel unhappy with themselves and the state of things. The Repenter would deeply, deeply, like to do things to alleviate this guilt. And their form of repentance involves trying to change their interpersonal habits and consumer choices so as to minimize their contribution to the broader social issue, and help the particular black people they interact with. In this way, by doing that sort of self-work, they hope to be able to live in a world that is admittedly unjust while making it that little bit better, and through such efforts be able to honestly maintain a positive self-image. So the Repenter invests in lessons on sophisticated etiquette around interracial interactions.56 They are keen for their workplace to celebrate diversity, and satisfying their desire for both a tolerable workspace and various consumer goods drives what commentators call 'woke capitalism'.57 They seek out and do their best to appreciate the work of non-white thinkers and artists.58 Sometimes all this can be rather ostentatious, and one rather suspects the moral kudos for being seen to do as much is playing rather too large a role in their motivations. But we need not be so cynical; these may well be the result of a heartfelt sense that such behaviour is required by justice, given the genuine guilt which their group membership and historical situation has placed upon them. Another common type among the white majority may be called the Represser. Much like the Repenter, they are aware of our broad historical situation, and not immune to pangs of guilt. But the Represser feels this is largely unreasonable—too much focus on the negatives rather than the positives, and prone to generating an irrational self-hatred in themselves and among their fellow whites. The Represser does not deny the central points of the earlier narrative, but they would like to change the focus away from the potentially guilt-inducing elements. Progress has been made and there are, after all, good things that have happened in our history too. They do not wish to return to de jure racism, but instead hold out hope for a society that is, in some sense, considerably more relaxed about race, and less prone to fixate on the historic crimes of white people. Such a society could do better at learning from the past, by considering it with an air of cool detachment. So far as we can now draw on the past's lessons, salient to the Represser is that the success cases for progress seem to be ones wherein race was previously central to how we organized things (say, immigration law) but is now de-emphasized or suppressed. So the Represser seeks out intellectual support for the idea that our narrative is altogether too cynical.59 They seek to implement colour-blind policies and norms that avoid drawing attention to race. And through their consumer choices they try to encourage media that embodies race-neutral virtues. In a sort of mirror image to the Repenter, one can sometimes get the impression that the wish to avoid discussing race is a second-best alternative for a person who simply dislikes guilt and wishes to remain ignorant of anything that may induce it,60 or, worse, harbours anti-black animus, but lacks the courage to explicitly challenge anti-racist social mores. But, again, it need not be, and the Represser strategy can be the result of a heartfelt conviction that the path to just unity goes via de-emphasizing differences and stressing the encouraging progress made in learning to get along. Before moving on to our third type, let's be clear about a few things. First, the Repressers and Repenters are diametrically opposed. The Repenter's response to our situation crucially depends on playing up exactly the things the Represser wants to play down. They are exactly opposed strategies in responding to a Du Boisian social problem. These are cases wherein one finds 'the failure of an organised social group to realise its group ideals, through the inability to adapt certain desired lines of action to conditions of life'.61 Repenters and Repressers alike are responding to their failure to realize racially egalitarian group ideals, but one wants to adopt lines of behaviour that acknowledge this failure and alleviate guilt and the other wishes to suppress discussion of the failure. It is near enough psychologically impossible for the same person simultaneously do both, and it is thus difficult for Repressors and Repenters to share spaces and resources. Second, the above two types by no means exhaust even elite majority group responses to the situation outlined in the narrative. For one thing, most people are simply checked out of politics. For another, (for example) explicit white supremacists would fit into neither group. But among the chattering classes of the whites, I believe the Represser and Repenter character archetypes capture many of the most important personas. Third, and finally, neither's strategy for responding to the narrative involves large-scale changes in the distribution of ownership of stuff, including housing and living arrangements. Repenters and Repressers are engaged in a fundamental conflict, but it is a conflict over how to psychologically manage the results of living in a materially deeply unequal society, not a conflict about how or whether to reduce that material inequality. Of course, the rest of us do not simply sit by and watch the whites duke it out among themselves. If nothing else, they still have ownership of the stuff and a democratic majority, so most of us are dependent on them for making a living. How, then, have the PoC intelligentsia—people of colour sufficiently engaged in politics to be tapped into the white culture war and the historical narrative underpinning it—responded to the opportunities and challenges presented thereby? With a dextrous entrepreneurial spirit! Which is to say, by cashing in. In institutions like academia, more dominated by the Repenter type, there has been the opportunity for mediocre if sharp-eyed young PoC intelligentsia to present themselves as bearers of black thinkers' insight.62 It is considerably harder to pull this off from within academia as an advocate for Represser views. But where there is demand, there will be supply. And there is a large audience keen for a black thinker to give voice to an intelligent version of the Represser narrative. Sufficiently talented black thinkers have been happy to oblige.63 Various media organizations and political groups likewise provide opportunity for similar pseudo-spokespeople for PoC intelligentsia catering to both Repressers and Repenters. The white majority's superior wealth easily translates into control over cultural institutions. They are the bulk of the consumers of newspapers and intelligentsia magazines, their children the largest racial demographic in elite colleges; they write the cheques that form the basis of the endowments, and tend, in fact, to be the ones sitting in positions of upper authority in newsrooms and university administration. And among this white majority there is demand for a PoC intelligentsia. It is a vital part of Repenter strategy for alleviating guilt that they listen to such voices. Repressers are not as tied to this strategy, but if one is troubled by accusations of racism (and these may come from within), the fact that black thinkers agree with your perspective will be salient and interesting. Those PoC intelligentsia who are able to catch and retain the attention of one of the white bourgeois factions thus gain access to their resources and the institutions they control for promoting their own view. They thereby attain a kind of leadership status which is not related to having the skills required for such a job,64 nor popular legitimacy with the people they purportedly speak for.65 A PoC intelligentsia thus emerges, without ownership of its cultural situation,66 and catering to white tastes while giving the impression of representing PoC perspectives. 'The actual work is for you to deconstruct the things within you: whiteness,' Rao said. 'Whiteness harms people of color, but worry about yourself. Stop worrying about us—that's paternalistic, too.' Bond echoed this sentiment. 'This idea that we, as white people, need to go out and make these big external actions—that's just white supremacy,' she said. 