A Rationalist Theory of Legitimacy
University of Oxford 2017
This is a digital version of my doctoral thesis. You can obtain a PDF version (which includes page numbers) here.
This website was created using pandoc, plus a number of jQuery/CSS tricks gleaned from, amongst other places, here and here, and lots of hope.
The background picture is from User Amadvr from the Wikimedia Commons, and stands under this license.
Almost everyone is ruled. That is, we are subject to political institutions which coerce, threaten, nudge, order, adjudicate, command, advise, demand, offer, take, redistribute, and influence us in uncountably many ways. Few if any other institutions shape our lives in the fundamental way political institutions do. This raises the natural question of what, if anything, can justify this kind of impact. Answering this question is one of the main tasks of a theory of political legitimacy, the contours of which I aim to provide in this thesis. My search for such a theory does not start from scratch, however. It is informed by four background commitments, or starting points: a moral theory which includes rights, and is focussed on the moral status of individuals; a rejection of anarchism—the intuition that something more than minimal political institutions can be legitimate; a belief that democratic and public justification theories of legitimacy fail; and lastly, the view that we have no general duty to obey the law.
I will generally not argue for the truth of these background beliefs, and at most sketch how we might defend them. Still, I should note that at least three of these background commitments are widely accepted in the literature. Most contemporary philosophers believe that people have rights in some form, and they also generally reject (anything less than) the minimal state. In addition, few philosophers any longer believe that there is a general duty to obey the law. So the results of this thesis should be interesting for wide swaths of philosophers. What precisely is entailed by each of these four background commitments will become clearer throughout this thesis, but it is useful to start with a general description of them to give a sense of the problems to come.
First, I believe that something in the vicinity of Kant’s moral philosophy is right. That is, I believe that each individual possesses a dignity and inviolability which entails that they never ought to be treated in certain ways, even if doing so is morally superior in certain outcome-oriented ways. In particular, individuals have moral rights: rights over their bodies, their minds, and the things they own; a right to determine their own lives freely without arbitrary interference from others; a right to do as you please, within certain limits, without being used for purposes you do not endorse, however benign or altruistic these purposes might be. Rights are not all there is to morality—virtues, for example, play a role in how we ought to live, and a fully developed moral theory will surely pay some attention to outcomes. But when it comes to thinking about how we ought to treat each other, rights form the core of morality. (These are, of course, not uniquely Kantian views, and you might arrive at the same conclusions via other means.)
These moral foundations are compatible with a wide variety of ways of fleshing out the details of a substantive theory of justice. Prominently, both Rawls and Nozick invoke Kant when they develop their theories of justice, but they reach thoroughly different conclusions. My sympathies are more with Rawls than Nozick—that is, with a liberal rather than a libertarian theory of justice—but it is not trivial to show that Kantian foundations favour one over the other, and I will not attempt to do so. In general, I will stay largely neutral on which particular theory of justice we should favour. However, in at least one respect Nozick is the more observant Kantian. Nozick realises that it is difficult to see how political institutions could be legitimate in the face of the strong, far-reaching rights that the Kantian ascribes to individuals. No matter how you fill in the details of a theory of individual rights, they work like trumps, ruling out non-consensual interference by others. This seems to entail that consent is necessary for legitimacy: political institutions can only rule over those who have agreed to be ruled by them. This theory I will call voluntarism about political legitimacy—or voluntarism, for short. Here we come to one of the first major issues in this thesis, which I will later call the Basic Problem: it seems that if we accept Kantian foundations, then we ought to be voluntarists.
Nozick offers an argument that his minimal state could be legitimate without many people’s, or even most people’s, consent. He describes a thought experiment in which a consensually formed “protective agency” becomes dominant on a given territory, and eventually extends its rule to everyone living on its territory, whether consenting or not, thus forming a state. But this argument appears to merely show that a minimal state could have arisen without consent under some favourable conditions, not that it actually did. At any rate, even if we grant that this gap can be bridged, Nozick’s argument only succeeds as an argument in favour of a minimal state. This is a state which is concerned with enforcing natural rights, protecting people from violence, and similar tasks. In one sense, this state will be anything but minimal. Tracking and undoing past injustices, for example, is a big and difficult task, and an institution wishing to tackle it will require a large number of employees, resources, and power. But while rather large in terms of sheer size, this state would still be a minimal state in a sense more relevant to us, insofar as its aims or functions are quite limited. A Nozickian minimal state could not permissibly engage in poverty relief, or in the provision of public health care, or in education, or in public funding of the arts and sciences, or in development aid, or in ensuring equality of opportunity, or any of the other functions modern states tend to fulfil.
This brings me to a second background commitment which will shape this thesis: a rejection of political institutions which are minimal in the functions they fulfil. One label for this commitment might be the welfare intuition: the functions modern welfare states fulfil are, in principle, legitimate functions for political institutions. I formulate this carefully, because the intuition should not be misunderstood as yet committing us to the welfare state. There might be political institutions very unlike states which legitimately fulfil the same functions as the welfare state. At any rate, I leave the welfare intuition on a vague level, so as to be neutral on how precisely its details ought to be filled in. You might be a full-blown egalitarian of the G. A. Cohen sort, or have a more limited vision for legitimate state functions. Any view which wishes political institutions to fulfil more functions than a Nozickian minimal state will qualify, which is why anti-minimalism is the better label for this commitment. Anti-minimalism claims that non-minimal states are in principle legitimate. It is compatible with thinking that existing states overreach in their extent, and that many of their current activities cannot be justified—indeed, I will suggest that this is very likely. Still, anti-minimalism at its core is the claim that at least some political institutions at least some of the time legitimately fulfil functions which go beyond the basic task of keeping public order. It should be obvious that anti-minimalism makes the Kantian’s Basic Problem significantly more difficult to solve. Not only do we need to show that non-consensual ruling is compatible with individual rights, we also need to show how some relatively elaborate and invasive functions of political institutions are compatible with individual rights.
The Failure of Public Justification and Democratic Theories
Let me turn to a third background commitment. Most contemporary philosophers are not voluntarists about legitimacy, and many do not even take the position seriously enough to argue against it. Neither are they anarchists or minarchists. Instead, there are two dominant approaches to the problem of legitimacy, based around public justification and democracy. I think that both of these theories fail. I will not argue this point at length in this thesis—largely for reasons of space—so let me briefly summarise my objections against both theories.
First, defenders of democratic theories argue that political legitimacy is based on democratic procedures. On such a view, the legitimacy of a decision is entirely or primarily a function of the procedures we use to make that decision. There are different ways to describe what conditions this procedure needs to fulfil—openness, fairness, equality, and so on—but such details need not interest us here. One of the main strengths of the democratic theory is that it accords well with an every-day picture of the origins of legitimacy. It is common to think that legitimacy must be based on “the people’s will”, “popular sovereignty”, or some such.
However, if we look at democratic theories of legitimacy in a cold, hard Kantian light, we find that they are thoroughly unsatisfying. The dilemma is simple. On the one hand, the democratic process might rest on some previous consent that we will accept the outcomes of the process, or it ends with each of us giving our consent to the outcomes. In either case, it is not the democratic process itself which ultimately bestows legitimacy, but still consent. On the other hand, if the democratic process does not end with consent, or is not itself based on previous consent, we must presume that some individuals do not consent to the procedure or its outcome. If this is the case, then it is unclear what the democratic procedure adds. There might have been a fair process in which 51% of humanity decided to take my kidney to give it to someone else. Perhaps even 99% have decided to do so. But if individuals are inviolable, each with a separate sphere of freedom with which we cannot interfere for the sake of collective purposes, then this does not change a thing.
Saying that democracy is insufficient for legitimacy is compatible with appreciating it in many other ways. Democracy is a good method to pool widely dispersed information, to hold rulers accountable, and to prevent power from being hogged by corrupt elites. Some form of democratic governance is likely to be the instrumentally best way to make political decisions. It might even be that democracy becomes part of our theory of legitimacy once we go further “downstream”: once we have figured out the fundamental grounds on which legitimacy rests, it might turn out to be an additional constraint that decisions ought to be made democratically. But I find it unlikely that democracy figures in our ultimate explanation of why political institutions are legitimate.
The second type of theory which dominates contemporary debate are public justification theories of legitimacy, or as I will call the view, justificatory liberalism. These theories are close cousins of social contract theories. Justificatory liberals claim that legitimacy is not based on actual consent, but hypothetical consent. The state can coerce you because you would agree to it, if you were morally reasonable and epistemically competent. Many people are neither, so we can ignore their lack of consent. While this view is based on idealising actual individuals, this is not the idea we find in the early Rawls’s political contractualism, or Scanlon’s moral contractualism, where the “reasonable people” are completely idealised people. For Rawls and Scanlon, the metaphor of contract is simply a heuristic device to tease out certain intuitions about justice or morality, and the moral contractors are pure creations of their philosophical theories. But the idealisations justificatory liberals impose are moderate. Our hypothetical counterparts are still recognisably us, though purged of our worst vices and errors. Thus, what different people could “reasonably accept” will still depend on their actual preferences and beliefs, and thus will differ across people. For example, what a Catholic and an atheist could reasonably agree to will not coincide.
This makes justificatory liberalism attractive, as it still appears to be sensitive to the beliefs, choices, and identities of actual individuals. This is a big part of justificatory liberalism’s initial appeal. Many justificatory liberals have argued that paying attention to what people could or would agree to, if moderately idealised, is a way of respecting them as free and equals. So this is a view which promises to deliver on some of the Kantian intuitions about the respect we owe to others, while also promising to avoid minimalism. For example, one hope is that we can justify the welfare state on justificatory liberalism because everyone would accept this type of state if they were appropriately reasonable.
I explored those issues in my B.Phil. thesis. The short result is that justificatory liberalism fails in both its promises. First, it does not actually escape minarchism (or perhaps even anarchism). If you look at the actual pluralism in modern societies, it is impossible to justify liberal policies to everyone. Again, at most a minimal state can be justified, one which pursues the most uncontroversial tasks, like upholding public peace. Justificatory liberals generally claim the opposite, and have offered a wide variety of arguments that their view does not have this implication—but I believe that none of these proposals can be made to work.
Second, the moral foundations of justificatory liberalism are unconvincing. Voluntarism provides us with a clear idea of what we ought to morally respect: the free and autonomous exercise of people’s wills. This is not to say that every act of consent is morally binding, as consent can be uninformed or involuntary. Still, where it is binding, we can tell a convincing story about the “moral magic of consent”, as Heidi Hurd has memorably put it. Justificatory liberalism can offer no such story. We do not respect actual individuals or their will when we focus on what their moderately idealised counterparts “could reasonably accept”. It is true that justificatory liberalism is sensitive to the beliefs and choices of individual in some sense, but it is not at all clear that it is sensitive in the right way. Again, this is a crude overview, and justificatory liberals have offered many arguments trying to defend their view; but I will gloss over these arguments and my replies to them here. Like democracy, public justification is not a plausible ground of legitimacy.
There is a last background commitment which further contributes to the philosophical quagmire. It is now widely accepted in the philosophical literature that it is unlikely that people have general obligations to obey their rulers. Generally speaking, rulers do not enjoy authority—the right to impose duties on the ruled. What precisely this result entails is a matter of some philosophical delicacy, to which I turn to in the next chapter. However, we can already suspect that this will make our aim of justifying non-minimal government even more difficult to achieve. After all, if governments cannot impose duties on their citizens, it is not immediately obvious how they could rule at all, or rule in a non-minimal way.
So we have encountered an impasse. Three major features which could ground legitimacy—consent, democracy, public justification—fail to justify anything more than the minimal state. In addition, the second and the third theory are morally unconvincing, at least seen in the light of a Kantian approach. It is at this dialectical point that my thesis picks up the loose ends. I defend a fourth theory, rationalism. Rationalism sees legitimacy as a merely derivative value. On such a theory, we should distribute the right to rule on the basis of whether doing so brings about desirable consequences. Rationalism, I will argue, need not be a consequentialist theory. In particular, the value I will put at the centre of rationalism, justice, is deontologically structured: it is a value we ought to promote, but also a value which imposes constraints on how we can promote it. In short, rationalism claims that the right to rule should be distributed such that justice is promoted best, insofar as possible within the constraints of justice. This theory stands in direct conflict with voluntarism, as it claims that consent is not necessary for legitimacy. Furthermore, unlike democratic or justificatory theories of legitimacy, rationalism does not even claim that a light-weight substitute of consent is necessary for legitimacy, such as hypothetical consent or a democratically exercised “general will”.
Setting aside public justification and democracy, we have a stark contrast between voluntarism and rationalism, which will guide us through this thesis. What reason do we have to focus on this contrast rather than some other? One main motivation comes from Patrick Riley’s history of social contract theory in his Will and Political Legitimacy, a book which has been unjustly overlooked by most philosophers. Riley’s main focus is on how the major authors in the social contract tradition—Hobbes, Locke, Rousseau, Kant—understand the notion of the will differently, which is less interesting to us. But another important observation by Riley is that, while these authors pay lip service to voluntarist language, they explicitly or implicitly temper the voluntarist element in their theories once push comes to shove, and some drop it altogether. Rousseau, for example, makes the “general will” central to the legitimacy of the state. But this notion oscillates between being the de facto combined will of all, and a normative standard of what the combined will of all ought to be. Rousseau’s indecision, in other words, is between a genuinely voluntarist standard of legitimacy, and a rationalist standard of legitimacy which builds on which decisions would be right. As Riley highlights, a focus on the latter is precisely what Hegel suggests in his criticism of social contract theory. Hegel argues that the human will is fickle and unreliable, and its exercise could not form the basis of a legitimate state. On the other hand, “[w]hat is right and obligatory is the absolutely rational element in the will’s volitions”. If we put this in other language, we can read Hegel as claiming that people consent for reasons, some of them good, some bad. Why not focus, Hegel suggests, simply on the good reasons, and cut out any reference to the will altogether? In that case, consent drops out as a basis for legitimacy.
A rough generalisation of this point is that much of the social contract tradition, and by extension I think, much of liberal thought, is concerned with reconciling two tenets: a commitment to autonomous, free willing as a foundation for legitimacy, and the thought that reason alone could establish and justify legitimate power. If this is true, we can see “will” and “reason” as the two fundamental poles liberal theories of legitimacy have tried to reconcile. Rationalism and voluntarism are theories which exclusively focus on one of these two elements, while we can see other theories as attempts at hybrid theories. For example, an appeal to moderately idealised consent is precisely an attempt to “realise some of the values of voluntary participation in a system of institutions that is unavoidably compulsory”. We can also pursue the contrast between rationalism and voluntarism into the present. Jeremy Waldron argues that the fundamentally liberal idea is that
a social and political order is illegitimate unless it is rooted in the consent of all those who have to live under it; the consent or agreement of these people is a condition of its being morally permissible to enforce that order against them.
Like the social contract theorists, Waldron starts from a moral ideal that makes consent central. But Waldron notes that this core liberal idea can be spelt out in two distinct ways, which he calls “voluntaristic” and “rationalistic” approaches to legitimacy. Waldron associates the former approach with actual-consent views and Rousseau, the latter with the hypothetical-consent tradition and Kant as well as Rawls. Waldron notes the radical difference between these two ways of approaching the topic:
When we move from asking what people actually accept to asking what they would accept under certain conditions, we shift our emphasis away from will and focus on the reasons that people might have for exercising their will in one way rather than another.
Again, voluntarism and rationalism emerge as the two clearest poles of the liberal tradition. In short, the voluntarism—rationalism contrast cuts to the core of a fundamental issue in liberalism. This gives us good reason to focus on these two theories as our main contenders.
Given this rich pedigree, we should not of course expect rationalism to be a new theory. Most prominently, it has been advocated by Allen Buchanan, whose central statement of his view is that
[a] wielder of political power (the supremacist making, application, and enforcement of laws in a territory) is legitimate […] if and only if it (1) does a credible job of protecting at least the most basic human rights of all those over whom it wields power and (2) provides this protection through processes, policies, and actions that themselves respect the most basic human rights.
I have some quibbles with Buchanan’s formulation—I reject Buchanan’s “satisficing” formulation, and I see no reason to limit ourselves to “the most basic human rights”. Buchanan is also not a Kantian, so for him the project starts from rather different foundations. There are other differences, but I will not greatly be concerned with them, partially for reasons of space. Instead, I will aim to shed light on some issues in rationalism which have been ignored by authors in the literature. In particular, I will argue that rationalism should be sophisticated—that it should aim to be only indirectly, not directly, applied.
Developing the structure and content of rationalism will be one of the major tasks of this thesis, so I leave the details to later chapters. Here I only wish to make some initial remarks. First, justice has a deontological structure: choosing certain means towards our ends is itself unjust. So when I say that justice should be “promoted best”, this means that rulers should seek to promote justice within the constraints given by justice. We will see that justice has a complicated structure—while it is useful to compare rationalism with instrumentalist approaches like consequentialism, it also bears important dissimilarities.
Second, I will not specify a worked out theory of justice in this thesis. Instead, my main concern is with teasing out the connections between justice and legitimacy. This is a question which is interesting (and difficult) enough by itself. This entails that we are sometimes left with relatively general statements about what rationalism might practically demand, but this is unavoidable.
Third, you might worry that consent itself is a demand of justice—that if rulers treated the ruled against their consent, this would constitute an injustice. This would collapse the difference between voluntarism and rationalism. Indeed, the following two positions are equivalent for practical, though not metaphysical, purposes:
Consent (ultimately) grounds legitimacy. Thus, consent is necessary for legitimacy.
Justice (ultimately) grounds legitimacy. Justice requires consent. Thus, consent is necessary for legitimacy.
To avoid this implication, I will group theories (A) and (B) together under the label “voluntarism”. Put otherwise, I will understand by rationalism only those positions which deny that justice generally requires the consent of the governed to their government.
A last remark concerns the label I have chosen. “Rationalism” is an unhappy label insofar as it has been used in a wide number of other meanings. We can find it as the opposite of empiricism in the history of philosophy—plotting, for example, Leibniz against Locke; as a term in the debate regarding moral motivation, being the position that we can have reasons for action unconnected to our desires; as a label for trust in our ability to rationally plan society on a large scale on the basis of abstract ideas, as opposed to more pessimistic, piece-meal conservatism; and lastly, as a generic label for any view that emphasises reason over emotion, intuition, or perception, and perhaps any view that opposes mysticism, tradition, and religious superstition. As it happens, I sympathise with all these types of rationalism—but for the purposes of this thesis, “rationalism” is not intended to carry any of these connotations.
This describes the set of philosophical problems this thesis aims to tackle. I will proceed as follows.
Chapter 1 starts by spelling out what legitimacy is. Unfortunately, the concept of political legitimacy is somewhat unclear, so some conceptual legwork needs to be done before we can even start our discussion. I explain why governments generally do not have authority. This will lead into a discussion of how legitimate governments without authority are possible. My aim is to show that this is a genuine conceptual and substantive possibility, and does not fail for some obvious objections which have been given in the literature.
Chapter 2 sets out one of the major puzzles of this thesis: if we accept a roughly Kantian picture of individual rights, then it seems that voluntarism about legitimacy is inescapable. That is, more-than-minimal governments could only be legitimate if they enjoyed our consent. This chapter describes the problem at depth, calling it the Basic Problem for rationalism. I then go through a set of unconvincing attempts to solve the problem.
Chapter 3 suggests that, if we are to solve the problem, we ought to look at theories of rights. I outline two major theories of rights, understood as theories of how rights can be justified, the interest theory and the status theory. My philosophical sympathies are with the latter, but it is useful to discuss the interest theory, given that it is the dominant account of rights in the literature. I spell out this theory in some detail, and argue that we can solve the Basic Problem if we accept it.
Chapter 4 turns to the status theory of rights, which I already describe in the previous chapter. Here a solution to the Basic Problem is significantly harder to find. At this point, I return to where I started: Kant, or more precisely, Kantianism. Kant’s political philosophy suggests that political institutions are necessary to fully bring our rights into existence. Because this is the case, political institutions do not conflict with us having rights. Far from the state violating our moral status, it is necessary to first bring a full recognition of this status about. This is a difficult argument, and there are some ways to make it badly. This chapter discusses how we might best pursue this point.
Chapters 1 to 4, taken together, aim at a defensive argument: to show that political institutions can, at least in principle, be legitimate without consent and authority, even if we accept the Kantian picture of moral rights. If this is correct, then Kantian foundations do not necessarily entail voluntarism. But this merely shows that rationalism is a possibility; it says little about what speaks in its favour, or how we should understand its structure. The last two chapters are given over to this more constructive task.
Chapter 5 starts from the general question of what value we should make primary in an approach to legitimacy which sees it as a derivative value. Instead of justice, we could focus on perfectionist values, or more undemanding values such as public order and stability. I resolve the issue by arguing that justice should be primary. I explain what we should understand by justice, and what it would mean to promote it best. I also show that the argument for the primacy of justice is closely linked to the Kantian argument in chapter 4.
Chapter 6 asks how we ought to apply rationalism. Directly following rationalism is often self-defeating: in complex, pluralistic societies, trying to promote justice on a case-by-case basis might overall promote justice worse when compared to following some other standard which does not directly reference justice. I explain how this idea can be cashed out by building on analogous work done with regard to sophisticated consequentialism.
This chapter fulfils two important tasks. The first is conceptual: I will define what it means to call a set of political institutions legitimate. The second is substantive: I will outline one of the main assumptions of this thesis, philosophical anarchism. These tasks are interconnected, as my answer in both cases will be that we should search for legitimacy beyond authority. On the conceptual side, many authors accept a close connection between three concepts: political authorities make the claim that those subject to it have a duty to obey; an authority is politically legitimate just in case those claims are true. Thus, legitimacy implies authority implies duties to obey. I will describe an alternative way to conceptualise legitimacy which breaks this link. On the substantive side, I outline a defence of philosophical anarchism. Philosophical anarchists deny that political institutions generally enjoy authority over its citizens, and that citizens generally have a duty to obey their rulers. (As we will later see, philosophical anarchists will allow some exceptions.)
The substantive question is the more interesting one. If rulers do not generally enjoy authority, what kind of political institutions (if any) can we expect to be legitimate in the non-obedience-entailing sense? This is a large question which will occupy us for the rest of this thesis. In this chapter, I will first outline what the concept of legitimacy is not (sec. 1.1), and what it is (sec. 1.2). I describe why philosophical anarchism is likely to be true (sec. 1.3). In the last two sections (sec. 1.4 and sec. 1.5), I offer a general account of how legitimate government without authority is possible. My aim here is to answer some standard criticisms which have been offered against such views. I end with a caveat (sec. 1.6): it will be much more difficult to defend the state if we accept philosophical anarchism.
What Legitimacy Is Not
There is no standard meaning of the term legitimacy. Sociologists, lawyers, philosophers, and ordinary people use the term to point to different properties of political institutions. Only some of these meanings are of philosophical interest. Let me briefly rule out some of the uninteresting alternatives.
First, we need to distinguish an empirical from a normative sense of legitimacy. A common way to talk about legitimacy, widespread especially amongst social scientists, is to understand it as the support a regime enjoys, or (in a circular-sounding definition) as the degree to which individuals believe political institutions to be legitimate. When a government is supported, or believed by its citizens to be support-worthy, is an empirical question. As such, we should leave it to the social scientists. Instead, I focus on the normative question: what kind of political institutions ought to be supported by individuals? Admittedly, facts about support are often relevant to normative theory. In most theories of legitimacy, including mine, a government can usually be legitimate only if it has de facto power: that is, if it can effectively implement its policies. For this, it usually relies on some degree of support by its subjects. However, actual support is not part of the conceptual core of legitimacy.
Second, “legitimacy” is often used in a generic normative sense, where it is synonymous with other words such as “right”, “proper”, “adequate”, or “binding”. This usage is too vague, however, to allow much theorising. Philosophers have made various attempts to make the concept more precise which I will describe in the next section. This makes it important to remember that “legitimacy” will take a technical meaning throughout this thesis which will not always track the vague usage of the concept. When we call a political institution legitimate in the technical sense, we are not necessarily commending it in one of the more generic senses.
There are three other common, normative usages of the concept we have to guard against. First, legitimacy ought not to be identified with justice. I will argue in this thesis that legitimacy is grounded on justice. However, this is a substantive claim: conceptually speaking, justice and legitimacy are not the same. Justice, depending on who you ask, is a matter of how we ought to distribute burdens and benefits, or what rights people have, or what they deserve. Legitimacy, on the other hand, is about who has the right or permission to rule. It is a substantive claim that the latter should be distributed on the basis of our ability to bring about the former. More generally, we should conceptually distinguish the legitimacy of a political institution from its good-making features: from the benefits or other desirable properties it has.
Second, legitimacy is often closely associated with legality. When a decision or policy is described as illegitimate in public discussion, often this is simply taken to mean that it was unlawful. In a philosophical context, however, there are reasons to resist this identification. We should be legal positivists regarding the law: it is an empirical question what is lawful or unlawful, while legitimacy is a normative property. Thus, some policy being legitimate need not entail it being lawful, and neither is the inverse true—though, as I will later argue, there are many reliable substantive links between the two.
A last possible source of conceptual confusion has to do with procedural norms. The language of legitimacy is closely associated with the language of legitimation, and you might think that the latter is conceptually primary to the former. In particular, calling a decision legitimate is sometimes taken to be synonymous with it being democratically decided, or decided in some other correct way. Again, you might claim that, as a matter of substantive moral theory, democracy is necessary for legitimacy. But it is not useful to make this conceptually necessary.
Moral Qualities of Political Institutions
By contrast, when philosophers talk about legitimacy, they usually understand it to be the right to rule. The advantage of this rendering is that we have a sophisticated philosophical apparatus to analyse what we mean by “right”, the Hohfeld schema. The Hohfeld schema allows us to distinguish four possible interpretations of what “a right to X” could mean: a liberty-right, a claim-right, a power-right, or an immunity-right. Unfortunately, there is no agreement amongst philosophers with which of these four, or which combination of them, we should identify the right to rule. The conceptual disagreement sits on top of a substantive disagreement. Should we normally see governments as being, normatively speaking, quite powerful entities? That is, entities which possess far-reaching authority and moral powers? Or should we see them as more limited? I will start by outlining the moral qualities which we could ascribe to political institutions. I return to the substantive question at the beginning of next section.
Privileges and Permissions
When we say that a government is legitimate, we might mean that it is allowed to rule. In the Hohfeld schema, this entails that we see legitimacy as a liberty-right or privilege. If a ruler has a liberty-right to rule, then they are under no duty to anyone not to exercise power, and citizens have a corresponding lack of a claim-right that the rulers not do so. Alternatively, ignoring the Hohfeld schema, we might say that some entity is legitimate in case it is permitted to rule. You are permitted to φ just in case it is not wrong for you to φ, or if you have no duty not to φ.
There is a difference between privileges and permissions which is almost always ignored by authors in the secondary literature. You might have a liberty-right to walk through the park—as no one has a right to demand that you do not—while not having a permission to walk through the park, as you might have a duty to be at home practising your piano skills. The point is that privileges belong to the language of what we owe to others—to what some have called “bipolar” morality. We have privileges against others. On the other hand, permissions belong to what morality requires of us, to “monadic” morality. Moral requirements in this category are not owed to anyone else. The connections between these two types of moral requirements are not immediately clear. In particular, consider the following claim:
You have a permission to φ (= no duty not to φ) just in case you have a privilege to φ (= no duty against anyone not to φ)
It is not obvious that this claim is true. For example, it might be that I have a privilege to waste away my money: I have no duty to anyone to be frugal. At the same time, it might be that I am not permitted to waste my money away, perhaps because doing so is not virtuous.
This technical difference notwithstanding, privileges and permissions share one important feature: they are both non-exclusive. Assume that you are permitted to eat chocolate. But having this permission, by itself, is compatible with everyone else having such a permission. It is even compatible with others having a right to stop and interfere with your chocolate-eating. The same is true of your privilege to eat it. In a chocolate-eating contest, you have a privilege to eat chocolate, but no right to stop others from eating yours. This shared feature of privileges and permissions allows us to group them together as the minimal moral qualities that political institutions can have. Any approach which identifies legitimacy with having merely these minimal qualities I call a minimal conceptual understanding of legitimacy.
The minimal approach is standardly associated by many authors specifically with coercion. On this account, legitimacy is the permissibility of using coercive power. But I see no convincing reason to further narrow the minimal approach in this way. Non-coercive political action can also be considered in terms of legitimacy. If a ruler puts up crucifixes in schools, or decides to raise the interest rate, we can meaningfully ask whether these policies are legitimate. Not all government action is coercive, and questions about legitimacy would remain even if rulers suddenly lost all their coercive powers. So we should construe ruling broadly: recommending, setting incentives, discouraging, and making law are all parts of this activity, as is coercing.
We might think that the minimal position is insufficient, on either conceptual or substantive grounds. We might claim instead that a legitimate ruler has not merely a liberty-right to rule but a claim-right. In the Hohfeld schema, claim-rights correspond with directed duties. If I have a claim-right to eat chocolate, others have a duty to not interfere with my eating, not to take away my chocolate, and so on. We can see that the proposal can be understood differently, depending on how you understand the content of the claim-right. One prominent suggestion is to see legitimacy as a claim-right to rule corresponding with duties of others not to interfere with, or undermine, political institutions. We can call this a protected liberty, as it is essentially a liberty-right to rule coupled with a claim-right that others not interfere. Let us call conceptions of legitimacy which see it as a protected liberty moderate approaches to legitimacy.
Advocates of the moderate approach are comparatively rare. That is perhaps because their position is an odd halfway house between the minimal approach and the standard authority-based approach I will discuss in a moment. Indeed, it is doubtful whether we can make the idea of “non-interference” plausible without it collapsing into obedience. What does it mean to not “interfere with” the activities of a political institution? For example, it seems that we do not interfere with the state in its tax collection efforts just in case we obey its tax laws. So the defender of the moderate view has some work to do to show how her position is intelligible. I will set it aside in the following.
Authority and Moral Powers
The classic way to understand legitimacy is to see it as the claim-right to be obeyed. I will reserve the label “authority” for this right. The reader should be warned, however, that there is no consistency in usage in the secondary literature. Many authors apply the label authority to entities which exercise de facto power, and who claim to make authoritative commands, even if there is no duty to obey those commands. These authors then call entities whose claims in this respect are true legitimate authorities. On that usage, the North Korean state has authority over its citizens, but its authority fails to be legitimate. On my usage, the North Korean state has no authority, but merely pretends to have it.
What is authority? We should see it as the ability to impose new moral duties on others through one’s willing. If Emily has authority over Finn, then Emily telling Finn to go to bed provides Finn with a strong reason, perhaps even a duty, to go to bed. This ability—to affect the duties of others, and in particular, the ability to create new duties—is what crucially sets apart authority from the minimal and moderate senses of legitimacy. It is important to dwell on this point, as it will be crucial to understand the debate between the philosophical anarchist and the anti-anarchist to which I turn later. Recent debates in the secondary literature have sharpened awareness that authority is one instance of a more general phenomenon, the possession of a moral power. A moral power is the ability to change the normative situation of others. By “normative situation”, I mean the duties and (Hohfeldian) rights that a person has, but we might widen it to include any moral reason one has. Let me explain. A paradigmatic way of gaining a moral power over others is through promising. Assume that I promise you to do you a favour. If you then ask me to help you paint your house, then I have a new duty to help you paint your house. I have this new duty on the basis of you willing it to be the case that I have this duty. In this sense, my normative situation has been changed through your will.
We need to contrast the exercise of a moral power with other ways in which our normative situation can be changed. One important way is that through acting, individuals can affect empirical facts in a way that indirectly changes what others ought to do. Mundane examples abound. Assume that I step in front of your car. You now have a duty to stop, or swerve around me, because you have a duty not to harm me. So there is a trivial sense in which I have changed your normative situation. However, I have not exercised any moral power over you: it is not through my willing that your duties changed. It is useful to have a label for this effect: let us say that I merely triggered reasons (or duties) that you already had.
Intuitively, there is a clear difference between these two cases. Various explanations of the difference between genuine exercises of moral power and mere triggering of pre-existing reasons have been proposed in the literature. We only need a very rough description of the difference for our purposes. Observe first that the genuine exercise of a moral power is normally based on our willing, more precisely, on our intending to change the normative situation of another person. In the promising case, I intend for you to help me paint my house. On the basis of me intending it to be so—plus some other facts, such as successfully communicating my intention—you gain a duty to help me. In the car case, I might have no such intention. So exercises of moral powers are linked to our intentions, at least in the standard case.
A second important feature of exercising a moral power is the directness of the link. When I stop in front of your car, I might in fact intend to stop you. I might have even intended to impose a duty on you to stop. And true enough, I succeeded in making it the case that you had such a duty. But this intention does not figure in the explanation of why this new duty on part of the driver arose. You would have had a duty to stop no matter what my intentions. So a more precise formulation would be that, when a person exercises a genuine moral power, the intentions of that person are (normally) an ineliminable part of the explanation of the resulting change in the normative situation of the other person.
I return to this distinction, and some of the issues surrounding it, in section 1.4.1. For now, we can use it to give a precise statement of the standard or maximal approach: on this view, legitimacy is the possession of authority, understood as the moral power to change the normative situation of citizens through one’s will.
Let us return to the conceptual issue. We have three suggestions for what legitimacy at its conceptual core is:
- a mere privilege or permission—the weak view;
- a claim-right to non-interference, but falling short of obedience—the moderate view;
- a moral power, in particular a right to demand obedience—the strong or standard view
Corresponding to these three conceptual choices, you can advocate different substantive claims. Call anarchism the claim that, relative to a given understanding of legitimacy, no actual political institutions are legitimate or can be generally expected to be legitimate. This provides us with three anarchist claims:
- Actual rulers do not generally have a privilege or permission to rule.
- Actual rulers do not generally enjoy a protected liberty to rule.
- Actual rulers do not generally have authority—they have no claim to our obedience.
The last position is usually called philosophical anarchism, though this is a rather awkward label: most philosophers through history who called themselves anarchists—such as Godwin or Bakunin—would find philosophical anarchism a rather emaciated and tame view in comparison to what they advocated. Perhaps a better term would be anti-authoritarianism, as the central claim is that governments have no authority—but this label also has unhelpful connotations, so I will mostly stick with the standard term.
This thesis is predicated on the assumption that philosophical anarchism is true. In other words, I presume that rulers do not generally have a moral power to change the normative situation of their citizens, and they can not generally demand obedience from them (claim 3). At the same time, I reject anarchism about minimal legitimacy (claim 1). I accept that many actual rulers enjoy a privilege or permission to rule. One immediate upshot is that I resolve the conceptual question in favour of the weak view. When I speak of legitimacy in the rest of this thesis, I mean by this term a permission to rule. This conceptual choice is entirely on the basis of convenience. As I believe that rulers generally do not enjoy legitimacy in the strong sense, it is useful to reserve the term for a moral quality we can expect political institutions to have (at least sometimes). But not much hangs on this—if you prefer a stronger conceptual understanding of legitimacy, just substitute “permission” wherever I speak of “legitimacy”. In short, I will call the moral power to impose duties authority, and the permission to rule legitimacy—a convention shared by some other authors.
Formulating Philosophical Anarchism
With this out of the way, let me turn to two questions: first, what precisely does philosophical anarchism claim? And second, why should we believe it? I start with the first question. There are in fact two formulations of philosophical anarchism:
Narrow Formulation. Political institutions do not generally have authority over their subjects.
Wide Formulation. Citizens do not generally have a duty to obey the political institutions they are subject to.
The narrow and wide formulations are almost always elided in the literature. This is because of an implicit assumption that if citizens have a duty to obey, then governments must have authority over them. But this assumption has recently been challenged by various authors, justifiably so. I describe the problem in a moment. For now, let me note some other features of the wide and narrow formulation.
First, both formulations are weakened by the qualifier “generally”. Even radical philosophical anarchists accept that governments have authority, and citizens duties to obey, in certain situations. Many government employees have duties to obey the government because they have voluntarily agreed to work for it. Note, however, that government employees must obey the government insofar as their specific role is concerned; their agreement to serve in a specific function is unlikely to entail a general duty to obey all government directives. As a second example, there are some emergency cases where following the commands of an authority is necessary to achieve some essential moral aim. In war time, following the commands of your government might be necessary to guarantee that we successfully resist an unjust aggressor who intends to kill us all.
Philosophical anarchists are sceptical about how prevalent such cases are. Still, they accept in principle that we have duties to obey in such cases. So there might be specific groups of people or circumstances in which governments enjoy authority, and we ought to obey. This is no great embarrassment for philosophical anarchism as long as these exceptions are sufficiently rare. Note that a charitable reconstruction of the opposite position, of a defender of the duty to obey, should also allow for exceptions. To use a famous example, if you are alone on an empty traffic intersection in the dark of night waiting at a red light, it seems you do not even have a pro tanto duty to stop, even if the law demands that you stop. So the difference between the philosophical anarchist and their opponent is gradual rather than binary: both allow some exceptions to their views.
Let us now turn to the difference between the narrow and wide formulation. One direction of entailment is unproblematic: if political authorities have a right to demand obedience, then citizens have a duty to obey. This follows directly from the correlativity of rights and duties in the Hohfeld schema. But the other direction of entailment does not hold: we could have a duty to obey which is not grounded on the authority of government. Stephen Perry calls this the reverse-entailment problem. To see one instance of this problem, consider the following case. Imagine that all citizens promise each other to obey Lucinda. In this case, each citizen has a claim against all other citizens that they obey Lucinda, and everyone is under a duty to obey Lucinda. But it is not the case that Lucinda herself has a right to be obeyed: she is merely a beneficiary of an independent agreement. If Miranda fails to obey Lucinda, then Lucinda has no claim for restitution or compensation. Lucinda cannot demand that Miranda obey her—only other citizens can demand this of Miranda. The point is that Miranda’s duty to obey Lucinda is not owed to Lucinda.