'This internal work is the hard work; it's the work that never ends.' They see the deconstruction of whiteness as a prerequisite for true anti-racist work. Jackson explained, 'Until you deal with your inner stuff—until you can say, "I'm coming from a place where I recognize that I have these thoughts, and I'm working on it"—everything else is performing.'67 And, more generally, there has grown up a small but profitable industry of racial self-help gurus, who extract money from Repenters in exchange for helping them alleviate guilt without challenging the fundamental material inequalities underlying the situation.68 Of course, it should be clear that the PoC intelligentsia are, to a certain extent, simply playing the hand life dealt them. As per our narrative, they do not tend be independently wealthy or come from family backgrounds that permit self-support. For the most part, they occupy institutions in which they are the minority, and the taste-makers and the persons with the ability to direct resources of those institutions will largely be drawn from the white bourgeoisie's chattering class, which is to say will very often be Repenters or Repressers. They have to eat, and which of them we see and hear from is filtered by the tastes of the white majority. If their behaviour seems disappointing, then one ought to consider the circumstances which birthed them rather than engaging in moralistic personal critique. The sad fact is that once patterns of racial interaction become entrenched, they become very hard for individuals to deviate from.69 We need a kind of structural change, one that changes the causes of behaviour embedded in the social environment we are all responding to, generating mass rather than individual behavioural changes.70 To gain a more concrete sense of how these archetypes present themselves in everyday life, let us imagine a hypothetical culture war racial flashpoint. Considering how the character archetypes thus sketched might respond to it will be instructive. So let us suppose that the nation is responding to yet another extra-judicial killing of a black man by a white police officer. What sort of behaviours should we expect from the character types given what we have said about them? The Repenter is driven overwhelmingly by guilt, a desire that their personal behaviour and environment not be associated with racism or racial social ills, but has no great desire to change the material basis of the society in which they live. A brutal reminder of black vulnerability to death at the hands of the state will thus first and foremost activate that guilt in a visceral way, leading to heartfelt distress. They will wish to signal their extreme disapproval of police behaviour and induce organizations they are associated with to do the same—for instance, perhaps their workplace could be made to issue a statement affirming that black lives matter, or some such, while committing to uphold anti-racist practices in its own behaviour. They will look to prominent PoC intelligentsia for intellectual leadership on the question of what should be done at a social level, but since they are largely drawn from society's upper echelons, they are liable to be uncomfortable at practical actions deemed extreme—that is, which could actually upset the ability of the police to perform the social function of protecting their lives and property. Hence, beyond sloganeering and support for piecemeal reform (supporting candidates who combine radical slogans with piecemeal reform policy will be especially tempting), Repenters are unlikely to engage in any sustained push for large-scale change. Guilt is assuaged so far as possible by their personal actions towards the good and the disassociation of their organizations from racist animus. They listened to PoC intelligentsia voices and made some efforts towards acting upon what they were told. When the next police killing causes outrage, the cycle may be repeated. The Represser is driven by a desire to avoid unnecessary guilt-inducing focus on race, to ensure there is recognition of progress made so we do not lose what is valuable, and to avoid any suggestion that they are personally responsible (or responsible qua white person) for the ills of black people they have never interacted with. Any framing of the police killing as racist, or which focuses on racial disparities in treatment by the police, is hence liable to annoy. If the Represser can be satisfied that there is no proof that specifically racial animus motivated the police officer, they will likely bemoan the public rushing to judgement on this point, and offer alternative (race-neutral) explanations for the police officer's behaviour. They will decry any comparison between present events and past explicitly racist atrocities such as lynchings, and highlight the absurdity of suggesting that contemporary police officers still behave anything like their equivalents in the bad old days of de jure racism.

There are no potential conflicts of interest relevant to this article. All relevant data are included in the article.

There are no potential conflicts of interest relevant to this article. All relevant data are included in the article.

Self-employment, which accounts for one in seven workers in OECD countries, has ceased to decline in recent decades and is now growing in these countries.1 It comprises shop owners, restaurateurs, consultants, lawyers, and an increasing number of unskilled workers in industries like construction and transportation, in which self-employment was unusual in the past.2 Yet these workers, the vast majority of whom do not employ others, have meagre access to social protection. In some countries, the self-employed are legally excluded from unemployment, sickness, and occupational injury benefits—the three areas of social protection that, retirement pensions aside, self-employed workers value most.3 And in other countries they become de facto excluded because, despite having statutory access to these benefits, the eligibility requirements to accrue them are tailored to waged work.4 Unprotected exposure to social risks makes self-employed workers, on average, three times more likely to become income-poor than their salaried peers.5 Over recent years, the working conditions and social protection of the self-employed have gained ground in public policy debates. A growing consensus exists among practitioners and policy-makers on the importance of strengthening the social protection of the self-employed in four ways. One is to offer them unemployment benefits in order to protect them from poverty in the event of bankruptcy. Another is to secure them access to insurance for occupational diseases and injuries. The third is to grant them adequate protection from day one in case of sickness. The fourth and final way is to give them access to benefits and measures that allow for a better work–life balance, including maternity, paternity, and caregiver leave. Moreover, as part of the implementation of the European Pillar of Social Rights, the European Commission has recently put forward a proposal that encourages EU member states to offer adequate social coverage to self-employed workers.6 The debate over these measures, however, has lacked normative input thus far. Even though self-employment was a central concern for pre-industrial thinkers like James Harrington, Adam Smith, and Thomas Paine, who conceived of it as an alternative to the ‘servile dependency’ that waged work involved, recent philosophers have largely overlooked it.7 The burgeoning philosophical analyses of work have paid scant attention to the nature and value of self-employment, nor to the normative significance that the distinction between self-employment and employee work may have for the legal and social protection that self-employed and waged workers are due. This neglect may be explained by the tendency in recent philosophy to embrace a corporatist paradigm of working relations—a paradigm that equates work with employment, usually in large firms, and neglects self-employment and other non-standard forms of work. This is clearly true of philosophers working in the Marxist tradition, who conceive of workers as a class of waged earners compelled to sell their labour force to capitalists.8 But it is also true of neo-republican philosophers, who typically focus on the domination to which workers may be subject under the managerial authority of their employers,9 and of liberal egalitarians, who have recently centred their attention on how subordination to an employer may be compatible with the basic liberties of employees.10 The corporatist paradigm has two serious limitations. First, it sits uneasily with the reality of existing labour markets, in which standard employment coexists with a growing myriad of non-standard forms of work, including self-employed, gig, and multi-party work. Second, it is blind to pressing moral questions that fall outside the scope of standard employment relationships, such as the moral obligations that firms might have towards independent contractors, the self-exploitation of mom-and-pop owners that may result from market competition, the risks that algorithmic management may impose on gig workers, and the protection that informal self-employed workers in poor economies may deserve. These problems affect a substantial portion of the global labour force, and they are not bound to fade away. It is time for the corporatist paradigm, then, to give way to a non-wage-centred view of work: one that appropriately accounts for self-employment and other forms of non-standard work. This article takes some first steps in this direction by focusing on the moral significance of self-employment. It seeks to explore the extent to which self-employed work may offer adequate opportunities to realize goods traditionally associated with waged work. Identifying such goods has become central to the debates over what makes (waged) work decent or meaningful, and worthy of protection as a result—with scholars and policy-makers offering a diversity of views of the pecuniary and non-pecuniary goods that work may deliver. For example, whereas economists tend to focus on earnings and benefits,11 sociologists and philosophers often centre on the non-economic dimensions of work, including autonomy, self-realization, and socialization.12 Such goods operate, in these analyses, as normative standards for assessing different work arrangements. Inspecting whether self-employed work may also realize them is critical, then, to assessing whether self-employment merits protection—without prejudice to the independent reasons we may have to protect it. If, for example, self-employment happens to deliver such goods, then we may have pro tanto reasons, whose stringency will depend on how successfully self-employment delivers them relative to employee work, to improve access to protection for the self-employed, making it easier for those with a preference for self-employment to go solo as a result. The article proceeds as follows. After offering a two-dimensional account of self-employment in Section II, the remainder of the article examines the link between self-employment and three of the core goods of work: autonomy, self-realization, and self-provision. Sections III and IV inspect the non-pecuniary goods of autonomy and self-realization, respectively. Work may no doubt provide a wider range of non-pecuniary goods—for example, it may often improve health, marriage prospects, and family stability compared to joblessness. Yet the reason for focusing on autonomy and self-realization is not just that these are among the core goods of work, as their salience in philosophical and sociological analyses suggests. It is also that they encompass other narrower goods—such as self-direction and self-sufficiency in the case of autonomy, and social contribution and recognition in the case of self-realization—which loom large, given their significance, in recent scholarly and political debates on the value of work. Section V turns to a pecuniary good: the economic capacity to fend for oneself and one's family—or self-provision, as I label it—in an attempt to capture what is normatively significant about earnings. The conclusions I draw, albeit tentative, favour improving the social protection of the self-employed. I first argue that self-employed workers are both less and more autonomous, as it were, than those who work for an employer, as self-employment may sometimes lessen the off-duty autonomy that people acquire through work. Yet it significantly increases workers’ autonomy within work. And I also argue that self-employment, although unnecessary for attaining self-realization and self-provision, is distinctively useful for securing each of these goods, under plausible, far from outlandish, conditions. Before proceeding, two caveats are in order. First, the arguments I develop here do not entail that self-employment should be promoted, in the sense that workers should be conditioned or induced to develop a preference for self-employed work.13 They rather seek to show that, under actual circumstances, yet also under an wide range of neighbouring ones, some individuals have reasons to engage in independent work, and that this decision deserves adequate social protection, so that those with a preference for self-employment have suitable opportunities to realize this preference without incurring unbearable risks. Second, although the arguments I develop apply both to the self-employed who work on their own and to those who employ others, it is worth teasing these two categories apart. For one thing, improving the social protection of solo self-employed workers (that is, those who have no employees) has greater urgency. This is not only because solo self-employed workers account for more than three quarters of the self-employed,14 it is also, and chiefly, because they are in a more precarious situation, in terms of earnings, financial security, and job stability, than self-employed employers. For another thing, social contributions, which are the counterpart of social benefits, tend to be higher in the case of self-employed employers, who are usually among the high-earning self-employed.15 Whom we should prioritize among the self-employed, and how we should fund their benefits, are questions to be addressed separately, once we have answered the question of whether the social protection of the self-employed merits improvement. The concept of self-employment is not neat. The definitions we find in our legal systems lack the consistency of those of waged or employee work. Take, for example, the 1993 International Classification of Status in Employment (ICSE-93), which the International Labour Organization has long endorsed. According to this classification, which is the standard that national statistical systems still most widely use, the self-employed are those workers who are not compensated with a salary. They instead work for profit, in the sense that their income directly depends on the revenue that their productive activity may generate, being more exposed to market risks than waged workers as a result. Compare, for example, the self-employed owner of a beer garden, who will lose her clientele and make no money on a rainy day, with her employed waiters, who will take home a day's wage despite having spent the day sitting on their hands. Yet an exclusive focus on the for-profit dimension of work, which the ICSE-93 exemplifies by limiting the self-employed to the residual group of workers who do not earn a wage, is insufficiently discriminating. It misclassifies a number of workers who, although working for profit, have much in common with wage labourers. One example of this misclassification is that of dependent contractors, who provide goods and services for a single client with whom they have a commercial agreement. Another example is that of contributing family members who work without a contract in family businesses run by the head of the family. Like employees, these workers are under the authority of someone else—the single client or the head of the family—who restricts their autonomy in many aspects of their work. A more nuanced definition of self-employment, which the International Labor Organization's (ILO) 2018 update of the International Classification of Status in Employment (ICSE-18) incorporates, considers a second dimension as well—namely, that of independence. According to the updated definition, a major difference between self-employed and employed workers is that the self-employed, unlike employees, are not subordinated to the authority of a boss. Self-employed workers are free to decide basic aspects of their work, such as when, how, and with whom to work, whereas employees are not. This is a definition that is increasingly used in various jurisdictions to reclassify bogus self-employed workers as employees, and to extend certain employee rights, like sick pay and paid holidays, to dependent contractors on the grounds that, despite their being legally classified as self-employed, they lack autonomy in how to run their businesses, as well as the authority to hire and dismiss staff.16 Considering these two dimensions—profit and independence—we can define genuine self-employed workers as those who earn their income by engaging in independent productive activity at their own risk.17 This allows us to distinguish the self-employed both from standard wage workers, who engage in dependent work for a salary, and from intermediate categories that sit uneasily between self-employment and waged work. These categories do not only include the above-mentioned cases of dependent contractors and contributing family members. They also include employers who, despite earning a salary from their own business, are independent enough to have a work experience that is closer to that of their self-employed peers. And they also include members of worker cooperatives who, despite having collective ownership and control of their business, are subordinated to the authority of the management of the cooperative, and enjoy less individual control over their working conditions than self-employed workers do.18 As an initial exploration of the nature and value of self-employment, this article sets aside the complexities that borderline cases, such as those of dependent and bogus self-employment, may pose. It rather focuses on whether, and on what grounds, the decision to engage in genuine self-employment deserves better protection than it currently has. In order to address this issue, the next three sections examine the value of self-employment as a vehicle for attaining three core goods of work: those of autonomy, self-realization, and self-provision. As noted above, these goods may not cover the full gamut of goods that can make work valuable. Yet it is hardly controversial that their absence entails serious work deficiencies, such that if, for example, self-employment happens to poorly