I will give a more worked-out example highlighting, amongst other things, the reverse-entailment problem in the next section. Let us ask for now which formulation of philosophical anarchism we ought to prefer. Most philosophical anarchists wish to deny that we have duties to obey, so we should read them to advocate the wide formulation. I will proceed on this assumption, though later I will raise some doubts whether the core of philosophical anarchism should really be identified with the wide formulation.
The Negative Case for Philosophical Anarchism
Let us turn to the second question: why believe philosophical anarchism? There are two primary reasons to believe that philosophical anarchism is true. First, all existing attempts to defend that we have a general duty to obey have failed. Second, there are principled reasons to doubt that political institutions enjoy authority. I will discuss each of these two sets of considerations briefly, though I should say from the outset that I do not aim for a full defence of philosophical anarchism.
Let me start with proposals to defend a general duty to obey the law. Three main families of views can be discerned, usefully distinguished by John Simmons. Transactional theories ground duties to obey in some specific, voluntary transaction. Associative theories claim that we have duties to obey because we stand in some morally significant relationship to our fellow citizens or the political institutions we are subject to. Lastly, natural duty theories claim that we have some natural duties which, when coupled with certain empirical facts, give us a duty to obey the state. Each of these theories suffers from well-known defects—defects which are so general that they transcend particular formulations of each theory.
Consent is a relatively uncontroversial way by which governments could gain authority. However, transactional theories have to face Hume’s simple observation that few people, if anyone, have explicitly consented to be ruled by their government. In response, transactional theorists have two moves available, both unconvincing. First, they can weaken what it means to consent. Locke, for example, appears to say that it is acceptance of benefits, or residence on the state’s territory, which amount to consent to the state’s authority. While it is true that many people have agreed in these weakened senses to the state, these forms of agreement can simply not ground duties to obey. Perhaps we have duties of fair play to obey if we have received certain benefits from the state. But at this point, transactional theories collapse into a natural duty account, and the notion of consent ceases to play a relevant role.
A second move is to go hypothetical, and claim that individuals would agree to the state, if only they were fully reasonable. This fails on the basis of Dworkin’s famous dictum that a “hypothetical contract is not simply a pale form of an actual contract; it is not a contract at all”. That I would agree to be subject to an authority under some circumstances does not make me subject to that authority. At this point, one sometimes encounters the claim that we should obey because we ought to consent to the state. But we ought to consent, presumably, because the state is good, or confers benefits, or because we have some other natural duties towards it: so the view again collapses into a non-transactional account.
Associative accounts fare no better. Admittedly, we sometimes have duties to obey others because of our relationship to them, even though we have made no explicit agreement to obey. The relationship of parents to their children might be such a case. But expanding this analogy to the citizen—state relationship is subject to a simple dilemma. On the first horn of the dilemma, our relationship to the state is in no way like that between a child and parent. Indeed, thinking about our relationship to the institutions which govern us in this way paints a deeply patronising picture, one which we should not accept as liberals. On the second horn of the dilemma, it is true that we stand in many weaker relationships vis-à-vis our fellow citizens: we share cultural, social and linguistic bonds. Some of these bonds might ground special duties towards our compatriots. But however you try to cut these duties, it is hard to see how you get a duty to obey out of this material. Merely qua being British, you have no duty to obey the British Heritage Society, and in the same way, no duty to obey British parliament.
This leaves natural duty accounts, which come in different flavours. On fair play accounts, you ought to obey the state because others obey it, and your failure to obey would free-ride on their contribution. You owe it to your fellow citizens as a matter of fairness to share the burdens of producing the goods the state brings about. There are two problems with this line of argument, one empirical and one principled. First, the general obedience of each individual is hardly necessary to bring about the goods the state produces. Second, and more fundamentally, it seems repugnant how others could force you into obedience, especially if it is very costly for you to escape the benefits of the state. Perhaps you benefit from the security brought about by a national army, but that hardly gives you a duty to conscribe into that army.
Another approach is Samaritanism. On this theory, each of us needs to obey the law because doing so is necessary to escape the horrors of the state of nature; we have a duty to assist others in escaping those horrors. There are at least two problems with this view. First, the account is empirically questionable. Insofar as obedience is required to escape the state of nature, this is at most so for a very narrow set of the state’s directives. Most of the state’s commands are utterly unnecessary for the purpose of fulfilling this duty. Second, the account fails the so-called particularity requirement. It is unclear why I ought to obey the directives of my government, rather than the directives of some other government, to fulfil my duties of assistance. Other natural duty approaches suffer from similar problems as Samaritanism. On a Rawlsian account, we have duties to obey because we have a natural duty of justice—a duty to support, promote, and uphold just institutions. Again, it is hard to see how this argument can provide us with a general duty to obey, and even less clear how it can provide us with a specific duty to obey my state.
The Positive Case for Philosophical Anarchism
These are all well-known points from the literature. A huge philosophical industry continues to produce new attempts to refine these positions. But my discussion should be enough to show that support for philosophical anarchism rests on strong and general considerations. There are also some positive reasons to favour philosophical anarchism, which tend to be less clearly developed in the literature. Let me highlight, again briefly, some of these.
We can start with an unconvincing argument, one that Robert Paul Wolff has famously given. Wolff starts from a purportedly Kantian claim that each of us has a duty to act autonomously—that is, a duty to weigh by ourselves the merits of any particular course of action open to us, and to act on the basis of such deliberation. Obedience, however, requires that we surrender judgment: it demands that we act merely because we have been commanded to act, independent from the merits of so acting. Thus, Wolff concludes, governmental authority is fundamentally incompatible with our autonomy. Given that we have a duty of autonomy, no government could possibly have a claim to our obedience. Note how sweeping the argument is: if correct, it would show a priori that governments do not have authority. We can see how implausible Wolff’s claim is once we focus on the more general case of people exercising moral power over each other. A standard way in which Anne can gain a moral power over Ben is if Ben makes a promise to Anne (“I am at your disposal on Tuesday!”). If Wolff’s argument were right, making promises would be wrong, as keeping a promise also requires me to “surrender judgment” in an important sense. But surely this cannot be right.
Still, there is something worthwhile which Wolff’s argument helps to bring out. While authority and autonomy are not incompatible in the drastic way Wolff described, authority surely diminishes our autonomy in some respect. If I have a duty to be at your disposal on Tuesday, then I am less autonomous in this respect. This problematic aspect of authority can be sharpened if we refocus on the account of moral powers I gave. John has a moral power over Marissa if John can impose new duties on Marissa through his will, through him intending that Marissa ought to have a new duty. This surely gives John great power. At the same time, it puts Marissa (assuming that she has no reciprocal power of the same form over John) at a disadvantage. It appears to make Marissa John’s inferior: she is now subject, in some respects, to John’s will. Because moral powers are such awesome qualities to have, we expect that there is a high justificatory burden for showing that anyone has such a power over anyone else. There is a heavy liberal presumption that we are equal in authority, and that by nature no one enjoys authority over anyone else. So philosophical anarchism is heavily favoured as a baseline.
A second observation in favour of philosophical anarchism closely follows from the first once we focus on how moral powers in general are justified. Marriage vows, for example, plausibly give rise to moral powers. A marriage vow gives you a power to impose duties—at least strong reasons—on your partner through your willing, from the mundane (“wash the dishes”) to the life-changing (“let us move to Australia”). There are some features of this case which make it unproblematic. First, the powers we gain on each other are mutual—you can make those claims, as can your partner. Second, we enter such relationships voluntarily. Third, we give up our autonomy to form a relationship which is in itself valuable. Political authority, however, does not share these features. We do not share mutual authority with the state—the relationship is entirely one-directional. Second, our subjection to the state’s authority is not standardly voluntary. Third, it is also difficult to think that being subject to the state’s authority is somehow part of forming a valuable relationship. We gain no love, or friendship, or other intimate bonds, from being subject to the state’s power.
These considerations certainly do not amount to a conclusive proof of philosophical anarchism. But the negative case, taken together with the positive case, strongly suggest that we should take philosophical anarchism seriously, and I will proceed on the assumption that it is true. This moves us to the next question: what comes beyond authority? In other words, which forms of political institutions, if any, can we still defend if those institutions generally do not enjoy authority?
There are two general kinds of answers. First, we might give pessimist answers: without authority, only very restricted kinds of political institutions can be legitimate (remember: in the sense of being permissible). Such a pessimistic answer is given, for example, by John Simmons. For him, non-consensual political institutions without authority can still pursue some restricted activities, simply because everyone would be permitted to pursue such activities. For example, a government without authority could still punish murderers, or protect the innocent from assault. It might also permissibly organise voluntary neighbourhood watches, and trade on the free market. However, such political institutions could not tax or coerce widely, or make binding law. They are simply private agents writ large.
Another response would be an optimistic philosophical anarchism, or using the alternative label I suggested, optimistic anti-authoritarianism. On this view, governments can do quite a lot even without authority. Once we rid ourselves of the outdated assumption that the state is like a parent or a military commander or a spouse, we discover that this is not the end for political institutions. Developing this case is the task of this entire thesis; for now, I pursue two initial aims in the rest of this chapter. First, I complicate philosophical anarchism by showing that, even if political institutions do not enjoy authority, we will still have various duties towards them. Second, many authors have too quickly dismissed optimistic anti-authoritarianism. For them, showing that government has authority is crucial to showing it to be legitimate. I argue that such objections are generally unsuccessful.
Duties without Authority
It is a commonplace to observe that philosophical anarchism does not entail that you should oppose the state. The fact that you have no duty to obey someone does not entail that you should seek to undermine them. (Think of your neighbour’s parents when you were a child.) A second point commonly made by philosophical anarchists is that even if you ought not to obey the commands of the state, you might still have a duty to conform with its commands, simply because the state mandates what you had a natural duty to do. You ought to conform with the state’s directive not to murder, simply because you ought not to murder; there is no great mystery to this duty. So if we subtract duties to obey from our duties, we are not left without any duties vis-à-vis the state’s directives. I will elaborate on this picture in this section to highlight what kinds of duties we are likely to have under philosophical anarchism.
The Example of Coordination
It is useful to start from a much-discussed example, that of coordinating action. Many of our natural duties can only be discharged, or can only be discharged conveniently, if we coordinate our behaviour. For example, I have a natural duty not to endanger others. But in many situations, I need to be able to predict the behaviour of others, and know their expectations, to avoid endangering them. If I am driving a car on the highway, I need to rely on a conventional system of traffic lights and signs, a shared symbolic understanding of turn and brake lights, an agreed-upon set of rules for determining the rights of way, and so on, to avoid harming others. If we have different interpretations of what a red light means, or who has the right of way in an intersection, we will both endanger each other. So in this instance, a natural duty requires that I coordinate my behaviour with others.
We can boil down this structure to its essentials by focussing on the case of Joe, who I assume to be a traffic coordinator on an imaginary intersection. Joe, by a set of generally understood hand gestures, signals to drivers which cars are allowed to proceed into the intersection, and which ought to stop. Assume that Joe is doing a decent job at directing traffic, and that there are no independently established conventions on rights of way. If both you and I arrive at Joe’s intersection, then each of us has a reason to follow Joe’s hand signals. Joe does not impose new duties on me, he merely changes how I ought to discharge duties I already had. From our definition of authority, it follows that Joe does not exercise authority over me: Joe has no moral power over me to change my normative authority through his will in the relevant sense.
The issue is complicated, so it is necessary to be precise. Joe’s directives have some classic features of authority. First, if Joe orders me to stop, then I ought to stop, independent of the merits of Joe’s command. Perhaps it is somewhat inefficient for me to stop now. But my duty to coordinate my behaviour is strong enough that I ought to conform to Joe’s directives as long as he is an overall tolerably efficient coordinator. That is, I ought to stop, drive, or turn, merely because Joe signals me that I ought to stop, drive or turn. So it might look as if I have a genuine duty to obey him. All this, however, can be explained on philosophically anarchist grounds without assuming that Joe has authority. That I have a duty to do what Joe says does not show that Joe had authority. Here, we can see the reverse-entailment problem at work.
We can bring out Joe’s lack of authority in two other ways as well. First, I do not owe my obedience to Joe. Assume that I ignore Joe’s signals, haphazardly driving through his intersection. I acted wrongly. But I did not wrong Joe, I wronged my fellow drivers. Joe can point out to me that I acted wrongly, but he is in no special moral position in this respect. A second point is to look at the relevant counterfactuals. Imagine a different society in which traffic coordinators are unnecessary: traffic is not that heavy, and there are some rough conventions in place regarding rights of way. If Joe stood on his intersection now, I would have no duty to obey him—though it might be polite or convenient to obey, of course. In this Joe differs from Joanne, whom, I assume, we promised to obey. We would continue to have a duty to obey Joanne’s traffic directives, even once the coordination problem has fallen by the wayside. This simple contrast shows that Joe’s intentions are an eliminable part of explaining why our normative situation changed. Indeed, if Joe was a weathervane which we used as an indicator for rights of way, he would change our duties in just the same way.
What is the upshot of all this? Joe enjoys no authority, in the relevant sense of having a moral power. He merely enjoys a liberty-right to direct traffic. However, because he happens to be a coordinator in an empirically salient position with morally important stakes, his directives happen to trigger duties in others. Call this a transmission channel: an indirect connection through which Joe’s commands cause duties in others to exist. This makes Joe’s case a paradigm of how a philosophical anarchist can explain individual duties without authority.
This is not the standard message which authors have taken away from coordination cases. Many have thought that it shows that citizens have a duty to obey, or that coordinators like Joe enjoy authority. These claims rest on an insufficiently precise analysis of what it means to possess a moral power. However, even if we accepted one point, namely that there is a duty to obey in this particular case, this would not be a great concession. First, because of the reverse-entailment problem, acknowledging that citizens have a duty to obey does not entail that government has authority. Second, we are only considering a single case. So even if the necessity of coordination can show that there is a duty to obey in some cases, it fails as a general explanation of a duty to obey. A third point goes back to the difference between the wide and narrow formulation of philosophical anarchism. I argued that one thing that troubles the philosophical anarchist is that some should possess moral powers over others. But on closer inspection, this supports the narrow formulation of philosophical anarchism, not the wide. There is something problematic in obedience insofar as it is grounded in others having moral power over us; but there is nothing problematic in obedience as such—or at least, obedience as such is much less problematic. This can be seen in Joe’s case. Many of the standard lines of objecting to obedience do not apply to his case, simply because our duty to obey him (if there is such a thing) is an emergent property from the nature of the coordination problem we face. This should alleviate many of the moral objections we might have.
I merely used the coordination case as one example. We need not solely rely on the coordination account to explain individual duties to conform with the directives of political institutions. There is a whole range of transmission channels through which government action translates into individual duties. Return, for example, to the issue of fairness. Assume Joe exercises a liberty-right in providing a public good. This might trigger fair-play-based duties that we do our share in some cases. This way, even though Joe can not directly impose duties on us, his actions indirectly transmit into individual duties to act in a certain way. Again, I owe no obedience to Joe.
The picture which emerges is the following. I will argue in later chapters that government has a permission to rule. However, except in rare cases it has no authority—that is, no moral power over the normative situation of those it rules over. It cannot change my normative situation at will. However, we all hang together in an intricate web of empirical relations. I feel the actions of political institutions through the waves they send through that net, and those waves will change what duties I have. Thus, through a complex set of transmission channels, many of us will have duties to comply with governmental directives across a decent range of cases. It is often useful, in the face of this complicated web, to pretend as if we had a general duty to obey. Conformity with the law might be a good rule of thumb, and law-abidance might be a virtue. It might even be useful to pretend that the law presents itself to us as a “seamless web”. But in the end, there is no general duty to obey the law, and generally political entities do not enjoy authority.
Law-Making without Authority
We can extend Joe’s case to law-giving as well. When Joe directs traffic, he does so by simple directives. But actual states attempt to coordinate behaviour by much more complex means—in particular, through law. Can law-making without authority be explained? Remember the multiple ambiguities surrounding the idea of authority. Our precise question is whether law-making can be explained without a power to change the moral situation of citizens. The first part of an answer is the observation that philosophical anarchists can easily acknowledge a power to change the legal situation of citizens at will. A government can give individuals legal titles, or legal duties, or find them in breach of legal requirements. This is not a mysterious power if we accept that there is a separation between legal and moral norms.
Imagine that Joe stops standing on his intersection. Instead, he develops a traffic code of his own, which he starts publicly announcing, telling people to abide by it. Joe has the power to change his code at will. Joe also has the power to find others in violations of his code as he sees fit. He can write letters saying others violated his code, or tell them that under it they have certain rights when they enter an intersection. Through Joe intending to make those changes, people’s Joe-code-related rights and entitlements change. None of this, however, translates directly into any change in our moral situation. Everyone is free to make up their own code, and tell people how they fare by it. But we have also established that Joe is a coordinating authority. Assume that we come to see that Joe’s code is useful and efficient—more efficient than a system of hand gestures. If compliance with Joe’s code becomes wide-spread, then it will eventually become morally relevant to me as well, as I ought to coordinate my behaviour with others. If Joe changes his code now, so will my moral duties and rights. But again, this is not a case of authority—not a case of exercising a moral power—but Joe merely triggers already existing duties.
It is here useful to comment on Arthur Applbaum’s concept of legitimacy, which shares some important similarities with my view. Applbaum develops a view on which legitimacy is the ability to “create and enforce nonmoral […] prescriptions and social facts”. It would be natural to rephrase this as the claim, wholly compatible with the minimal conceptual approach, that legitimacy includes the moral permission to exercise legal power. However, Applbaum rejects this possibility, claiming that his position does not reduce to minimalism. But his reasons for rejecting this move are confusing. He writes:
[(1)] the creation of a legal right or duty by a legal power can change the normative situation of a subject of the legal right or duty by changing his moral rights and privileges, powers, and immunities even when no moral duty to obey the legal power has been generated. [(2)] Having the power to change these moral statuses is moral power.
Claim (1) is correct: creating legal rights and duties can change our moral rights and duties, through the various transmission channels I described. But the inference from this to claim (2) rests on a mistake. From the fact that someone has the capacity to cause changes in the moral situation of others we cannot infer that they have a moral power to do so in the strict sense. This is precisely the difference between mere reasons-triggering and the exercise of a moral power I sketched in section 1.2.3. As far as I can see, Applbaum offers no other reason to believe that claim (2) is true, or that (1) entails (2).
Doubts about Minimal Legitimacy
The picture of duty transmission provides us with the ability to answer some objections that have been brought against anti-authoritarian optimism. Let me turn to some of these.
A Pale Shadow
We can start with Christiano’s dismissal of minimal legitimacy, found prominently in his article on legitimacy in the Stanford Encyclopedia of Philosophy. Christiano starts with the following two claims:
[(1)] in the case of an authority as merely justified coercion [in our words, merely permissible coercion], the subjects’ reasons for obedience are merely their desires to avoid punishment. […] [(2)] [it] is the case that subjects have duties but those duties are not essentially connected to anything in the authority. The subjects instead act more in accordance with reasons that are independent of the authority when they obey the authority.
The first claim is false, the second misleading at best. First, there is no reason to think that in the absence of the state having authority, individuals only have prudential reasons to conform with the state’s directives, or merely act to avoid punishment. As we saw in the coordination case, a host of genuinely moral reasons will combine to give individuals reasons to conform with the state’s directives. The second claim is a bit more on track. It is true that subjects’ duties are not essentially connected to anything in the authority. Subjects’ duties are only contingently connected to the state’s directives. In this respect, Christiano’s description is true. But what, at any rate, is the problem with this? Christiano continues:
[(3)] Such a society does not engage the subjects as moral persons, it merely attempts to administer the activities of persons so as to bring about in a morally justified way a desirable outcome. […] [(4)] To the extent that a political society is best when it involves the mutual recognition and affirmation of the moral status of each person, the kind of society that involves merely justified coercion of some by others is a pale shadow.
Christiano also provides us with an example. He compares a legitimate state without authority to the gaolers in a prisoner of war camp. The gaolers can justifiably use coercion to constrain their prisoners, but they have no claims against them.
Christiano’s third claim is puzzling. Why would a minimally legitimate government not be able to engage its subjects “as moral persons”? Assume that you and I are strangers. We have entered no contracts, made each other no promises, and have no other communal bonds or political relationships. In this situation, none of us has authority over the other. What stops us, in this case, from respecting each other as moral persons? In fact, respecting each other as moral equals becomes more difficult once we introduce moral powers, precisely because moral powers always introduce some inequality. If our relationship is that between landlord and tenant, or boss and employee, one party has a moral power that the other lacks. If anything, would a society not be more equal if it lacked such relationships in the political context?
At any rate, perhaps Christiano’s main concern is with coercion, as his language and the example of the prison camp suggest. If a concept of minimal legitimacy implied that states were essentially like prison camps, then this would indeed be a strong objection. Gaolers and prisoners are enemies, not friends or allies. There is no reason to think, however, that this is a necessary feature of relationships under minimal legitimacy. Think back again to Joe’s case of coordinating traffic. People there are not enemies; they are co-operators, if perhaps somewhat anonymously. Minimally legitimate governments need not engage their citizens merely through the instruments of whip and shackle.
Combat and Competition
Let me turn to a related worry which might stand behind Christiano’s objection. Perhaps he worries about moral combat, situations where morality plots us against each other. Moral combat is a situation in which X can permissibly force Y to φ, and Y can permissibly resist X’s attempt, and is allowed to use force in resisting. In the classic example, two shipwreck survivors are fighting for the plank which will save only one of them from drowning. In our context, moral combat would be a situation where government permissibly coerces its citizens who are permitted to fight back. Our situation would be like that of a prison camp, only worse: the gaolers try to forcibly hold the prisoners in, the prisoners try to forcibly escape, and all act permissibly.
A weaker idea is that of moral competition. If X and Y morally compete, then there is one function, F, which can only be fulfilled by one party, and it is permissible for both X and Y to attempt to F. For example, sellers in a market are all permitted to try to sell their goods to the same buyer. In our context, moral competition would mean that non-political agents are permitted to fulfil the same functions as the state. For example, Thomas Senor claims that, if the state did not have a claim-right to rule, then the Boy Scouts of America would be permitted to punish a car thief, as would be the police. Ryan Windeknecht similarly objects that if states did not enjoy authority, then Wal-Mart would be as permitted to provide disaster relief as the US government.
What should the philosophical anarchist reply? As I outlined the minimal concept of legitimacy, it does not include a claim-right against interference or even resistance, or a claim-right to exclude others from fulfilling some function. So we should accept that moral combat and moral competition are conceptually possible on the minimal view. However, it is a separate question altogether whether they are empirically likely. Focus on Joe’s case. When Joe first steps on the intersection and starts directing traffic, he is merely permitted to do so—he has no special rights against interference. But assume that Joe over time establishes himself as traffic coordinator. Assume that Lucinda envies Joe, and wishes to take over his role. She starts distracting him, stands in his way, gives competing signals to drivers, and tries to establish a different traffic code. However, Lucinda would not be permitted to do so, because she would undermine and destroy the coordinating conventions established by Joe, compliance with which establishes a valuable good.
So there is no moral competition between Joe and Lucinda: Joe is allowed to direct traffic, Lucinda is not. But crucially, our explanation does not invoke a claim-right of Joe to exclude Lucinda. Again, our explanation is indirect: Joe has changed Lucinda’s non-moral situation in a way—he has triggered certain duties—such that she is not permitted to intervene. In a next step, we would need to show that Lucinda’s duty to abstain from interference is also enforceable against her by others. If Lucinda interferes with the provision of an important public good, then perhaps everyone is allowed to stop her from doing so. If this is the case, then Joe would have a permission to exclude Lucinda from regulating traffic. There is of course no guarantee that this transmission is automatic. If Joe’s effectiveness as a coordinator starts to fade, then at some point Lucinda will be permitted to compete with Joe. We can exploit the same general model of explanation to alleviate doubts about Senor’s and Windeknecht’s cases. For example, it is impermissible for the Boy Scouts of America to punish car thieves, because doing so would undermine the activities of a central coordinator, which sets unitary rules and enforces them; it is likely that we have fairness reasons to leave enforcement to such institutions. This in turn makes it impermissible for the Boy Scouts to punish car thieves.
A critic might pursue two replies. First, she might argue that even the conceptual possibility of moral combat and moral competition is troubling. But I fail to see why it would. Indeed, in some cases they are eminently sensible. If Wal-Mart could indeed efficiently provide disaster relief, then it seems permissible for it do so. The second reply is simply to challenge the philosophical anarchist on the details. Can the philosophical anarchist really explain that only the state can do X, or that citizens cannot resist the state imposing Y, etc.? An answer to such cases would need to rely on a fully developed account of the transmission channels through which individual duties are propagated. The answer of the philosophical anarchist will be partially revisionary, and partially reconstructive. But I see no principled reason why such an answer cannot be given.
The Argument from Analogy
The next objection I wish to consider comes from Christopher Wellman. He writes:
When it comes to slavery, it would be highly implausible to adopt a stance of ‘philosophical abolitionism,’ insisting that slaves have no moral obligation to obey their masters but denying that slavery is necessarily impermissible. No one would endorse such a muted form of abolitionism because the same feature of slavery that explains why slaves are not obliged to obey their masters (i.e., nonconsensual coercion) equally explains why it is impermissible to enslave others. Thus, it seems at the very least awkward to support the analogous position regarding political states […] which insists that citizens have no moral obligation to obey the law but denies that there is anything morally problematic about imposing this law in the first place.
Wellman’s argument, put simply, is the following: the institution of slavery is unjust. For this reason, slaves have no duty to obey their masters. But the same reason—that the institution is unjust—also pushes against considering the overall institution to be legitimate in the more minimal sense, as being permissible. The same is the case, Wellman suggests, for government. If you believe that we have no duty to obey your government, then the same reason supports the conclusion that government is illegitimate. If this argument it correct, then optimistic anti-authoritarianism is in trouble.
Wellman’s imagined “philosophical abolitionism” is indeed nonsensical. Because slavery is unjust, it is impermissible. However, I rejected philosophical anarchism not because I claimed that political institutions are generally unjust. I claimed that it is mysterious how political institutions could have the moral powers that they profess to have, moral powers which are usually only gained by agreement or promise. I also went through a series of proposals in favour of a duty to obey, all of which I rejected. But none of my objections rested on the ground that government is unjust, or in some other way morally objectionable. So my rejection of the authority of the state was based on quite specific grounds, which cannot be leveraged into a general rejection of legitimate political institutions. Consider an analogous case. You might advocate anti-authoritarian education. That is, you reject that children ought to obey their parents. You reject this claim for specific reasons: you think, for example, that an authority-relation between parents and children is patronizing or has negative effects. However, this claim does not commit you to reject parenting altogether. There is nothing in general patronizing about parent-child relationships.
We can, of course, construe Wellman’s objection as the more straightforward charge that the state is unjust, or morally flawed in some other respect. If this were true, then it could not be legitimate in the minimal sense. I have not yet provided an argument that the state is not unjust—I turn to this charge in the next chapter.
Punishment without Authority
Another objection comes from Thomas Senor. He starts with the
Punishment Principle. If A punishes B for violating a natural duty, without having a claim-right to do so, then A acts unjustly (i.e., impermissibly).
If the punishment principle was true, anti-authoritarian optimism would again be in trouble. Punishing is one of the main activities of the state; so a state without authority would be fairly limited. However, the principle is not particularly plausible. It contradicts a powerful intuition that everyone is entitled to enforce the natural law without any special right. According to this intuition, it is permissible for X to coerce Y, even without X having authority over Y, if Y has wronged someone else.
Senor provides us with another example which is supposed to show how government without a claim-right to coerce is problematic:
If [A] steal[s] [B]’s car, that does not give [C] the right to take [A’s] boat in retribution, even if taking [A’s] boat is a perfect punishment for [A] having stolen [B’s] car.
What is troubling about the given example is that C profits from punishing A, even though it was B who had a special claim for compensation. Thus, C’s intervention, unless authorised by B, wrongs B. So what is wrong with C’s intervention is not that she punished A without a right, but that she violated B’s special right in the process. We can use this to roughly state a new principle,
Narrow Punishment Principle. If
then C acts unjustly (i.e., impermissibly).
- C punishes A for violating a natural duty A had against B,
- C has no authority over A or B,
- B has not authorised C to punish A, and
- B was in a position to authorise C to punish A,
Although this principle is more limited in scope, it still poses a problem for the idea of a permission to rule without authority. For example, the state regularly punishes people who have wronged other people. The victims have generally not authorised the state to punish the perpetrators. Thus, the state is infringing the victims’ rights to punish the perpetrators themselves. This would make the state acting in such ways impermissible.
You might reject that victims have such rights. However, we can generalise Senor’s objection. If people have certain rights—of which an exclusive right to punish wrongdoers is only one—and the state has no authority, then it seems that the state wrongs its citizens when it punishes them. This point is independent from whether you agree with Senor’s proposed punishment principles. Ronald Dworkin, for example, writes:
though obligation is not a sufficient condition for [permissible] coercion, it is close to a necessary one. A state may have good grounds in some special circumstances for coercing those who have no duty to obey. But no general policy of upholding the law with steel could be justified if the law were not, in general, a source of genuine obligations.
There are different ways to read what Dworkin is aiming at, but the fundamental worry is clear. If the state has no authority over its citizens, how can it coerce them? Private agents are certainly not allowed to coerce others in the many ways the state does. A standard example to bring out this worry is by focussing on the issue of taxation. How can the state forcibly take money from its citizens if it enjoys no authority over them? Thus, we can see, the problem is not so much a lack of authority, but rather the fact that other people have rights. What each of us can permissibly do is constrained by those rights. The argument, in a rough form, can be boiled down to the following:
- People have rights—e.g., a right to punish perpetrators themselves.
- You (usually) need authority to permissibly infringe those rights.
- Thus, a state without authority (usually) acts impermissibly.
- Thus, a state without authority (usually) acts impermissibly.
This is the most serious charge against optimistic anti-authoritarianism, which will concern us for the next three chapters.
The State and Its Alternatives
I have so far written generically about “rulers” and “political institutions”, not states. There are several reasons to prefer this focus. First, insofar as legitimacy is the right to rule, there is no principled reason to think that there is a special theory of state legitimacy, rather than a general theory of legitimacy which extends to all types of political institutions. The current theory applies equally to states, super-national institutions such as the EU, and sub-national institutions such as a supreme court, insofar as all these institutions rule in a relevant sense. This brings us to the question what it means to rule. On any plausible view, the right (or permission) to rule will be a bundle of separate rights (or permissions). A ruler might have some of these without having others. Following the classic scheme of distinguishing three branches of power, we can roughly distinguish
- legislative rights: the right to make laws;
- adjudicative rights: the right to resolve conflicts between individuals;
- executive rights: the right to enforce the law.
Modern states claim for themselves all three rights at the same time, but this is an historical contingency. In the Roman empire, for example, the state would adjudicate legal conflicts, but enforcement of verdicts was generally left to individuals. We can also imagine a system where there is a unitary source of law, but individuals can choose between different, competing courts when it comes to adjudicating legal disputes. So it is by no means necessary that all three kinds of rights are held by one agent.
There are other features of the (modern) state which, historically speaking, set it apart. One is that the state does not only claim to have the three kinds of rights mentioned, but also to have a monopoly on them. Monopoly rights are second-order rights to exclude all others from exercising some function. A right to a monopoly on legislation, for example, is a right to exclude all others from making law for a certain group of people or territory. States typically claim to have a monopoly to make any kind of law, and to exclude others from making law in any area of human activity. But claims to have monopoly rights are not a timeless feature of political institutions. The medieval Holy Roman Empire, for example, was a complicated power-sharing agreement between the emperor, his arch-ducal electors, princes and independent cities, the church, and others. None of these institutions could claim to have unrestricted monopoly rights to any of the three central functions of government.
A last feature which the medieval example brings out is that modern states generally make first-order claims and monopoly claims with regard to a certain territory. That is, states stake out separate territories and then claim to have the exclusive right to rule with regard to everyone on that territory. Again, this is a contingent feature of political institutions. In medieval times, authority was often person-based or role-based, rather than territory-based. Guilds, the church, and universities, for example, had (non-monopolistic) rights to rule over their members, but there was no generally defined territory belonging to those authorities.
All this goes to show that the prevalent question in the literature—what justifies the state?—can be broken down into a series of more specific questions, at least:
- what justifies a ruler’s legislative rights?
- what justifies a ruler’s monopoly on legislative rights?
- what justifies a ruler’s adjudicative rights?
- what justifies a ruler’s monopoly on adjudicative rights?
- what justifies a ruler’s executive rights?
- what justifies a ruler’s monopoly on executive rights?
There are other questions we might ask, and we might break these seven down into more fine-grained ones. The important point is that we might be able to justify some rights to rule of a given ruler, but not all of them. Monopoly rights in particular are generally quite hard to justify, and I admitted that rulers will often be in a situation of moral competition—that is, they will not have a monopoly on the right to rule. This is an important limitation in the aims of this thesis. My focus will be on whether non-consensual ruling can be justified; but the kinds of non-consensual ruling which can be justified in this sense might fall significantly short of what we normally recognise as a state, especially insofar as its monopoly claims are concerned.
You might think this confusing, as I also accepted non-minimalism as a commitment for this thesis. But justifying non-minimal ruling does not entail justifying the state. The anti-minimalist idea is that political institutions are not limited to the most basic functions of keeping order. What the anti-minimalist is opposed to is the night-watchman state, and what she wishes to defend are political institutions which are engaged in (say) redistributing resources. Anti-minimalism in this sense is a substantive claim about legitimate functions that political institutions can and ought to fulfil.
Indeed, a defender of the night-watchman state, if understood to be a genuine state with the features I outlined, might have a more extensive vision of political institutions than an anti-minimalist in this respect. Too often, those with a non-minimalist vision for the functions of political institutions too easily align their project with statism. But the one does not require the other. Return to the medieval case. Imagine a set of overlapping, non-consensual institutions, each of which has only limited rights to rule. Each of these authorities fulfils a small part of the functions of the modern welfare state, but cannot be genuinely seen as a state in its own right. However, the combined impact of this dispersed set of institutions might still be the same. To put the point differently, when the Holy Roman Empire transformed into separate political entities more clearly identifiable as states, we can imagine that the impact of political institutions felt by individuals need not have been greater or smaller. Other things being equal, there is no reason for the anti-minimalist to care for one type of political organisation over the other.
One might object that it is impossible for political institutions to fulfil the more expansive functions the non-minimalist wishes for without also having an extensive set of rights to rule, and in particular, monopoly rights. But even our very cursory look at the history of political institutions suggests that this is not generally true. If, on the other hand, the organisation of political institutions in the form of states were indeed necessary to achieve certain important benefits, an instrumentalist approach to legitimacy is well-placed to take this into account. If a system of territorially based states with strongly recognised monopoly rights best promotes justice, then the rationalist is likely to endorse such a system. But establishing the antecedent of this conditional is a big ask.
In short, defending the legitimacy of non-minimal, non-consensual ruling is a difficult task; defending non-consensual states is even more difficult. I will only tackle the former. Nonetheless, it is often useful to focus our discussion on states. I do so simply because states are the best-known type of political institutions. The majority of philosophers have also written about the state. However, we should keep in mind that I provide no general defence of the state, and merely use it as an accessible example.
The Basic Problem
Individuals have rights, and there are things no person or group may do to them (without violating their rights). So strong and far-reaching are these rights that they raise the question of what, if anything, the state and its officials may do.
Nozick’s opening lines of Anarchy, State and Utopia summarise succinctly the problem which will concern us for this chapter and the next two. I will call it the Basic Problem. Its outlines are simple: if people have rights, these rights heavily constrain what rulers might permissibly do. If rulers do not enjoy the consent of their subjects, or have not in some other way gained authority over them, much of what they do violates the rights of individuals. This makes them illegitimate. Thus, any interesting form of ruling must be based on consent to be legitimate. This would invalidate the rationalist approach I advocate.
The task in this part of my thesis will be to disarm this problem. My ultimate solution will depend on a claim about the nature of our rights (ch. 4)—in other words, we have to go into quite deep theoretical grounds to escape the problem. But let us not get ahead of ourselves. This chapter has two aims. First, I will state a precise version of the Basic Problem which allows us to categorise attempts to escape it. Second, I will reject several proposals to escape the Basic Problem one can find in the secondary literature—this will take up the rest of the chapter.
The Silence of the Liberals
Before we start, I wish to tackle one serious misconception: namely, that the Basic Problem is only a problem for libertarians. It has been a long-debated question in libertarian circles whether a minimal state is compatible with the strong ownership rights that libertarians ascribe to individuals, with libertarians being split roughly in the middle. At the same time, a discussion of the problem has been curiously absent in “high liberalism”. What I have in mind is philosophy done in the egalitarian, Rawlsian tradition, exemplified by philosophers such as Ronald Dworkin, Amartya Sen, and G. A. Cohen. High liberals accept that individuals have rights, and they also accept that rights act as constraints on what can be done to individuals. But while one can easily find high-liberal discussions of how rights constrain the state in specific ways—say, with regard to the promotion of religion, or in its pursuit of perfectionist aims—there is little to no awareness that individual rights might undermine the state quite generally.
One reason for the high liberals’ neglect of the Basic Problem might be the implicit assumption that the problem only arises if you accept a Lockean theory of individual rights, the one favoured by libertarians. On a Lockean theory, all important rights are property rights. On a rough statement of this view, property is taken up through original acquisition subject to some proviso constraint, and all permissible interactions afterwards rely on voluntary transfers between individuals. It is indeed difficult to see how the state could legitimately arise on this picture without consent. However, high liberals reject the major elements of the Lockean story. They reject that all rights are property rights, and they are likely to acknowledge that individuals have welfare (“positive”) rights. In addition, the property rights high liberals assign to individuals tend to be more restricted.
However, a Lockean account of individual rights is not necessary to get the Basic Problem off the ground. Instead, we can begin with an appealing liberal idea, that individuals should be masters over their own affairs. That is, they should be in control—they should have a wide range of freedoms regarding how they can lead their lives. This sounds rather abstract, but is brought out by a wide range of areas of individual morality. A paradigm case is sexual morality. It is a hallmark of liberal morality that (simplifying somewhat) sex is permissible just in case it is consensual, and impermissible otherwise. Liberals believe that, as far as sexual morality is concerned, individuals have strong rights against interference, and consent is the major (if perhaps not the only) dividing line between permissible and impermissible interference.
Liberals think the same in the area of contracts. Individuals should be in control of whom they associate with, what agreements and contracts they enter, and so on. Examples of how consent is crucial can be multiplied at will. Whether it comes to sex, friendships, contracts, occupations, or marriages, liberals believe that individual consent is central. As Heidi Hurd puts it, consent turns “a trespass into a dinner party”, “a battery into a handshake” and “a theft into a gift”. Nothing in these observations commits us to a Lockean theory of rights. However, when taken together, they suggest that individuals have a general right against having their freedom restricted without their consent. It is easy to see how this gets the Basic Problem started. Rulers interfere with our lives in myriad ways. No private individual could permissibly have this kind of impact on us without our consent. Thus, while rejecting the Lockean theory of rights will change some of the goalposts of the debate, it does not at all eliminate the Basic Problem.
Beyond rejecting the Lockean theory of rights, it is hard to say what the high liberals’ general strategy to avoid the Basic Problem is or would be. One hint is contained in Thomas Nagel’s review of Anarchy, State and Utopia, where Nagel dismisses Nozick’s version of the Basic Problem. One major part of Nagel’s criticism is indeed a rejection of Nozick’s Lockean theory of rights—Nozick’s “moral intuitions seem wrong” when he assigns to individuals absolute property rights. More interestingly, Nagel also claims that rights “limit the pursuit of worthwhile ends, but they can also sometimes be overridden if the ends are sufficiently important”. So Nagel appears to rely on a public goods defence of government: there are certain public goods, and they are of such importance that they outweigh, or implicitly constrain, the rights of individuals. I suspect that an implicit reliance on the public goods defence explains the silence of most high liberals regarding the Basic Problem. But a defence of this strategy in the high liberal tradition remains sketchy or implicit. In short, high liberals should take the Basic Problem much more seriously than they generally have.
Stating the Problem
Let me turn to a more precise statement of the dilemma that Nozick hints at. The Basic Problem consists of six premises which, taken together, entail political anarchism:
||Individuals have strong moral rights not to be coerced, dispossessed, harmed (etc.) against their will.
||Generally, individuals have not transferred, waived, or abandoned their moral rights, or forfeited them through wrongdoing, or lost them in some other way.
||Generally, rulers do not have the moral power to change or void these rights.
||Non-minimal rulers systematically coerce, dispossess and harm (etc.) individuals in non-consensual ways.
||Rights can never be permissibly infringed, except to avoid catastrophically bad outcomes.
||Generally, non-minimal rulers are not needed to avoid catastrophically bad outcomes.
||Generally, non-minimal rulers lack legitimacy unless they have been consented to.
Let me describe these claims in depth.
The Shape of Moral Rights
Our starting premise is that individuals have rights:
||Individuals have strong moral rights not to be coerced, dispossessed, harmed (etc.) against their will.
When I say that these rights are “moral”, I mean this in a minimal sense. A moral right is one which we have independently from any political, legal, conventional or social structures, and independently from whether anyone believes that we have it. Calling them moral in this way does not entail that everyone has the same moral rights, or that our rights cannot change. I also do not wish to make any metaphysical commitments as to what ultimately grounds moral rights.
Furthermore, this premise does not rely on any specific theory of what rights we have. It is meant to be acceptable across a wide range of theories of rights. Let me describe some of the rights any such theory will ascribe to individuals. A first set of rights I briefly described above is a general set of rights against non-consensual interference. Individuals enjoy rights over their own bodies and minds. They enjoy the right not to be harmed, threatened with harm, constrained, manipulated, or forced in other ways unless they have committed wrongs which would merit such treatment, or consented to such treatment. Second, most theories of justice agree that individuals enjoy property rights over themselves and their possessions. This is a claim which is attractive quite independent from a Lockean theory of rights: we think individuals are free to exchange, destroy or transfer property as they see fit in a wide range of cases. Rulers, insofar as they take our property, violate those rights. Third, and a bit more controversially, many theorists think that individuals have rights to enforce their own rights, as we saw in Senor’s objection. If someone violates my rights, then I, as the victim, have a prerogative of demanding punishment, restitution, or apology, and I have the power to waive punishment as I see fit. The state, on the other hand, usually demands the right to pursue and determine punishment, even to waive it, without the victims’ consent.
There are two main ways, focussed on premise (1), to escape the Basic Problem. First, we could dismiss the idea that people have rights altogether. Some theorists have taken precisely this way out. Sonu Bedi observes that rights constrain what democratic majorities can legitimately do. This is counterintuitive, he claims. Thus, we should not think that individuals have rights. This is a radical way to escape the Basic Problem. It is not at any rate one I wish to go, as a commitment to a rights-based morality is one of the fundamental background claims of this thesis. However, there is a more moderate possibility. We can accept that individuals have rights, but argue that the shape of their rights is such that political institutions can permissibly operate. Perhaps our rights have “holes” in them which allow political institutions to operate. This is one of the main strategies I will consider later in this chapter, and ultimately endorse in chapters 3 and 4.
Facts about Transactions
Moral rights are not static—we can transfer, waive, abandon and forfeit them. Regarding these possibilities, the following is a plausible empirical claim:
||Generally, individuals have not transferred, waived, or abandoned their moral rights, or forfeited them through wrongdoing, or lost them in some other way.
To assess this claim, we would need to know under what conditions people give up their rights in these ways: I turn to the details in section 2.3. But at least prima facie, claim (2) is plausible.
The third premise is simply a statement of philosophical anarchism:
||Generally, rulers do not have the moral power to change or void our moral rights.
If rulers had authority, then they could exercise a moral power to make some of our rights void. Assume John has a right to play with his toys. If Lila has genuine authority over John, then her command that John go to bed voids (at least temporarily) that right. Along the same model, governmental authority could be used to explain governmental legitimacy in the face of individual rights. But as I argued in the last chapter, rulers do not generally have this kind of power.
The next claim conceptualises what rulers do:
||Non-minimal rulers systematically coerce, dispossess and harm (etc.) individuals in non-consensual ways.
This assumption is partially an empirical observation: the state coerces, taxes, punishes, gives rules which would otherwise not apply to us, restricts our movement, and so on. This premise, like premise (1), is light on specifics. It expresses a commitment to a non-minimal state, but the details could be sketched out in different ways. The reason for proceeding in this way is to be able to consider a general version of the problem which is not limited to any particular view of the legitimate functions of a non-minimal state.
The Strength of Moral Rights
So far we have established that individuals have rights, that they conflict with non-minimal ruling, and that they have not been waived or forfeited. As our next premise, we need to connect the infringement of a right with impermissibility:
||Rights can never be permissibly infringed, except to avoid catastrophically bad outcomes.
Two separate ideas are expressed by this claim. The first is Nozick’s idea that rights are, at least in their primary role, constraints not goals. Goals are aims that we pursue, such as our own happiness, or everyone’s happiness, or cultural enrichment, or some such. On the other hand, a constraint—I avoid the more cumbersome label “side constraint”—“is a principle that says that, whatever one’s goal(s) might be, one may not pursue it/them by the means named by the principle”. When we say that individuals have a right to free speech, we are first and foremost saying that there are certain ways in which we should not treat people, such as preventing them from speaking. Seeing rights as constraints rules out any type of view which included rights merely as some aim to be evaluated in a metric of evaluating consequences—e.g., it rules out Nozick’s imagined “utilitarianism of rights”.
Saying that rights are constraints, however, does not yet establish the relative strength of these constraints vis-a-vis other moral considerations. It could be that constraints are easily overcome, like (say) rules of etiquette. It might be that if the goals you pursue are important enough, rights-based moral constraints no longer apply or are overridden. The second idea expressed by claim (5), then, is that rights are of significant strength—that they are “trumps” in some important sense. If something is a trump, then it enjoys stringent, perhaps even lexical, priority over other conflicting values. According to claim (5), outside certain extreme circumstances, there are no goals which are important enough to invalidate a constraint. I discuss some attempts to weaken this assumption in section 2.6.
Claim (5) does not express the idea that rights are absolute. It allows that we can infringe rights to avoid catastrophic outcomes. We might be allowed to shoot down a passenger airplane taken over by terrorists, or you might be allowed to break into my house to save a child from a fire. So to complete the argument, we need one last claim:
||Generally, non-minimal rulers are not needed to avoid catastrophically bad outcomes.
A more precise formulation is that non-minimal rulers, insofar as they rule non-minimally, are not necessary to avoid catastrophic consequences. It might be that some of the activities of a non-minimal ruler are necessary to avoid disastrous consequences, but that would only justify those activities. A recent suggestion is to escape the Basic Problem by denying claim (6). Samaritans argue that political institutions can permissibly rule because the alternative would be disastrous. I discuss this proposal in section 2.3.
If we take (1) to (6) together, we reach a devastating conclusion about the moral status of non-minimal ruling:
||Generally, non-minimal rulers lack legitimacy unless they have been consented to.
This is political minarchism, the claim that (at most) minimal political institutions are legitimate. The conclusion does not rule out that there are some local cases of legitimate exercise of power, or that an otherwise illegitimate ruler might act permissibly in some particular instances. After all, some individuals might have waived or forfeited some of their rights, some rulers might have authority some of the time, and some forms of ruling might be necessary to avoid catastrophic outcomes. But such instances will be rare, and we will expect that large parts of the activities of modern, liberal states turn out to be illegitimate.
To avoid this conclusion, we must reject or modify one of the six premises of the Basic Problem. There is not much point in denying premise (4)—which merely describes what non-minimal rulers do—and I already argued for philosophical anarchism, premise (3). This leaves us with four alternatives to escape the Basic Problem: (a) arguing that individuals have given up or forfeited their rights, (b) endorsing Samaritanism, (c) modifying the strength of moral rights, and (d) modifying the shape of moral rights. I will discuss these four proposals in this order.
Facts about Transactions
There are some transactions through which the rights of individuals are temporarily or permanently set aside, to the effect that political institutions would not infringe them if they acted in certain ways. I group these types of transactions together as giving up rights on the one hand, and forfeiting them on the other.
Giving Up Rights
Individuals have the power to waive, transfer, and abandon their rights. When you waive a right, you decide not to insist on it, but you retain it in principle. In such a case, the waiver gives another person a permission to infringe your rights in a particular instance, though more permanent waiving is possible as well. For example, you waive your right to bodily integrity when you allow a doctor to operate on you; you waive your right to exclusive use of your pen if you loan it to your friend. In a transfer, on the other hand, you confer a right on someone else. If you sell your piano, you transfer your property rights in it to someone else. Lastly, you might abandon (or alienate) a right. If you discard a book into a waste bin, for example, you are understood to give up all your claims to the book, and anyone can claim the book for themselves.
There are difficult debates on whether all rights can be transferred, waived, or abandoned, but I will assume that we are capable of waiving, transferring, or alienating our rights at least in such a way that a non-minimal state could become legitimate. These three types of transaction have similar moral effects. Showing that a transaction of one of these kinds has taken place would show that the state is permitted to rule over the individual in the relevant respect. Therefore, I will group them together, and speak generically of individuals “giving up” their rights.
The next question is under which conditions individuals give up their rights. The standard answer is that we give up rights only through individual consent which is free, deliberate, and explicit. However, in the world we live in, the overwhelming majority of individuals have not consented to the state in any such way. But perhaps there are forms of consent which fulfil weaker criteria and still count as binding. Thus, the crucial task becomes to find forms of consent which are sufficient for giving up a right.
First, it is plausible to think that consent does not need to be explicitly uttered. If a doctor tells you before an operation “shake your head if we should not proceed”, then a failure to shake your head can be read as consent. Similarly, if the mayor asks “no objections?” in a council meeting, and no one raises an objection, this can be read as agreement to the point discussed. In both cases, some non-verbal behaviour counts as consent. But this does not help the defender of the state, as our situation vis-à-vis the state is importantly dissimilar. There has been no point in time at which the state or other political institutions have given us an explicit opportunity to dissent, and where observers would have clearly identified our non-dissent as a form of consent.
So we need an even wider sense of what counts as consent, and some such cases appear to exist. If you and I enter a boxing ring, then I implicitly give up my right against being harmed in some ways. In some jurisdictions, unmarried couples gain the same legal rights and duties as married couples if they live together for a certain amount of time; their legal situation is as if they were consensually married. We might try to exploit these analogies to argue that we have tacitly consented to the state. Perhaps residence on the state’s territory, profiting from its benefits, or submission to its rules amount to tacit consent which is sufficient to count as waiving our rights against non-interference.
What makes these cases work? Turn first to the boxing case. Entering a boxing ring, in a certain way under certain circumstances, is understood by all reasonable observers to express the intention to engage in a boxing fight. Being harmed is a clear risk of engaging in such a fight; as long as we think you are decently clear-minded, and you enter the ring voluntarily, we thus think you accept the risk of being harmed. This is as good as explicit consent, for three reasons: (1) There is a set of conventions governing this case, known to all participants, which allow us to interpret one form of behaviour (entering a boxing ring) as expressing a certain intention (a willingness to engage in a boxing fight). Similarly, if you order a meal in a restaurant, you are conventionally understood to agree to pay for it. (2) The intention to engage in the relevant kind of activity is generally understood to be linked to an intention to waive certain rights and immunities in the process. That is, if you intend to engage in a boxing fight, you are also understood to not see certain possible harms done to you as wrongs. Similarly, if you express an intention to have your hair cut, you are commonly understood to intend that you are okay with others touching you. Lastly, (3) people who enter boxing rings usually do so voluntarily. They are—barring extreme circumstances—not forced to fight, and can avoid stepping into the ring easily.
The example of marriage-like partnerships fits this structure as well, though the plausibility of the argument already starts to fray. Living together, sharing financial means, and raising children can reasonably be read as intending to form a lasting partnership (element (1)). Intending to form a lasting partnership can be seen as taking on certain rights and duties, such as a duty to care for each other in case one’s partner falls on hard times (element (2)). Lastly, in the standard case entering such bonds is voluntary (element (3)).
But even if we can make a weakened form of consent explanation work for these two cases, the analogous argument with regard to political institutions fails. First, living on the state’s territory, or submitting to the rules of the state, is compatible with a whole range of intentions. We might submit to the state out of fear; reside on its territory because this is where we always lived; comply with the state’s rules because we profit from such compliance, but not because we consider ourselves bound by them. There are no clear, established conventions, as in the case of a boxing fight, to interpret these kinds of behaviour as expressions of an intention to give up certain rights. Second, most people do not live on the state’s territory, or submit to its rules, voluntarily, as they have no reasonable alternatives. This is one of Hume’s famous points against Locke. Leaving their state has, for most individuals, enormous social, financial and cultural costs.
Mike Otsuka has sketched the conditions which would need to be fulfilled for us to have reasonable alternatives. If those conditions were given, Otsuka argues, tacit consent would be sufficient to establish the legitimacy of the state. First, individuals who do not want to be ruled need to have the opportunity to “withdraw” onto a territory over which they have exclusive rights. Second, rulers need to promise (monetary) compensation for those who do not wish to be ruled for various other disadvantages they suffer. Lastly, there would need to be a “diverse range of choices of political societies which occupy the full range of political, cultural, and urban-to-rural possibilities to which people tend to be attracted”. I think Otsuka’s proposal still relies on an unrealistically cosmopolitan picture of the ability and willingness of individuals to move between societies; but we can set those worries aside. Even if we assume that his proposal is correct, it should be enough to see that our current world is nothing like it. So habitual submission cannot be taken to indicate that we have waived moral rights against interference.
Another way in which rights can be lost is by forfeiting them. We forfeit our rights if we wrong others, or if we are a serious threat to them. If you threaten my life without good cause, then I can permissibly injure you in self-defence, without thereby counting as infringing your rights. Similarly, we can permissibly imprison murderers, as they have forfeited their right to freedom. This suggests another way to escape the Basic Problem: if we have all forfeited certain rights against government interference, it would be possible for government to operate. The problem with this claim is that, at least on an intuitive level, most of us have not wronged others, at least not to the degree necessary that we would have forfeited most of our rights. What is your crime that the state can so freely interfere with your life?
Still, some authors have explored this way of arguing for the state. What you need for the argument to work is the claim that individuals would wrong others if they remained in the state of nature. On some ways of reading Kant’s position, this is his argument, which I will discuss in section 4.1. Another theorist who offers such an argument is Massimo Renzo. He starts from the assumption that we have a right of self-defence against unjust threats. Second, he claims that submission to the state’s authority is necessary to avoid the Hobbesian evils of the state of nature. Thus, he claims, people who do not submit to the state are unjust threats, and forfeit their rights against interference. This explains how the state has a right to enforce the law against them even without their consent. The state has a right to force us to submit to its rules in much the same way as we have a right to prevent immediate threats to our lives.
However, a crucial inference in Renzo’s argument is based on a mistake. We can accept, for argument’s sake, that most people’s submission to the state is necessary to thwart the Hobbesian evils of the state of nature. But this is insufficient to show that any particular individual is an unjust threat. Consider the following example. If I did not contribute my share to organising a neighbourhood watch, many houses might get burgled, as the watch would be spread out too thin. But this does not show that I am a threat to others. The burglars are the unjust threat, and they have forfeited certain rights. So Renzo needs to appeal not merely to a principle on which we can act in self-defence against unjust threats; he must rely on a wider principle according to which we can also coerce those who fail to prevent unjust threats by others. It is not at all clear, however, why we would forfeit rights in such cases.
Renzo’s perspective on the problem here is skewed because he compares our situation under the state to that of twenty people in a rowboat. He assumes that we can escape to the life-saving shore only if we all row in a coordinated fashion. It sounds plausible that if I refuse to row in this situation, then I pose an unjust threat to others in the boat. This gives the rowers a right to force the non-rowers to do their share. But as I emphasised in the previous chapter, it is implausible to think that this is the general situation in which we are vis-à-vis our fellow citizens and the commands of the state. Almost no political situation is such that our non-compliance is directly a threat to the survival of others.
At any rate, the principle of self-defence would only apply in limited circumstances, where there is an immediate threat to limb and life. So if this line of argument can be used at all, it only justifies a minimal state. Such a state is limited to the bare necessities of keeping peace, but nothing more. We could not invoke self-defence, for example, to justify taking possessions from our neighbours to improve distributive justice, or to fund public schools, or to ensure equality of opportunity for everyone. We need to appeal to some other principle to justify those activities.
Renzo’s view has some similarities with another view, Samaritanism, the main defender of which is Christopher Wellman. First, Wellman argues, we have a duty to rescue and protect other people from serious harm. Second, the state is necessary to avoid such catastrophic outcomes. Without the organising hand of the state, Wellman claims, individuals in modern societies would be threatened by disorder and chaos. Third, the costs imposed by the state on individuals are reasonable. Wellman argues that in such situations a general principle applies, which I reconstruct as
the Rescue Principle. If non-consensual coercion of some person is necessary to fulfil important duties of rescue to save others from disastrous consequences, and the costs imposed on that person are not unreasonable or disproportionate, then such coercion is permissible, even if such coercion infringes one or more of the rights of that person.
If we take these assumptions together, we get the result that the state can permissibly infringe the rights of its citizens. Advocates of Samaritanism have also tried to develop Samaritanism as an argument for the state’s authority, but I will not pursue this part of the argument.
The Rescue Principle has strong, independent foundations in moral thought. Its appeal can be brought out through simple examples. If it is necessary to take your car to drive a severely injured person to the hospital to save their life—perhaps we are in the middle of the wilderness, and your car is the only viable means of quick transportation—this will be permissible without your consent or even against your dissent. But the Rescue Principle does not allow just any kind of interference. In particular, it entails that rights-infringing rescue efforts must pass a three-part test: (1) there needs to be an enforceable duty of rescue, (2) infringing the rights of some other person needs to be a strictly necessary means—not just a convenient or efficient means—to achieve rescue, and (3) the costs imposed on other individuals need to be reasonable, and proportionate to the aim pursued. We are not permitted, for example, to force you at gunpoint to drive the injured person to hospital, even if this is the only way to achieve our aims, and even if you are under a duty to drive the person to the hospital.
This is an exacting test when applied to state activity. In view of these restrictions, Wellman claims that Samaritanism will only make a minimal state legitimate, as few state functions fulfil all three requirements. This strikes me as correct. Samaritanism will probably allow the state to organise, and through forced taxation finance, a police service and a basic justice system, emergency health care services, military defence, and relief for natural disasters. However, the principle is unlikely to allow the state to regulate trade, provide full public health care, fund public infrastructure, control borders, redistribute wealth to increase distributive justice, or interfere with people’s lives in many other ways.
Even those areas the state would be allowed to organise under Samaritanism would look crucially different from how they are organised now. The state would only be allowed to provide a police force as long as citizens do not freely organise an alternative. If a voluntary neighbourhood watch started to provide security in some neighbourhood, the state would be under a duty to retreat from this area, and to cease to forcibly tax its inhabitants in support of its own police force. This would be the case even if the neighbourhood watch would be less efficient and well-organised than the state’s police, as long as they uphold basic public order. In short, the state is a decidedly second-best option on the Samaritan view. This entails that it should always attempt to diminish its own footprint, and in the long run aim for itself to disappear completely. If you repeatedly find yourself using other people’s cars to drive people to hospitals, then something is going very wrong: you are under a duty to change something in your approach so that this happens less often, or preferably not at all. And if you have to do it, you should always look for ways to keep your impact on the rights of others as small as possible. In summary, the Samaritan argument might give us a minimal state under some conditions. But it is an unlikely argument to give us any more than that.
Extending the Rescue Principle?
Candice Delmas disagrees. She argues that a much wider state could be justified on Samaritan premises. There are two parts to Delmas’ response. First, she argues that actual individuals living in modern states are still subject to wide-spread violations of their rights—for example, one in six women experience attempted rape. She also explores the situation of African Americans and Native Americans in the United States, who according to Delmas are constantly exposed to great risks even though they live in otherwise stable political societies. These groups still effectively live in a state of nature, and thus require rescue. In the face of this situation, she claims that
[t]he state’s most effective way to adequately protect individuals’ safety, at least in the two cases examined here, is to secure equal access to health, education, employment, and the protection of the law. All of this suggests that a minimal state is not sufficient to secure individuals’ basic safety, and that, beyond the enforcement of criminal law, Samaritanism can justify a variety of social and political measures designed to empower individuals.
A second suggestion by Delmas is that we ought to widen the interpretation of Samaritanism. The Rescue Principle should cover more than security from attack. She states the “Samaritan mission” as “[providing] vital benefits and [removing] some of the obstacles to a meaningful human life”. We have a duty to provide everyone with a minimally meaningful life. With the scope extended in such a way, a much wider area of state activity becomes legitimate:
Samaritanism can thus justify raising taxes to fund disaster relief, health care and disability aid, and to ensure the less privileged and most vulnerable members of society have their basic needs met, so long as these taxes do not impose unreasonable burdens on constituents.
Poverty, in particular, makes living a meaningful life difficult. So poverty relief will be required under Delmas’ modified Samaritanism. In summary, Delmas argues that closer attention to what rescuing individuals from danger would require, and a widening of Samaritanism will make the liberal social welfare state legitimate.
With regard to Delmas’s first point, we can grant that Wellman is overly optimistic about the security that individuals enjoy in actual states. Delmas is correct in claiming that many people, especially members of minorities, are still exposed to injustice in otherwise peaceful societies. Thus, we should not think that duties of rescue are fulfilled once we have established general political stability. We can also accept that “equal access to health, education, and employment” are good long-run means to “empower individuals” and to save them from those injustices. If we pursued those political aims, we would alleviate the structural reasons of why some individuals are highly vulnerable to injustice.
However, this looks less and less like a problem of rescue. Instead, these activities amount to solving a complicated social problem. Imagine that you notice that in your area people repeatedly have medical emergencies, but no way to cover their health care costs. You are under a duty of rescue to help those in emergencies you come across. But it is implausible to think that your duty of rescue requires you to resolve the systematic causes which lead to people’s inability to cover their health care costs. You might well have such a duty, but this is better understood as a duty of charity, or perhaps as a duty of care, or some other duty.
Thus, the gist of Delmas’ response is on her second proposal, to widen the content of Samaritanism. Many liberals find it plausible that we have the kind of positive duties Delmas describes—that is, a duty to provide others with the opportunity to lead a meaningful life. The ability of the state to provide the material conditions necessary for such a life might well figure prominently in its justification. But in the current context, we are asking whether acknowledging such a duty helps us to solve the Basic Problem; and here Delmas’ modified Samaritanism is less promising. The Samaritan duty of rescue and Delmas’ duty of providing people with opportunities for a meaningful life are very different in their strength and scope. In particular, they differ when it comes to infringing the rights of others. The importance of rescue allows me to take your gun to defend a third person from an unjust threat. But I would not be allowed to take your gun to give it to a neighbourhood watch intended to prevent further attacks. The point is that, even if there are the kinds of duties Delmas highlights, these do not give us the sufficient permission to infringe the rights of others. So this reply does not help us with the Basic Problem.
The Structure of Rights
Samaritanism claims that we can infringe individual rights where the alternatives would be catastrophic. The problem, however, is that the exceptions this view allows are spread too thin to allow for political institutions which fulfil any significant functions. I will now focus on views which argue that we can permissibly infringe rights even in less-than-catastrophic circumstances. On these views, rights are weakened in some important sense. But before we can discuss such views, we need to note one difficulty regarding the structure of rights.
Take a case where breaking your leg is necessary to save n people’s lives. One of your rights is plausibly at stake here—your right to bodily integrity. Settle on an n which you find large enough such that you agree that it is permissible to (non-consensually) break your leg. There are two ways to think about this situation. First, we might argue that your right to bodily integrity covers this case, only that it is overridden. On this view, rights express pro tanto constraints: they compete with other rights and other moral values. The scope of different rights overlap, and can command us to choose different courses of action. Standard terminology introduced in this context is to distinguish the morally neutral infringement of a right from the violation of a right, which is always wrong. The infringement theorist argues that not all infringements are violations. If you extend such cases beyond Samaritanism, then you deny claim (5).
In the second way of thinking about the case, it is not the case that your right is infringed, because your right to bodily integrity did not cover the case in the first place. It was rather that your right to bodily integrity should have been better understood as a highly specified right—the right “not to have your leg broken unless doing so was necessary for some morally important aim”. In particular, you had a right not to have your leg broken unless breaking it could save n lives or more. On this way of thinking, rights are like Swiss cheeses with complicated holes—but if a right applies, its force is absolute. Theorists of this variety, whom we can call the specificationists, assume that rights cannot conflict. For the specificationist every infringement of a right is a violation.
So when we talk about “weak” rights, this is ambiguous between these two views. On the infringement theory, we weaken rights by holding their scope constant, but by introducing additional circumstances under which they can be permissibly infringed. Assume that we wish to decrease n to some lower m. The infringement theorist would hold on to the claim that you have a right to bodily integrity. But it is now the case that this right is outweighed—permissibly infringed—already in cases where m lives are at stake. On the specificationist theory we could not say that, as there is nothing which could outweigh a right if it applies. Instead, to weaken a right we need to narrow its scope. The specificationist would say that the specification of the right has changed: we were wrong in thinking that you had right1 (the right to not have your leg broken unless that was necessary to saving n lives): you only had right2 (the right to not have your leg broken unless that was necessary to saving m lives). Our general talk about a right to bodily integrity is imprecise, and hides this more ragged structure which each particular right possesses.
The two views see rights quite differently. To use a metaphor, there are two ways how we can fence in some area while keeping it accessible to bypassers. First, we can keep a relatively wide area enfenced, but make the fences low in some parts. This allows people to climb over those fences in some places, and pass through the area. This is how the infringement theorist sees rights. Alternatively, we can keep the fences high and unclimbable, but leave a small, complex passageway through the area. This is the specificationist’s view.
The distinction complicates many of the following discussions. Many arguments about the shape of rights (claim (1)) can also be read as arguments about the strength of rights (claim (5)) and vice versa, depending on which view of the structure of rights you favour. Most authors are not clear on what view of rights they have, and accordingly could be interpreted either way. I start with some views which are more naturally interpreted as weakening the strength of rights, before I turn to strategies which are more naturally read as being concerned with shape.
Let us consider rejecting premise (5). Such positions deviate from Samaritanism in that they accept that rights can be permissibly infringed even in cases where the consequences are less than catastrophically bad. Perhaps this was the type of position Delmas was aiming for. I consider two such suggestions, but reject both.
Presumptions Against Coercion
A first suggestion is that rights merely provide us with pro tanto reasons not to take some action—that is, they establish a presumption against interference, but if the value we can achieve by infringing the right is sufficiently important, then this presumption can be overcome. We briefly saw that this claim was part of Nagel’s answer to Nozick; Allen Buchanan’s rejection of libertarianism relies on a similar idea.
This position runs into three problems. First, infringing a right generally calls for compensation or restitution, or at least apology. Assume that it is a permissible rights-infringement that I take some of your money to build an orphanage. Let us stipulate that this is not a rescue case: the orphanage would make life for orphans better, but the situation for them would not be dire otherwise. However, even if this is a permissible infringement, I incur some duties against you. For example, I should give back the money at some later point if I can; if I cannot, I should try to make it up to you in other ways. By analogy, for example, we would expect that if a government taxes you, then it incurs a duty to pay you back at some later point, or make it up to you in some other way. Thus, on the current view government would systematically owe its citizens compensation.
One might reply that government already compensates us through the various benefits it provides—an idea to which I return to below. Even if we accept this thought, however, there is a second problem. Infringements of rights, even if they are permissible, should be generally avoided. If I can finance my orphanage in some way other than taking your money, then I ought to do so. Infringements, while not necessarily wrongs, are morally subpar: they should be the exception rather than the rule. Remember that we are currently considering a position on which the state is permitted to operate because the importance of its functions allows it to infringe rights. Note that such infringement would be systematic. So while we could not say that the state systematically wrongs individuals, it systematically engages in a morally subpar activity. This immediately casts it in a bad light. The permission of the state to operate would be a stained permission. We would continue to consider a legitimate state deeply problematic. As in the case of Samaritanism, we should actively work towards its dissolution, to replace it with an arrangement which is not rights-infringing. Indeed, I struggle to find another example of an acceptable social institution which is based on systematically infringing rights.
The defender of the view of rights as merely pro tanto prohibitions could reject the feature of rights my objection rested on—namely, the assumption that infringing a right is morally subpar. Imagine that the state aims to build a railway from London to Manchester. Doing so would have great social benefits. Assume that all of the people owning the land along the proposed track have voluntarily sold their land—except Blue, who holds the property rights to the last remaining piece of land. She has some interest in that plot—she uses it to herd her cows, say—but this interest is not essential: she could easily herd her cows elsewhere. Many find it plausible that even if Blue has property rights to the plot of land, we can permissibly and non-consensually take the land away from Blue as long as we adequately compensate her for the loss—for example, by paying her market prices for her property. On this position, respecting a right requires us to either not infringe it or to pay the required compensation. We now see both of these options as equally morally desirable. If we choose the second option, our action is not morally subpar as long as we adhere to the correct compensatory requirements—this is the crucial difference to the position I considered in the last subsection.
If such a view were true, justifying political institutions appears to become easier. The non-minimal state could infringe the rights of individuals if it pays individuals adequate compensation. You might object that while this makes the state possible, it makes it impracticable, given the enormous financial means which would be necessary to compensate citizens. This issue can be alleviated if we add that compensation for rights infringements need not be monetary. Instead, the compensation might be in terms of services provided: the state might infringe a right, but recompensate the individual by extending the state’s protection to her.
Some theorists reject compensatory understandings of rights straight out of hand. On conceptions of rights in which they are seen as protecting our autonomy, the idea that we could permissibly infringe them as long as compensation is paid is anathema. I agree with this idea—rights are not prices we can choose to pay; they are based on recognising the choices of individuals. I describe this idea at greater depth in the next chapter. For now, let us note that many philosophers find it plausible that in some cases, some of the time, we can infringe rights if adequate compensation is paid. However, even if we accept this idea, the compensatory understanding of rights does not carry us far enough to show how the state could be legitimate.
First, it is clear that the scope of any principle of permissible compensation must be limited in some way. Assume that Green is in the same situation as Blue. However, in Green’s case, the land is not used for herding cows, but is the Green family’s ancestral home lands, or perhaps has high religious significance for her—it is Green’s version of the Temple Mount, say. It is plausible to think that it would be impermissible for us to take Green’s property from her in such a case, even if we compensate her. Green has an important interest in making central life decisions for herself. If we take away the house from Green, she loses this ability regarding one area of her life, the ability to choose where one lives. There is no way we can give this ability back to her. This suggests a first way in which compensation principles are likely to be limited. They will allow permissible infringements only where such infringements do not concern major life decisions of individuals. We think it permissible to ignore Blue’s lack of consent because her decision concerns no such major issue. We tend to think that, if nothing of importance is at stake, then the community should not be held hostage by the unreasonableness of some.
There are other ways in which compensatory principles are likely to be limited. Extrapolating from the example of Blue, we can expect that compensation finds its moral place in market-like situations: cases where (1) the stakes are comparatively low, or where the goods at stake are interchangeable and people have no attachment to any particular token of the good we take away; where, in addition, (2) there are accepted standards for evaluating goods and agreed-upon methods of compensation (e.g., market prices), and lastly, (3) where everyone’s situation would be improved if we allowed unilateral rights infringements.
The situation individuals face with regard to political institutions fulfil none of these requirements. One crucial feature is how all-encompassing the influence of political institutions is: they shape virtually every aspect of our lives. They do much more than take away a cow meadow here or there. Modern states intensely shape our lives through their legal, political, educational, punitive, social, welfare, cultural and tax systems. So the stakes are anything but low. Second, there is no clear, certainly no widely accepted, standard for compensating individuals for this type of interference. The idea of compensation already becomes muddy when we turn away from objects with clear monetary values. But what is adequate compensation for being forced to live one’s life under the non-consensual sway of an institution like the state? Lastly, it is not clear that, if we disregarded the dissent of a few, everyone would be better off. In the original case of Blue, one ground for thinking that it would be permissible to take away the meadow from her is that she would profit from the deal—she is simply unreasonable for rejecting to sell the meadow. It is not at all clear, however, whether this is true in general.
In short, even if we accept that sometimes rights can be permissibly infringed if adequate compensation is given, this does not help us in escaping the Basic Problem. The political problem of justifying the state is too dissimilar to the kinds of market situations in which compensation principles plausibly apply.
The last type of response is to play around with what rights individuals possess. Now we can return to Nagel’s review of Nozick. He writes that most people “will find no echo” of the intuition that we have the strong self-ownership rights that Nozick thinks we have. The best interpretation of this claim is that Nagel thinks that individuals have very limited ownership rights. In particular, we do not have full rights to our pre-tax income. So if the state taxes us, it does not infringe our rights. Nagel explicitly defended this claim many years later in a book co-authored with Liam Murphy.
I wrote in sec. 2.1 that weakening property rights is not enough—there are many other rights that get us started on the Basic Problem. But Nagel’s remarks on property suggest an answer to the Basic Problem which is based on developing a substantive theory of justice where individuals have limited rights in general. There will be many ways to pursue this strategy, so there is no catch-all objection which would apply to all of these different approaches. We can focus on one proposal, Samuel Scheffler’s. Scheffler, also in a response to Nozick’s Anarchy, State and Utopia, argues for a view of natural rights on which “individuals have a right to whatever liberty is necessary to insure a reasonable chance of living a decent and fulfilling life”. Thus, “people’s natural rights are satisfied and exhausted once their minimum welfare and liberty needs are met”. Scheffler explicitly argues that this picture of rights avoids Nozick’s conclusion. Taking resources from the rich, for example, does not diminish their ability to lead decent and meaningful lives, and thus should not be seen as a rights infringement. Similarly, if the state threatens or coerces you to bring about benefits for others, this also does not necessarily undermine your ability to lead a decent life.
Scheffler’s proposed weakening would allow for a legitimate, non-minimal state. It is, however, difficult to see what speaks in favour of his account of rights. Scheffler notes, for example, that his view would entail that punching someone in the nose does not violate any of their rights, as they retain the ability to lead a decent and fulfilling life. Scheffler argues that there are other moral injunctions, not based on rights, which forbid people to punch each other’s noses. That is true, but the judgment delivered by his view is still highly implausible—it certainly seems that punching people would count as a rights violation. More generally, it is hard to see what principled motivation there is for limiting rights in this way. Why should rights only be concerned with protecting “decent” lives? One intuition which supports this thought, of course, is the idea that rights are a subsidiary, and not particularly important part of morality. Perhaps we need rights language merely to insulate individuals from the worst excesses of democratic majorities. But this just moves us away from one of the background assumptions of this thesis, the idea that rights form a crucial part of morality. Anyone who takes the idea that individuals have rights seriously should be suspicious of Scheffler’s quick down-grading of their scope.
But while Scheffler’s response is not worked out enough, I think he points us in the right way. We need to look for a principled way to revise our theory of rights so as to answer the Basic Problem. The most principled way to argue about the shape of rights is by focussing on different theories of rights, specifically theories of what justifies rights. I will look at such theories in the next two chapters.