secure them, it then becomes inimical to good work—and the reasons to protect it accordingly diminish in stringency. Analyses of autonomy at work tend to focus on the extent to which workers exercise self-direction or control over their working conditions, including their schedule, tasks, and workload. Karl Marx famously diagnosed the alienation of the proletariat on the grounds that the capitalist division of labour dispossesses workers of control over production.19 And Max Weber warned of the impact that the bureaucratization of economic activities and organizations could have on workers’ control over their jobs.20 These analyses home in on workers’ degree of autonomy while at work. But the significance of autonomy at work is also prior and subsequent to work performance. It is prior because the decision to take one particular job instead of another can be more or less autonomous. And it is subsequent because one of the rewards of work is the off-hours autonomy that we may acquire through it. This section inspects the extent to which self-employment may foster or stifle these three dimensions: (a) autonomy within work, (b) autonomy in one's choice of work, and (c) autonomy through work.21 As we will see, the self-employed are more autonomous than their salaried peers in some, yet not all, of these dimensions. Autonomy within work is the main motivation for engaging in independent work, which most self-employed workers cite.22 Self-employment grants self-direction: it allows workers to decide which work to perform, when to perform it, and under which conditions. This discretion sharply contrasts with the diminution of autonomy that waged work, by its very nature, entails.23 In entering an employment relationship, workers agree to subordinate themselves, while on duty, to their employer's authority. The specific impact of such an agreement on employees’ autonomy no doubt depends on their bargaining power, the peculiarities of the employment contract, and the way in which managers exert their authority. An engineer with abundant job alternatives, no need to clock in, and a hands-off boss surely enjoys more autonomy on the job than an unskilled courier with few outside options, tight delivery times, and an intrusive manager. Yet it is also useful to distinguish between ideal and non-ideal scenarios in this respect. In non-ideal circumstances, in which labour law is defective or defectively enforced, bosses’ authority can significantly undermine individual autonomy. Workers aplenty undergo capricious rescheduling, verbal abuse, sexual harassment, and other forms of abusive management.24 These practices harm them in various ways, one of which is the curtailment of their autonomy within work. And these practices do so not just to the extent that workers actually suffer them. The mere possibility of being abused for what managers may see as a failure can also undermine employees’ self-direction as to basic decisions about, for example, when to undertake pending tasks or how to address customers. Self-employment is valuable and worth being better protected and promoted, then, because it shields workers from abusive management by removing its cause: subordination to managerial authority.25 It is hardly surprising, then, that self-employment increases in jurisdictions where labour regulation is loose and employers can more easily abuse their staff.26 But one could argue that the superiority of self-employment disappears, in this respect, when legal protections effectively protect workers from abusive authority.27 So it is important to see how in ideal scenarios, in which such protections are in place and are complied with by employers, self-employment may still foster autonomy within work, in the following two ways. First, under existing employment law, employers retain the prerogative to direct workers in all the aspects that remain unspecified in the employment contract and by labour legislation—whose content is inevitably incomplete, and desirably so for efficiency reasons.28 Such residual authority permits them to make final decisions on shifts, promotions, overtime, and so on.29 In ideal circumstances, employers would not use such authority to purposely trespass on employees’ interests by, for example, reallocating them against their will or assigning them overtime on their birthday. Yet workers’ autonomy may suffer from the mere fact that, whenever they disagree about what to do, the employer's judgement will always prevail over their own. Think of a mechanic whose boss tells her to fix a customer's car in a particular way. If she disagrees with her boss about what the best solution for the car is, she will be compelled to act against her judgement. And if, in contrast, she happens to agree, then she will not be able to revise her judgement without great cost, including the risk of being dismissed. Second, although some wage workers may enjoy substantial autonomy within their work—for example, they may be able to decide when they work and whether to do so remotely or on site—there are certain economic liberties which the self-employed have which they do not have (or have only to a lesser degree). Such liberties include rights to negotiate prices with customers, to self-brand, to secure autonomous capital investment, as well as other liberties required for operating in the market as a freelancer or independent producer.30 Employers can no doubt decide to bestow such liberties upon their employees. For example, the owner of a tailor shop can let her employed tailors design suits and negotiate their price with customers. When this occurs, employees enjoy greater option freedom, in the sense that they gain access to a wider range of options than otherwise.31 But such freedom is subject to the will of the employer, who can always restrict, with no reason offered, some such options or rescind them altogether. The employer, by contrast, enjoys agency freedom, in the sense that her economic liberties are legally protected from others’ interference.32 In sum, only self-employment ensures that economic freedoms are enjoyed in a modally robust way. An important objection is that self-employed workers, despite not being subject to managerial authority, may see their autonomy limited by their dependence on suppliers, creditors, and customers, whose wills they have to satisfy to keep their businesses going. The economic liberties of a self-employed tailor are meagre, so the objection goes, if market competition compels her to take orders she dislikes or put in more hours than she would rather work.33 There is a kernel of truth to this objection. Yet it also misses two relevant considerations. The first is that we should distinguish personal dependence, as it exists between employees and their employers, from dependence on the aggregate, as it exists in the market. The self-employed rely on many, sometimes hundreds or thousands of anonymous customers, yet not on any particular customer.34 Although this dependency may restrict their autonomy in various ways, as the objection rightly points out, it does not subject them to the kind of domination that the employment relationship often involves. The critic could insist, however, that the self-employed are sometimes subject to personal dependencies, such as when they rely on a single supplier or purchaser with market power.35 Here is where the second consideration comes in—namely, that the kind of personal dependencies to which self-employed workers may be subject emerge from horizontal market relationships, which are qualitatively different from vertical employment relationships.36 To illustrate, suppose that the self-employed tailor has no choice but to rely on a single supplier who sets abusive prices. The crucial point is that, although she may end up having to work longer hours to pay the supplier, the supplier cannot direct her as to how to tailor ballgowns, or when to clock on and off. Horizontal dependence is surely not innocuous. Yet it does not entail submission to the command and supervision of an alien authority.37 After a long working day, the self-employed tailor goes home having to report to no one but herself. Let us now turn to autonomy in one's choice of work, which importantly differs from autonomy within work. Workers may, as a result of an autonomous choice, take jobs that involve little autonomy within their work. And workers may also take jobs that involve more autonomy within their work than they would rather have, as when a journalist who wishes to be hired by a newspaper is bound, because there are no openings in media outlets, to freelance. In sum, autonomy within work and in one's choice of work, albeit often related, are distinct. How autonomous one's choice of work is depends on two conditions. The first is objective, and requires that occupational freedom be secured. Occupational freedom entails the capacity to choose aspects of work that are central, such as with whom, by which means, and to what ends one works. Sabine Tsuruda has recently argued that managerial authority, when bestowed solely on employers, is