Theories of Rights
In the previous chapter, I discussed various attempts at making rights compatible with non-consensual political institutions. I argued against several solutions, but also suggested that there might be a more principled way to attack the problem. If we look at general theories of rights, I will argue, we can discover whether and how the Basic Problem can be resolved. This is the task I pursue in this chapter and the next.
By a theory of rights I understand a theory which explains in general how we can justify the existence of a moral right. Alternatively, we might ask what grounds rights. This is language significantly stronger than that of “justifying”, and it introduces complicated metaphysical issues—e.g., how rights supervene on natural features. While interesting, such questions would carry us away from our main task. I focus on what moral arguments establish that a right exists, and we can abstain from commenting on the more difficult metaphysical and metaethical issues. Similarly, there is a related debate concerning the nature or function of rights. This debate is not about what rights there are, but what a right is when people have it. I will also stay away from this literature. One might object that we cannot undertake to justify rights before we know what it is we are justifying. But it is a mistake to think that we need a worked-out theory of the nature of rights to be able to have some idea of what they are. From the last chapter, we know that they are demands against others which entail weighty constraints; that their infringement is morally subpar, and triggers demands for compensation and restitution; and we can add, from the first chapter, that rights can be described through the Hohfeld schema. We have an intuitive, if not very precise, grip on what they are before any deep philosophical analysis.
I will consider two well-known ways to justify rights. The first is the interest theory, which justifies rights on the basis of how they serve our interests. The second theory justifies rights as the appropriate moral response to some feature of the right-holder, and I will call this the status theory. Both are broad families of views, and there is significant variety in how the details of these theories can be worked out. Those details, however, are only interesting to us insofar as they affect the Basic Problem, and we will see that they usually do not. Why focus on both theories? First, the two theories are by far the most prominent approaches in the literature. So it is independently interesting to know whether we can solve the Basic Problem on either of them. Second, I also believe that the prominence of these two theories is earned: they are the most plausible accounts of how rights are justified. In particular, both theories are individualistic theories of rights: both justify rights on the basis of some genuine feature of the individual. Kantian moral philosophy as I have sketched it is committed to individualism in this sense: it takes individuals as the fundamental building blocks in moral theory. Rights express the essential protection-worthiness of individuals. Alternative theories of rights, such as rule-consequentialism or a natural law approach, are not individualist in this sense, and thus we can exclude them as candidates from the beginning.
I think that a version of the status theory is true. But by focussing on both theories, we can offer an ecumenical argument that does not rely on defending this controversial claim. In short, the principled solution to the Basic Problem I suggest is the following:
||Either the interest theory or the status theory is true.
||On the interest theory, individual rights are shaped in a way which allows a non-minimal state to operate.
||On the status theory, the same is true.
||Individual rights are shaped in a way which allows a non-minimal state to operate.
In this chapter, I provide the first leg of the argument—that is, I argue for claim (2). I first describe the interest theory (sec. 3.2), before I show this result (sec. 3.3). I also discuss the outlines of the status theory (sec. 3.4), and sketch how the Basic Problem is harder to solve in this context (sec. 3.5). I continue discussion of the status theory into the next chapter.
The Interest Theory
The interest theory justifies rights on the basis of interests. Tasioulas offers the best and most elaborate version of the interest theory available in the literature. He describes the following four-step procedure as to how an interest theorist justifies a right:
||For all human beings within a given historical context, and simply in virtue of their humanity, having X (the object of the putative right) serves one or more of their basic interests, e.g. interests in health, physical security, autonomy, understanding, friendship, achievement, play etc.
||The interest in having X is, in the case of each human being and simply in virtue of their humanity, pro tanto of sufficient importance to justify the imposition of duties on others, e.g. to variously protect, respect or advance their interest in X.
||The duties generated at (ii) are feasible claims on others given the constraints created by general and relatively entrenched facts of human nature and social life in the specified historical context.
||All human beings within the specified historical context have a right to X.
The great appeal of the interest theory is the idea that claims about rights can be broken down into three simpler elements: (i) a theory of interests, (ii) a theory of importance, and (iii) a theory of feasibility. We will see that this initial impression is somewhat misleading. I will develop what I take to be the best version of the interest theory on the basis of Tasioulas’ four-step account, but we will see that the theory ends up being more complex.
Tasioulas, like many other authors, develops his account as a theory of human rights, while I am interested in moral rights. On some views, this is just a notational difference. You might use both “human” and “moral” (and also the label “natural”) to denote those rights that we have independent from social and legal conventions. But while Tasioulas accepts that rights are pre-conventional moral rights, he adds a further qualifying condition: something is a human right, he writes, if we have it “simply in virtue of [our] humanity”. Again, we might take this to be a rephrasing of the claim that human rights are moral rights. But Tasioulas is best read as suggesting that human rights are universal: if something is a human right, then everyone everywhere possesses it. A second feature we tend to associate with human rights is that they are particularly important rights. If we accept either of these two features of human rights, then they turn out to be a (proper) subset of moral rights.
However, from the background of the Basic Problem there is no reason to limit ourselves to human rights in this sense. All moral rights act as constraints on governments, whether they belong to the narrower subset of human rights or not. So in the following, I will focus on interest theories as theories of justifying moral rights in general. I discuss Tasioulas’ account as if it was offered as an account of moral rights. Another point is that I will limit my discussion to basic moral rights. There might be various instrumental reasons to assign non-basic right to individuals to protect their basic rights. But the justification of non-basic rights will follow some instrumental schema, and so need not adhere to Tasioulas’ four-step procedure.
Let us begin with step (i). The crucial feature of the interest theory is that it starts from an account of human well-being and justifies rights on that basis. In his formulation, Tasioulas focusses on “basic interests”. This entails that not all interests will be taken into account when we justify rights. Instead, only a subset of interests is even considered relevant and gets passed onto stage (ii) of Tasioulas’ test. Different moral theories might carve out this subset differently. Some will restrict themselves to the needs that individuals have (which I assume to be a subset of interests), or the basic interests we have in being capable to lead decent lives, or personhood-related interests. However, it is hard to see on what basis we make this selection. After all, one could leave the filtering of the relevant interests to the second stage, on which interests are considered according to their importance. The reason to already filter interests at the current stage is presumably that some interests are not even in principle candidates for translating into rights. Perhaps, for example, an interest in being loved could never ground a right, no matter its strength.
Tasioulas, due to his focus on human rights, focusses on the interests that we have “merely in virtue of our humanity”. We can think of those as universal interests that everyone has. This is closely connected to two other features of Tasioulas’ view. Tasioulas presumes that the relevant interests are objective components in our well-being: their goodness does not depend on us acknowledging them to be good. So something can be in our interests without us desiring it, or knowing it to be in our interest. Lastly, Tasioulas like many other interest theorists prefers a coarse-grained account of interests. The interests which ground rights are interests in integrity, or family life, or freedom of speech—so pretty general interests which cover a wide variety of situations.
Once we move away from human rights, it is hard to see why the interests which ground rights would need to fulfil any of these three features. Let me highlight the universality assumption in particular. Focus on the case of the Amish. The Amish have no (strong) interests in democratic participation, or free public universities, or a publicly funded police force. However, they have an interest to be free from the influence of modern technology, to educate their children communally, and so on. Given their lifestyle, these are very strong interests, and so seem prime candidates for translating into various rights. But these are not universal interests everyone has, so we could not ascribe them any rights in this respect. The flipside of the problem is that most of us have interests which the Amish do not have—an interest in publicly funded universities, say. So these interests could also not ground rights, as they are not universal. I take it that these are counter-intuitive implications for an interest theorist.
Interest theorists standardly make a couple of theoretical moves to avoid these implications. Many theorists suggest that the interest theory is built on claims about what is “typically”, “generally”, or “usually” in the interest of individuals. Tasioulas argues that the interests underlying an interest theory are “standardized”. By this he means that in the “specification [of an interest] one may abstract from some variations among individuals by focussing on the standard case of an ordinary human being living in a modern society”. This explicitly allows that we can ignore exceptions to the rule such as the Amish. So when we encounter a right that does not serve any of my interests, that is no counterexample by itself, as the right might still serve a “standardized” interest.
The problem with this move is that it undermines the individualistic nature of the interest theory. An interest theory which included such a modifier would not take individuals seriously; it would merely take the “normal” or “representative” person seriously. It does not take seriously the Amish, or anyone who deviates from the statistically average. There are of course practical reasons to stick to such simplified categories: insofar as the law is a blunt instrument, we have reason to focus on the interests of standardised people. But there is little reason to restrict our account in this way when we look at moral rights, where our primary interest is on getting the moral facts right. In the following, I will assume that the interests which ground moral rights need not be universal. So we cannot avoid paying close attention to the interests of each individual. This makes it possible that there are differences in our basic moral rights. I discuss some of the implications of this aspect of the interest theory in sec. 3.3.3.
Another line of response to save the idea that interest-based rights are universal is by tinkering with the account of interests. We might, for example, move to an objective account of interests on which everyone has the same basic interests. We might provide a list of such interests through an objective list account of well-being. The plausibility of this approach is closely linked to the question of how finely we individuate interests. There is perhaps some level on which we can plausibly describe the Amish and their secular brethren as having the same objective interests—e.g., an interest to live self-determined lives. But the plausibility of this account vanishes as soon as we move to a more fine-grained individuation of interests. In that case, it strikes me that the most plausible case to be made is that what makes life go well for the Amish is different from what makes life go well for those living in secular society. We should acknowledge this claim, even on an objective theory of well-being.
We can now turn to stage (ii). Here we show that
The interest in having X is […] pro tanto of sufficient importance to justify the imposition of duties on others, e.g. to variously protect, respect or advance their interest in X.
Here we can see how the interest theory is committed to individualism. An individualist theory of rights is one that justifies rights on the basis of genuine features of each individual, not with reference to some aggregate property. As Waldron puts it,
[an interest] theory of rights bases its commitment on the good to each individual, taken one by one […]. A’s right to free speech is based on the importance of A’s interest in speaking, B’s right is based on the importance of B’s interest […] and so on.
On the interest theory, individual rights are justified on the basis of the importance of each individual’s interests. At first sight, there does not appear to be any weighing against the interests of others—we simply establish that a right exists on the basis of the importance of the interest underlying it. There is no reference here to the interests of other people at all. Other authors have also emphasised this feature and even praise it as one of the main advantages of the interest theory. Rainbolt calls it “direct justification”, and gives the following useful example:
The right not to be assaulted is justified by something like the fact that being hit generally is not in one’s interest […]. In this case, it seems that the feature of the individual is a sufficient reason for the obligation. The justificatory work is done by the feature of the individual. If one asks, ‘Why shouldn’t I assault Tim?’ a plausible answer is, ‘It hurts him.’
One immediate worry is that direct justification leads to an implausible proliferation of rights. I have a strong interest to be provided with someone else’s kidney, assuming that I need one. So on these grounds, I would have a right to be provided with a kidney. The case easily generalises to other body parts, material resources, political power, and so on—it seems I would have quite extensive rights to the world around me. This is hard to believe. In addition, I also have a strong interest in not being forced to provide kidneys to others, based on an interest in my bodily integrity. Assuming that everyone has such an interest, this would lead us to assign a right to everyone not to have their kidneys forcefully taken. The conflict with the right to be provided with a kidney is easy to see. Direct justification would not merely proliferate rights extremely, it would also set us up for a massive conflict between different people’s rights.
This problem arises if we do not take into account the burdens associated with assigning a right. Each right corresponds with a duty (or some other Hohfeldian incident). It is often in people’s interests not to have those. I have an interest in your kidney, but when compared to the burden on you of taking your kidney, it is not clear whether this interest is strong enough. If the interest theory is to be at all plausible, its advocates must make some comparisons between the benefits of having a right and the burdens of having the corresponding duties. Contrary to the surface appearance of Tasioulas’ test, he is aware of the problem. Tasioulas wishes to include the burdens of assigning a right in step (iii) as part of a general estimate of what is “feasible”. We should reject a set of duties, he tells us, which would be “excessively” burdensome or which “would be [too] burdensome in relation to our capacity to realize other values, including human rights”. But the label “feasibility” is misleading. It is certainly feasible that I have a right to your kidney. The problem with this right is not that it is infeasible, but that it is unjustified. Compare: it is feasible that we kill people for violating traffic laws; it just would not be right.
So in step (iii) of the procedure, the interest theorist needs to include some form of weighing between the interests of the possible rights-holder and the corresponding duty-holder. We need, in short, an account of how burdens and benefits across different people are to be weighed. The interest theory must avoid making comparisons on the basis of aggregate quantities; otherwise, it risks collapsing back into a form of utilitarianism. Instead, the interest theorist must explain how comparisons can be made which respect individualism. What should replace aggregation is an extremely difficult philosophical issue, and I will provide no solution here. The important upshot for our purposes is that direct justification for the interest theory cannot be defended, at least not in the sense that interests directly translate into rights merely on the basis of their strength.
One question we still have to briefly tackle is whether the interest theory is supposed to be read in a consequence-oriented way, or in a purely a priori way. Tasioulas speaks of a right “serving” an interest, and this is language I have also generally followed. However, there are two interpretations of this phrase:
- An interest is served through a right just in case, were the right socially or politically recognised, this would promote (or protect, or in some other way positively affect) the interest.
- An interest is served through a right just in case the right expresses the appropriate moral reaction to the interest.
Interpretation (A) focusses on what causal impact a right has, while (B) does not. If we interpret the interest theory along the lines of (A), then it is a consequentialist theory, but in the minimal sense that assigning rights is some function of the expected consequences of the legal or social recognition of some right. (As we saw, those consequences will not be evaluated in a straight-forward aggregative way.)
Which of these interpretations we give to the interest theory has some impact. Consider the following scenario: we live in a world in which unbridled capitalism, and a system in which rights to social welfare are neither legally nor socially recognised, best serves everyone’s interest; on the other hand, if we legally or socially started to recognise such a right, everyone would be worse off. On interpretation (A) of the interest theory, we could not justify a (moral!) right to social welfare, as the right would not actually serve the relevant interest. But on interpretation (B), we might be able to justify such a right. We might claim, for example, that recognising a right to resource provision is the appropriate moral reaction to the strong interest we have in our own welfare, even if its legal or social recognition would be detrimental.
Almost all advocates of the interest theory intend their theory to be read along interpretation (A). This is suggested by the common causal language of a right “serving”, “protecting”, or “promoting” (etc.) our interests. However, there is one difficulty in understanding (A), as there is no obvious interpretation of what it means for a moral right to cause anything. Moral rights just are; they do not have any causal impact. Of course, the individual or collective recognition of moral rights has a causal impact. If individuals accept that people have a moral right and then act in accordance with this belief, the interests of individuals will be served, promoted, thwarted, undermined, protected, and so on. But this brings us to the next problem. There is no unique interpretation of what it means to recognise a right. Consider Anna’s moral right that no one step on her toe. There are several interpretations of what it would mean for this right to be recognised:
- There is a social convention not to step on Anna’s toe (social recognition);
- People believe that Anna has a moral right that no one step on her toe (epistemic recognition);
- There are institutions who reliably punish those who step on Anna’s toe (political recognition);
- Some combination of different degrees of (i) to (iii).
It matters which of these interpretations we choose. Anna’s interests might be promoted and protected to different degrees depending on whether her rights are socially, epistemically, or politically recognised, and to what degrees. The interest theory will sometimes deliver different results given different accounts of recognition. So to complete our account of the interest theory, we would need to fully specify what it means for a right to be recognised. It is not immediately clear which specification we ought to choose—none is clearly privileged over the others. Most authors probably assume some kind of legal-cum-social recognition, and this is also what I will presume.
Voluntarism and the Interest Theory
I will add some further amendments to the interest theory as we go. But enough of the theory has emerged that we can ask what practical implications it has for the shape of our rights with respect to the Basic Problem. Here is an intuitive idea: if our rights are based on interests, then they will not rule out forms of government which, while non-consensual, promote our interests. An interest-based set of rights which did would be self-defeating: it would stop us from achieving the very purpose for which we assign rights. It is this idea which I will develop in this section.
We can start with the first element of the interest theory. There are two general types of interests which are relevant when it comes to political institutions. We can somewhat schematically distinguish them as interests in being free, and interests in being provided with certain resources. We can expect that individuals differ in what specific freedoms and resources they have an interest in, and how strong those interests are, but until section 3.3.3 I will presume that interests across people are relatively homogenous.
There are some theories of well-being on which our interests in freedom greatly outweigh interests in resources. At its most extreme, consider a broadly Nietzschean theory of well-being on which creating your own, fully independent life is the most significant element in your life going well. If this is true, having to live under a government—or for that matter, being subject to the social or political power of anyone—is one of the central ways to make your life bad. If you believe in such a theory of well-being, then attempts to justify government become hard if not impossible. Another claim which leads to similar results is that independence is not by itself a good, but that we can develop into flourishing, virtuous people only if we are free from the social control of others, that is, only under conditions of anarchy. These theories do not merely claim that we have an interest in freedom; they claim that we have an overriding interest in being free from the control of others.
There are several points to be made against this theory. First, we can observe that many people have lived under modern governments happily, and have not thought of government interference with their freedom as a major detriment to their well-being. Even where many citizens think that government makes their life worse, most people do not object to the very fact that the state exerts power; they object to the particular policies that their government has pursued. There are some people, of course, who love their own freedom so much that they flourish only if they live outside central social control altogether. But they are the exceptions to the rule. These are observations the Nietzschean can hardly dispute. So she owes us a sociological error theory which explains why most people suffer from a type of false consciousness when they perceive their lives as going well while they are subject to government. I am sceptical whether such a theory can be provided. We can certainly be mistaken about how well our own lives are going; and we can even imagine that large groups of people are mistaken about their own well-being. But the Nietzschean theory would make virtually everyone everywhere mistaken.
Independently, it is hard to see what would be so bad in being subject to government. We are not asking whether it would be wrong or unjust if you were subject to government control—we ask whether it would be necessarily bad for you. We must clearly separate these dimensions. If we do, a large part of the appeal of a Nietzschean theory of well-being vanishes. There is no question that we have some strong interests in being free—in being able to shape and create our own lives. Political institutions, especially coercive ones, have various negative effects on us. They have a tendency to stunt critical, independent thinking, and to make us overly anxious and dependent on those in power. Even if we accept these as inevitable negatives, living in political society has various positive benefits. It is difficult to see why we could only flourish in an anarchist society.
Let us move on to a second point. You might reject the Nietzschean theory of well-being, but accept that we have strong interests in living our lives freely. We need not nail down the idea of freedom. It might contain both “negative” and “positive” elements: we have an interest in finding no obstacles in the pursuit of our most-valued aims, and we have an interest in having the ability to reach those aims. A first observation is that entities who could make us unfree are numerous and not limited to political institutions. In the absence of government, we might find our liberty restricted through our family, our clan, local leaders, our church, and through other economic, social and military institutions beyond our control. It might turn out that an anarchic system of social organisation is considerably worse than a system with centralised government control in this respect. While an anarchic society might not have one agent who monopolistically restricts our freedom, it might have a wide and dispersed set of agents who restrict our freedom such that we are overall less free. This suggests a simple defence of political institutions: we have no right against their interference because they promote our overall freedom.
This brings us to an issue with the interest theory I have not yet tackled. Assume that you have a freedom-related interest in not being taxed. On the basis of that interest, we assign to you a right not to be taxed. Assume that you also have an interest in no one forcing you to work for them. So we assign to you a right that no one can force you to work for them. You have a whole range of freedom-related interests, so you will end up with a large set of rights against interference. However, there is no guarantee that this set of rights serves your overall interest best. The point is that if government was allowed to interfere with some of your freedoms, it might be able to create more freedom for you in other ways. For example, if we allowed political institutions to take some of your property, it can finance a police force which might make you freer in other respects.
This is a simple aggregation problem. It can be resolved by not deriving rights interest-by-interest, but by looking at the entire set of rights we ascribe to individuals. We can also throw in the corresponding duties, and so look at the whole set of rights and duties that individuals have. I see no reason why the interest theorist should resist this move. Different interests of the same individual combine into a sense of how that person’s overall life is going. Insofar as the interest theorist cares about interests one-by-one, it would be absurd not to also pay attention to how people’s overall lives are going. An interest theory which focussed on one interest at the expense of a person’s overall well-being would be absurd. If we furthermore add the idea that the interests and rights of individuals can differ, a quite complicated picture emerges. Schematically:
||For individual A, XA, YA, ZA, … would serve some of A’s basic interest(s); for individual B, XB, YB, ZB, … would serve etc.
||A’s interest in XA, YA, ZA, … is important enough to justify imposing duties on others; B’s interest in XB, YB, ZB, … is important enough to justify etc.
||Giving A the set of rights and duties RA, giving B the set of rights and duties RB, etc., best balances burdens and benefits as relating to A’s interests in XA, …, B’s interests in XB, …, etc.
||A has rights and duties RA, B has rights and duties RB, etc.
This statement of the interest theory is much more complex than the simple version we started out with. I should note in passing that one of the most appealing features of the theory, its apparent simplicity, has now disappeared. At any rate, the central move we can make inside this enhanced theory should now be clear. We have an overall interest in our rights being limited in certain ways, on the plausible assumption that political institutions can often serve our overall interest. This is especially true if we take into account interests in resource provision. But even if we limit ourselves to interests in being free, as I have done so far, we can get the argument going.
The outcome of this argument is not that we have no rights against governments whatsoever. For example, we have interests in not being tortured, and an interest to choose our own sexual partners and friends. Political institutions could not permissibly infringe those rights. What the argument shows, however, is that certain of our rights are limited. For example, it is likely that not having a right to our full pre-tax income is in our overall interest, because not having that right allows beneficial political institutions to operate. Difficult questions need to be answered—e.g., how optimistic we should be regarding political institutions, and how you weigh interests in freedom against interests in resource provision. But it is likely that across a wide range of cases, that trade-off will come out in favour of political institutions.
This strategy relies on the assumption that having a government generally is in our all-things-considered interest. You might doubt this claim, and argue that modern states tend to be corrupt and inefficient. Alternatively, you might think that a minimal (or no) state is preferable, as a society largely unconstrained by political institutions will lead to both greater wealth and individual freedom. Large parts of the anarchist literature are concerned with sketching how a stateless society, or one with a limited state, would be economically preferable in just this way. I cannot fully engage with this literature here. At any rate, it is clear that a number of political institutions are not in the interest of those who are governed by them—for example, the North Korean state. What shall we say about such cases?
This brings us to the question of how the interest theory ought to deal with exceptions. Political institutions might be generally in our overall interest, but this government might not be, or not in this respect. How should the interest theory react to such exceptions to a general rule? This is intimately linked to the question of how fine-grained we ought to be in our application of the interest theory. I have already suggested that we should apply it on an individual-by-individual basis, and not with reference to the interests of the “typical” or “standard” individual. However, it is not clear whether the same type of pressure exists to consider interests on a fine-grained issue-by-issue basis. We might think that basic moral rights are quite general in their content, and so we might choose a comparatively coarse-grained way of individuating issues. In that case, it is unlikely that we will have general rights against government interference, because it is unlikely that it is generally in our interest to have such strong rights. Assume, however, that you assign rights on a more fine-grained basis. In that case, North Koreans would likely have a right against being interfered with by the North Korean government, over and beyond other general rights such as a right not to be tortured. If this is the case, then there is no universal solution to the Basic Problem. In some cases, non-minimal political institutions would violate our rights, in others they would not. What would emerge from such a fine-grained application of the interest theory is a highly specified right: a right not to be subject to certain bad political institutions, such as the North Korean government; but this right would not extend to beneficial government such as (say) the Swedish government.
Let us turn to another type of exception. Not being subject to government, while not in everyone’s overall interest, is in the overall interest of some. To return to the case of the Amish, they have a strong interest in being free from most kinds of interference by the US government. They prefer a communal, agrarian life which is designed as a rejection of modern society. As another case, assume that there are some genuine Nietzscheans: people who see themselves as self-creators who find any type of obstacle imposed on them by others unbearable; who would rather starve than be helped or coerced by government or anyone else. The interest theorist should accept that people whose psychology is structured in such ways have a prima facie claim to be free from government interference.
These interests need not translate into rights, however. First, even if some individuals have these interests, they still need to be balanced against the interests of others (in step (iii) of the justificatory procedure). We could give the Nietzscheans a wide-ranging set of rights to be free, but that might prove to be very disruptive to the operation of the state. This in turn is likely to set back the interests of the majority who are not die-hard anarchists. My guess is that we will find some compromise in which we give committed anarchists rights to opt out of some aspects of the state. Perhaps, for example, we require them to abide by basic rules which ensure public order, such as traffic rules and property law. We are likely to also require them to pay some basic amount of tax. At the same time, we will allow them to avoid the more involved aspects of the state, such as public health insurance, contribution to redistributive policies, jury duties, and so on.
Thus, political institutions will have some semi-independents living on their territory to whom only a subset of the public rules applies. The state will be permitted to treat these independents in some ways, but not in all ways in which it can treat the majority of its citizens. There are various practical problems in implementing such schemes. It might be easy to exempt the Amish from some forms of state control, as they are an easy-to-identify group which is clearly separate from the rest of society; but less homogenous groups are much harder to deal with. Is it enough if you claimed to be a Nietzschean to be allowed to opt out of the majority of legal rules? If we allowed self-proclaimed Nietzscheans to pay a lower tax rate, many of the rich might suddenly discover that they were Nietzscheans all along. Still, it strikes me that existing states are not creative enough in drawing up solutions to accommodate dissenters.
A Natural Conclusion
Be that as it may, these exceptions do not show that rights justified through an interest theory would undermine the legitimacy of the non-minimal state in general. Instead, they show that such a state could not exercise its powers indiscriminately; it will be required to make exemptions for particular groups. In summary, a complicated picture emerges if we accept the interest theory. Most individuals in most types of circumstances will not have rights against taxation and other types of government interference. This leaves space for non-minimal political institutions to operate. Thus, if we accept the interest theory of moral rights, including the various modifications of it I suggested, then we can solve the Basic Problem. If you step back from the argument I have provided, this is not in the end greatly surprising. The interest theory is a reductive theory which grounds rights on interests. We should expect that certain kinds of non-consensual government are permissible, as long as they are in the interest of the majority and do not impose unreasonable costs on the few. In fact, under many circumstances the interest theory justifies a claim-right of some individuals that a non-minimal state be established, a social welfare state in particular. This follows straight-forwardly from the interest that some individuals have in such a state to exist. Far from supporting voluntarism, the interest theory points us towards a rationalist theory of legitimacy.
This completes the first leg of my solution to the Basic Problem. I think a solution based on the interest theory is the one most high liberals will favour. I have presented this argument because the interest theory is so widely accepted, not because I myself accept it. I believe the status theory to be correct, which I will now describe (though not argue for). As we will see, if you accept this theory solving the Basic Problem is significantly harder.
The Status Theory
The status theory is a non-consequentialist theory of rights. It does not ground rights on their ability to serve some aspect of individuals, such as their interests, freedom, or autonomy. Rather, the idea is that rights are the proper moral response to a feature that individuals possess. To put the contrast differently, the interest theory sees rights as instruments to promote something we find valuable in individuals; the status theory sees rights as expressions of the respect we owe to something we find valuable in individuals. So while an interest theory might say that we ought to promote the autonomy of individuals, the status theory tells us that we ought to respect the fact that individuals are autonomous entities who lead their own lives. The difference this makes should be clear. If the focus is on promoting autonomy, we can make the kinds of weighing moves I suggested in sec. 3.3.2. But respecting autonomy does not allow such weighing in an obvious way. Indeed, a plausible thought is that respecting the autonomy of people entails not overriding or ignoring their choices, even if doing so would promote their overall autonomy better. If so, justifying the state becomes much harder.
The feature, or set of features, that we ought to respect in individuals can be various. In line with the general strategy of this thesis, I will abstain from providing a substantive account. I will generically speak of the “moral status” of individuals we ought to respect, to stand in for whatever the precise feature is one’s preferred theory singles out. But let me take note of some candidates. A first category of candidates are metaphysical descriptions of the nature of individuals. We might say, for example, that individuals are autonomous persons with separate lives to lead, or entities endowed with free will; or beings who are capable to respond to reasons, which have the power to set ends and pursue them rationally. A second type of category would focus on a more naturalistic description of individuals—e.g., that humans are social animals vulnerable to predation, or entities who pursue their own self-interest, or who are committed to certain ground projects and aims. A third category consists of moralised descriptions of individuals. We might focus, for example, on personhood as a ground for rights, or dignity, or the worth of persons and their inviolability. The general tendency in recent moral philosophy is to pursue the third type of approach, and tell a metaphysically shallow story about why individuals possess dignity or some type of inviolability.
As a second step, a status theory needs to specify what respecting the relevant feature of individuals entails. I will work with a very broad-brush notion, speaking of the “appropriate moral reaction” to a feature of individuals. The appropriate moral reaction to entities with dignity, for example, is to treat them respectfully, to sustain them and assist them in the face of threats to their dignity, and so on. The appropriate moral reaction to an entity with autonomy is to respect their freely made choices, to not undermine their capacity to act autonomously and restore it where possible, to create conditions in which autonomy can flourish, and so on. As you can see from these descriptions, a status account need not lead to merely negative duties and rights. Sometimes the appropriate moral reaction to some feature of individuals is to bestow some benefit on them, or to help them develop some capacity, or to protect them from certain threats.
In short, the justificatory argument in a status theory proceeds along the following lines:
||Individuals are entities with moral status—e.g., who are autonomous persons with an ability to set and pursue aims.
||The appropriate moral reaction to entities with moral status of this kind is to never do X to them (or always do Y, or promote Z).
||Individuals have a right that we never do X to them (or always do Y, or promote Z).
Let me discuss some of the features of this scheme. First, while the interest theory is remarkably flexible in its ability to account for the varying interests that individuals have, the status theory at first sight looks rather rigid. It starts from highly abstract features that all individuals share, and thus appears to entail that everyone everywhere has the same basic moral rights. Would it not have problems to account for the Amish or the Nietzscheans I discussed above? This would be a misleading impression. The status theory is flexible in at least two respects. First, the feature on which we ground rights will often be a general capacity, such as a capacity for reason or autonomous action. But we owe respect not only to this capacity, but also to its exercise. The Amish, for example, exercise their autonomy by choosing to live in the small communities they do. As a result, respecting the Amish’ commitment to this form of life is an important way to respect this more general capacity—and obviously, they exercise this capacity differently from a secular cosmopolitan. Second, what the appropriate moral reaction to a capacity is often depends on cultural and social background facts. This is particularly clear with dignity. A part of respecting the dignity of others is not to make them, or regard them as, social inferiors. What this practically requires is largely determined through social and cultural conventions, and thus will greatly differ across various societies.
Still, it should be admitted that the status theory will tend to assign rights to individuals with much more general content than the interest theory. One worry is that these rights will be too general. Take the issue of freedom of speech. Whatever exactly we choose as the ground of our rights, it seems that the status theory can only give relatively bland advice. We might say, for example, that respecting that people have minds of their own entails that everyone should be able to speak their minds. But surely we are interested in more specific results—what about hate speech, pornography, subliminal advertising, financial contributions to political campaigns, or libel? Some have suspected that the status theory has either nothing to say about these issues, or delivers just vague generalities. Part of the response to such objections is in the details. We would need to show that (for example) a dignity-based theory of speech rights can deliver both adequate and specific results—say, with regard to hate speech. Alternatively, we can accept that a status theory often delivers indeterminate results. So far as moral theorising is concerned, it might provide no clear guidance for a number of cases. But this raises the question whether our moral theory, all by itself, really needs to deliver such determinate results. As I will argue in the next chapter, the vagueness of our moral rights entails that we need political institutions which provide our rights with more specific content.
A more fundamental objection concerns the idea of the “appropriate moral reaction”. You might suspect that this notion is doing much of the argumentative work in the status theory. What explains that respect for autonomy requires that we assign individuals rights to freedom of speech? It seems that the status theorist is reduced to a “just-so” explanation. These worries are particularly strong if we start from a moralised description of the respect-worthy features of individuals. Consider a status theory which focusses on the inviolability of individuals. Such a theory will tell you that you ought not to sacrifice the few in favour of the many as you would otherwise disrespect their inviolability. This theory reshuffles the justificatory issue to the claim that individuals are inviolable. However, that looks to be an unsatisfying answer—you likely want to know why individuals are inviolable, and the theory provides no answer to this question. This is a particularly pressing problem for types of status theory which start from an overly moralised account of the moral status of individuals. Other theories will run the danger of being uninformative much less. Still, some uneasiness might remain about the notion of what counts as the “appropriate moral response”. In this respect, the status theorist can only plead guilty by association. Any deontological theory claims that there are some ways in which we never ought to treat individuals. At some point, the explanation of this fact will come down to some bedrock claim regarding the nature of morality. We might try to describe the idea of the appropriate moral response in other language, and we can motivate it through a series of examples. But it is hard to see what a deeper explanation would be that we should treat (or not treat) people in some ways because they are certain kinds of entities. In this respect, the status theory is not more or less mysterious than any other deontological theory.
With regard to the interest theory, I argued that we miss important trade-offs if we assign rights on an individual-by-individual, issue-by-issue basis. Can the same be said for the status theory? We certainly need to take all individuals at the same time into account when we assign rights on the status theory. For example, the dignity of individuals might give us reason to assign rights protecting potential victims from hate speech. At the same time, the fact that individuals have “minds of their own” gives us a reason to allow for broad rights to freedom of speech including, perhaps, a right to utter hate speech. So here we have two ways to think about individuals which pull into different directions. It is not immediately clear what the right response is—whether we should put more emphasis on the dignity of the victims of hate speech, or the autonomy of the perpetrators.
At any rate, the current conflict is not the type of trade-off case we considered previously in the context of the interest theory. We cannot trade off respect for one respect-worthy feature in one individual against respect for another (or the same) respect-worthy feature in another individual. What we would instead need is a fine-tuning argument: we would need to argue that respect towards persons with both dignity and autonomous agency favours one solution over another.
Even if we set these problems aside, there are other ways in which a status theory needs to be fine-tuned. In the context of the interest theory, I highlighted that it might be in our overall interest not to have certain rights. Is there any equivalent of this thought in the status theory? In some ways, there is. Focus on the status of inviolability. Assume that there is a set of prohibitions regarding some individual that forbid sacrificing or harming that individual for the greater good. There could be more or fewer such prohibitions, allowing fewer or more types of trade-offs. To the degree that more such prohibitions apply to the individual, let us say that she enjoys a higher degree of inviolability. If individuals enjoy a high degree of inviolability, then it will be impermissible in many cases to use individuals non-consensually to achieve collective goals. For example, I could not force you to contribute towards rescue efforts benefitting third parties.
However, if we expand the scope of inviolability in this way, we at the same time reduce another form of status that individuals have, which we might call “saveability”. You are saveable in a particular situation if other people have a duty to save you, and they have a license to coerce others into contributing towards such help. The extent of your saveability is given by how many situations there are in which other are allowed to save you. The inverse relationship between inviolability and saveability should be clear. The greater the extent to which people are inviolable, the smaller the extent to which people are saveable, and vice versa. Again, this is not the type of consequence-oriented trade-off which we encounter in the context of the interest theory. This case does, however, point towards a way in which a status theory needs to be fine-tuned, in that it needs to adjudicate between the different types of moral status individuals have. It is not enough to say that individuals are inviolable; we need to specify more precisely the extent to which they are inviolable rather than saveable.
Voluntarism and the Status Theory
The Status Theory provides a bigger challenge to rationalism. If we accept certain claims about individuals—in particular about the importance of their freedom, will, and independence—then the appropriate response to these features appears to be to assign to them far-reaching rights against interference. This would lead us to a voluntarist position. Is there any way we can avoid this implication?
The Finetuning Argument
One attempt is to go down the fine-tuning road, and argue as follows: if we provide individuals with strong rights against interference, then we pay strong respect to the idea that they are independent and free individuals, and that everyone should be master of their own life. Call this the sovereignty of individuals. Sovereignty expresses the idea that individuals are akin to small-scale sovereign states who possess strong prerogatives over their own affairs. But we should not be so quick as to jump from this to a voluntarist conclusion. If individuals had more restricted rights against interference, we would respect them more in some other way. In particular, if we allow a non-consensual welfare state to operate, this expresses respect for the idea that individuals pursue aims in their lives, and that having the ability to achieve these aims is an important element in leading a meaningful life. On this picture, individuals are the kinds of beings it is worthy to support, so let us label this support-worthiness. Assume that we accepted a distribution of rights and duties which allowed the welfare state to take from some to give to others. This would make individuals less sovereign. But it would put greater emphasis on the idea that individuals are support-worthy.
I sympathise with the Finetuning Argument. It is strange that sovereignty should be the primary thing we should care for in individuals. Respect for people requires us to both take into account their independence and autonomy, but also their vulnerability and how their lives are going. However, the Finetuning Argument is severely limited in its dialectical reach. Those with a more libertarian bent will simply remain unconvinced. They are likely to stress that it is indeed our sovereignty which is the primary feature of individuals that we should respect. It is hard to provide arguments on this level which convince the other side, as we seem to have reached normative bedrock pretty quickly. So there are good reasons to look for another argument which transcends these limitations—that is, an argument that does not commit us to a particular fine-tuning of different types of moral status.
The Anti-Paralysis Postulate
Another, closely related argument is Eric Mack’s Anti-Paralysis postulate. Mack states it as follows:
[W]hen working out the detailed specification of person’s rights, one is to avoid specifications that systematically morally preclude individuals from exercising their rights or from conducting their lives in ways that a specification of their right is supposed to protect. The intuitive idea that the detailed specification of rights – insofar as that is the job of philosophical reasoning – must be guided by the purpose for which rights are to be recognized as a crucial dimension of morality.
As an example, Mack mentions a right to engage in risky activities—i.e., activities which risk injuring others. If we had no right to engage in such activities, virtually nothing we do would be permissible. Driving a car, for example, routinely risks severely injuring others, even if probabilities might be small. If we had rights against being exposed to any kind of risky activity, driving a car would be morally wrong. Rights would leave each of us with little liberty. But this, Mack claims, would be self-defeating: rights are supposed to make us free to pursue our lives; this is why we accept them into our moral theory in the first place. Thus, according to Mack, we do have a right to engage in activities which bear some risks that others are harmed.
Mack suggests that Nozick should have appealed to the Anti-Paralysis postulate when arguing for the legitimacy of his minimal state. The problem for Nozick is that the minimal state non-consensually coerces some independents who are living on its territory. It forces those independents to contribute to the minimal state; but that seems a clear violation of their rights. With regard to those independents, Mack claims that Nozick should have argued the following:
in virtue of the public goods feature of rights protection, rights-bearers would be substantially less able to coordinate so as to protect their rights if their rights are understood to preclude their being required to contribute to the funding of protective schemes. If that is the case, a plausible specification of their rights would not include a right protected by a property rule against being required to contribute to such a scheme.
In other words, the minimal state can permissibly coerce independents, because otherwise we would be paralysed: without the state, we would spend much of our time trying to protect our rights, and unable to enjoy their exercise.
The Anti-Paralysis principle is intuitively convincing. It expresses a methodological commitment to reflective equilibrium: we will want the fundamental moral grounds of why we assign rights to align well with the particular shape of rights we assign. The trouble for the Anti-Paralysis argument, however, is much the same as for the Finetuning Argument: it will have little dialectical power against those who are not already convinced. In the context of the state, the question is precisely what would count as objectionable forms of paralysis, and what would not. Anarchists will just dig in their heels and reject that a state-of-nature solution is paralysing. This is not to say that the Anti-Paralysis postulate is entirely without usefulness. Later, I will appeal to it in a different context (sec. 4.2.2). Similarly, I do not suggest that the Fine-tuning Argument is wrong; just that its dialectical power is limited. The Kantian argument I will explore in the following chapter aims to be compatible with a wider set of substantive theories of rights.
The Kantian Solution
Justice (Gerechtigkeit) is holy, but only through law (Gesetz) the Right (Recht) is first constituted.
We are still stuck on the problem of how we can combine two central ideas: that individuals have strong moral rights, and that a non-minimal state can legitimately operate. In search of a solution to this problem, Kant’s political philosophy is of particular interest. Kant combines a deontological approach to morality with a rationalist approach to legitimacy. Kant’s main idea is expressed in the epigraph: there is an intricate connection between political institutions and justice. We require political institutions to stand in fully rightful relations to each other. Thus, far from political institutions conflicting with our moral status, they are necessary to bring it about, in a sense to be described. This is a far-reaching, non-trivial claim about the nature of our rights.
In this chapter, I motivate a Kantian argument along these lines—I will spend no time on Kant exegesis, and do not claim that the argument I develop would be acceptable to Kant. First, I describe one way how one might aim to advance the Kantian argument (sec. 4.1), based on the right to independence, which I think is unconvincing. However, there are some central ideas we can glean from this account, on the basis of which I offer an improved argument (sec. 4.2). Afterwards, I tackle the details (sec. 4.3), limitations (sec. 4.4), and implications (sec. 4.5) of the argument.
The Argument from Independence
I start with the mainstream way in which the Kantian argument is usually given, as an argument from independence. The rough outlines of this argument are decently clear. At its core is the claim that we have one, and only one, basic moral right: a right to “Freedom (independence from being constrained by another’s choice), insofar as it can coexist with the freedom of every other in accordance with a universal law”. This is a quite sweeping right. It entails that freedom can only be curtailed for the sake of other people’s freedom. Given such a strong right to be free, we would expect a voluntarist approach to legitimacy. This would be a standard Lockean way to proceed from this basis. But a key idea stops the Kantian from going down this route. This is the belief that in a state of nature I am exposed to the arbitrary and “unilateral” interference of other people with my freedom. So we remain unfree in the state of nature.
The best-explored example, both by Kant and in the secondary literature, are property rights. Kantians accept, like Locke and the libertarians, that I am allowed to acquire parts of the world as my own property. Lockeans set some type of constraint (“proviso”) on acts of original acquisition. As long as your acquisition respects this proviso, your acquisition is rightful. Thus, there is a set of moral rules which regulate property acquisitions and transfers. As long as you act in accordance with these principles, no other agent’s authorisation or consent is needed for you to rightfully acquire property. In principle, political institutions are not required on this picture at all, at least as the moral rights themselves are concerned.
The Kantian departs at this juncture. The crucial idea is that any acquisition of property in the state of nature is problematic: if I declare this piece of land mine, then I unilaterally impose my will on all others, and thereby restrict their freedom. If I successfully acquired land, I would impose new duties on everyone else—for example, I make it impermissible for others to trespass on my property without my consent. This diminishes the freedom of others, as it constrains them in what they can permissibly do. Thus, the Kantian concludes, the rightful acquisition of property in the state of nature is impossible, as any such acquisition would violate our natural right of freedom.
On the flipside, moving to a civil condition can remedy this defect. A civil condition is one in which we stand under a shared authority which determines what is ours through public laws, and assures us of our rights through coercive means. Kantians add that such a civil condition can only be brought about through political institutions like the state. The law given by such institutions expresses the unified will of all; so being subject to these rules does not violate our right of independence. If I acquire property in a civil condition in accordance with the publicly set rules, then I do not impermissibly diminish anyone’s freedom. Rather than the state being in conflict with individual rights, as the Lockean claims, having a state is the only way to avoid systematically violating the basic moral rights of others. The tables have been turned.
This argument does not justify any kind of coercive political institution. Governments which do not fulfil some minimal conditions fail to bring about a civil condition. But I set discussion of those conditions aside. Schematically, the argument is the following:
||We have a basic moral right to be free from the arbitrary interference of others.
||Only under non-consensual political institutions (of a certain kind) are we free from the arbitrary interference of others.
||Non-consensual political institutions are necessary to avoid violating a basic moral right.
||If political institutions are necessary in such a way, then they are legitimate.
||Non-consensual political institutions (of a certain kind) are legitimate.
We must not mistake this argument for a Hobbesian or Samaritan argument based on the benefits of leaving the state of nature. In a Kantian context, we can assume the state of nature to be a benign and peaceful situation in which people generally cooperate and adhere to basic moral rules. Instead, the argument from independence is an a priori argument: it makes the much more controversial claim that a certain kind of entity is required in order for us to act rightfully. So the argument does not rely on outcome-focussed assessments at all.
The argument is tied to a specific claim about the content of our rights (claim (1)), a form of Kantian Republicanism. According to Kantian Republicanism—again, I will ignore whether this is Kant’s position or not—we have a fundamental right to be free from the unilateral interference of other people. There are non-Kantian forms of republicanism which make similar claims. Some contemporary republican philosophers have claimed that freedom is not being dominated by others, where domination is understood as the arbitrary interference of others. This allows us to make a similar argument that the state is necessary to free us from the domination of others.
There are four reasons why I reject the argument from independence. First, the overall strategy of this thesis has been to remain neutral, as far as possible, on substantive questions of justice. It would be preferable if we had a solution to the Basic Problem which was compatible with a wide range of moral theories (even if not all). The current argument, however, ties us to a specific claim about the content of our rights.
Second, it is hard to see how the state is both necessary and sufficient to free us from the unilateral interference of others. The offices in real political institutions are filled by people who are corrupt and biased, who exercise power arbitrarily, and who abuse the state apparatus for self-interested purposes. This exposes the argument to a dilemma. On the first horn of the dilemma, we might say that corrupt state officials act in a merely unilateral way. In that case, public institutions are not sufficient to free us from the domination of others. While the state frees us from the domination of our fellow citizens, it exposes us to the domination of another group of people, state officials. If this were the case, the state would infringe our right to freedom, and we would have made no progress in our argument. On the second horn of the dilemma, we might say that corrupt state officials do not act in a unilateral way. But in that case it becomes hard to see why the state is necessary to free us from the domination of others. Private individuals in the state of nature are as biased and self-interested as state officials. Thus, there is pressure to think that it is also possible for them to act in a non-unilateral way. More could be said on behalf of the Kantian republican, but I am sceptical whether any attempt to avoid this dilemma succeeds.
Third, the argument has a number of hard-to-swallow implications. As Kant has famously noted, the argument entails a general prohibition of revolution. Once we are in a civil condition, any extra-legal attempt to replace the existing legal system with another must be seen as a unilateral exercise of the revolutionaries’ will, and thus wrong—unless, per impossibile, everyone agrees to revolution. So we would not be allowed to switch legal systems in extra-legal ways. Another awkward implication of the argument is that we do not merely need a state for the civil condition to be realised, but a world state. After all, if we entered into a civil condition merely with fellow Frenchmen, then attempts to acquire property would still unilaterally impose duties on Germans. Only if French and Germans entered into a single state, and thus subjected themselves to the same unified will of all, could this be avoided.
Lastly, Kantian Republicanism does not rest on plausible moral foundations. There is no convincing sense in which we have a general right to be independent from the arbitrary or unilateral will of others. We interfere with each other all the time—there is surely no unqualified right against interference as such. So everything hangs on the qualifiers “arbitrary” and “unilateral”. The problem is that it is hard to see what a non-circular, non-question-begging account of arbitrary interference looks like. Think about it in terms of the state. We must say that the interference of others with us in a state of nature is arbitrary, whereas the interference of the state is not. It is hard to see how we could draw this line without already presupposing what we aim to show—i.e., that the state can be morally justified.
The Argument from Publicity
Fortunately, there is a different way to advance the argument which gets rid of any reliance on Kantian Republicanism. One of the main ideas contained in the argument from independence is that publicity is important to our rights. A useful slogan to summarise this idea is that justice must not only be done, but that it must be seen to be done. If our rights are not publicly recognised—if there is not a shared framework of rules which governs the content and distribution of rights and duties—then they remain incomplete in an important sense. This idea is not wedded to any particular view on the content of rights. Instead, it is a general claim about the nature of our rights. I will argue:
||Whatever the content of our moral rights, for those rights to be fully realised they need to be public.
||Certain kinds of political institutions are necessary for our rights to be public.
Taken together, these two claims entail that our rights can only be fully realised under political institutions. In this modified argument, the fault with the state of nature is not that people are wronged in it, but that their rights are incomplete. This changes how we ought to complete the rest of the argument, a task I turn to in sec. 4.2.2. But first I wish to comment on the crucial notion of a right being “realised” completely or incompletely.
The Realisation of Our Rights
We can return to the example of original acquisition. Assume that I stake off a portion of land in the state of nature. I presume now that such acquisition by itself does not wrong others. But the right you have gained under these circumstances lacks some important features. First, there is no guarantee that others will recognise your claim to original acquisition. We can expect disagreement over the correct rules of acquisition, even if everyone shares a broadly Lockean mind-set. There will also be disagreements over factual matters which underlie the application of moral principles. Furthermore, there is no way how you could publicly register your claim. There is no court or authority where you could field complaints against others who violate your property rights. Lastly, when it comes to enforcement, you have to rely on your own wits or on the help of others. Taken together, these aspects of the state of nature make it inconvenient: they make the enjoyment of your rights less secure, and give you reason to wish for political institutions. But the Kantian intuition is that these shortcomings also amount to a more fundamental flaw. If your rights are lacking in these ways, then your rights remain merely provisionary, or as I will say interchangeably, incomplete. In a state of nature, you should not be thought of as having the relevant property rights in a full sense.
In this context, it is useful to distinguish three dimensions along which we can think about rights. Assume you have a property right in a house you built. Two questions to ask are the following:
- do others infringe this right?—e.g., do they damage or enter your house without your consent?
- beyond non-infringement, do others respect and promote your right?—e.g., are they willing to help you to protect your house against damage or entrance by third parties?
What I have in mind with regard to the second question is the degree to which individuals fulfil secondary, rights-related duties. For example, the strict correlate of a right to bodily integrity is a duty not to maim or harm. But there is also the weaker duty to prevent third parties from maiming or harming others. Failing to prevent others to do so is not itself a rights-infringement, so it does not fall under heading (i). Still, the fulfilment of these additional duties is an important aspect of caring about rights.
However, I am interested in a third, yet independent dimension:
- to what degree is your right realised?—e.g., to what degree is the content of your right determinate and public? is there a public authority which ensures you in your right against others, and resolves conflicts regarding it? are your rights recognised in the moral community you are a member of?
The answer to question (iii) is independent from our answer to the other three questions. First, the realisation of a right is distinct from its non-infringement. People in the state of nature might be well-disposed and never trespass on your property. Still, the Kantian claims, your right is lacking something essential: a public recognition that only political institutions can add. For example, if you had a conflict with your neighbour over the precise boundaries of your property, there would be no recognised procedure you could use to resolve this conflict. You might still be able to “get your right” through bargaining with or coercing your neighbour. For the Kantian, this is a lack in your right, independent from its non-infringement.
The realisation of a right is also different from respect for it. In addition to not trespassing, people in the state of nature might defend your property claims against third parties. There might be a neighbourhood watch which helps protect your property. In general, the state of nature might be sociable and peaceful. What is objectionable, from the Kantian point of view, is that there is no public recognition of your rights. Your rights continue to rely on the haphazard, dispersed and vague attitudes of others, rather than being certified to you as a matter of public right. As in the argument from independence, this is not a practical problem, one stemming from the inconveniences of the state of nature. Instead, the current argument rests on a genuine metaphysical claim about the nature of rights. It claims that, in the absence of certain social facts obtaining, we cannot be said to fully have moral rights. The relevant social facts include, in a sense I specify in the next section, the public recognition of persons as rights-holders.
We can contrast this account with two alternatives. According to Benthamite scepticism, there are no moral rights, but only legal rights. There are the rights that particular legal systems assign to us, but such rights are purely conventional. The Kantian picture opposes this view. We have genuine moral rights before any legal conventions arise. These rights are grounded in the status we possess as human beings and form a core part of morality. However, certain social facts need to obtain such that this status can be said to be adequately brought about. On the other hand, a natural law interpretation of moral rights claims that we have moral rights independent from any social facts obtaining. People might disagree about the natural law, or have no political institutions implementing it; but this does not in any way seriously undermine or diminish that we have those rights. The Kantian agrees with the first half—we have provisionary rights in the state of nature, and their existence is independent from social facts obtaining—but disagrees with the second. Whether we can be said to fully have moral rights depends on certain social facts obtaining.
Completing the Argument
This, then, is the crucial idea behind the Kantian argument: there is a structural incompleteness in our rights in a state of nature, and political institutions are necessary to remedy this incompleteness. I have already noted the first two steps in the argument:
||For our rights to be fully realised, they need to be publicly recognised.
||Certain kinds of (non-consensual) political institutions are necessary for our rights to be publicly recognised.
||Certain kinds of (non-consensual) political institutions are necessary to fully realise our rights. (from 1*, 2*)
But you might shrug your shoulders at this point: what is so important about our rights being fully realised? You might be morally indifferent to whether they are, especially given that I have distinguished this issue from infringing and respecting a right. So we need an additional premise explaining why the realisation of rights matters. It is at this point that we can connect the current argument back to the status theory. I argued that the appropriate moral reaction to individuals having a certain type of moral status is to never treat them in certain ways. We can now add the idea that, as long as individuals do not have fully realised rights, we fail to adequately react to them as entities with moral status. I will provide one way to motivate this idea in the next section. Let us see for now consider how this idea fits into our overall argument. We can argue
||Realising the rights of persons is a necessary part in the appropriate moral response to their moral status.
||Certain kinds of (non-consensual) political institutions are necessary for the appropriate moral response to what kind of entity persons are. (from 3*, 4*)
We can now turn to a second problem. I acknowledged that people have rights in the state of nature. They are merely provisionary rights, but still rights in the sense that we could not permissibly violate them—that is, they still impose weighty constraints on our behaviour. If this is the case, you might object, then we still have not solved the Basic Problem. Even if non-consensual political institutions are necessary (as (5*) claims) for some purpose, they still stand in conflict with provisionary rights we have in the state of nature. It is in this context that we can return to Mack’s anti-paralysis postulate. As a reminder, the principle claims that
when working out the detailed specification of person’s rights, one is to avoid specifications that systematically morally preclude individuals from exercising their rights or from conducting their lives in ways that a specification of their right is supposed to protect.
We can modify the principle for our purposes. The shape of our rights should be drawn in a way such that they allow for their full realisation. Rights express the idea that individuals possess a type of valuable moral status. It would be self-defeating if morality was designed in a way that systematically stopped us from realising that kind of status. So we can invoke the anti-paralysis postulate: our moral (provisionary) rights do not in fact stand in the way of political institutions. We have no general provisionary rights against the interference of political institutions—if anything, we have a claim that political institutions be established. Schematically, this allows us to complete the argument as follows:
||If some task or function is necessary for the appropriate moral response to the moral status of persons, then we have no right against this task or function being fulfilled. (Modified Anti-Paralysis Postulate.)
||Certain kinds of (non-consensual) political institutions are permissible. (from 5*, 6*)
This was a long argument, but the basic idea is simple enough. The public recognition of rights-holders as rights-holders by political institutions is required to fully respect them. Thus, far from political institutions standing in opposition to a rights-based morality, they are required by it. Through the idea that morality is structured to avoid self-defeat, this allows political institutions to operate.
The Idea of Publicity
Let me put some flesh on some of the details of the argument. We need to show in what specific sense our rights are incomplete in the state of nature. My strategy will be to make this idea plausible by discussing three interlocking ideas: the notion of moral community, the idea of determinate rights, and the importance of adjudication.
We can start with Feinberg’s Nowheresville. Feinberg imagines a world in which there are no rights. There are moral requirements which govern how we ought to treat others, and those requirements impose the same prohibitions and permissions that a rights-based morality would. However, Feinberg argues, people in Nowheresville are missing something important—while some actions are wrong in Nowheresville, we can never wrong other people. We stand in no moral relations: there are no moral claims or demands we can make on each other. Rights, particularly claim-rights, are important precisely because they fill this gap. Rights enable us to make claims against each other. They bestow an important role on us, that of being a claimant. Without this role, we would in some sense be merely patients, not agents, in a moral theory. It is easy to connect this back to the status theory: only if we give individuals rights do we adequately respond to the fact that persons are moral agents, not merely moral patients. Only a moral scheme which contains the role of rights-claimant has a chance to adequately reflect what kind of entity individuals are.
We can modify Nowheresville to provide support for a stronger point. Assume that the Nowheresvillians have moral rights. But imagine that there is no unified practice of claiming. Most people either do not accept that morality is rights-structured, or they have widely divergent ideas about the content and structure of valid claims. Any particular claim you could make against others is with overwhelming certainty going to be ignored or rejected by those around you. There is no shared practice of claiming, but at most vague and dispersed ideas about rights exist.
The Kantian claims that this modified Nowheresville is not a great advance over Feinberg’s original Nowheresville. Without a roughly shared and unified idea about which claims are correct and which are incorrect, we cannot be genuinely said to have the status of claimant. Having moral rights without a society which acknowledges them is like being rich in a metal which no one accepts as currency. In this respect, Nowheresville is not merely a thought experiment. You can think of any period in history where there was no social recognition of the fact that people had rights and moral status. In such societies, you could have claimed that some property rightfully belonged to you, or that your body was inviolable, or that you were allowed to speak your mind freely, but no one would have recognised those claims—in some societies, those claims would not even have been found intelligible. For you to be a claimant of rights, you need not merely be in a position to truthfully make them as a matter of moral theory; you also must be in a position to make them as a matter of social and political reality.
Another way to put the point is that modified Nowheresville lacks a developed moral community. If we are members of the same moral community, then we accept the same ground-level rules about how we ought to treat each other. In such a community, I can expect my rights to be reliably recognised. I can expect the claims I make to be taken seriously, even if I will not always succeed in making them. A moral community will still contain wide and deep disagreements between individuals about how we ought to treat each other. But a moral community has mechanisms to deal and contain those disagreements, and to adjudicate conflicts between people who disagree about their claims. Furthermore, the norms of a moral community regulate our interactions as a whole: they balance and coordinate our rights, claims, liberties, moral powers and so on into a system. We can normally expect that only written, institutionally developed rules can successfully fulfil such a role in the face of the complexities and scale of modern society.
The norms of a moral community are not generally based on consent, but some weaker attitudes will normally be required. The majority of the members of a moral community must be willing to uphold, and act in accordance with, the norms of that community, at least in principle and in its rough outlines—otherwise, it ceases to exist. We are here lead back to the importance of coordination. Without political institutions to coordinate the content of shared moral norms, we will often fail to form a moral community. But certainly no general obedience to all rules is required, and the existence of a moral community is compatible with serious contestation about the rules—e.g., in the form of civil disobedience. In much the same way, if there stops being widespread agreement on the rules and vocabulary of English, then at some point we stop being speakers of the same language; at the same time, there can be a variety of accents and styles of English which do not threaten its existence as a single language.
This brings us to another objection. Is it not possible that the relevant type of moral community can arise without political institutions, through the invisible hand of social and cultural forces? After all, there are language, sports, and etiquette conventions without any institutions which set and enforce them. So why do we need political institutions to constitute a moral community? A first part of a reply is that standing in some form of community in these other respects is not morally important. Consider the case of etiquette. It is convenient for the Nowheresvillians to know that their fellow citizens follow the same rules when it comes to the direction in which the port is passed at posh dinner tables. Having the same rules in etiquette eases interaction, and reduces social friction. But there is no fundamental moral requirement that we form an etiquette community, or that we become members of the same language or sports community. If Nowheresvillians follow fundamentally different etiquette conventions, speak different languages or follow competing sport teams, social interaction will often be awkward, difficult, and prone to misunderstanding. But there is no deep moral loss in such cases.
The second part of my answer is partially a concession. We can imagine circumstances where political institutions are indeed not necessary to make us part of the same moral community. Assume that Nowheresville is a small, tight-knit community in which people have deeply shared cultural, social, and moral beliefs. Assume furthermore that there are long-established, well-known, and generally accepted conventions which provide individuals with a clear idea of what claims they have against others and the community. We can also assume that there is some procedure for resolving disagreements over this shared code—say, a convent of village elders. If we fill in enough of the detail, it seems plausible that the Nowheresvillians form a moral community in which individuals are publicly established as right-holders. Other things being equal, the shared norms of this society deserve our respect, and there is no need to introduce political institutions.
So it is not strictly speaking true that political institutions are necessary to make our rights complete. However, the exceptions to this rule are rare, and limited to tight-knit, small-scale societies. Modern, pluralistic societies are too complex that decentralised, spontaneous mechanisms could generate social conventions which are strong and determinate enough to provide us with the kind of moral community we need to realise our rights. Thus, a more precise statement would be that political institutions are necessary in modern, pluralistic societies. I will proceed with the argument without always explicitly mentioning this caveat.
Another lens through which we can approach the issue is by focussing on the indeterminacy of provisionary rights. Assume that in the state of nature I build a house, and start cultivating some of the surrounding lands. This gives me provisionary rights in the house and the land. The existence of these moral rights can be established merely through moral reasoning. We can tell some story about the importance of self-ownership, and how this connects to a general moral permission to acquire and possess property. But such moral reasoning greatly underdetermines many aspects of the right I am thought to have acquired. Do my claims extend only to those lands I am cultivating at the moment? Or land I tend to cultivate, but for now I merely fenced in? What about land I used in the past, but now no longer need? And so on. The precise boundary of my rights are not clearly established.
The Kantian believes that your rights are not merely hard to know, or difficult to enforce, in this case. There is a genuine sense in which your rights are indeterminate. A useful analogy is to compare provisionary rights with indeterminate shapes—that is, a shape which has fuzzy boundaries. In the state of nature, we have both rights and moral status. Some forms of moral behaviour are clearly forbidden by our provisionary rights, and some clearly allowed. But there is a broad, hazy border where your rights neither clearly apply nor clearly fail to apply. The idea of self-ownership, for example, could be spelt out in different ways, and the questions I raised could be given different answers on that basis.
Why is determinacy in our rights important? Again, we can point to a host of empirical inconveniences which are associated with having indeterminate rights. For example, indeterminate rights are generally less valuable to their holders as they provide for fewer specific protections against the behaviour of others. But again the Kantian stresses that there is also a more fundamental issue. We ought to respect you as a moral agent by providing you with the status of a claimant. But if you are only capable of making very general and vague claims against others, then we fail to provide you with the appropriate type of status. With regard to owning property, for example, it is not enough to be able to find one’s property acquisitions and transfers to be recognised in some vague sense. To respect you as an owner, you must also have the ability to make reasonably determinate ownership claims. There must be clear ways available to you to acquire and transfer property, to demand restitution for property damages, and so on.
It is useful to add that we are interested in the determinacy of shared rules. The rules you might apply might be very determinate, and the rules I apply might be very determinate. However, this does not establish that we are subject to a shared and promulgated set of rules. Promulgation I understand to be quite minimal: a set of rules is promulgated in case that everyone knows the rules, or could come to know them with reasonable effort. Slightly more extensively, promulgated rules are those about which there is common knowledge: knowledge of the form that I know the rules, and I know that you know the rules, and you know that I know that you know the rules, and so on. In complex, pluralistic societies, only political institutions can make shared rules which are both promulgated in this sense and sufficiently determinate. A government, for example, will say how property boundaries are to be determined and where they lie, how property transactions between individuals can be made and enforced, how conflicts are to be resolved, and so on. On this basis, individuals come to have a determinate idea of what is theirs; and if their political institutions are reasonably effective, they can also expect that these rules will be generally known by others.
We should ward off two misunderstandings. First, the Kantian does not demand, or aims for, full determinacy (if this is even a coherent notion). No actual legal system is fully determinate. It might even be the case that legal reasons systematically underdetermine legal outcomes in hard cases. So hoping to dispel all indeterminacy in a legal system is an illusory hope. Second, there are different kinds of principles we might use to delineate property boundaries, and so to resolve indeterminacy. The Kantian is not committed to the claim that all of these ways of resolving indeterminacy are equally good. Some solutions might be fairer, more efficient, more principled, more liberty-preserving (and so on), than others.
You might be sceptical with regard to this story if you merely think about property rights, or rights such as rights to your own body. Are these rights not decently determinate, even in a state of nature? I am not convinced by this response. We tend to overestimate how determinate moral rights really are because most of us are used to living in modern societies with sophisticated legal systems. When we think of moral rights, we tend to think of them implicitly as the legal rights we already have, but with the state and its institutions mentally subtracted. Thinking about them in this way is a great mistake. Rights which people have in conventional societies without centralised authorities tend to be very vague, and they are often not more than pointers to some method of community mediation, or some abstract affirmation in philosophical textbooks.
The issue can be brought out more clearly by focussing on a special class of rights. Many moral theories accept that people have welfare rights, which we can understand to be rights to be provided with some level of material resources. If we had such rights, they would entail positive duties in others to provide us with such resources. In line with the overall approach of this thesis, I take no stance on whether people have such rights, or to what extent. What I want to highlight is that indeterminacy worries are very strong in this context. Indeed, one sometimes finds formal arguments lodged against the notion of welfare rights. A common critique alleges that welfare rights lack determinate holders of the duties which correspond to those rights. The argument continues that, because all rights have determinate duty-holders, welfare rights cannot be rights.
I agree with the first part of this critique: welfare rights in a state of nature are gravely, perhaps catastrophically, indeterminate. They lack clearly specified duty-holders, and their content is severely underspecified (What kind of resources have people a claim to? How many? etc.). But with regard to the second part of the criticism, the Kantian will turn it on its head: far from showing that welfare rights are not rights, it merely shows that rights remain provisionary in a state of nature. That, however, does not undermine that people generally have those rights. Only under a public scheme that specifies and clarifies the content of welfare rights, and which allocates the duties which correspond to those rights, can we be said to fully have them. Because welfare rights are genuine rights, but remain incomplete in the state of nature, there is a general duty to escape it. This will make the step from a state of nature into a civil condition much more dramatic: welfare rights in a state of nature are so indeterminate as to be virtually dormant; only under public institutions do they first see the light of day in a genuine sense.
Let me highlight a third way in which we can see the importance of political institutions. As brought out in the example of Nowheresville, rights are closely linked to a social practice of claiming: we raise, make, reject, answer, ignore, face, adjudicate (and so on) claims. So what is as important to rights as their substance is a regulated process by which we can make them. If you own some property, for example, and you are to be considered its rightful owner, it is crucial that you have an attendant right to fairly and publicly defend your claim to this property in the face of challenges. Just as with positive rights, it is hard to see how this procedural aspect of our rights could be satisfyingly realised in a state of nature.
We can focus on a simple case. Assume that we live on an archipelago consisting of many small islands. Each island is just big enough to support one family or group of cohabitators; by convention, each island is assigned to one such group. This convention, plus some others regarding trading and travelling between islands, are generally known and accepted. In this archipelago world, indeterminacy in the substance of our rights seems a less pressing problem. But what if there is some conflict? You claim, for example, that I entered your island at night and stole some precious fruit. I deny that I did. There is no issue with the determinacy of the rules: they give a clear verdict. Furthermore, the response of our moral community would be clear—if only we agreed on what happened. Our conflict calls for some form of resolution. But how should we resolve our disagreement in the state of nature? Which procedure should we follow in resolving this conflict? Which rights do we have inside any such procedure? These issues have no clear answers in the state of nature, which for the Kantian points to yet another flaw.
Returning to the concession I made above, under very benign circumstances it might be possible for an anarchic society to have social conventions in place which satisfyingly regulate conflict resolution as well. Perhaps there are some island elders which are generally accepted as the go-to people in case we have a conflict. It is not enough, however, for there to be such generally accepted adjudicators. These village elders could not just make pronouncements at will; they would need to establish a complicated procedure which ensured some form of fairness in the whole process. I am again sceptical, on merely empirical grounds, how much we can expect such procedures to arise in modern societies.
A Tentative Conclusion
Let me summarise. In section 4.2, I spoke somewhat vaguely of the public recognition of our rights. In section 4.3, I put some flesh on this idea through the overlapping ideas of moral community, determinacy, and the procedural element in rights. In each case, I aimed to show that (i) in the state of nature, some important feature of our rights is missing, and (ii) in modern societies, only non-consensual political institutions can remedy this lack. Other arguments in the literature go into a similar direction, and could be used to further bolster the Kantian argument—e.g., the idea of moral coordination, or the importance of assurance, or the claim that political power ought to be “fiduciary”.
We have identified a quite fundamental aspect of our rights which makes political institutions necessary. This still underdetermines which type of institutions we ought to pursue. It is clear that not all institutions are equally good in this respect. Some are more inefficient in ensuring the public recognition of our rights, and some fail in this task altogether. More work would need to be done to discuss these differences. For now, we can focus on the big result: if the Kantian argument succeeds, then we can see how it can be both true that individuals have moral rights, merely on the basis of their moral status, and that governments can permissibly operate even though they do not enjoy general consent, or authority over anyone. This solves the Basic Problem.
Let me turn to some limitations of the argument. The Kantian argument establishes the legitimacy of political institutions insofar as they are necessary to fulfil some task related to fully realising our rights. This arguably includes whatever else is necessary to fulfil that task, but not more. For example, the argument I have given establishes that it is legitimate for political institutions to establish public courts which adjudicate conflicts where there are no such courts in place. This in turn legitimises some further tasks: e.g., it will be necessary to finance the courts in some way. So we also have an argument for the legitimacy of limited taxation, for the legitimacy of establishing a hierarchy of court officials, and so on.
We can immediately see an important limitation of the argument. There is some amount of taxation necessary to finance an appropriate court system, but the same political institution would not be permitted to tax some additional amount to use for other purposes. Showing political institutions to be legitimate in one way does not show them to be legitimate in all ways. How does this accord with the non-minimalist commitment I outlined in the introduction? The Kantian argument, as I have developed it so far, shows a number of functions of political institutions to be legitimate, such as a pursuit of the rule of law. More abstractly speaking, political institutions are allowed to take the steps necessary such that we form a moral community. But we are yet one step away from having an argument for the legitimacy of (say) welfare provision. If we wish to use the Kantian argument to show this function to be legitimate, we would need to argue that, unless the state provides welfare to its citizens (on at least some level), our rights are not fully realised. I will sketch how we might pursue such an argument in the next chapter. Specifically, we would need to argue that unless our rights are promoted (on at least some adequate level), they cannot be said to be fully realised. So if you think that people have positive rights to welfare, then promoting these rights is a legitimate function of political institutions, and whatever is necessary to fulfil this function—i.e., the redistribution of resources.
However, this makes many results tentative. Precisely because the Kantian argument is not an empirical argument from the benefits of the state, but a complicated argument about the nature of our rights, it is difficult to make. It will not be possible to defend all the functions of political institutions, and we might be required to give up some much-cherished anti-minimalist intuitions about legitimacy. It is unclear, for example, how public funding for arts and culture, insofar as it is financed forcefully via taxation, could be shown to be legitimate on the Kantian argument.
The Implausibility of Monopoly
We can also return to the question of the state and its alternatives. Which bundle of rights to rule are political institutions likely to have on the Kantian picture? I distinguished legislative, adjudicative and executive rights, and second-order rights to a monopoly on these functions. The argument I have provided in this chapter goes some way towards showing that legitimate states will have legislative and adjudicative rights. They have the right to make law and to resolve conflicts amongst people. Both activities are necessary to make our rights public. In the next section, I will also argue that promoting rights is a legitimate task of political institutions, and they will need some executive rights to fulfil this task.
Could we use the Kantian argument in favour of monopoly rights? We would need to argue as follows: for our rights to be fully public, there need not only be institutions which make, adjudicate and enforce public rules; there also needs to be a single institution which fulfils all of these functions without having any competitors. It is doubtful, however, whether this is true. For rules to be public and determinate, they surely cannot be contradictory—that is, there cannot be confusion over whether, for example, the boundary between two properties should be drawn this or that way. Legal rules must be internally consistent, but it must also be clear which rules applies to a given case if there are several, competing rules. But that does not require that there is a single author of these rules—even less so a single adjudicator or enforcer—as long as there is appropriate clarity regarding which rules apply when. Thinking back to the medieval model of political institutions, in one respect we might be governed by church law, in another respect by imperial law, and in another respect by guild law. As long as the division of responsibility between these different bodies of law is decently clear, it is not obvious why we should think this morally problematic from a Kantian perspective.
Thus, it is likely that political institutions must sometimes allow other institutions aside from themselves to rule over the same people—though as I have just emphasised, such competition must happen in a regulated way. This is compatible with most potential competitors having a duty not to compete with political institutions. Where existing political institutions operate well and efficiently, and where competition would only undermine and disrupt existing social conventions, any attempt to rule by newcomers might be morally wrong. But again, this result does not hold universally. In short, while the Kantian argument is often explicitly offered as an argument for the state, it is unlikely that it succeeds as such an argument insofar as monopoly rights are concerned—at the very least, significantly more philosophical work would need to be done to show that it does.
This already points towards some of the practical implications of the Kantian argument. Let me close this chapter with discussing some others.
First, you might suspect that one upshot of the Kantian argument is that democracy is necessary for political legitimacy. The idea is simple: for us to stand in the appropriate kind of moral community, each of us needs to have an equal say in collective matters. While there are pressures towards a democratically organised system inside the Kantian view, much depends on the details. For us to form a moral community, we need to stand in a publicly organised framework of shared rules. However, much of that can be done without democratic governance. Many of the crucial ideas I invoked—e.g., that we require promulgated, determinate rules—do not require democracy. But neither does the argument I have provided exclude democracy. On some ways of cashing out the details of the status theory, the appropriate moral response to persons might indeed be to give everyone an equal say. We can leave this question open, to be resolved by a substantive moral theory.
A second question I already raised concerns revolution. Kant famously claimed that we owe absolute obedience to rightful institutions. I have already given reasons why we should be philosophical anarchists, so I set these concerns aside. Still, you might have some worries about regime change under the Kantian argument. Revolutions involve a return to a state of nature, if only temporary. Would the argument from publicity not forbid such changes? In the process of revolution, we would diminish the rights individuals have—we would, at least briefly, reduce already established rights to merely provisionary rights. Would the current view imply that revolution is thus always wrong?
While I argued that this is entailed by the argument from independence, the same is not true of the argument from publicity. First, not all political institutions publicly realise our rights. Our moral situation vis-à-vis tyrannical regimes is as if we still lived in a state of nature. So at least these types of regimes do not provide an obstacle to revolution, though there might be other reasons to oppose revolution. More importantly, remember the different dimensions on which we think about rights. It is always wrong to violate a right. But nothing I have written suggests that it is always wrong to diminish the degree to which our rights are realised. Assume that a state slowly disintegrated, more and more failing to publicly constitute us as right-bearers. This process need not involve that anyone is wronged. Similarly, if you advocate revolution, that process need not wrong anyone.
In short, there is no rights-based constraint against revolution—this is the crucial difference to the argument from independence. At the same time, the Kantian argument from publicity does establish a general, non-absolute presumption against revolution: insofar as revolutions diminish our rights, and re-introduce a flaw political institutions were meant to resolve, we have a reason to be sceptical of them. Kantians in general will prefer legal, public ways of changing the norms of a society over extra-legal, non-public ways.
A third question concerns global justice. I have spoken about us standing in a moral community, and that we have rights in the full sense only if we stand in the same moral community. But what are the boundaries of this moral community? Would the Kantian argument entail that there should be a global moral community?