incompatible with occupational freedom.38 It undermines its epistemic conditions, she argues, because employees cannot make fully informed choices about jobs whose conditions the employer can alter in unpredictable ways. If correct, this reasoning provides a rather straightforward argument that self-employment opportunities should be available without having to incur unbearable risks. But it also seems to exclude the right to submit oneself to the managerial authority of an employer as an instance of occupational freedom.39 Tsuruda's argument proves too much, given that, by the same token, our freedom to marry would be undermined by the decision to marry someone with greater power to change the terms of the marital relationship as the marriage unfolds—because, for example, she is better educated and more affluent. But this is implausible, for such a decision can perfectly well result from an autonomous assessment. And the same is true of the decision to work for a hierarchical firm. Indeed, some workers may rather avoid the responsibilities that working for a more horizontal firm—or running their own business, for that matter—may involve. What matters in order to preserve people's freedom, in choosing their spouse as much as their occupation, is that ample opportunities to decide the kind of relationship they want to get into are on offer. The opportunity to avoid putting oneself under an employer's thumb is special in this respect—not only because of the effects that, as discussed above, employers’ residual authority can have on workers’ lives, but also because, as Kolodny observes, in modern societies hierarchical relations of authority, barring those between the state and its subjects, are exceptional.40 Given that work is one of the few places in which people are subject to other people's will, we should be particularly wary of the conditions under which such subjection occurs. Self-employment is valuable and worth protecting and promoting, then, because it contributes to ensuring that the choice of work of those who wish to avoid managerial authority is autonomous, and that those who submit to such authority do not do so because they have no alternative. The second condition that autonomy in one's choice of work requires is subjective—namely, that work be chosen because, or partly because, its end product is in line with workers’ personal goals. A job that is simply chosen in order to escape poverty, or to fulfil others’ expectations, is heteronomous.41 As a matter of fact, it must be noted, the decision to become self-employed often fails to meet this condition, as many become so out of mere necessity. This is true of developed economies, in which a decent portion of the self-employed—one in five in the EU—would rather work for an employer.42 Yet it is particularly true of developing economies, where self-employment, which accounts for nearly half of the workforce, mostly results from job scarcity.43 None of this entails, however, that in such cases self-employment undermines the subjective condition of autonomy in one's choice of work. First, what these empirical observations show is merely that, absent good job opportunities, distinguishing when self-employment is autonomously chosen and when it is not—when workers are pulled or pushed into self-employment, as organizational psychologists and labour economists often put it—becomes harder. Second, even when good jobs are available, self-employment expands one's opportunities to tailor one's work to fit one's goals. It allows a baker to produce patisserie, and a freelance journalist to write articles, free from the constraining direction of a boss. We will come back to this point in the next section. For now, it suffices to note that, when opportunities for self-employment are scarce—for example, because the social protection of the self-employed is meagre—it is harder for some workers to align their work with their personal goals. Finally, consider how working solo, rather than for someone else, may impinge on workers’ autonomy outside work. The greater flexibility that self-employment offers no doubt improves such autonomy in one sense, as it enables workers to adjust their working schedules to other activities and responsibilities, such as taking care of their children, walking their dog, or taking yoga classes. Many women, for example, cite the possibility of combining work with their care duties, thus avoiding the parent-unfriendly schedules that wage jobs often involve, as an important reason for becoming self-employed.44 But self-employment also poses various challenges to the autonomy that workers enjoy after they clock off—starting with the burdens that running a business, unlike working for one, involves. When it comes to bookkeeping or meeting administrative requirements, such burdens can be significant. And although the self-employed can mitigate some of them with professional assistance, those who cannot afford such services are likely to spend numerous evenings agonizing about their taxes and books, with significant costs not just of time, but also of mental energy. Second, there is the individualized responsibility that the self-employed—who risk not just their income, but also their capital investment in running a business—bear for the success of their economic activity. In order to keep their business afloat, the self-employed end up taking on more undesired work, and thus enjoying less leisure time, than wage workers do.45 And the pressure they endure, even when off-duty, to find new ways of branding their business, to offer new products, and to attract new clients often makes them constantly think about work, rendering their leisure of poorer quality. Finally, there is the financial insecurity that results from low earnings combined with higher exposure to market risks. More than 70 per cent of the self-employed, for example, earn less than the median employee.46 Under such conditions, sudden illnesses, quiet spells of work, and bad market luck can lead to serious income dips, even to bankruptcy. Further, when self-employed workers lose their shirt, they not only lose the means to realize their aims, they also tend to rely on family, friends, and other informal safety nets, which may create dependencies that may further harm their autonomy. One reason why people enjoy working, when they do so, is that it brings self-realization, which I shall now analyse. Before inspecting the ways in which self-employment may be conducive to self-realization, however, let me briefly consider why work, unlike other neighbouring activities, offers a privileged context for accessing it. An activity is a potential source of self-realization if it has two central features.47 First, it involves the exercise and development of valuable skills and capacities that people choose to develop. Second, its execution lends itself to positive appraisal by others. Thus, private activities that involve the development

Journal of Political PhilosophyVolume 31, Issue 2 p. 158-176 Original ArticleOpen Access A merely national 'universal' basic income and global justice Martin Sticker, Corresponding Author Martin Sticker [email protected] Department of Philosophy, Cotham House, University of Bristol, Bristol, UKSearch for more papers by this author Martin Sticker, Corresponding Author Martin Sticker [email protected] Department of Philosophy, Cotham House, University of Bristol, Bristol, UKSearch for more papers by this author First published: 15 March 2023 https://doi.org/10.1111/jopp.12289AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL In this article, I raise a justice problem for a universal basic income (UBI) if implemented in one or a few nations, but not globally. I raise this objection from the perspective of someone who is ultimately sympathetic to a UBI, even a merely national one. My argument specifically problematizes the unconditionality of a UBI and the unprecedented benefits for those who receive it. UBI advocates themselves often draw attention to this unconditionality in order to justify the greater moral appeal of a UBI over other schemes. Thus, my argument has particular relevance for those sympathetic to a UBI. I argue that, in response to the justice problem I raise, a merely national UBI should be accompanied by compensation for non-UBI-recipients who contribute to what I will call the UBI's 'effectiveness'. In Section I, I explain what I mean by a merely national UBI, and in Section II, I develop a justice or exploitation problem that a merely national UBI faces. In Section III, I argue for a compensation scheme for non-recipient contributors. In Section IV, I clarify the scope of my argument, and in Section V, I discuss how my argument relates to a potential global UBI. I UNIVERSAL BASIC INCOME A UBI is income paid irrespective of willingness to work, of wealth, of other sources of income, and of personal circumstances. If a UBI is pitched at a sufficiently high level, it realizes a number of goods for its recipients: it frees them from the most basic forms of economic insecurity without