My answer is twofold. First, if our interactions with foreigners are not governed by shared public rules, then our rights are indeed incomplete in some respect. Assume that Nowheresville is split into two, North and South. Each half has political institutions which provide public rules, but there are no shared rules across both Northerners and Southerners. If I am a Northerner, then my rights are still provisionary with regard to Southerners. This is a deplorable lack. We can go through the same examples as given above. What if I think someone from the other half has been stealing from me? In this case, there is no regulated, determinate, public process through which I could raise and resolve this accusation. Independent from the inconvenience this brings, we should say that this shows a lack in my rights.
So only under a world government are we all subject to the same public rules, and thus members of the same community. But we should not misunderstand what this entails. The argument does call for a global set of public conventions which regulate the rightful relations between people. However, these global conventions need not be provided by a world government—that is, an institution which directly exercises power over the entire globe. Assume that the northern and southern governments in divided Nowheresville agree on a set of determinate rules about the reach of their jurisdictions. Both governments agree on rules how cross-border disputes ought to be resolved. Under such circumstances, a northern Nowheresvillian and a southern Nowheresvillian are subject to different laws. But they still stand under the same public rules in some sense—say, the public rule that Northern rules apply in the North, and Southern rules in the South. Returning to the global case, the Kantian argument requires that there be some globally shared conventions which regulate the rightful interaction between citizens of different states. However, this does not require a world state or world government.
The Primacy of Justice
The primary result from the last three chapters is that a rights-based morality does not require us to be voluntarists about political legitimacy. Put otherwise, rationalism is compatible with individuals having strong moral rights. However, this is merely a compatibility result. It does not yet give us much in terms of a positive case in favour of rationalism, or an idea about how that position works. Luckily, we do not have to start from zero to provide such a case. In the last two chapters, I sketched the structure of a rights-focussed morality. We encountered rights in their negative role, as constraints on political action. But I will argue that rights are not merely constraints. They also form the central ingredient in an aim we ought to pursue. Promoting this aim, justice, is the primary virtue of political institutions. We should arrange our political institutions such that justice is promoted best, within the constraints allowed by justice itself. This is the guiding thought behind rationalism which I sketch in the last two chapters of this thesis.
Ideally we would want some form of “proof” for rationalism. But as with all normative views which are as close to moral bedrock as rationalism, it is hard if not impossible to come up with anything resembling proof. However, there are four less ambitious philosophical tasks we can undertake to motivate the position:
- we paint a clear picture of the underlying value(s) behind the view, and why that value should be considered so important;
- we describe the structure of the theory—how its different parts hang together, as well as which philosophical commitments it entails;
- we show how the theory can account for intuitive judgments about legitimacy, both regarding general principles and specific case judgments; and
- we answer possible objections to the theory.
In this chapter, I tackle the first task. In the next chapter, I explain the logical structure of sophisticated rationalism and some of its moving parts. This addresses task (ii) and goes some way towards tasks (iii) and (iv). Unfortunately, space restrictions do not allow me to address the whole range of issues with regard to tasks (iii) and (iv).
The aim of this chapter, then, is to describe the guiding ideas behind rationalism. The best way to approach this task is by trying to outline clearly the fundamental moral picture which underlies the view, and how it forms a consistent whole with other commitments, including those I have outlined in previous chapters. I will start by briefly describing the wider family of views rationalism belongs to (sec. 5.1), and some aspects of formulating it (sec. 5.2). The majority of this chapter is then given over to defend the primacy of justice. I explain both “primacy” and “justice” (sections 5.3 and 5.4, respectively). I then consider Waldron’s defence of the primacy claim (sec. 5.5), which I amend in some important ways (sec. 5.6). I then discuss some upshots of the argument (sec. 5.7).
Legitimacy as a Derivative Value
Rationalists claim that political legitimacy should be assigned to, or withheld from, policies and political institutions such that justice is promoted best. “Promoting justice” here is shorthand for “promoting justice within the constraints imposed by justice”. Rationalism is part of a broader family of approaches to political legitimacy which see it as a derivative value. Advocates of such views claim that political legitimacy should be distributed exclusively on the basis of what best achieves some substantive value or set of substantive values.
The fundamental intuition behind this view is clear enough. There are some desirable ways how we wish our societies to be—e.g., a free, equal society with autonomous, flourishing individuals living under determinate, shared rules of justice. We value the ways such a society could be in themselves—that is, they have final value. On the other hand, legitimacy is a derivative (i.e., non-final) value. Political institutions should be chosen in a way which gets us closer to types of society which have final value. The value of political institutions, and more broadly, how we distribute the right to rule, is entirely derived from this aim. We do not value any particular way of organising political institutions particularly highly in itself.
There are different types of derivative value. The most well-known kind is instrumental value. Something has instrumental value if it is a (causal) means to bring about something that has final value. Money is a means to bring about happiness, and thus has instrumental value, but we are unlikely to value it in itself. Instrumentalism about legitimacy thus is the claim that we value political institutions only insofar as they bring about valuable outcomes. Another type of derivative value is constitutive value. Something has constitutive value if it is a constitutive part of something which has final value. For example, the iron bars which make up the Eiffel Tower have constitutive value, insofar as we believe that the Eiffel Tower itself has final value. Still, the value of the iron bars is derivative, not final: we do not value them for themselves, but only for what they together constitute. We can now see that rationalism is not a type of instrumentalism. According to the Kantian argument, political institutions have constitutive value. The full realisation of our rights is constituted, not caused, by their public recognition through political institutions.
Still, it is useful to group instrumentalism and rationalism together, as both claim that legitimacy is a merely derivative value. For advocates of both positions, questions about legitimacy are not fundamental questions in political philosophy. The fundamental normative questions concern values such as justice, well-being, moral status, freedom, and equality. Questions about legitimacy arise as secondary questions once we have answered these questions. A philosophical position which started from the value of legitimacy would put the cart before the horse. Some political realists, for example, have argued that political legitimacy is a value independent from justice, and some have gone so far to argue that legitimacy is prior to justice, and sometimes outweighs it. Both instrumentalism and rationalism oppose this position.
Another assumption shared by both instrumentalism and rationalism as I will understand them is substantivism: what primarily matters for legitimacy is bringing about the right distribution of benefits, burdens, and rights (and so on). Think about ways how health care could be provided in the United Kingdom. The instrumentalist and the rationalist will start from the idea that there are some ways in which health care is best provided. That is, there is some quantity and distribution of health care services, benefits, burdens, and so on, which such theorists thinks to be morally the most desirable. For the moment, I speak of a “desirable” distribution quite generically, to leave open whether this is determined on the basis of justice or some other value. For the instrumentalist as well as the rationalist, whether a particular distribution of health care benefits is desirable is an objective matter—it is independent from what individuals think about the matter. Some individuals might feel unjustly treated if we withhold certain health care resources from them. But whether we ought to give those resources to them is almost always independent from their view on the matter. Substantivism rules out a couple of other ideas. Whether some distribution is desirable does not in general depend on how we have brought this distribution about. In particular, whether we have democratically chosen a distribution or not does not generally affect its desirability. Furthermore, what counts as a desirable distribution has nothing to do with what is justifiable to people, but can be determined objectively.
If legitimacy is a derivative value, and the value it derives from is a primarily substantive value, then it is clear from the outset that procedures will play no role in our approach to legitimacy on a fundamental level. This does not yet entail the unrelenting instrumentalism about democracy of someone like Richard Arneson. Democracy might play a role as a secondary value. It might sometimes make a small difference to the desirability of our health care system whether we have collectively or democratically or transparently chosen it, or not. But admittedly, even such a weaker type of anti-proceduralism is a controversial philosophical commitment. As I warned in the introduction, I do not have the space to defend substantivism about justice adequately.
Alternatives to Rationalism
If you see legitimacy as a derivative value, it is yet an open question what other value or set of values it is derived from. Rationalism claims that this value is justice, but other views have also been put forward. Judith Shklar, for example, exhorts us that liberalism should make “the evil of cruelty and fear”, as opposed to the pursuit of justice, “the basic norm of its political practices and prescriptions”. On the other hand, perfectionists tell us that “people live together not for justice but for pursuing better lives”. Roughly speaking, these two alternatives present a more minimalist and a more maximalist vision of the appropriate aims for political institutions. Realists like Shklar emphasise achieving order and public peace as the fundamental goals of political institutions. The pursuit of justice is seen as dangerous and misguided by them. Perfectionists object to rationalism from the other end: justice is seen by them as a niggling and merely remedial value, and political institutions which only, or primarily, aimed at justice as too restricted in their ambition. Rationalism need not always conflict with these alternatives, insofar as promoting peace and well-being might be requirements of justice, or pre-conditions of achieving justice. But they will not always be, so we need to decide.
Rationalism is a combination of seeing legitimacy as a derivative value, and a claim about the fundamental, final value in politics. Schematically,
Legitimacy as a Derivative Value. Political legitimacy should be distributed such that the best type of political society is brought about, as seen from an objective, substantive perspective.
Primacy of Justice. In the evaluation of political institutions, justice has lexical priority over other values.
Justice-First Rationalism. Political legitimacy should be distributed such that, first and foremost, justice is promoted best.
We can see now that the defence of rationalism falls into a defence of legitimacy as a merely derivative value, and a defence of the primacy of justice. I will be concerned with defending the latter in the rest of this chapter.
Varieties of Rationalism
Before moving on, I wish to highlight some features of the formulation of rationalism I have chosen. There is a tendency in the literature to discuss straw-man versions of rationalism or its instrumentalist cousins because of a failure to consider the wealth of ways in which this position can be formulated.
First, the formulation of rationalism I chose is indifferent between anarchism and statism. If anarchism is the best way to promote justice, then we ought to distribute legitimacy in that way—that is, we should consider no institution or policy to be legitimate. Put otherwise, “nobody gets the right to rule” is one way of distributing the right to rule. Unfortunately, some formulations of rationalism, and more generally, some ways of stating the problem of legitimacy, obscure this insight. For many philosophers, the question of legitimacy is the question of which “constitutional essentials” we ought to choose. But this already narrows our choice to types of intentionally chosen institutional structures; anarchism, where there are no constitutional essentials, is by definition excluded. Similarly, we should choose a formulation of rationalism which is not biased in favour of the state, and compatible with favouring non-state political institutions.
Second, the above phrasing of rationalism does not yet single out any particular level of legitimacy. We can assess the legitimacy of individual political decisions, of rulers, of institutions, and of whole constitutional frameworks independently; different rationalisms formulated on different levels will give different results. The prevalent view in the literature is a trickle-down picture of legitimacy. This view is in fact so dominant that many authors treat it as an obvious feature of legitimacy which merits little discussion. On this picture, questions of legitimacy are primarily questions on a macro-level: about the legitimacy of the “basic political structure”, or the constitutional essentials, or the political decision-making process. The legitimacy of individual laws and policies is seen as a secondary question on this picture. Either their legitimacy reduces entirely to questions of macro-legitimacy—e.g., a policy is legitimate if it has been made under a legitimate basic political structure—or their legitimacy is heavily determined by facts about macro-legitimacy. In the latter case, authors usually admit that grave injustice could defeat the legitimacy of particular policies, but believe that the macro-legitimacy of a political system establishes a heavy presumption in favour of the legitimacy of particular policies. This view is expressed, for example, by Rawls:
A legitimate procedure gives rise to legitimate laws and policies made in accordance with it […]. Neither the procedures nor the laws need be just by a strict standard of justice, even if, what is also true, they cannot be too gravely unjust. At some point, the injustice of the outcomes of a legitimate democratic procedure corrupts its legitimacy, and so will the injustice of the political constitution itself. But before this point is reached, the outcomes of a legitimate procedure are legitimate whatever they are.
In the face of the prevalence of the trickle-down view, it is tempting to incorporate it into our formulation of rationalism. But it is in fact an open question whether this is the best way to proceed. We might often promote justice (or other values) best by deviating from a constitution or a decision-making process. In fact, the rationalist has reason to prefer the precise opposite of the trickle-down view: a bottom-up view on which we evaluate legitimacy policy-by-policy, and the legitimacy of political institutions is merely a function of their tendency to make legitimate policies. I leave the question open as a theoretical matter, though I will argue in the next that there are good reasons to favour a moderately trickle-down picture on the practical level.
Third, rationalism does not tell us that we should choose the rulers or institutions which are best at promoting justice. Instead, it tells us that we should choose institutions such that justice is promoted best. This is a slight, but crucial difference. Rationalism—and in the same vein, instrumentalism—should not be confused with the claim that the wise, or the philosophers, or the elites should rule. Unfortunately, this confusion is common. David Estlund, for example, opposes the “correctness theory” of legitimacy, and he spends considerable effort on arguing against what he labels
The Authority Tenet. “The normative political knowledge of those who know better is a warrant for their having political authority over others.”
Estlund argues that the Authority Tenet is false, though the details of his argument are not important for our purposes. Estlund appears to think that the Authority Tenet is a commitment of his instrumentalist opponent, which he uses as a foil to his own view. But in that case, Estlund is attacking a straw-man version of instrumentalism, or at least an excessively narrow-minded version of the view. Rationalism, for example, entails the Authority Tenet only if you accept a further claim, namely:
If we put those who have superior normative political knowledge into power, then justice is promoted best.
But this claim is not necessarily true, and clearly false in many realistic circumstances. First, knowledge about justice is not practical skill in promoting it, as anyone who knows philosophers can attest. We would not generally want philosophers to be rulers, even if we value their knowledge. For this reason, rationalists will not usually recommend that we make philosophers kings or queens. Second, even where some have superior skill in promoting justice, putting them into positions of power would not necessarily promote justice best. Making a donkey or a fool king might promote justice best—assume that doing so bears some religious or cultural significance which would pacify our society. In such cases, we ought to choose the donkey or the fool over our most skilled politicians (other things being equal). There are various practical correlations between knowledge, skill, and the promotion of justice. So it is true that rationalists will be more favourably inclined towards rulers which are knowledgeable or skilful. But such a commitment is not a fundamental element in rationalism or instrumentalism.
Lastly, I have provided a maximising formulation of instrumentalism. Allen Buchanan, on the other hand, offers a rationalist position which does not require that justice is promoted best, but merely that it is promoted well:
A wielder of political power […] is legitimate (i.e., is morally justified in wielding political power) if and only if it (1) does a credible job of protecting at least the most basic human rights of all those over whom it wields power and (2) provides this protection through processes, policies, and actions that themselves respect the most basic human rights.
That legitimacy is a threshold concept is another idea which is commonly repeated in the secondary literature. Many philosophers hold that political institutions are legitimate as long as they are decently just, but believe that them being sub-optimal does not undermine their legitimacy. As with the trickle-down view, this idea is rarely argued for. In an instrumentalist or rationalist context, it is not obvious that we should prefer a threshold formulation. In fact, a maximising formulation seems more natural—why should we settle for political institutions that are suboptimal in the pursuit of the type of society we want? I will presume that on the fundamental level, rationalism will be a maximising doctrine.
Formulating the Primacy of Justice
In this section and the next, I describe what advocating the primacy of justice amounts to, before I turn to substantive defences of the claim.
The most famous statement of the primacy of justice is provided by Rawls:
[(1)] Justice is the first virtue of social institutions, as truth is of systems of thought. [(2)] A theory however elegant and economical must be rejected or revised if it is untrue; likewise laws and institutions no matter how efficient and well-arranged must be reformed or abolished if they are unjust. […] [(3)] The only thing that permits us to acquiesce in an erroneous theory is the lack of a better one; analogously, an injustice is tolerable only when it is necessary to avoid an even greater injustice.
A first question concerns the scope of the primacy of justice. Rawls identifies justice as the first virtue of social institutions. This is quite radical. It would entail, for example, that justice is the first virtue of the family. That does not seem particularly plausible. In the following, I narrow the scope of the primacy thesis to only cover political institutions. Another point regarding scope is that the primacy of justice is a claim about which features of institutions are most desirable. It does not assert that justice is the first virtue in the private life of individuals. Justice does not always trump all other reasons for individuals, such as reasons of friendship or personal commitments, in deciding what to do. This is a misunderstanding of the primacy thesis one sometimes encounters.
Second, we need to say how strong the primacy of justice is. We need to be cautious in distinguishing between respecting the constraints of justice, and promoting justice. Rawls is plausibly read to be concerned with respect for justice, and his claims (1) and (3) are convincing in this regard: if an institution violates rights, we should consider it to be illegitimate, at least in those respects. This follows simply from seeing rights as constraints. The more interesting question, however, is whether the promotion of justice also enjoys priority over other values. A first formulation of this idea would be
Strong Primacy of Justice. In the evaluation of political institutions, the promotion of justice takes lexical priority over other values.
There is also a weaker way in which we can construe the primacy of justice. This reading assigns to justice conditional priority:
Moderate (Conditional) Primacy of Justice. In the evaluation of political institutions, the promotion of justice takes lexical priority over other values up until a certain level of justice is achieved.
To see how these two formulations differ, compare the following four sets of institutional arrangements:
||Degree to which justice is promoted
||Degree to which other values are achieved (e.g., efficiency)
If all four arrangements S1-S4 were feasible options, the strong primacy of justice tells us that we ought to choose S1, no matter what other values are achieved. If only S3 and S4 were feasible options, we should choose S3. The moderate reading agrees that, in the choice between S3 and S4, we ought to choose S3, as achieving justice is the first task of political institutions. However, when it comes to the choice between S1 and S2, we might be allowed to choose S2 on the moderate view depending on how we weigh the additional gains in justice in S1 against the losses in other values. There is no simple argument as to whether we should prefer the strong or moderate primacy thesis. I return to the issue below.
The Concept of Justice
I have clarified the options we have regarding the scope and strength of the primacy thesis. The last, crucial question is the following: what is justice? Most authors take it as obvious that we have an intuitive grip on the concept of justice. Cohen, for example, tells us that he understands justice to be what Nozick and Rawls disagree about. He assumes that this is enough to tell his readers what he talks about. All the interesting questions to be asked for Cohen, and most other philosophers, are about when and why some political arrangement is just. In Rawlsian terminology, the debate is about which conception of justice we should choose, not about the concept itself.
This exclusive focus on conceptions is unjustified. It is far from clear whether different contemporary authors discuss the same issue when they talk about justice. Luck egalitarians, for example, focus on the justice of distributions. Libertarians, on the other hand, see justice as a matter of enforceable duties. Given this, it is in principle possible to be a libertarian with a theory of justice understood as enforceable duties, while at the same time accepting a luck-egalitarian theory of justice understood as desirable distribution. There would be no conceptual contradiction in these two commitments, though there are likely to be various tensions as a matter of moral substance.
So a pressing philosophical task is to clarify what we mean by justice. I should say from the outset that I think that the word “justice” is used too vaguely in ordinary usage to decide the matter on purely linguistic evidence. I will outline three competing ways to understand justice, before I describe my own in the next section. I notice in passing problems with defending the primacy thesis with regard to these other concepts.
Justice as a Wide Value
Justice is often understood to be a wide value—for example, as encompassing all of interpersonal morality, or all perfect duties, or everything which it is right to do. On the wide interpretation of justice, the primacy of justice becomes:
In the evaluation of political institutions, what is right takes lexical priority over other considerations.
It is hard to disagree with this claim, as it is close to trivial. What values could compete with justice in this sense? The defenders of stability or communal perfection presumably wish to say that it is right to pursue stability or perfection—so they do not disagree with the primacy of justice after all. Perhaps the primacy of justice in this general sense is tantamount to the primacy of the right over the good. A rationalism based on this kind of primacy might be true, but it is uninteresting because it rules out so little. There have been other attempts to explain the concept of justice on similarly general levels. Some philosophers have argued, for example, that the core of the concept is provided by formal equality—the idea that like cases should be treated alike. I find no such way of understanding the concept useful, and set them aside.
Justice as Everyone’s Due
A second understanding of justice goes back to Justinian’s famous dictum that justice is the “constant and unceasing determination to render everyone their due”. Many modern theorists still pay lip service to this formula. We might understand what is “due” to a person as what it is “fitting”, “appropriate” or “right” to give to a person. In that case, the current definition collapses back into one of the broad understandings just outlined. A more substantive interpretation understands what is due to individuals as what individuals deserve. Desert is some way of proportionally rewarding people for their achievements or efforts. On this reading, the primacy of justice becomes
In the evaluation of political institutions, giving people what they deserve takes lexical priority over other considerations.
This claim has the advantage of being more substantive than the one we considered previously. It is clear, for example, that stability and aggregate well-being are values different from desert, and compete with it. But while substantive and interesting, this claim about primacy has little initial plausibility. It is not clear, for example, why we should always give additional resources first to those who deserve them, rather than those who need them. It is hard to accept that desert should take priority over any other value. Any rationalism with this concept of justice at its heart would be unconvincing.
Justice as Fair Distribution
Another common way to understand justice is to see it as fairness, especially as the fairness of distributing certain burdens and benefits. Much contemporary liberal theorising is built around this sense of justice. There is some danger again for this definition of justice to collapse back into the wide sense. When we hear distribution, we think of income and wealth, but philosophers like Rawls also include rights, duties, and the “bases of self-respect” amongst the things to be distributed, and we might further expand the list to well-being, opportunities, liberties, and so on. The problem is that, as Iris Young remarks, “[a]ny social value can be treated as some thing or aggregate of things that some specific agents possess in certain amounts”. To avoid this criticism, we must settle on a reasonably specific list of distribuenda, such as Rawlsian primary goods. Once we have settled on such a list, the primacy of justice as fair distribution is not trivial. A perfectionist, for example, denies that the appropriate distribution of primary goods always takes precedence over any other concerns. The primacy of justice on the current reading then is the claim that
In the evaluation of political institutions, the fair distribution of a specified list of goods takes lexical priority over other considerations.
It certainly matters that things in our society are distributed well. But should we always prefer fairness in distribution over other values—such as efficiency, or stability, or aggregate well-being? I find it difficult to think that we should. At any rate, let me turn to the concept of justice I prefer.
The rationalism I advocate has a different concept of justice at its centre. On this way of understanding the concept, justice is about what rights we have against others. There is some connection of this concept of justice to the other three. First, rights are an important determinant of what is just in the wide sense, but they do not exhaust that wider sense. For example, it is right to combat pollution and climate change, but our duties to do so are only partially explained by rights. Second, rights are related to desert, but there is no necessary connection. Philosophers might deserve more public recognition, but they have no right to it. Third, individuals have valid claims to certain benefits, or against having certain burdens bestowed on them. But a distribution being fairer does not necessarily imply that anyone has a right to it obtaining. Inversely, sometimes individuals have a claim to be treated in some ways without this being an issue of distributional fairness.
Justice as an Aim
In the last two chapters, we have encountered rights as constraints. Insofar as rights impose constraints, they enjoy lexical priority over other values. For example, it would be wrong for the state to torture me in the pursuit of some good, even if this is the promotion of justice itself. However, in the following I am not interested in justice as a constraint, but justice as an aim. It is important to keep these two roles of justice apart. Justice as constraint tells me what it would be wrong for me to do others; justice as aim describes what the right type of society is I should try to bring about.
Justice, in short, appears in our moral scheme as both constraints on the means we are allowed to choose for our aims, and as an aim we ought to pursue. To see this, return to the status theory. Individuals have some moral feature, such as autonomy. It is appropriate for us to respect this feature. For example, we should not interfere with the autonomous choices of individuals. This leads us to an autonomy-based constraint, expressed as the right of individuals not to be interfered with in certain autonomous choices. However, the same feature of individuals also enters into an aim we ought to promote. For example, we should aim to promote, within the constraints of justice, a society in which the interference with individuals’ choices is minimised as far as possible, and in which individuals find the opportunity to exercise their autonomy.
On this basis, we can ask how we should flesh out the details of justice as an aim. We can start with a simple proposal of what it means to promote justice. There is an intuitive sense in which we can speak of particular interactions as just or unjust, precisely insofar as they violate or do not violate a right. There might also be positive rights—that is, rights that certain states of affairs obtain. If this were the case, then there is another immediate sense in which a particular state of affairs obtaining, or failing to obtain, is just or unjust. If someone’s needs are met, this would be just; if not, unjust.
However, such particular judgments do not yet give us a clear idea of what it means for large-scale entities to be just, such as an entire society across time. One way to make such comparisons would be an aggregative proposal, on which we rank the justice of a society simply by counting the number of rights violations in it. A moment’s thought reveals that this will not do. We think that different rights violations matter to different degrees—e.g., a violation of a minor property right matters less than a violation of bodily integrity, such as maiming someone.
On a more complex aggregative picture, the violation of different rights are given different weights, and the measurement of injustice becomes a weighted aggregation function. There are various epistemic and practical problems with finding out the correct weights, but let us set those aside. It is also likely that we will care about distributional patterns. It will matter in our assessment whether everyone’s rights are violated slightly, or whether a few people’s rights are violated greatly. On this basis, we might construe a complicated index of how just different societies are.
If we accept either the simple or the complex aggregative picture, then the justice of a society is a function merely of the pattern of rights violations in it. But justice as an aim is different from justice as a constraint. In construing a measure of the former, we are not slavishly bound to only work from materials taken from the latter. In particular, we can take into account aspects of our societies beyond rights violations. Let me describe some of these additional ingredients.
Kinds of Respect
First, rights can be respected in better and worse ways. Rights come with a hard core of duties which we need to fulfil to not count as infringing them. But there is also a larger belt around these core duties which demand that we respect these rights in other ways. If you have a right to your pen, for example, my core duty is to not steal or destroy it, and doing so would wrong you. But there are duties relating to this right beyond these. If I do not stop someone else from stealing or destroying your pen, I do not infringe your right, but I plausibly fail in a duty I had—I respect your right less than I should have.
Transposed to the level of a whole society, compare two societies with an identical pattern of rights violations. But in one of these societies the duties belonging to the wider protective belt of these rights are better respected, while in another they are less respected. Focus, for example, on property rights. Imagine that one society has a culture in which property is given enormous social and institutional attention. People are more respectful and cognizant of the fact that this pen (house, car, body) is yours. The second society has no such features. Many of the rights violations in this second society could have been easily stopped by bystanders who felt no duty to come to the victim’s help. If we care about rights, we should say that the first society is more just than the second, even if the pattern of rights infringements is precisely the same.
As another example, take a society in which the rights of a certain minority are rarely infringed, but the majority has generally low respect for those right-holders. “Honour” killings of women, or police killings of black Americans in the United States, for example, are comparatively rare in absolute terms. What we find problematic about these cases is not merely the occurrence of such killings, but also the broader culture of disregard for the rights of women and minorities which they express. This culture is not merely bad, it is unjust. If we want to capture this, we need to go beyond mere patterns of rights infringements.
Intentional and Accidental Violations
From the point of justice, there is a distinction between a tsunami taking a hundred lives, and a suicide bomber taking the same number. The former are not rights violations, while the latter are. It does not directly matter to justice whether people die prematurely; it matters whether people kill people. More relevant in the current context, there are also differences in how a right can be violated. Some killings are accidental—e.g., when I trip over you and injure you in the dark. We will want to draw a difference between accidental and non-accidental violations of rights. Compare again two societies with the same pattern of rights violations. In one society, these violations are not based on malice—they are based, say, on unforeseen or unintended consequences of our choices. In a second society, rights are violated intentionally and deliberately, perhaps even as a matter of central social planning. We should say that these societies are unequal in terms of justice, even if the distribution and amount of rights violations in them are equal.
The accidental/non-accidental distinction points us towards the importance of individual agency in general. It matters not only that individual rights are violated, but also how their violations are connected to the intentions and beliefs of the perpetrators. This is not that surprising if we think back to the Kantian argument. It matters to justice whether we are part of the same moral community—whether we respect each other as entities with moral status. Some ways of violating a right matter more from this perspective than others. For example, some rights violations are degrading. They are not merely harmful, but also intended to publicly dehumanize their victims. Where a rights violation has such a communicative role, we care about it much more than we do care about other, equally violent crimes which are done without such intent.
We have seen that, when we construct a measurement of justice as an aim, we will not stop merely at the pattern of rights violations, but will take other features into account as well. What precisely these features are, and what their relative weights are, is for a substantive theory of justice to resolve—a theory of justice as an aim.
The Rights That Matter
A second question is whether in fact all rights violations matter to justice as an aim. As I argued previously, rights are grounded on our interests or our moral status. We are considering justice, on the other hand, as the first virtue of political institutions. But it is possible that there is a gap between the two: not every interest which grounds a right, and not every aspect of our moral status which grounds a right, are a proper concern in politics, and should enter into how we assess the justice of our societies. In short, not all rights necessarily enter into justice as an aim.
Consider the following examples. My ill mother has a right that I visit her at the hospital. My chess club has a right that I inform them if I cannot compete for them. My students have a right that I return their work to them on time. Assume for now that these are genuine rights. But you might think that violating these rights does not give rise to injustice—at least not to the kind of injustice which we think to be relevant to the assessment of political institutions. Two societies which are equal in all respects, except that in one ill mothers remain unvisited and students receive their essays back too late appear to be the same in terms of justice.
There are three possible lines of reply. First, we might deny that the language of rights properly applies to these examples. We might say, for example, that I owe it to my mother to visit her in some weaker sense of “owing”, but deny that she has a right that I visit her. I do not find this option too plausible. Rights can be quite mundane. Students do have a right to have their essays returned on time, even if a weak one, and certainly not a human right. Second, we can accept the outcome, and say that there are such rights, and that they are a matter of justice. We might say that of the two societies described, one is indeed a tiny bit more just than the other. We might add that this is a difference which we usually do not care about, being swallowed up by more important issues. This is also an awkward theoretical commitment. Even if the issue is of little practical relevance, it seems wrong to use the language of justice in such cases.
Lastly, we can restrict the rights which are relevant to justice in one way or another. We might say that while the rights described are indeed rights, they are rights irrelevant to justice as an aim. One way to do so is to focus merely on enforceable rights. An enforceable right is a right which has a correlating duty that can be enforced by others—e.g., if I have an enforceable right that you pay me £100, then I can forcibly take the money from you in case you fail to pay me. Classic social contract theory claims that when we enter the state we give up our enforcement rights to the state, which in turn is tasked with enforcing our rights. This is of course a fiction—we have not actually given up our rights in this way—but it describes an influential way to think about the functions of the state. If ill mothers have an enforceable right that their children visit them, social contract theorists would argue, political institutions should ensure that this right is enforced. This, we might think, is what political institutions are for.
However, the rationalist might well deviate from this widely held assumption—after all, she has no particular ties to the social contract tradition. She might claim that not all rights, not even all enforceable rights, should be of concern to political institutions. One possible motivation behind this move is the following: even if my mother has an enforceable claim that I visit her in hospital, this is not the sort of claim that is of political, or public, concern. Whether I visit my ill mother or not is irrelevant to the wider social relationship we have to each other as citizens. Think back to the Kantian argument. It matters whether there is a public social recognition of everyone as entities with moral status. So it matters, for example, whether there is a social recognition of the claims to property that others have. But it might not matter in this respect whether we visit ill mothers, or hand back student essays on time. Drawing the dividing line between “public” and “private” matters in a satisfying way is of course difficult—it might turn out that enforceable and publicly relevant rights closely overlap. The important point here is merely that the rationalist is not bound by any simple view that all rights necessarily matter to justice.
Waldron’s Argument for Primacy
We can now consider how we might defend the primacy of justice in this sense. I start by considering a set of arguments by Jeremy Waldron. Waldron does not advocate the same concept of justice as I do—his is closer to a distributional sense of justice. Still, we can easily adapt his arguments to fit our purpose. I will argue that Waldron’s argument for primacy fails, but also that it points us in the right direction. I pick up the pieces in the next section and offer an improved argument.
The Assimilation Strategy
Waldron describes two strategies to defend the primacy claim which work in tandem. The first I label the assimilation strategy. This strategy is to argue that, for any alleged competitor to justice, we can show that wherever this value appears to threaten the primacy of justice there really is no conflict, as both values demand the same. Imagine that someone claimed that national security competed with justice, and sometimes overrode it. The assimilation strategy operates in two stages. First, we show that the value of national security can be broken down into benefits and burdens to individuals. National security can be analysed in terms of many separate people individually being safer. Second, once we have disaggregated this value, we ask “justice questions” about it. We show that the distributional profile of individual security benefits is itself demanded by justice. So there is no actual conflict: national security is subsumed under justice. This strategy does not entail that national security and justice never conflict, or that national security dissolves into justice. It merely claims that in cases where it appears as if some non-justice value trumps justice, there is no real conflict. Modified to our context, the strategy is to say that when justice appears to conflict with other values, people in fact have rights-based claims that we bring about the other value.
Consider the following counterexample to the primacy of justice by G. A. Cohen, which I slightly modify. Focus on people’s property rights. We could use two policing schemes to ensure respect for those rights. First, we could have a lean but highly efficient police force which ensures decent respect for property rights (scheme 1)—but some crime such as theft and robbery will remain. Alternatively, we could have a very large police force which ensures that property crimes are virtually extinct (scheme 2). The latter system, however, would be an economic drag on our society. On first sight, this looks like a straightforward conflict between justice and efficiency. So it seems that rationalism would tell us to choose scheme 2. However, if costs are high enough it seems more intuitive to prefer scheme 1.
We can use Waldron’s assimilation strategy in response. We can observe, for example, that a loss in efficiency will often also be a question of justice. The larger police force of scheme 2 must be financed in some way, and this is presumably done through taxes. But individuals have rights to their property, and taxing them beyond a certain degree might violate their rights. A larger police force also has other hidden costs relevant to justice: it might be more intrusive, and threaten to interfere with civil rights in other ways. Lastly, richer societies are often more peaceful; the wealth gains in scheme 1 might indirectly promote justice, even if our police force is smaller. In short, there are many justice-internal reasons to doubt that scheme 2 best promotes justice. The assimilation strategy suggests that, on a closer look, justice might recommend scheme 1.
The assimilation strategy is also powerful when it comes to various realist competitors to rationalism. We can briefly return to Judith Shklar’s liberalism of fear. This view, she tells us, focusses on cruelty and fear as the basic evils that political institutions are meant to prevent. Its central focus is on “physical well-being and toleration”. She suggests that our main priority in politics should be on avoiding cruelty, civil strife, disorder, and tyranny. Waldron’s assimilation strategy suggests the appropriate response. First, if Shklar is correct in her empirical-historical claims regarding the prevalence of cruelty, then a rationalist, too, will care greatly about stability. Before justice can be promoted, we need to ensure circumstances in which this is possible. Thus, stability is a question of justice, avoiding cruelty even more obviously so. To this degree, Shklar’s liberalism of fear does not contradict rationalism at all. Second, assume that our political institutions successfully eliminate the threat of cruelty and fear to a reasonable degree. We could then either invest more in promoting justice, or (as Shklar’s liberalism would seem to suggest) more in stability. If this were the case, it is not at all clear why we should prefer the second.
However, there are two limitations to the assimilation strategy. First, at a crucial point the assimilation strategy relies on the claim that justice makes the same demands as some other value. But without having a fully developed theory of justice, we do not yet know how plausible these claims are in detail. It is similarly unclear what Waldron’s optimism in this respect is based on. Second, we cannot universally rely on the assimilation strategy. If we claimed that in all interesting cases, justice and other values coincided in their demands, justice stops being an independently recognisable value. But I rejected such global interpretations of justice (and so does Waldron). So we must allow that justice and other values at least sometimes pull into genuinely different directions. To tackle such cases, we need a different strategy.
The Primacy of Individuals
We can now turn to the second strategy suggested by Waldron. He writes:
A different pattern of analysis is required for competition between justice and institutional virtues that concern genuinely collective goods. Here the argument is more direct and controversial: We say that justice has primacy because how individuals are doing simply matters more than how things go with collective entities.
Assume for the moment that the cultural health of a community—i.e., the amount to which there is a flourishing artistic and cultural scene in Britain—cannot be reduced to a distributional profile, but is a good which genuinely pertains to the collective entity Britain. When presented with a conflict between communal flourishing and justice, Waldron would assert the primacy of the latter because individuals are primary.
However, it is hard to imagine goods which cannot at least partially be broken down into distributional effects for individuals. So for most cases, we will use a combination of Waldron’s two strategies. Assume that the cultural health of a community has a reducible and a non-reducible part: there is a sense in which a community is culturally thriving if the individuals belonging to it thrive. But we might also say that a community can be culturally thriving or be in cultural decline even if the situation of individuals remains the same. Waldron would then argue that insofar as the cultural health of a community is reducible to its distributional profile, it does not override justice because it does not conflict with it (assimilation strategy); and that insofar as it is not reducible to effects on individuals, we think it secondary to justice (primacy of individuals).
In short, Waldron’s argument is to split values into values pertaining to individuals and values pertaining to collectives, and argue for the primacy of the former. However, the problem is that this argument fails to establish the priority of justice in particular. It would establish the primacy of justice if justice were the only value pertaining to individuals; but this is in no sense true. Even a utilitarian can argue that promoting well-being is an individualised value—after all, she is interested in promoting the well-being of individual people. So asserting the primacy of individuals does not rule out enough.
An attempt to further narrow the field is provided by Waldron when he adds that justice is not only a value pertaining to individuals, but that it is
individualistic in a double sense; justice is assessment of individual outcomes by individualized criteria. We assess how A is faring in the distribution of some good by reference to the intrinsic importance of certain facts about A, such as A’s needs or deserts or merits or A’s basic moral standing as a person.
However, not even this narrowing is sufficient to establish the primacy of justice. Desert and merit, for example, are also doubly individualistic in Waldron’s sense. Fair distribution also arguably is. So at the very least, the current argument does not select which of these concepts of justice enjoys primacy. We might also be able to reconstruct some non-justice values as doubly individualistic, such as certain perfectionist values. So even this refined version of the primacy of individuals cannot ground the primacy of justice (understood in our sense) in particular.
Another problem I note in passing is that Waldron has done little to establish the lexical priority of justice, though he is explicit that this is the priority he aims for. Consider the following analogy. Humans are more morally important than animals. But from this it does not follow that any benefit to humans, however small, outweighs any benefit to animals, however big. Similarly, saying that individuals enjoy primacy over collectives does not yet establish that any value pertaining to individuals enjoys lexical priority over any value pertaining to collectives.