conditions such as a work requirement,1 affords so-called 'real freedom', the freedom to choose the way recipients want to live,2 and allows recipients to exit unattractive jobs and even the job market altogether.3 If a UBI realizes these goods for a recipient, I will say that the UBI is effective. In the next section, I will argue that it is unjust if a merely national UBI is made more effective by the exploitative labour of those who do not receive a UBI and thus cannot opt out of their unattractive jobs, lack the freedom to live as they want, and are subject to basic economic insecurity. A UBI is currently often presented as a radical yet feasible solution to pressing political and social problems, such as poverty,4 the demeaning nature of conditional welfare schemes,5 costly and overly bureaucratic welfare-state regulations,6 the drive for ever more growth and the resulting environmental destruction,7 the fact that much necessary care work is currently not financially rewarded,8 and that workers are increasingly being replaced by robots and computers.9 For my purpose, two things are especially significant. Firstly, a UBI is frequently presented as a non-utopian proposal that, from an economic perspective, could be implemented in some countries now or in the near future.10 Secondly, a global UBI, by contrast, would require major political, economic, and social changes worldwide. This is currently rather utopian, though perhaps 'no longer a mere pipe dream'.11 A UBI, insofar as it is intended as a solution to some of the most pressing issues of our time and one that is financially and politically feasible now or in the near future, will be one implemented in only a few and, most likely, relatively wealthy countries.12 This raises the question of how a merely national UBI would be impacted by and, in turn, affect the current unequal und unjust global order. II THE GLOBAL JUSTICE PROBLEM OF A MERELY NATIONAL UBI The problem I raise in this section is a more pressing version of the already well-known exploitation or free-rider objection to UBI. My problem is concerned with the exploitation specifically of non-nationals or non-residents, who contribute to but do not benefit from a merely national UBI. The literature pays surprisingly little attention to injustices between different nations that a merely national UBI might amplify. In fact, discussions of exploitation and unfairness within the UBI literature take place almost exclusively on an intra-national level.13 The original exploitation objection is inspired by Rawls's claim that Malibu surfers are not entitled to public funds.14 The (in)famous surfer is someone who could contribute to society either via paid work or in other socially beneficial ways, but, instead, spends all his time surfing. While the surfer potentially poses a problem for all government schemes founded on cooperation between citizens, the objection is particularly pressing for a UBI, as here benefits are explicitly not conditional upon willingness to cooperate, for example by seeking paid employment. UBI advocates see this unconditionality as a feature, not a bug.15 Nonetheless, in the case of the surfer, a UBI seems to clash with 'a widely accepted notion of justice: it is unfair for able bodied people to live off the labour of others'.16 This objection is sometimes phrased in terms of unfairness, sometimes in terms of exploitation or parasitism, since the surfer lives off the labour of others who (supposedly) do not receive adequate compensation for the services they perform. I take it that the underlying intuition is that it is unjust if one person's lifestyle is subsidized by the labour of other persons and there is no reciprocity in this subsidy, even though the first person could reciprocate. There is a straightforward response to the surfer problem. Take Lazy (the surfer) and Crazy, who works hard and whose income and other taxes help fund a UBI that allows Lazy to spend all of his time surfing. Crazy cannot reasonably complain about this, since, if he finds the lifestyle of Lazy more appealing than his own, he could give up his job and become a surfer himself. After all, an effective UBI affords him the real freedom to live this way. If Crazy does not give up his job, then this is because he chooses the life of work and higher income—in which case both Lazy and Crazy live the life they chose, and the UBI is working out well for both. Thus, Lazy is not exploiting Crazy, since Crazy can withdraw his labour and is afforded the same economic security and freedom to live the life of his choosing as Lazy. I find this response to the original surfer problem persuasive.17 Things would be different, however, if someone in Crazy's position could not choose to give up their job and become a (financially adequately supported) surfer, because they (I'll call them 'Exploited') are neither citizen nor resident18 of a country that pays a UBI. Given the globalized world that we live in, Exploited might well perform work that benefits Lazy as well as Crazy. Such work can be a subsidy of their lifestyles if Exploited's labour is cheaper than the labour of those in the same strong bargaining position as Lazy and Crazy to demand higher pay or better working conditions. After all, Lazy and Crazy have the freedom and security to opt out of work altogether. If it is unjust that someone's lifestyle is subsidized by the labour of another person and there is no adequate reciprocity in this subsidy, then a merely national UBI suffers from a justice problem with regard to the subsidies Lazy and Crazy receive from Exploited and the lack of reciprocity. In the debate about the fairness of a UBI there is much concern about non-contributors, such as the Malibu surfer, receiving benefits. There is much less concern about people who do contribute, but do not receive benefits because they are of a different nationality or live in a different country.19 The latter, it seems to me, constitutes the most pressing justice problem. Let me spell this out in more detail. A merely national UBI excludes certain people as recipients, even if they contribute to the effectiveness of the UBI in that country. They can make such a contribution by subsidizing UBI recipients with their cheap labour, which increases the relative value of the UBI for recipients, as it makes products and services more affordable than they would be if everyone received a UBI. This impacts what and how much those living off a UBI can afford, the opportunities open to them, and how attractive it is for them to exit jobs (or to reduce hours), and thus whether the UBI realizes the goods that it promises, and to what extent it does so. These subsidies for UBI recipients would be, in one sense, not dissimilar from currently existing conditions. After all, the citizens of developed countries already benefit from global inequalities, poor working conditions, and low pay in developing countries where people are forced to work to meet their basic needs. However, there currently exists a work requirement even in wealthy countries, whereas there would be no such requirement for those enjoying an effective UBI. It is reasonable to assume that if non-recipient contributors were not forced to work for their subsistence, because they also enjoyed a UBI that frees them of the need to work for their subsistence, then some services and goods that are currently offered or produced cheaply would become more expensive. The subsidies would likely (at least in part) disappear, and the UBI thus become less effective for recipients. A prime example for exploitation, that subsidizes the citizens of wealthy countries and that many individuals in developing countries would be unlikely to put up with if they were in a position to opt out, is so-called sweatshop labour.20 I take it that sweatshop labour is a relatively uncontroversial example of exploitation.21 The low pay and bad conditions in sweatshops are (part of) the reason why this work is profitable for their owners and why this mode of production can benefit end-consumers. Even after the introduction of a global UBI (or a merely national UBI in countries with sweatshops), there may be people willing to manufacture the type of products formerly produced in sweatshops, but they would presumably need stronger financial incentives and would be able to demand better working conditions that would make these products more expensive. Now, imagine that Exploited works in a factory that builds surf equipment and, due to poor working conditions and low pay for her and her colleagues, this equipment is significantly cheaper than equipment produced by Lazy's fellow citizens who enjoy a UBI and lack strong financial incentives to work in a surfboard factory for low pay. In an extreme case, the only surf equipment Lazy can afford from his UBI is equipment produced by Exploited. Without this cheap supply of equipment, it would not be possible for Lazy to live the pure lifestyle of a surfer and he would have to take up (part-time) employment to afford his equipment. A UBI then allows Lazy to live the way he wants only because of the labour of Exploited, who, unlike Crazy, does not have the option to withdraw her labour and live off a UBI. Even if Exploited does not manufacture the specific products that Lazy needs to live the pure lifestyle of a surfer, as long as she does manufacture products or render services that Lazy or Crazy want or need, and does so more cheaply than if she were able to opt out of work, the UBI that Lazy and Crazy enjoy goes further in terms of purchasing power than it otherwise would. Part of the value of a merely national UBI for recipients thus stems from the fact that this UBI is merely national and that it still leaves in place sources of cheap labour elsewhere. One might respond that my empirical assumption that a global UBI would dry up sources of cheap labour and drive up prices is questionable. After all, one of the potential advantages of a UBI is that it removes financial disincentives to take up low-paid or precarious work; disincentives put in place by misjudged tax and welfare systems that leave people financially better off on welfare than in work or that effectively impose high taxes on low incomes. People might still want to work to top up their UBI and they would not need a subsistence wage in return. Thus, a UBI might even make certain products and services cheaper. I think that intrinsically rewarding or meaningful work might become cheaper, if people enjoyed a UBI and could choose their occupations without having to earn a living. However, it is also plausible to assume that recipients of a subsistence-level UBI would be unlikely to take up a substantial amount of sweatshop-type labour, unless these jobs were paid significantly better and/or working conditions much improved (and would thus no longer qualify as sweatshop labour). After all, one of the main benefits of a UBI is that it grants recipients the freedom to decline or exit unattractive jobs and even escape 'forced labor market participation' altogether.22 What, if any, jobs would UBI recipients want to decline if not poorly paid, dangerous jobs, involving repetitive labour, coercion, and health risks? Thus, it is plausible to assume that many currently cheap products would become more expensive if, globally, everyone enjoyed a UBI.23 This would reduce the living standards of people in high-income countries subsisting on a fixed income in the form of a UBI, and might even reintroduce the necessity to work. The effectiveness of a UBI in providing basic financial security and allowing recipients to live as they wish depends in some measure on cheap (and exploitative) labour by those not enjoying a UBI. III COMPENSATION FOR NON-RECIPIENT CONTRIBUTORS My argument above notwithstanding, a merely national UBI would be an effective tool to combat many forms of injustice within that nation. After all, even in wealthy nations many citizens are in precarious and vulnerable financial positions that employers, corporations, loan sharks, and the government can exploit. A merely national UBI, if pitched at the right level and financed in the right way, would serve justice within that society, and it would therefore be premature to dismiss it. There is a parallel case that demonstrates how defenders of a UBI can respond to the justice problem raised in the previous section: namely, the ethical debate about the harmful climate impact of Alaska's Permanent Fund Dividend (PFD). Since 1982 the US state of Alaska pays an annual dividend—$1,114 in 2021—to all its residents. The PFD represents the largest example of a UBI paid over a prolonged period (albeit not at subsistence level). Moreover, the Alaska financing model, taxing natural resource appropriation, is attractive to UBI advocates, since many other financing models, such as income tax or VAT, disproportionately affect working people and/or people on low incomes.24 Steve Winter, however, argues that the 'PFD makes participating recipients complicit with grave wrongdoing because of its connection to the oil industry, a practice responsible for 150,000 deaths per annum'.25 Winter's argument is, in short, that the oil industry involves intrinsic wrongs, and that the Alaskan PFD makes recipients complicit in intrinsic wrongs. This complicity is itself wrongful. Winter does not think that his objection is 'devastating' for the Alaska model.26 He proposes 'to combine basic income payments with sufficient funding to stop the killings'.27 Since the oil industry28 does not kill anyone directly, but rather costs lives due to indirect effects such as disease, extreme weather events, and other foreseeable effects of global warming, adequate preventive and remedial measures could be taken to remove the reason to object to the Alaska model. The Alaska model would then have two components: 'Funding derived from natural resource appropriation would provide monies for a basic income and to prevent the deaths that would otherwise occur due to climate change'.29 This latter scheme would not have to prevent all deaths that occur from climate change, as Alaska is globally only a relatively minor source of oil. It would merely have to mitigate 'those negative effects (we are focusing on unjustifiable deaths) that could be attributed to the industry in question'30—meaning, specifically the Alaskan oil industry. Winter calculates that averting 405 killings per year would be sufficient to 'discharge Alaskans' "fair share" of responsibility' for the wrongful killings.31 He also emphasizes that this move is applicable to all models of financing a UBI based on harmful resource extraction.32 He does not, however, generalize his argument to other ethical challenges for a UBI. The general intuition behind Winter's idea is that those who benefit from unjust arrangements have stringent obligations to compensate victims. This principle is not limited to a merely national UBI made (more) effective by exploitation. In the next section, I will discuss whether it makes a moral difference if exploitation specifically props up a merely national UBI. For now, I will show that a plausible reply to the problem I discussed in the previous section is to take up a modified version of Winter's suggestion of providing what he calls 'avert-funding'. I will phrase my own proposal in terms of compensation, specifically for those who contribute to a merely national UBI but do not receive a UBI. To better understand this proposal, let us look at four salient points of comparison between the problems of the Alaska model and the global justice problem from the previous section. Firstly, Winter stresses that 'Alaskans are not guilty of' the deaths resulting from the oil industry.33 Mere complicity in a wrong is less severe than committing the wrong itself. This point is important, since avert-funding or compensation usually would not be an adequate response to wrongs that an agent herself commits. If I murder you, I cannot plausibly compensate for this by donating money to buy medicine that saves someone else's life. Compensation schemes are appropriate for wrongs which I am complicit in or benefit from, but of which I am not the perpetrator. The perpetrator's primary duty is to refrain from wrongdoing, not to offset the negative effects or avert harm elsewhere. Winter suggests that complicity in the Alaska PFD is 'equivalent in veniality with wearing sweatshop clothing'.34 This is quite a contentious claim, since being complicit in killing is commonly considered much worse than complicity in various forms of exploitation. Presumably, Winter's point is here not about the impact of killing and exploitation on victims, but about illustrating the structure of moral responsibility. The idea is that agents are not actively harming anyone themselves, but they are complicit. Importantly, this is already the case for the globally affluent who do not receive a UBI, but who are buying sweatshop products, making them complicit in exploitation. There would, however, be an important sense in which merely national UBI recipients' complicity had an additional dimension. As I suggested in the previous section, if it were not for cheap foreign labour that benefits the citizens of wealthy nations, a merely national UBI might not be sufficient to cover the basic needs of recipients, and thus to perform some of its most beneficial functions. Moreover, even if a merely national UBI covered basic needs without cheap labour from abroad, its relative value increases due to this cheap labour. UBI recipients' real freedom is subsidized by the labour of those forced to work for their subsistence. Secondly, Winter thinks that mere benefiting from wrongful actions is not sufficient for complicity in an ethically relevant sense. He stresses that '[t]he wrongfulness exposed by the complicity argument does not depend on benefits but rather upon wrongful agency'.35 By contrast, the problem I bring up draws on both wrongful agency and benefiting. Wrongful agency, for instance, obtains in consumer choices when we knowingly buy products produced under exploitative conditions. Moreover, we benefit from wrongful