Expanding Waldron’s Argument
There are two lessons to take away from the failure of Waldron’s argument. First, the assimilation strategy is useful as part of an overall defence of the primacy of justice. Second, the primacy of individuals is not enough to provide us with an argument for it. But from the previous chapters, we have two distinct ideas which suggest ways how we can improve the argument. The first is the suggestion that persons enjoy moral status, the second is Kantian institutionalism.
Moral Status and Primacy
Let me return to the idea of moral status first. In his own defence of the primacy of justice, Rawls suggests that moral status is crucial:
[e]ach person possesses an inviolability founded on justice that even the welfare of society as a whole cannot override.
This is a useful reminder. It is not simply that individuals enjoy primacy—as have now seen, this claim is too vague—it is that individuals enjoy primacy because, and insofar as, they are entities with moral status. This is a much more promising starting point for our argument. However, an immediate problem that we encounter is that this only establishes the primacy of justice as a constraint. It does not yet establish the primacy of justice as an aim. We could respect Rawls’s dictum while still prioritising the promotion of welfare over the promotion of justice, as long as we stay within the constraints of justice.
We might at this point assert, somewhat bluntly, that individuals have a duty to promote justice. I have above established that justice is an aim we ought to pursue, so it is plausible to also think that there is a duty to promote this aim. On this basis, we might provide the following argument:
It is the task of political institutions to discharge individual duties.
This task has lexical priority over other tasks.
Individuals have a natural, enforceable moral duty to promote justice.
Promoting justice has lexical priority over other values.
There are some important limits to this argument, however. First, a duty of justice is one duty amongst many—we also have a duty, to some degree, to promote the welfare of others. It is unclear why our duty to promote justice would take precedence over other duties. So the current argument establishes justice as one aim of political institutions, but it fails to establish it as an overriding aim. Second, a natural duty of justice will be comparatively weak. Individuals have a duty to prevent the most egregious injustices close to them, especially if doing so comes at little cost. However, this duty is quite limited in its scope and strength. If preventing injustice would entail significant sacrifices for ourselves, it is unlikely that we have a duty to prevent it, even if we might have a duty to do so. If political institutions are meant to fulfil our duties, then the result would be a relatively undemanding rationalism. But it strikes me that the duties that individuals have to promote justice and the duties that political institutions have to promote justice differ in strength. The current argument cannot explain this asymmetry.
So we still need some additional argument in favour of the primacy of justice. At this point, we can go back to the Kantian argument from the last chapter. The central idea of that chapter was that we need political institutions to fully realise rights. In particular, we need political institutions to be publicly recognised as persons with moral status. I also argued that this task is of great importance: we would fail to adequately respond to individuals as right-holders if we failed to have political institutions with the relevant features.
This already goes some way to show that certain justice-related functions of political institutions have lexical priority over other functions. Take the typical Kantian function of establishing a coordinated, reasonably determinate system of legal rules—in short, the task of establishing the rule of law. On the Kantian argument, our rights remain incomplete without the successful implementation of such a system. This would be a form of not adequately responding to the type of entity that persons are. So I take it that establishing the rule of law, together with fulfilling some other Kantian functions, is a lexical priority for political institutions.
This can best be seen by returning to the archipelago state of nature I sketched previously—i.e., a state in which our rights are indeterminate and insecure, and where no settled moral community exists. Assume that in such a state of nature a centralised political institution arose. But the first thing this institution does is not to establish the rule of law—that is, to determine through transparent, determinate law what rights each individual has. Instead, it starts by pursuing perfectionist aims. Before it is even clear what rights anyone has, it starts a large-scale redistributive program of education, publicly funded culture, and redistributive welfare provision. For the Kantian, political institutions of this type fail to take the moral status of individuals appropriately into account. Making individuals better off is one duty we have. But if we accept a moral theory focussing on the autonomy, independence (and so on) of individuals, then our first duty is to ensure that full respect is paid to these features of individuals. And if the Kantian argument is right, then this entails first bringing about, amongst other things, the rule of law.
So we already get a sense of how certain justice-related tasks enjoy lexical priority over others when we evaluate political societies. However, it is compatible with these observations that a government ought to (i) not violate rights, (ii) establish the rule of law, and (iii) promote aggregate well-being, in this order of priority. So we are still missing the result that the promotion of justice takes lexical priority over other, non-justice-related pursuits. A claim to bridge this gap is the following: political institutions fail to publicly realise us as entities with moral status unless they are designed such that the promotion of justice is their primary aim. In other words, a political institution which pursued aims (i)-(iii) would actually fail to adequately pursue aim (ii). In short, the argument is
People are fully morally recognised as entities with moral status only under certain kinds of political institutions. (Kantian Institutionalism)
If the promotion of justice did not enjoy (at least) weak primacy in our choice of political institutions, then we would fail to fully morally recognise the moral status of persons. (Linkage)
The promotion of justice is the (at least weakly) primary aim in the choice of political institutions. (Primacy of Justice)
We can see how linkage is plausible again by considering the implications of denying it. Consider a state which did not have at its primary aim the promotion of justice—where justice is understood in the way I outlined previously. Sometimes, such a state will give priority to aggregated well-being over the promotion of rights-based justice. Such a government, for example, might prefer spending money on education, arts or sports—pursuits which I assume cannot be assimilated into justice—while willingly accepting an underfunded court or police system which allows significant injustices to occur. This is not a merely theoretical possibility. Many existing democracies spend large amounts of resources on various projects in infrastructure, the military, or culture, while their justice, court and prison systems could use increased funding.
Could a government with such priorities be said to be fully committed to the task of constituting us as entities with moral status—as equals in the same moral community? This idea strikes me as implausible. If a state does not stop, for example, racially motivated killings of blacks, or does nothing to diminish a significant threat of rape to women on its territory, then we cannot say that this state fully, publicly recognises the moral status of these citizens. I am not saying that in failing to protect women from rape the state violates their rights, though some theories of justice might see such a failure to protect women as a direct violation of their rights (in which case the argument becomes even simpler). Rather, the claim is that in such a situation such individuals cannot fully be said to have rights. This claim goes beyond what I have argued in the previous chapter. There I merely argued that we need political institutions that fulfil legislative and adjudicative functions to be fully realised as right-bearers. The crucial new idea at this point is that our rights also remain incomplete if we are not also assured of our rights. Roughly, you are assured of your rights if you are protected from any immediate threats to have your rights violated by any third parties. To have assurance in this sense does not mean to be free from all risks of injustice—like the demand for fully determinate rights, this would be an impossible demand.
It is a difficult question to outline the precise contours of assurance, but the basic idea should be clear enough. Assurance is necessary for the full realisation of our rights; this task, in turn, has lexical priority over other tasks. So before political institutions can legitimately pursue any other aims, they must reasonable assure every one of their subjects in their rights—in short, it is a lexical priority that they promote justice up to a certain degree.
Stepping back, let me describe in a simplified way what kind of argument I have given. We start from a picture of the moral universe on which the most important feature is that we are persons with minds and lives of our own—there is a certain secular sanctity, an inviolability we call moral status, to each individual person. The radical Kantian claim is that public, political institutions are needed to fully respect, and bring out, this feature of individuals. While it is possible to respect others in a state of nature in some sense, we can fully respect them only if we live with them under shared institutions, under public laws and rules. This full respect, in turn, is only fully realised if the institutions we live under not merely respect our moral status, but take active steps to protect and promote it. This shows that promoting justice is the primary aim of political institutions.
There are some strands of argument left open from the beginning of this chapter, in particular regarding the scope and strength of the primacy thesis. The current argument gives us some hints as to how we can answer these questions.
The Scope of Primacy
According to the Kantian argument, it is the function of political institutions to constitute us publicly as entities with moral status, and for that reason justice is primary. This argument gives an elegant explanation for why the scope of the primacy claim is limited to political institutions: because only those institutions are necessary to publicly constitute us as equals, while others institutions are not necessary in the same way. As a contrast, consider the family. It is important for living together as a family that we are recognised as equals in some sense. But this is one function of the family which coexists with other aims, such as to provide the members of a family with an environment of love and care, and one which they can flourish as individuals. So while justice might well be a virtue of the family, it is unlikely to be the primary virtue. Similar arguments apply to other kinds of social organisation. A chess club is a private form of organisation. Its primary purpose is to further the chess-related aims of its members. So there is no presumption that justice is the virtue of this form of social organisation.
What about justice as a virtue of individuals? The Kantian argument can also explain why justice is not the first virtue of individuals. In our individual behaviour, we have to respect the rights of others. There might also be duties that we promote justice (as discussed in sec. 5.6.1). But on the basis of the Kantian argument, there is no reason to believe that these are the only or overriding demands on individuals. In fact, the Kantian picture suggests a certain separation of labour. It is the function of political institutions to promote justice, and to provide a public framework within which can be said to have fully realised rights. If political institutions are decently successful at fulfilling these tasks, then this lifts the burden from individuals to do the same, and frees them to pursue their private aims.
The Strength of Primacy
What about the strength of the primacy claim? As I claimed above, the argument establishes at least weak (conditional) primacy. We cannot count political institutions as being committed to individuals as rights-bearers if these institutions do not remedy grievous injustices where this is feasible. The issue is more difficult if certain levels of justice have been achieved. Consider a state which promotes justice decently—that is, which does a good job at promoting respect for rights, and which ensures basic rights for everyone. It would be feasible for the state to use more resources to promote justice, but these resources could also be used to address non-rights-based concerns, such as the general welfare. It is not clear whether a failure to use additional resources to protect our rights in this situation would undermine the public realisation of our rights. It might well be that a government is allowed to let some rights violations to continue (the example is not one where the government itself violates rights, but where others do), and weigh this against the promotion of general welfare.
At any rate, note that the moderate and weak primacy of justice start to converge in their practical demands the higher we set the relevant threshold of what we count as doing a “decent job” at promoting justice. On plausible versions of the weak primacy this threshold will be quite high. We live in societies which still allow wide parts of the population to be routinely exposed to fundamental wrongs. No one who cares for the importance of individuals and their moral status can remain unmoved by these injustices. So in many cases, the moderate and weak framings of the principle converge in their demands. I take this to be true for most Western societies.
Either way, what we have established is a justice-first rationalism, not a justice-only rationalism. Perfectionist, utilitarian, and other non-justice considerations can play a secondary role in our choice of political institutions. Some might think that this is insufficient: they will want a more thorough anti-perfectionism build into rationalism, on which justice is the only thing political institutions are admissibly concerned with promoting. We would need an additional argument to establish this claim, for which I do not have the space here. Note, however, that a status-based moral theory provides us with the materials for such an argument. If the autonomy and independence of individuals are the core moral features we ought to respect, then it is likely that additional constraints on state action fall out from our moral theory, such as a prohibition of paternalist or perfectionist policies.
Justice as Distribution
We can also return to some rivals of the current view, such as justice understood as distributional fairness (sec. 5.3.2). The Kantian argument can be used to show the primacy of rights-based justice, but it fails as an argument for the primacy of distributive justice. Assume that we could make the distribution of resources in our society more fair, or in some other way more desirable; but assume that we failed to do so. Assume this is a case where no one has any rights to those additional resources. In that case, then this is not a failure to recognise others as beings with moral standing. Some other kind of moral failure might be involved in this case, of course. But political institutions which fail to achieve the fairest distribution do not fail in their central task, which is to recognise individuals as valuable entities with moral status.
On the Kantian moral picture, then, distributions are morally epiphenomenal when it comes to justice as an aim: they matter only insofar as they matter to our rights. The Kantian argument cannot be easily modified to provide an argument for the primacy of fair distribution. The same set of remarks apply to other values which compete with justice, such as minimalist and perfectionist values. Governments might often have the opportunity to make us safer, or better off, or our societies more peaceful and flourishing. Insofar as these are not questions of rights, however, governments need not pursue these aims with any kind of lexical priority.
In subsection 5.4.3, I argued that we should reject an aggregative way of constructing a justice measurement. The Kantian argument provides us with some reasons why this is not so surprising after all. It matters that our moral status is publicly recognised by political institutions. One important way in which this is done is by minimising the number of violations which occur. But there is no reason to suspect that this is merely a matter of quantity, even if weighted appropriately. To return to one of the examples, we can see why it matters whether rights are intentionally or merely accidentally violated. If the realisation of our rights depends on their public recognition, then we will care less about accidental rights violations in evaluating justice than we care about knowing and intentional violations. Thus, the Kantian argument gives us a general reason to expect features beyond the mere quantity or distribution of rights-violations to matter when we construct a rights-based measurement of justice as an aim.
The Rights That Matter
Another issue I touched on is the question which rights matter to justice and which do not (sec. 5.4.4). The current argument also gives some support for the answer I gave there: not all our rights are public rights—that is, some rights need not be publicly affirmed by political institutions to constitute us as beings with equal moral standing.
Return to the example of a student’s right to have their essays returned on time. Assume that we want to say that this is a right. Does it, however, matter to justice? No: the protection and promotion of this right is not relevant to recognise us as individuals with moral status, and to fully constitute us as rights-bearers. If we contrast two societies which are equal in their respect for rights, only that in one students continually get their essays back late and in another they do not, there is no relevant difference as to how much we partake in justice. Similar remarks can be made with regard to the other examples I have given.
If the last chapter is correct, then legitimacy is a value derivative of justice. But this is a very general claim. Consider an analogously wide claim, that rightness is based on goodness. This claim, consequentialism, only starts to gain shape once we fill in various details. We need to say, for example, whether we ought to be act- or rule-consequentialists, satisficing or maximising consequentialists, and so on. Similar questions arise for rationalism. We can ask whether we should apply it on the level of individual decisions or general rules, whether justice ought to be best or merely well promoted, and so on. In this chapter, I deal with only one question out of the thicket of these issues, but the one which is the most important: how directly or indirectly are we to apply rationalism? I will offer a more precise statement of this question as we proceed. For now, I want to approach it intuitively by highlighting a problem about rationalism you might have already anticipated, which is that rationalism seems a highly counterintuitive way to think about legitimacy.
Mind the Gap
There are some central elements which we—both laymen and philosophers—ordinarily expect in a theory of legitimacy. First, procedures seem central to legitimacy. When you think about legitimacy, it is likely you will think about legitimation: processes of bestowing, creating, and transferring legitimacy. Almost all complaints about a lack of legitimacy of the European Union, for example, are complaints about a perceived democratic deficit of the EU. More broadly, many philosophers think that it is crucial to legitimacy whether decisions have been taken in the right way. It matters not merely what decisions have been made, but whether they have been made by the correct, representative institutional bodies. Second, you are likely to think that legal rules play an important role for assessments of legitimacy. In a theory of legitimacy, most theorists expect some mention of courts, legislatures, legal procedures, and so on. A natural thought is that political decisions need to adhere to certain legal forms to count as legitimate. Third, claims about legitimacy will be strongly guided by historical considerations. For the US government to be a legitimate government, it needs to have a continuous and ongoing relation to its citizens and its territory. The Canadian government might be better in ruling the US, but it would be an unjustified usurper. Most theorists would reject it as illegitimate straight out of hand.
In short, there are three major ingredients in everyday thinking about legitimacy: democracy, law, and history. But strikingly, these play virtually no role, or a much reduced role, when we think about justice. First, procedures play some role in justice—sometimes, justice might require that we use some procedure. And there might be some cases of pure procedural justice, where we cannot know what justice demands antecedently to some procedure. But usually we think that justice makes non-procedural demands—or at least so I claimed. Second, while justice requires some elements of a legal system, and even though justice and the rule of law are close cousins, our thinking about justice puts much less emphasis on legality. It does not usually matter to justice whether we have filled in the correct forms, and whether the right legal formalities have been observed. Lastly, while our conception of justice is partially guided by historical concerns, there are many aspects of justice which are unconcerned with history. The fact that some government has promoted justice well in the past does not guarantee that it will continue to do so; whether something promotes justice has to be evaluated continually anew.
In short, there are some procedural, legal and historical features to justice, but they play a minor role overall. On this basis, it should be clear that there is a significant gap between the determinants of legitimacy and the determinants of justice. We can exploit the gap to quickly come up with counterexamples to rationalism. Assume that Canada invades the United States, and in violation of legal precedent, democratic procedure and international law, starts to rule over Americans. We can assume that the Canadians promote justice much better than any American government could. Let us also presume that the transition does not itself involve any rights violations, and that costs associated with this transition are in general low—there is no civil war or general uprising against the new Canadian rulers. Insofar as the promotion of justice is concerned, we should prefer the Canadians to rule the United States. Thus, if justice underlies legitimacy, then there are strong pressures to say that the Canadians’ occupation and their subsequent rule are legitimate. But many theorists will find this implication unacceptable, and some will even regard it as a reductio of rationalism altogether. For someone’s rule to be legitimate, these objectors will argue, they cannot be an usurper. Power must be gained in a democratic and legal way to be legitimate.
We could exploit the gap between justice and legitimacy in many other ways to provide tempting counterexamples to rationalism, and we will encounter more throughout this chapter. The rationalist has two main strategies available to answer such counterexamples. First, she could accept that her theory is a revisionary theory of legitimacy: she could accept that legitimacy should be assessed in a way very different from how we currently assess it—for example, with much less focus on procedures and much more focus on substantive outcomes. On this view, we are mistaken in how we commonly think about legitimacy. Note that, as part of advocating philosophical anarchism, I am already committed to the claim that a variety of widespread normative ideas about political institutions are mistaken, in particular claims about authority—so this thesis already contains many revisionary commitments.
Ideally, we would not only reject common claims about legitimacy and authority, but also be able to offer a debunking explanation: a sociological, cultural or evolutionary theory which explains why people have generally held mistaken ideas about legitimacy. I think such a theory could be offered. At the core of such a theory would be the idea that governments for centuries have used their positions of power to convince their subjects that they enjoy wide-ranging authority and rights to rule, whereas in fact they do not. A belief in legitimacy focussed on procedures and law, and less on outcomes and justice, helps those who are invested in existing hierarchies of power, as it makes easier for them to assume the mantle of legitimacy for their own aims. Such beliefs in turn interact with a number of other misleading ideologies and psychological phenomena which work in favour of governments—e.g., a close identification of society and nation and state, a secular sacralisation of democracy and the law, the aesthetic and social appeal of power, and a general overestimate of the abilities of government to improve the human condition. But even sketching such a theory is far beyond the scope of this thesis.
The second, conciliationist strategy goes into the opposite direction. On this strategy, we aim to show that on closer inspection the implications of rationalism align well with standard ways of thinking about legitimacy. Understood properly, rationalism does not demand that we radically alter our practices or drop well-cherished intuitions. With regard to the Canadian usurpers, we might point out for example that Canadian usurpation is likely to have great transition costs—indeed, a Canadian occupation might lead to a full-blown civil uprising—or that usurpation often simply is forbidden by justice, as it involves violating rights-based constraints.
These two responses exist on a scale and can be combined. You might be revisionary about some of our beliefs regarding legitimacy, and more conciliationist about others. Indeed, it strikes me that a combination of the two strategies will work best in defending rationalism. There are a range of empirical circumstances which we can imagine in the Canada/US example. We should first use the conciliationist strategy and argue that, for a wide range of realistic circumstances, rationalism considers a Canadian invasion indeed to be illegitimate. On the other hand, if circumstances are benign enough, we should simply accept that Canadian occupation is legitimate, and that our anti-usurper intuitions about legitimacy in those cases are mistaken.
Rationalism is too easily associated with revisionism, and then dismissed because its implications seem so strong. But rationalism has much greater conciliationist potential than commonly thought. I will sketch a general, powerful conciliationist strategy, one which is inspired by a similar strategy used by consequentialists. This rests on seeing rationalism as a two-level view. On the fundamental or critical level, we only care about justice. But there are many good reasons to care about other features on the applied level—to care about procedures, legality and history as aspects of legitimacy. How convincing this move is depends on whether we can give a plausible sketch of the two levels and their interaction.
I start by describing the move that conciliationists make in debates surrounding consequentialism (sec. 6.2). I then argue that a closely analogous strategy can be made to work for rationalists (sec. 6.3), and sketch some of the details (sec. 6.4 and 6.5). I turn to various objections in the last section (sec. 6.6).
In a simple formulation, consequentialism claims that
||An action is right just in case it leads to the best outcomes.
(A) is a criterion of rightness—that is, (A) gives necessary and sufficient conditions for when an action is right or wrong. But (A) need not play any role in our conscious thinking about morality, or in our character dispositions, aims, beliefs, public rules, intentions and so on. To paraphrase Derek Parfit, the aim of a moral criterion like (A) “is not to be believed, but to be true”.
We need to distinguish a criterion from a moral decision procedure. A decision procedure is a body of rules or advice which is intended to provide agents with guidance on what to do. One way to state (A) as a decision procedure would be
||Make a list of available actions. Estimate what consequences each action would lead to, and assess them according to their goodness. Then do the action which leads to the best consequences.
The crucial insight of consequentialists is that if you accept (A), you need not endorse (B). The reason for this is that following a decision procedure like (B) is itself a moral decision, which might not be best given the circumstances—so it might be wrong to follow according to (A). For example, estimating what is best given your alternatives takes time, and spending such time has opportunity costs. Going to the party might be the objectively best thing to do, as compared to sitting at home doing the consequentialist calculations of whether one should go. The consequentialist should say that it is right to go to the party, and wrong to waste your time with consequentialist calculations. Similar remarks apply to other aspects of our lives. When I talk about decision procedures, I do not limit that term to explicitly formulated algorithms. There is also the question, for example, what character dispositions we ought to cultivate. It might be best to cultivate non-consequentialist character dispositions, such that I am disposed to be kind, trusting, generous and promise-keeping. Individuals who scrutinise every particular interaction with others on the basis of the utilitarian calculus might produce worse results than someone who has a general habit of trustworthiness.
Sophisticated consequentialism claims that there will usually be a gap between the criterion of rightness and the right decision procedure. A gap is an instance where the algorithm, procedure, or character dispositions (etc.) which a moral theory recommends differ in their content from the criterion of rightness of that theory. There are many empirical facts which make gaps likely in a consequentialist theory. It is useful to briefly mention some of them, as they will also be relevant to sophisticated rationalism. First, there are epistemic constraints on human agents. It is implausible to assume that we can calculate consequences in all circumstances, even to a rough degree. Always keeping your promises is a good rule of thumb. It might be good to have relatively strong and lasting psychological commitments to keep one’s promises, and consequentialism might ask us to develop these.
Second, human psychology is subject to many other constraints. The demands of consequentialism, if we are fully conscious of them, might paralyse us and undermine our long-term ability to do any good. Our natural impulses of altruism and our energy to help others might be limited, and we might need to be “selfish”, or at least partial to those around us, to some degree to keep psychologically healthy. This, too, favours establishing a character, and following decision procedures, which are at most moderately consequentialist.
Third, it seems to be in the logic of certain goods that we cannot achieve them if we intentionally aim for them. Being spontaneous or falling in love are useful non-moral examples. In other cases, it might be possible but difficult. Consequentially motivated calculators might find it difficult to sustain truly fulfilling, genuinely intimate relationships. So it is better if they intentionally set aside consequentialism when they make friends.
Fourth, everyone following the logic of maximisation might lead to dilemmas where the overall result is worse for everyone. For example, if everyone knows that I will break a promise if I believe breaking it achieves the greater good, people might not make promises to me in the first place. But it is valuable in general that the practice of promising exists, as it allows cooperation in complicated settings, especially market systems. So it might be best if I dispose myself to act in a way which does not evaluate particular promises according to the expected value of keeping them.
These effects, when taken together, make it likely that the dispositions, desires, intentions, algorithms, thoughts, and psychological structure of an average agent should not be guided by consequentialist concerns in most situations. This allows the consequentialist strong hopes for answering many of the proposed counterexamples to their view. Where it looks as if consequentialists cannot care about their friends, or cannot keep promises, or cannot give special weight to their own projects, we can use the distinction between criterion and procedure to argue that in most circumstances first appearances are deceiving.
Another strength of sophisticated consequentialism is that it follows organically from materials purely internal to consequentialism. We can argue
||Agent A ought to maximise the good. (Consequentialist criterion of rightness)
||Because of limitations in A’s psychology and other factors, A following some procedure P with non-consequentialist content brings about more good than trying to maximise the good directly. (Empirical assumption)
||Agent A ought to follow P. (Non-consequentialist decision-procedure)
There is nothing awkward, in other words, if we separate criterion and procedure in consequentialism. It is the most natural and consistent application of the view.
Furthermore, it is useful to note that the decision procedure which a consequentialist recommends can itself have a highly intricate structure—that is, there might be several levels within a decision procedure. Consider the following advice with regard to friendship,
In every-day situations, be partial to your friends. However, you should from time to time check whether your dispositions to be partial are really in line with achieving the best outcomes. For this, do some rough estimates whether your current dispositions lead to acceptable outcomes. If they do not, you should relevantly revise your first-order dispositions.
This rule asks the agent to operate on two different levels, and sometimes switch from one to the other. But while the second-order level in its content is closer to a consequentialist criterion of rightness, it is still a simplification and approximation of that level. Even expected utility calculations are likely to deviate from what actually promotes utility best, assuming an objective criterion of rightness. So consequentialism can prescribe a complicated, hierarchically ordered form of psychology to agents. But we should not confuse these levels with the criterion—procedure distinction; these are different levels within a decision-procedure.
Interesting debates on the criterion—procedure distinction turn not so much on whether consequentialists should be sophisticated, but to what degree and how. Some consequentialists believe that the procedures we ought to follow are close to genuine act-consequentialism. Modern-day “effective altruists” tend to think so. For other consequentialists, the life of the consequentialist agent will contain little explicit consequentialist thinking. On its logically most extreme, self-effacing consequentialists will claim that we ought to actively disavow consequentialism, and act and believe in terms of some other theory.
Standards of Assessment
Before turning to the parallel application of this move to rationalism, there is one wrinkle I would like to add. In addition to criteria of wrongness and decision-procedures, there is a third important type of entity in moral theorising. A decision procedure is meant to be applied by an agent from the first-person perspective. It describes the motives and character that I should cultivate, and the rules and procedures that I should use in making moral decisions. But we also use moral principles in a third-personal role to assess the behaviour of others, without a direct connection to action-guidance. We wish to know whether others behaved well or badly. In particular, we want to know which reactive attitudes such as blame or praise we should take regarding certain actions, agents or institutions. Principles which regulate such evaluations I label standards of assessment—or standards, for short. One example of a standard would be, for example,
||Blame S for doing A just in case doing A, amongst the options available to S, did not lead to the best outcomes.
(C) mirrors (A), our criterion of wrongness. But quick reflection shows that there is no guarantee that an act-consequentialist will advocate (C). Again, choosing to blame someone is itself a decision, and following (C) might not be best according to (A). In fact, it looks like a terrible way to assign blame. We virtually always fail to do what is best. So all of us would be blameworthy almost all the time. If we assigned blame in this way, people will be discouraged from acting morally, if the practice of blaming does not lose coherence altogether. This is likely to be worse than some alternative, more tolerant method of assigning blame. So it is unlikely that we ought to blame according to (C) if we are consequentialists.
The general reasons which allow for a gap between standards and criterion closely mirror the ones I outlined in the previous section regarding procedures and criterion. Standards are meant to fulfil a practical function for real-world agents—they are meant to guide our real-world reactive attitudes. But we suffer from mental and moral imperfections, limited time to reflect, and limited access to information. Because a standard of blame has to be useable, it cannot invoke the same categories we use on the level of criterion. The independence of standards and decision procedure is a bit more difficult to see. We might think that an agent should be praised just in case she has conscientiously followed the correct decision procedure, and blamed otherwise. But there is no guarantee for this convergence in consequentialism. It might be that there is an advantage in blaming someone who has followed the correct procedure perfectly.
An Initial Objection
At this point, I want to dispel the initial objection that the distinction between criteria, procedures and standards only arises in the context of consequentialism. In non-consequentialist theories, you might argue, there is either not even a conceptual difference between these different entities, or at least no substantive difference. This would be a mistake. The distinction is both conceptually and substantively relevant in non-consequentialist views. Revisiting some of the simple cases given above is enough to drive the point home. There are some situations where we need to immediately act. If I only have a few seconds to save a drowning child, I should save it now, unthinkingly. No plausible moral theory asks us to spend the time working our way through the correct moral theory first.
This is a simple case, but it establishes that the distinction is relevant to non-consequentialist views. The crucial question is again not whether non-consequentialist views allow for a gap, but how much of a gap they allow. Consequentialists are often fine with very radical forms of disconnect between criterion and procedure. Many non-consequentialists have objected to a wide gap between criterion on the one hand, and procedures and standards on the other hand. They have argued that such gaps would lead to objectionable forms of schizophrenia. I will argue for a relatively moderate gap, one which should be acceptable to most non-consequentialist theorists. At any rate, I discuss some of these objections later, in sec. 6.6.
The rest of this chapter builds on the insights from consequentialism to develop an analogous sophisticated rationalism. Rationalism claims that legitimacy is based on best promoting justice. We can now see, on the distinctions available to us, that this is ambiguous between the following three claims,
Criterion. Some policy or institutional framework is legitimate just in case it best promotes justice.
Decision Procedure. Rulers should aim to do what best promotes justice.
Standard. When individuals assess legitimacy, they should do so on the basis of what best promotes justice.
I will argue that the sophisticated rationalist is only committed to the first claim; she does not necessarily advocate the second or third. In fact, as I will shortly argue, the sophisticated rationalist is likely to believe that there is a gap between decision procedures, standards and criterion across a wide range of realistic circumstances.
Before we start, there are some differences between sophisticated rationalism and sophisticated consequentialism which it is useful to highlight. First, decision procedures in morality are addressed towards individual agents. The natural equivalent in politics would be rulers as addressees. A theory of legitimacy in this sense would be a “user’s manual”: it would be a body of advice telling rulers how they can permissibly rule. There is a long tradition of political philosophy taking this form, as advice to young impressionable princes. Alternatively, a decision procedure could address citizens in terms of their role as citizens. It would then constitute a “citizen’s manual”: such a theory of legitimacy would be a body of action-guiding advice telling individuals whether to obey or disobey government demands.
In the moral context, manuals of these two types will be important insofar as they provide guidance to individual agents. But questions of legitimacy are arguably more often raised from the third-personal perspective. That is, one of the main functions legitimacy plays is in the mind of citizens who third-personally assess whether the policies and rules they are being subjected to are legitimate. So standards play a much bigger role in thinking about legitimacy than procedures. In fact, I will set aside questions about how rulers and citizens should act entirely, and focus on the gap between the criterion of legitimacy and standards of legitimacy. This is not to say that there are no interesting questions to ask with respect to these other questions.
A moral standard of assessment guides our reactive attitudes towards some agent or action. A political standard of assessment guides our (i.e., citizens’) assessment of some policy, ruler, or body of law. It indirectly influences whether we believe that we owe allegiance and support to these entities or resistance. While the standards by which individuals assign moral blame might be shared and their content public, this is not a primary feature of those standards. You could follow some standard A to assign blame to others, and I could follow some different standard B. There is no pressing need to coordinate our reactive attitudes. However, in political assessments it is often important that we reach some, at least rough, convergence. I presume that standards of legitimacy will usually take the form of social conventions: that is, they are public rules and mechanisms which people share, and the content of which is public knowledge. A very simple convention, for example, would be to regard a political decision as legitimate when it is made by a constitutionally validated source.
Second, I earlier highlighted that a moral decision procedure can itself have a complicated internal structure in which we move back and forth between different levels. The same is even more applicable to standards of legitimacy. The legal system, for example, is an intricately structured system of rules, offices, positions and procedures. Legitimacy assessments in this system are based on complicated social and legal transactions.
Third, consequentialism is a purely teleological view: it only knows aims, but no constraints on the pursuit of those aims. There are, however, constraints within rationalism. This puts limits on what type of standards and procedures the rationalist is allowed to use. We could not use, for example, a convention on which we consider some policy to be legitimate which violated individual rights, even if doing so would promote justice best. Another implication is that we can never entirely put thinking about justice in the background. We might not always aim to directly promote justice, but we must always respect it. So we need to be aware of which constraints there are and when they apply.
An Example: Normative Positivism
Despite these differences, there is much that the sophisticated rationalist can copy, mutatis mutandis, from the sophisticated consequentialist. What interests me most in the political case is a gap between the criterion of legitimacy and the conventional standard of legitimacy. Sophisticated rationalism claims that the latter will often differ in content from the former.
We can best see this applied in a particular case. Consider the following challenge to simple (i.e., non-sophisticated) rationalism. If legitimacy assessments were based on direct appeals to what best promotes justice, so the objection goes, politics would be difficult, contentious, subject to instability and strife, open to abuse and corruption, and so on. So politics conducted along rationalist lines would have all kinds of negative effects. I want to focus on one particular version of such an objection, Jeremy Waldron’s argument for “normative positivism”. Positivism is a claim about the nature of law. Positivists think that legal validity—whether something is law or not—depends merely on whether certain social facts obtain, usually whether there is some complex recognition of a legal rule by public officials. So there is no necessary connection between law and morality (the famous “separation thesis”). An associated claim is that legal validity can be explained through its sources: something is law just in case it has been decided in the legally correct way.
Waldron advocates positivism in this sense, but on the basis that we ought to be positivists, not that positivism is true as a matter of theoretical analysis. He argues as follows. Waldron stresses that in politics certain issues need to be authoritatively, and somewhat unanimously, settled. We need to have legal forms of interacting with each other which are broadly recognised by everyone. This means that law needs to be both univocal and transparent—it needs to provide a unique solution to conflicts, and this solution must be rationally accessible to most people. (That a Kantian sympathises with these ideas should be clear.) Waldron then adds that finding such solutions under conditions of reasonable disagreement about justice is challenging. In particular, he argues, a non-positivist legal system which made legal validity directly depedent on the moral merits of the cases before it would fail to adequately settle conflicts, precisely because we reasonably disagree about the matters under debate:
law must be such that its content and validity can be determined without reproducing the disagreements about rights and justice that it is law’s function to supersede.
Instead, Waldron suggests, we should choose to be legal positivists—we should make the question of whether something is law independent from its moral merits. Waldron’s claims are about legal validity, but we can extract from his remarks a challenge to a rationalist approach to legitimacy as well. The value of legitimacy has a settling function: it is supposed to settle certain important political matters we face in our societies. Legitimacy can fulfil this function only if it does not rest on controversial, factional beliefs. But justice-based rationalism appears to rest precisely on such beliefs. So rationalism would lead to unstable and unsettled political situations which undermines the very point of trying to find legitimate political solutions.
A similar criticism has also been advocated by Charles Larmore. Larmore also starts from the importance of disagreement. In light of such disagreements, he argues, “moral ideals” regarding human flourishing and justice are “part of the problems of political life, rather than providing the basis of their solution”. That is, if we develop a theory of legitimacy, we ought to look for ways to deal with our disagreements, not to reproduce them again. For this reason, we should turn away from thinking of political philosophy as applied moral philosophy which appeals to justice “as a purely moral ideal”, because such ideals are contentious. Larmore’s alternative proposal is that we ought to conduct political discussion in terms of the value of non-moralised “political justice”, though the details of that proposal are not particularly interesting to us.
Once we have the structure of sophisticated rationalism in place, we can see that Waldron’s and Larmore’s arguments need not worry the sophisticated rationalist, though they certainly provide a dire problem for the non-sophisticated rationalist. Sophisticated rationalism does not demand that every-day politics is conducted through direct appeals to truths about justice. If a politics based around appeals to justice leads to justice being promoted worse, then we should not conduct politics directly on the basis of justice, but some other basis—perhaps on the basis of positive law, as Waldron suggests, or on the basis of non-moralised “political justice”, as Larmore suggests. In other words, the sophisticated rationalist can accept the Waldron-Larmore argument as an argument for normative positivism on the level of standards, while holding on to the idea that what ultimately matters to legitimacy is only justice. It should be obvious what the reconciliationist potential of this strategy is: even though the internal grammar of legitimacy and justice are so different, the rationalist need not dismiss many of the standard features of thinking about legitimacy such as its focus on legal validity and procedure.
A General Argument
The case for normative positivism is one instance of a general argument, analogous to the argument in favour of sophisticated consequentialism:
||A policy or set of institutions is legitimate to the degree that it best promotes justice (within the constraints of justice). (Rationalist criterion of legitimacy.)
||Because of limitations in our society, following some non-rationalist standard (which stays within the constraints of justice), such as normative positivism, better promotes justice than trying to promote justice directly. (Empirical assumption.)
||We should assess legitimacy by applying that standard. (Non-rationalist standard of legitimacy.)
Just like sophisticated consequentialism, sophisticated rationalism is a natural development of rationalism once we have the criterion/standard distinction in place. This is worth emphasising. The conciliationist strategy I suggest is not an ad hoc fix, born out of a desire to account for some intuitive judgments.
The second premise in this argument is empirical, so there is no guarantee that there will be a gap between criterion and standard. However, many of the relevant empirical facts which enter into the second premise are quite robust. Some of these are similar to facts that I invoked in the context of sophisticated consequentialism. We must make political decisions under epistemic constraints and with a limited time budget. Knowing what best promotes justice in modern, complicated societies is, even for people who agree in their moral values, an enormously complex task. We are often better off using simplified rules of thumb for this task rather than aiming to promote justice directly.