structures, if we receive a UBI the effectiveness of which is enhanced by exploitation. Wrongful agency and benefiting raise several complicated issues concerning voluntariness. After all, one might argue that you are not morally responsible for benefiting from something that you have no say in—for instance, because your government transfers money to your bank account every month whether you want it or not.36 One way to address the global justice problem I brought up in the previous section would be to give potential recipients the opportunity to opt out of the UBI, an opportunity that exists for the Alaska PFD.37 However, unlike a compensation scheme, opting out of the UBI would not leave non-recipient contributors materially any better off. Moreover, it might still be a problem if certain (already privileged) people can opt in or out, an opportunity denied to non-recipient contributors. Finally, if people opted out of the UBI, then this might necessitate the reintroduction of welfare schemes that secure their subsistence if they become indigent. This would presumably leave unsatisfied those UBI advocates who emphasize the benefits of the universal coverage of a UBI.38 Thirdly, Winter maintains that 'Oil exploitation is not wrongful in itself', but that the production of greenhouse gas, which causes additional deaths, 'is a normal component of the industry'.39 In the case of exploitation of people, as in sweatshop labour, early deaths, due to workplace accidents, are fairly abnormal, but other impacts, such as negative long-term health consequences for workers, lost time with family and friends due to long shifts, lost opportunities, and so on, are inextricable features of the industry. The exploitation I am concerned with is not an unintended by-product, but essential to the high profit margins of sweatshop labour. Thus, the problem I raised is, in one sense, more worrying than the rather indirect effects of the oil industry. It is, I take it, plausible that someone who benefits from exploiting people has a more stringent duty to compensate victims than someone who benefits from resource extraction that, as a normal or even necessary side-effect, is also harmful to some people. However, my proposal does not hinge on this specific intuition. Fourthly, there is an important sense in which a merely national UBI, as well as the Alaska PFD, are both innocuous. In both cases the victims (of climate change or exploitation) are not made worse off by the redistribution of revenues in the form of a merely national UBI. The victims of climate change would die even if a private corporation pocketed Alaska's oil revenue. Analogously, whether consumers do or do not receive a UBI leaves sweatshop workers' working conditions and pay untouched, or at least there is no immediate connection here. This might well impact the stringency of duties of compensation that those benefiting from their labour have. In fact, I think that, since a merely national UBI does not make victims of exploitation worse off, a compensation scheme is the right middle ground between, on the one hand, opposing a merely national UBI on ethical grounds and, on the other hand, leaving the exploitation problem of a merely national UBI unaddressed. Winter proposes avert-funding that would save as many lives from the effects of climate change as the Alaska oil industry is responsible for taking, and he presents a clear number for this: namely, 405 lives. By contrast, it is very difficult to quantify how many people would have to be exploited for a merely national UBI to be effective, and this would greatly depend on the circumstances of the recipients as well as the contributors, and on overall political and economic conditions. It is therefore impossible to determine a concrete number of people that an avert mechanism must protect from exploitation. Furthermore, unlike those killed by climate change, the current victims of exploitation are still alive.40 In addition, whereas the effects of climate change are random, in the sense that we do not know which specific individuals will die or otherwise suffer from it (though we do know that some groups and regions are much more likely to be affected than others), there are concrete individuals whose work would benefit the recipients of a merely national UBI. These non-recipient contributors are owed something, and it would not be adequate to merely combat or avert exploitation somewhere, if we can at least approximately identify these contributors. My proposed compensation scheme, therefore, is not a form of avert-funding, but rather funding that aims to benefit specifically non-recipient contributors to a merely national UBI. This can take the relatively general form of benefiting anyone from whose exploitation recipients of a UBI benefit. For instance, fair-trade practices or even higher standards could be made legally mandatory and appropriate resources could be allocated to enforce these standards.41 We can also imagine more targeted measures that specifically benefit workers in developing countries manufacturing products and providing services crucial to the citizens of the country that introduced a merely national UBI, insofar as it is possible to single out specific factories or regions where this labour is performed. One such targeted measure could be direct money transfers to workers who have worked a set number of years or who have become unable to work. Another could be to relax immigration restrictions for non-recipient contributors, so that it becomes easier for them to become residents or citizens of the country that implemented the merely national UBI, and to set up funds to help them make the journey and settle in. This, of course, should not happen only once, but for each successive generation of non-recipient contributors replacing those who have retired or migrated. There are, admittedly, major problems of application here, since global supply and production chains are complex and often cross more than one border. Moreover, one might wonder whether our conception of who contributes to the effectiveness of a merely national UBI should include not only workers who manufacture cheap products, but also people who produce food and other necessary goods and services for these workers, or who perform unpaid reproductive labour that allows workers to produce goods and services as cheaply as they do. I cannot address these problems here. Maybe the best we could do would be to start with a few cases of exploitative labour that clearly benefit many UBI recipients and then extend the compensation scheme gradually, according to our best understanding of global supply chains.42 To clarify the scope of my scheme, I should note that my concern is not poverty and inequality—as is, for instance the focus of Thomas Pogge's Global Resource Dividend43—but exploitation that contributes to the effectiveness of a merely national UBI. Not all of the globally poor would contribute to a merely national UBI, and not all non-recipient contributors are among the world's poorest. My proposal leaves open the prospect that there are additional duties to compensate, because global poverty is due to structures put in place by the globally affluent and from which the affluent benefit regardless of whether they enjoy a UBI. In fact, I believe that my argument does not reveal an entirely new duty to compensate, but rather that it amplifies pre-existing duties to compensate for global injustices. IV AMPLIFYING PRE-EXISTING DUTIES It seems relatively uncontroversial that there are duties to compensate for exploitation or to try to reset (international) relationships so that they become non-exploitative. This is so regardless of whether nations have implemented a merely national UBI. This raises the question of what exactly is the role of a merely national UBI for my argument.44 At the very least, my argument functions as a prod to make global justice issues more central to the debate about the prospects and drawbacks of a merely national UBI. This is potentially already quite powerful, since some of the most persuasive arguments for a UBI are explicitly moral ones. A UBI is presented as something that individuals are owed as a matter of justice qua entitlement to the commons,45 or as a matter of social and economic justice.46 If there are significant moral reasons for adopting a UBI, then any discussion of the moral advantages of a UBI should consider how it might impact and interact with unjust global structures. My argument also provides additional reasons for compensating exploited workers. It thus amplifies the duty that privileged Westerners already have to compensate those from whose exploitation they benefit. How strong these additional reasons are, and to what extent a pre-existing duty is intensified, depend on why one thinks that a UBI is better than other ways of distributing resources (some of which were obtained unfairly). Those who advocate for a UBI assume that it offers concomitant goods or benefits that other schemes, such as conditional welfare provision, participation income