Another robust fact, mentioned by both Waldron and Larmore, is the fact that politics is characterised by deep and lasting disagreement. Our ability to pursue justice usually depends on the willingness of others to coordinate with us in their behaviour. If we want to achieve the various preconditions important to justice—peace, stability, public order, and so on—we have to rely on their cooperation. It is because of these features that we settle on clear, transparent procedures—democracy, law, public authority, and so on—which bypass our disagreements about justice. This almost always gives us a reason to set aside explicitly rationalist standards of legitimacy. The difference to philosophers like Waldron, of course, is that sophisticated rationalism sees no deeper moral significance in reasonable disagreement. Unlike justificatory liberalism, it thinks of reasonable disagreement as a practical, not a moral, challenge.
The Applied Standard
The example of normative positivism gives us an idea of how sophisticated rationalism will approach practical problems. In this section, I will further sketch the contours of the applied standard of legitimacy a rationalist is likely to advocate.
Desiderata for Standards
The project of developing applied standards of legitimacy is a complex problem partially based on hard-to-obtain empirical information. However, designing such standards is not purely empirical. It will be guided by the criterion of legitimacy we accept, and the way we will select the applied standard of legitimacy will also take into account normative desiderata. Let me highlight some principled desiderata which the applied standard ought to fulfil.
First, the applied standard ought to be accessible. That is, ordinary people should, with reasonable effort, be able to know what it demands. This imposes constraints on how complicated the standard can be, and how many features of a situation it takes into account. The standard should also be reasonably determinate: it should avoid delivering too many ambiguous results. It is for this reason that we expect the standard to focus on clearly discernible, public information—such as democratic procedures, laws, and other official trappings of institutions.
Second, different people in a given society, when applying the same standard, should converge on roughly equal judgments. This has to do with the coordinating function of legitimacy mentioned earlier. Legitimacy is unlike blame in this respect. It is not terribly important that we all agree on who is to blame and how. But convergence on judgments of legitimacy is often crucial. This is not to say that everyone needs to agree with the verdicts of the standard. Returning to the example of normative positivism, it is important that everyone knows what the law is; it is of less importance that everyone agrees that the law is just.
Third, the standard of legitimacy in some important sense needs to be realistic. That is, it must be a standard which we can expect most people to follow and apply. Closely associated with this desideratum is Williams’ demand that attempts at legitimation must “make sense” to people. “Making sense” again does not imply that the standard is acceptable to everyone. But it requires that most people see the proposed standard as a live option given their concrete social, cultural and historical conditions.
A fourth desideratum is that the applied standard should still provide citizens with reasons in some sense, though as we have seen, these need not be moral reasons. The applied standard should not be based on brute appeal to power or force, threats, or other forms of manipulation. The operation of the law, for example, is rationally accessible by citizens, and therefore fulfils this desideratum.
Fifth, the applied standard should not license or require actions which are clearly repugnant from the point of view of the critical level. In particular, the standard should not require us to actively support rulers and decisions which are paradigm cases of injustice. The applied level might well, however, require us to condone some suboptimal institutions. One thing it is useful to highlight at this point is that there is a difference between actively thinking in moral terms which are repugnant from the point of view of justice, and failing to always think in terms of justice. Think back to normative positivism. Positive law will often operate in terms of categories and concepts that we do not find in our theory of justice, and the law might make distinctions which are unknown in it. At the same time, the law does not operate in categories which are directly objectionable to justice.
Sixth, the applied standard should not be based on systematic deceit about the critical level. It should not be based on making the critical level a Sigdwickian esoteric morality, or on telling Platonic noble lies to some part of the population. This also excludes systematic self-deceit about the ultimate basis of legitimacy. Ideally, the critical/applied distinction should be publicly knowable by everyone, though I do not take this to be a necessary requirement.
More Conciliatory Results
Working out which standards fit these desiderata is a complicated process. But the idea that we have to “work out” the standards is in some sense already a misconception. It is not up to philosophers and social scientists to design standards of legitimacy from scratch. Any designed set of rules by which to assign legitimacy which philosophers could come up with is likely to fail the first three standards of legitimacy. A much better approach is to start from what we already have. Most societies have established conventions in place by which legitimacy is assessed. In fact, this is a tremendous understatement: the system of legal rules, constitutional arrangements and democratic procedures which characterise modern societies are gigantic cultural achievements. These are also already conventions which many people accept, and which they will be hard to move away from radically.
So our thinking about standards of legitimacy will not (usually) be thinking about how to design something from scratch, but rather about how we ought to tinker with an existing system of institutions and conventions. This provides another conciliationist result. The rationalist does not need to exhibit the hopeless naiveté of the Enlightenment planner. Legitimacy, for the rationalist, is not an exercise of dreaming up perfect societies in the ivory tower. The terms in which it demands we operate are the cultural terms of the historical here and now. Admittedly, there is a whiff of planners’ arrogance wafting through the work of some historically influential instrumentalists about legitimacy, such as Plato. Be that as it may, this is not a mistake the sophisticated rationalist need to commit.
Another conciliationist result is that on a practical standard of legitimacy historical continuity will count for something, even if it does not on a theory of justice. Consider the question, already raised in section 6.1, whether usurpers can be legitimate. Does someone who comes to power as part of a military coup enjoy legitimacy? On a simple rationalist theory, the way people come into power does not matter greatly. Indeed, if transitions are cost-free and involve no injustice, and a change in power promises better results in terms of justice, simple rationalists will positively welcome coups, usurpers, conquerors, occupiers and the lot. Buchanan sees this possible implication; in response, he tags on a non-usurper clause to his own instrumentalist principle of legitimacy, requiring that legitimate rulers promote justice well and are not usurpers. But theoretically speaking, introducing such a clause is awkward, and reeks of fixing rationalism ad hoc.
However, there are many good reasons to be sceptical of allowing legitimacy to usurpers. First, a standard of legitimacy which grants legitimacy to power-grabbers incentivises such behaviour. It is likely to result in more rapid and frequent changes in power, with the attendant costs of civil unrest and instability. Second, this system allows private judgment to play a big role. In its practical implementation, it sends a message to powerful elites that they can attempt to gain power if they think this would be preferable in terms of justice. But the opinion of small elites in this respect is likely to be skewed and unreliable. Third, there are good historical reasons to distrust usurpers and the rectitude of their intentions. So all in all, it makes good practical sense to follow a standard of legitimacy which favours existing institutional arrangements and discourages usurpers. Such a standard of legitimacy would usually assign legitimacy only to those rulers and institutions that have acquired their position in the prescribed legal way, and shown some continuity in ruling.
We can also see some of the more radical potential of two-level rationalism in this example. Just as a “free for all” convention regarding usurpers will be quite detrimental, so will be a rigid anti-usurper convention. If we never allow existing institutions to be toppled or overcome through extra-legal means, we would often be stuck with clearly suboptimal institutional arrangements. In the same way, it is hard to see why we should always feel bound by constitutions, or super-majority requirements for constitutional change, or existing agreements on separation of powers. We should not ignore existing legal arrangements lightly. Still, rationalism in its two-level form will always make us sceptical of any rules which allow no exceptions. Rationalists will not assign to rules the level of sanctity which some non-rationalists do. So some unlawful usurpers will be counted as legitimate by the rationalist, and might even find her active support.
Interlude: Public Justification
It is helpful at this point to discuss Jonathan Quong’s view. Quong is a political liberal, but one who uses a professed rationalist starting point. Quong’s fundamental view is that the state is necessary to fulfil our duties of justice. However, Quong rejects a “basic view” according to which legitimacy directly depends on justice. For our purposes, we can interpret him as rejecting a non-sophisticated rationalism on which criterion and standard are identical. Quong opposes this view because “[p]ervasive reasonable disagreement about justice[,] combined with the basic view of legitimacy, would undermine the liberal state’s capacity to act effectively”. By now we are familiar with this type of argument through our engagement with Waldron and Larmore. What is more interesting is Quong’s positive suggestion. He writes:
The natural duty of justice should not be understood as directing us to support only those institutions that are perfectly just. Rather, it directs us to support and comply with those institutions that are reasonably just, and by that I mean those institutions for which good arguments can be made by reference to values or principles acceptable to all citizens conceived as free and equal […].
Rephrasing Quong’s position in my terms, he claims that on the level of the standard of legitimacy we ought to be political liberals. That is, in making practical decisions about legitimacy we should appeal to the idea of what other people could reasonably accept or reject, because this is the best way to fulfil our duties of justice given the empirical constraint of reasonable disagreement. If this argument is correct, then political liberalism makes a surprising return, this time on the level of standards.
However, I find the idea that public justification would be part of a practical standard of legitimacy thoroughly implausible. First, note that public justification, just as much as “public reason” or the idea of what “people could reasonably reject”, is a philosopher’s brain child. Ordinary citizens do not usually think about legitimacy in these terms. Working out what these concepts require takes considerable philosophical energy and time. This is likely to violate the accessibility and realism desiderata. Whatever standard of assessment for legitimacy we use, it should make sense to most people. Political liberalism does not. Furthermore, note how little concrete action-guidance political liberalism provides. The idea of reasonable acceptance is not very helpful in determining whether particular laws and policies are legitimate. But we want a standard of legitimacy to play a practical, evaluation-guiding role: it should enable us to make precisely those kinds of judgments. Lastly, remember Waldron’s criticism that law based on justice would reintroduce the disagreements that law is intended to avoid. Politics based on public justification is open to a similar criticism. There is tremendous disagreement about what is reasonably acceptable to others, what is in accordance with public reason, and so on. There is little reason to think that a public reason standard of legitimacy would help us to overcome or diminish disagreement in this respect—it just reintroduces a different type of disagreement.
In short, a standard of legitimacy based on public justification is not a convincing, realistic candidate. Public justification is simply too high-level philosophical. Plausible ingredients in a standard of legitimacy will be much less philosophical—they will focus on concrete institutions, procedures, conventions, rules, political offices, and so on.
We have seen that criteria and standards of legitimacy are likely to come apart to some degree. However, rationalist thinking about legitimacy has some active role to play in our societies. So our next task is to sketch how the two levels interact. The analogous task for the consequentialist is to describe the moral psychology of a sophisticated consequentialist: of someone who fundamentally is a consequentialist, but has non-consequentialist character dispositions and follows non-consequentialist decision rules at the same time. Some of the goalposts in the current context again change. Our focus is not on the psychology of a single individual, but rather on how a non-rationalist convention of assessing legitimacy can coexist, inside a given society, with diverging opinions about how to achieve justice. I will first describe a natural model how one might understand the connection, which I then modify.
A first model comes to us from Hare, who thinks that the applied procedure is the one which we use under normal circumstances. The idea is that the applied procedure is the one we follow if we make every-day decisions. The critical level is triggered from within the applied level if we encounter a particular conflict, a conflict which we cannot solve on the applied level itself. On this conception, the applied level has a type of escape clause embedded in it. It might take, for example, the following structure:
If you encounter a problem which the current applied rules cannot clearly solve, or where following the current rules has clearly disastrous results, switch your level of moral deliberation from the applied level to the critical level.
Call this the Fallback Model. On this model, our psychology is hierarchically ordered. It combines a normal mode of operating with a switch to a critical mode in decisive moments. Think of it as a computer program which combines a day-to-day algorithm with a troubleshooting mode whenever the algorithm encounters a bump.
More details could be filled in, but we can already see a number of problems, especially if we try to transpose this model into the political realm. First, disagreement and conflict are the normal condition in politics. If we fall back onto considerations of justice as soon as we encounter a disagreement on the applied level, the latter seems to be of little help. Second, the Fallback Model suggests that there is a deeper and a more shallow level in our psychology. But we might worry that this psychological hierarchisation is unstable, or that it involves a kind of worrisome schizophrenia—objections I will discuss at greater depth in the next section. Third, even if the promotion of justice might be indirect, we must ensure that we stay within the constraints of justice. There needs to be some constant connection to the critical level to check whether we are.
However, the Fallback Model is not the only way how we can think about the connection of the critical and the applied levels. Another model is to use the critical and applied standards in parallel. Assume, for example, that you are in a court. In this context, you assess legitimacy by following Waldron’s normative positivism. But this does not require that you set aside, or forget about, the critical level. You can continue to evaluate the law under discussion in terms of the critical standard—in terms of what promotes justice best. When it comes to making a decision, however, you need to prioritise one level over the other. Which one you ought to prioritise will depend on the context, and whether you occupy some official role. If you are a judge, for example, you are normally required to assess the legitimacy of a law by its sources, as normative positivism demands. On the other hand, if you discuss the same matter in the public sphere as a private citizen, you might evaluate the law according to its tendency to promote justice best. The society I am imagining, then, has reached a certain amount of transparency about the relationship between justice and legitimacy. We all agree to “play the legitimacy game” as a preferable second-best in the face of the various empirical obstacles we are facing in our society, chief amongst them the pervasive fact of disagreement. We accept that there is a need to settle certain important collective questions, and in the face of this need we prioritise the practical level in a variety of contexts. But we continue to think about our society in terms of justice, and debate how legitimacy ought to be assigned in the pursuit of that aim.
Call this the Parallel Model. On this model, the critical level is exercised in parallel to the applied level. We constantly review and put under scrutiny the practices of legitimacy we follow from the point of view of justice; there is no trigger which makes us consider this level only sometimes. In many contexts we follow the practical standard, even if an assessment from the critical point of view differs. We already know, in principle, how this separation works. We are used to think that appealing to law and procedures is apt in some contexts and missing the point in others. If you ask whether the parking ticket you got yesterday was legitimate, you are asking a question on the practical level, and the appropriate response will be in terms of the legal authority of the person who has issued the ticket. If you ask whether the constitutional system under which the ticket was issued is legitimate, on the other hand, you will usually be understood to have asked a question on the critical level. That the constitution takes a certain legal form seems beside the point in those debates; we should understand the question as a question about the justice of the constitution.
The advantage of the Parallel Model is that it does not require us to hide from ourselves our different commitments. We are not consequentialists one moment, and committed friends the next. Sophisticated rationalism is compatible with us to knowingly and openly operating two different kinds of standards, without us trying to forget about the critical standard, or putting it in some other way into the psychological background.
Let us turn to some objections to sophisticated views. We need not make these up from scratch. Given that sophisticated rationalism is modelled on the basis of sophisticated consequentialism, we can look at some of the objections to that view, and see whether they also apply mutatis mutandis in the current context.
A first objection is that sophisticated consequentialism will “usher itself from the scene”—that is, it stops being a distinct position. The idea is that if consequentialism coincides in its judgments with non-consequentialism, then we should not call it consequentialism any longer. Similarly, if rationalism in practice looks like a non-rationalist view, we should stop considering it a distinct view. So we might raise the
Indistinctiveness Objection. If rationalism—in particular, if the promotion of justice—has no place in the applied deliberation of citizens about legitimacy, then it ceases to be a distinct, interesting position.
This objection misses the mark badly. First, even if sophisticated rationalism had no distinct practical implications, it would signal an important difference in how we think about legitimacy—it would make a difference to what is explanatorily more fundamental. This would at least have intellectual import. Second, it is simply not true that rationalism has no effect on the practical deliberation of citizens. The Parallel Model is very far away from self-effacing consequentialism which might indeed be indistinctive. On sophisticated rationalism, some important debates about legitimacy will be held, and even decided, in terms of promoting justice. There is no principled guarantee that rationalism will leave things untouched, as we saw in our brief discussion of usurpers. All this makes a clear difference to actual practices of legitimacy.
A second objection is more serious. A standard objection against sophisticated consequentialism is that the kind of psychological structure it demands of an agent is overly demanding, if not straight-forwardly impossible. For example, on the one hand consequentialism requires us to see our friends “from the point of view of the Universe”; at the same time, it encourages us to have emotional attachments to our friends based on loyalty and partiality. But, the charge goes, it is impossible for an average agent to uphold such contradictory commitments in a psychologically stable way. This objection is based on “ought implies can”: we simply cannot exhibit the kind of sophistication that the consequentialist demands; thus, any position which demands it is false.
If we transpose this objection into the political realm, and focus in particular on normative positivism, we can state it as the
Instability Objection. It is impossible for a society and its members to remain committed to normative positivism on an applied level, and rationalism on a critical level, in the long run.
The objection is that one of the two levels will encroach on, and ultimately win over, the other. If in some contexts we think about legitimacy in terms of promoting justice best, will we not lose patience with applied standards which make no reference to justice? This is in the end an empirical question. It is hard to think that there is an a priori obstacle to such a society. One important observation which should be cause for optimism is that our societies are already used to the coexistence of different kinds of moral vocabulary in different contexts—that is, a “justice” vocabulary in some contexts, and a “law” or “procedure” vocabulary in other contexts, and a “morality” or “virtue” vocabulary in yet another. We do not find it hard or confusing to say that a decision was legitimate but unjust.
Let me return to one of the previous examples. Could a society publicly accept and consistently act on Waldron’s normative positivism for rationalist reasons? Assume that we collectively decided that, because of the disagreements about justice we have, the best we can do is to make legitimacy a mix of legal validity and democratic procedure. This, we argue, is most likely to promote justice best in the long run, as seen from each individual conception of justice. So we apply legal standards when we assess legitimacy. Would such an arrangement break down over time? It is not clear to me why it would. One reason why such an arrangement could be decently stable is because we need not have a split personality to be committed to this scheme; we could be perfectly transparent about it. The critical/applied split would not be a secret, and we would not need to be systematically mistaken about the structure of our fundamental commitments, nor need we hide them from others.
When one talks about stability, one might have a stronger, Rawlsian requirement in mind. One might think that a political theory is subject to a stability test: if widely accepted in a society, it needs to be able to generate and sustain its own support over time. The entire third part of Rawls’s Theory of Justice is committed to showing that his theory of justice satisfies a stability constraint of this type. Rawls does this by developing a philosophical anthropology: he tries to show that on plausible philosophical assumptions about human motivation, rationality and the human good, people in a well-ordered society will remain committed to his theory of justice over time. A stronger objection to sophisticated rationalism would be that it fails a stability test of this kind. A sophisticated rationalism, we might argue, could not support itself over time.
My answer to this line of objection is twofold. First, I am optimistic that a similar philosophical anthropology could be given to accompany sophisticated rationalism. But admittedly, more philosophical work would need to be done to fully sketch such a philosophical anthropology. Second, perhaps a society guided by sophisticated rationalism would, over the long run, suffer from irremediable tensions. But here we can return to the assimilation strategy I sketched earlier. If there is foreseeable instability by following a particular standard of legitimacy, and this instability affects our ability to promote justice, sophisticated rationalism will take this instability into account. If the instability is not a matter of justice, however, then the rationalist will just dismiss it as not being a primary concern.
Even if the sophisticated consequentialist’s commitments to both friendship and consequentialism can be stably combined in her psychology, another objection is that such an agent suffers from a form of repugnant schizophrenia. The moral life of such agents lacks integrity, an integration of different values into a whole life. We can rephrase the objection in the context of rationalism and normative positivism. Every sophisticated rationalist is committed to justice, and regards justice as the ultimate foundation of legitimacy. At the same time, they “play the legitimacy game” which focusses away from what would promote justice best. This seems repugnant. Let us keep the label:
Schizophrenia Objection. A society which is committed to promoting justice on the critical level, and to normative positivism on the applied level, is objectionably schizophrenic, even if stable.
A first problem with the schizophrenia objection is that many ways of stating it are simply question-begging. We might think that consequentialists are repugnantly schizophrenic because their psychology requires them to value the wrong things, or value them in the wrong way. For example, what we might object to in the psychology of the consequentialist is that they do not intrinsically value their friendships, or do not value them in the appropriate way. But the consequentialist can simply reject those charges as question-begging. Similarly, I suspect that many particular ways of making the Schizophrenia Objection against rationalism run the danger of begging the question against it. If the objection were, for example, that the rationalist is not truly committed to legal procedures, or does not truly accept the intrinsic value of democracy, the rationalist could accept those charges, but reject them as question-begging. The rationalist starts from the assumption that there is no (important) intrinsic value in democracy. This might of course be wrong; but simply denying it without further argument begs the question against the rationalist.
There are two other reasons to think that the Schizophrenia Objection has little power in the current context, even if it might have against consequentialism. The first is again to note that the power of the Schizophrenia Objection varies with both the degree and the kinds of psychological disconnect that a sophisticated theory requires. The objection is strongest where we are required to value things one way “deep down”, but then deceive ourselves into valuing things another way on the everyday level. So the objection works strongest on a hierarchical model like the Fallback Model, which I rejected in favour of the Parallel Model. In addition, the substantive gap between critical and applied level in sophisticated rationalism is much smaller. What is terrifying about the sophisticated consequentialist is the idea that deep down they do not value their friends and loved ones in any special way at all, while at the same time they appear to pretend that they do. What we find disturbing about this kind of character is the almost absolute inversion of the type of valuing involved on these levels. However, the sophisticated rationalist is not quite such a bone-chilling character. Their rejection of usurpers, and their willingness to follow the law, might not quite go as deep as that of others. But surely this is not on the level of psychopathy we suspect in the consequentialists. What is more, there is some continuity between the critical and the practical level. Legal reasons which bear on a case, for example, often mirror moral reasons; for example, the categories in which we assess guilt in criminal trials (intention, harm, etc.), even if they do not perfectly track justice, have much in common with how we evaluate cases morally. The rule of law is in some ways a close cousin of justice. So it is not that the rationalist on the practical level responds to features of the political life which are entirely alien to us on the critical level.
What emerges from this thesis is a picture of political legitimacy which is radical in many respects. According to rationalism, we ought to be justice fanatics: it should be the first thing on our mind when we think about political institutions. Legitimacy is not an independent value, but one which is purely derived from justice, both as a constraint and as an aim. This changes dramatically how any philosophical work on legitimacy ought to be done. The interesting, fundamental normative questions to ask will be in a theory of justice; which policies and rulers we ought to consider legitimate turns out to be a result of combining such a theory with a number of empirical facts and feasibility constraints.
Furthermore, those political institutions which rationalism deems legitimate will be limited in several ways:
These institutions will not generally enjoy authority, or a power to impose duties on their citizens (sec. 1.3);
These institutions are not generally states—they do not normally enjoy a monopoly to rule, and under some circumstances they must allow that other institutions compete with them (sec. 1.5.2, 4.4.2);
These institutions are bound by the moral rights of individuals, which constrain what they can do and which cannot be overridden easily (ch. 2)—in particular, political institutions might need to allow semi-independents to live on their territory (sec. 3.3.3);
These institutions will only be allowed to fulfil certain functions, especially those relating to establishing the rule of law and the promotion of justice, but it is not immediately clear whether all functions modern states fulfil can be justified (sec. 4.4.1).
One thing these limitations have in common is that they entail that we should dismiss a monolithic picture of legitimacy. We are used to think about legitimacy as an overarching property of political institutions, which is powerful both in terms of scope and strength. If the state is legitimate, then it enjoys exclusive authority over everyone inside a given territory, in almost all areas of politics and with very few exceptions. Add in the trickle-down approach to legitimacy (sec. 5.2), and states turn out to demand for themselves vast amounts of power. Rationalism undermines this picture drastically. It suggests that legitimacy is a limited property which come in small patches, often discontinuous and localised, in highly complex shapes. We will be required to pay much greater attention to the particulars of each society, individual, political decision, and function that an institution fulfils. Some political decisions will be legitimate in some respects, but there will be no overarching argument which justifies all political institutions as such, especially not the state.
At the same time, this thesis also provides some reasons to be optimistic about the legitimacy of political institutions:
While political institutions do not enjoy authority, there are various transmission channels which will translate into individual duties to conform with their commands, and duties to abstain from competing with the activities of legitimate institutions (sec. 1.4);
Political institutions are compatible with strong, moral rights (ch. 3-4);
Political institutions do not play a merely instrumental role in promoting justice, but a constitutive role in realising our rights (ch. 4)—indeed, in this respect they are indispensable in modern societies;
How we best ought to pursue justice is highly sensitive to social and cultural constraints. Sophisticated rationalism will often recommend that we pursue justice indirectly, through legal and democratic procedures (ch. 6).
These ideas moderate some of the more radical potential of rationalism. Rationalism is unlikely to demand that we abolish existing political institutions altogether, especially not through violent revolution (sec. 4.5.2, 6.4.2)—rather, we ought to transform and modify them. Rationalism does not usually entail that we ought to design new political institutions from scratch, or that the philosophers should rule (sec. 5.2), or that everything should always be oriented towards the promotion of justice (sec. 6.3.3).
Many of the details of rationalism, and in particular how radical its demands will turn out to be depend on how you flesh out the details of a theory of justice. I have hinted at the general shape I think such a theory will take: it will be an individualist, status-based theory (sec. 3.4) which emphasises both the sovereignty and the support-worthiness of individuals (sec. 3.5.1), and which accepts that individuals have welfare rights (sec. 4.3.2). But beyond these rough contours, much more would need to be done to establish the content of such a theory.
My thesis also leaves a series of other lacunae which we would need to fill. For example, to assess how much moral competition will be allowed, we would need to fully describe the transmission channels through which individuals come to have duties to conform with the state (sec. 1.4.2). Furthermore, we would need to provide a full list of the Kantian functions that a state could legitimately fulfil (sec. 4.4.1). Lastly, much more empirical work would need to be done to assess which applied conventions of legitimacy are most likely to promote justice (sec. 6.4.2).
These are challenges to tackle in future work. What I hope to have established is not a theory of legitimacy which is “ready to use” in the face of various applied issues. Rather, my aim was to develop a theory which provides a general framework of thinking about legitimacy—a theory which on the one hand allows us to locate legitimacy as a value in the broader philosophical landscape, and on the other hand provides us with a useful structure to think through political problems.
Alexander, Larry. 1985. “Pursuing the Good–Indirectly.” Ethics 95 (2): 315–32.
Anderson, Elizabeth. 2014. “The Fundamental Disagreement between Luck Egalitarians and Relational Egalitarians.” In Distributive Justice and Access to Advantage: G. A. Cohen’s Egalitarianism, edited by Alexander Kaufman, 21–39. Cambridge: Cambridge University Press.
Applbaum, Arthur Isak. 2004. “Legitimacy in a Bastard Kingdom.” John F. Kennedy School of Government Center for Public Leadership Working Papers, Spring 2004, 73-94.
———. 2010. “Legitimacy without the Duty to Obey.” Philosophy & Public Affairs 38 (3): 215–39.
Arneson, Richard. 2003. “Defending the Purely Instrumental Account of Democratic Legitimacy.” Journal of Political Philosophy 11 (1): 122–32.
———. 2005. “The Shape of Lockean Rights.” In Natural Rights Liberalism from Locke to Nozick, edited by Ellen Frankel Paul, Fred Miller, and Jeffrey Paul, 255–85. Cambridge: Cambridge University Press.
Audard, Catherine. 2007. John Rawls. Montreal: Mcgill-Queens University Press.
Bader, Ralf. “Counterfactual Justifications of the State.” Manuscript. Available at http://users.ox.ac.uk/~sfop0426/Counterfactual%20justifications%20(R.%20Bader).pdf
Băiaşu, Sorin. 2014. “Kant’s Justification of Welfare.” Diametros 39: 1–28.
Bales, Eugene. 1971. “Act-Utilitarianism: Account of Right-Making Characteristics or Decision-Making Procedure?” American Philosophical Quarterly 8 (3): 257–65.
Bedi, Sonu. 2009. Rejecting Rights. Cambridge; New York: Cambridge University Press.
Beetham, David. 2004. “Political Legitimacy.” In The Blackwell Companion to Political Sociology, edited by Kate Nash and Alan Scott, 107–16. Oxford: Blackwell.
Bird, Colin. 2014. “Coercion and Public Justification.” Politics, Philosophy & Economics 13 (3): 189–214.
Boardman, William. 1987. “Coordination and the Moral Obligation to Obey the Law.” Ethics 97 (3): 546–57.
Brandt, R. B. 1988. “Fairness to Indirect Optimific Theories in Ethics.” Ethics 98 (2): 341–60.
Buchanan, Allen. 1984. “What’s So Special about Rights?” Social Philosophy and Policy 2 (1): 61–83.
———. 1999a. “Recognitional Legitimacy and the State System.” Philosophy & Public Affairs 28 (1): 46–82.
———. 1999b. “Rule-Governed Institutions versus Act-Consequentialism: A Rejoinder to Naticchia.” Philosophy & Public Affairs 28 (3): 258–70.
———. 2002. “Political Legitimacy and Democracy.” Ethics 112 (4): 689–719.
———. 2004. Justice, Legitimacy, and Self-Determination: Moral Foundations for International Law. Oxford: Oxford University Press.
Caney, Simon. 1991. “Sandel’s Critique of the Primacy of Justice: A Liberal Rejoinder.” British Journal of Political Science 21 (4): 511–21.
Celano, Bruno. 2013. “Publicity and the Rule of Law.” In Oxford Studies in Philosophy of Law, edited by Leslie Green and Brian Leiter, vol. 2, 122–44. Oxford: Oxford University Press.
Ceva, Emanuela. 2012. “Beyond Legitimacy. Can Proceduralism Say Anything Relevant about Justice?” Critical Review of International Social and Political Philosophy 15 (2): 183–200.
Chan, Joseph. 2012. “Political Authority and Perfectionism: A Response to Quong.” Philosophy and Public Issues 2 (1): 31–41.
Christiano, Thomas. 1999. “Justice and Disagreement at the Foundations of Political Authority.” Ethics 110 (1): 165–87.
———. 2008. The Constitution of Equality: Democratic Authority and Its Limits. Oxford: Oxford University Press.
———. 2013. “Authority.” Stanford Encyclopedia of Philosophy (Spring 2013 edition), http://plato.stanford.edu/archives/spr2013/entries/authority/.
Claus, Laurence. 2012. Law’s Evolution and Human Understanding. Oxford: Oxford University Press.
Coakley, Mathew. 2011. “On the Value of Political Legitimacy.” Politics, Philosophy & Economics 10 (4): 345–69.
Cohen, G. A. 2008. Rescuing Justice and Equality. Cambridge, Mass.: Harvard University Press.
Coleman, Jules, and Brian Leiter. 1993. “Determinacy, Objectivity, and Authority.” University of Pennsylvania Law Review 142 (2): 549–637.
Crisp, Roger. 1992. “Utilitarianism and the Life of Virtue.” Philosophical Quarterly 42 (167): 139–60.
———. 1997. Routledge Philosophy Guidebook to Mill on Utilitarianism. London: Routledge.
Cruft, Rowan. 2004. “Rights: Beyond Interest Theory and Will Theory?” Law and Philosophy 23 (4): 347–97.
———. 2006. “Against Individualistic Justifications of Property Rights.” Utilitas 18 (2): 154–72.
———. 2010. “Kamm and Miller on Rights’ Compatibility.” Ethical Theory and Moral Practice 13 (4): 393–401.
Delmas, Candice. 2014. “Samaritanism and Political Legitimacy.” Analysis 74 (2): 254–62.
Dworkin, Ronald. 1975. “The Original Position.” In Reading Rawls: Critical Studies on Rawls’ A Theory of Justice, edited by Norman Daniels, 16–52. New York: Basic Books.
———. 1986. Law’s Empire. Cambridge, Mass.: Belknap Press.
Ebels-Duggan, Kyla. 2012. “Kant’s Political Philosophy.” Philosophy Compass 7 (12): 896–909.
Edmundson, William. 1998. Three Anarchical Fallacies: An Essay on Political Authority. Cambridge: Cambridge University Press.
———. 2002. “Social Meaning, Compliance Conditions, and Law’s Claim to Authority.” Canadian Journal of Law and Jurisprudence 15: 51–67.
———. 2004a. An Introduction to Rights. Cambridge: Cambridge University Press.
———. 2004b. “The Duty to Obey the Law.” Legal Theory 10 (4): 215–59.
———. 2010. “Political Authority, Moral Powers and the Intrinsic Value of Obedience.” Oxford Journal of Legal Studies 30 (1): 179–91.
Enoch, David. 2014. “Authority and Reason-Giving.” Philosophy and Phenomenological Research 89 (2): 296–332.
———. 2015. “Against Public Reason.” Oxford Studies in Political Philosophy 1: 112–42.
Estlund, David. 2008. Democratic Authority: A Philosophical Framework. Princeton: Princeton University Press.
Fabre, Cécile. 1998. “Constitutionalising Social Rights.” Journal of Political Philosophy 6 (3): 263–84.
Feinberg, Joel. 1970. “The Nature and Value of Rights.” Journal of Value Inquiry 4 (4): 243–60.
Feldman, Fred. 2012. “True and Useful: On the Structure of a Two Level Normative Theory.” Utilitas 24 (2): 151–71.
Finkelstein, Claire Oakes. 2001. “Two Men and a Plank.” Legal Theory 7 (3): 279–306.
Finnis, John. 1979. Natural Law and Natural Rights. Oxford: Clarendon Press.
———. 1989. “Law as Co-Ordination.” Ratio Juris 2 (1): 97–104.
Fleischacker, Samuel. 2004. A Short History of Distributive Justice. Cambridge, Mass.: Harvard University Press.
Flikschuh, Katrin. 2008. “Reason, Right, and Revolution: Kant and Locke.” Philosophy & Public Affairs 36 (4): 375–404.
Franks, Benjamin. 2010. “Anarchism and the Virtues.” In Anarchism and Moral Philosophy, edited by Benjamin Franks and Matthew Wilson, 135–60. Basingstoke: Palgrave Macmillan.
Freeman, Samuel. 2001. “Illiberal Libertarians: Why Libertarianism Is Not a Liberal View.” Philosophy & Public Affairs 30 (2): 105–51.
Gardner, John. 2000. “The Virtue of Justice and the Character of Law.” Current Legal Problems 53 (1): 1–30.
———. 2012. “How Law Claims, What Law Claims.” In Institutionalized Reason The Jurisprudence of Robert Alexy, edited by Matthias Klatt. Oxford: Oxford University Press.
Garthoff, Jon. 2010. “Legitimacy Is Not Authority.” Law and Philosophy 29 (6): 669–94.
Gaus, Gerald. 1994. “Property, Rights, and Freedom.” Social Philosophy and Policy 11 (2): 209–40.
———. 1996. Justificatory Liberalism: An Essay on Epistemology and Political Theory. Oxford: Oxford University Press.
———. 2010. “Coercion, Ownership, and the Redistributive State: Justificatory Liberalism’s Classical Tilt.” Social Philosophy and Policy 27 (1): 233–75.
Goodin, Robert. 2007. “Why Social Justice Is Not All That Matters: Justice as the First Virtue.” Ethics 117 (3): 413–32.
Greenawalt, Kent. 1987. Conflicts of Law and Morality. Oxford: Oxford University Press.
Greenberg, Mark. 2014. “The Moral Impact Theory of Law.” Yale Law Journal 123: 1288–1342.
Green, Leslie. 1983. “Law, Co-Ordination and the Common Good.” Oxford Journal of Legal Studies 3 (3): 299–324.
———. 1988. The Authority of the State. Oxford: Clarendon Press.
———. 2007. “The Duty to Govern.” Legal Theory 13 (3-4): 165–85.
Griffin, Christopher. 2003. “Democracy as a Non-Instrumentally Just Procedure.” Journal of Political Philosophy 11 (1): 111–21.
Griffin, James. 2008. On Human Rights. Oxford: Oxford University Press.
Guerrero, Alexander. 2012. “Democracy and Legitimacy.” Ph.D. thesis, New York University. http://www.alexguerrero.org/storage/Final_Dissertation.pdf.
Hare, Richard Mervyn. 1981. Moral Thinking: Its Levels, Method, and Point. Oxford: Clarendon Press.
Harris, Edward. 1991. “Fighting Philosophical Anarchism with Fairness: The Moral Claims of Law in the Liberal State.” Columbia Law Review 91 (4): 919–64.
Hershovitz, Scott. 2012. “The Authority of Law.” In The Routledge Companion to Philosophy of Law, edited by Andrei Marmor, 65–75. New York: Routledge.
Hodgson, Louis‐Philippe. 2010a. “Kant on Property Rights and the State.” Kantian Review 15 (1): 57–87.
———. 2010b. “Kant on the Right to Freedom: A Defense.” Ethics 120 (4): 791–819.
———. 2012. “Realizing External Freedom: The Kantian Argument for a World State.” In Kant’s Political Theory: Interpretations and Applications, edited by Elisabeth Ellis, 101–34. Pennsylvania State University Press.
Horton, John. 2010. “Realism, Liberal Moralism and a Political Theory of Modus Vivendi.” European Journal of Political Theory 9 (4): 431–48.
———. 2012. “Political Legitimacy, Justice and Consent.” Critical Review of International Social and Political Philosophy 15 (2): 129–48.
Huemer, Michael. 2012. The Problem of Political Authority: An Examination of the Right to Coerce and the Duty to Obey. New York: Palgrave Macmillan.
Hunt, Lester. 2015. Anarchy, State, and Utopia: An Advanced Guide. Oxford: Wiley.
Hurd, Heidi. 1996. “The Moral Magic of Consent.” Legal Theory 2 (2): 121–46.
———. 1999. Moral Combat. Cambridge: Cambridge University Press.
———. 2005. “Why You Should Be a Law-Abiding Anarchist (Except When You Shouldn’t).” San Diego Law Review 42: 75–84.
Hyams, Keith. 2004. “Nozick’s Real Argument for the Minimal State.” Journal of Political Philosophy 12 (3): 353–64.
Irwin, Jones. 2010. “A Well-Being Out of Nihilism: On the Affinities Between Nietzsche and Anarchist Thought.” In Anarchism and Moral Philosophy, edited by Benjamin Franks and Matthew Wilson, 208–25. Basingstoke: Palgrave Macmillan.
Kagan, Shelly. 1991. “Replies to My Critics.” Philosophy and Phenomenological Research 51 (4): 919–28.
———. 2000. “Evaluative Focal Points.” In Morality, Rules, and Consequences: A Critical Reader, edited by Elinor Mason, Dale Miller, and Brad Hooker, 134–55. Lanham, Md.: Rowan & Littlefield.
Kamm, Frances. 2007. Intricate Ethics. New York: Oxford University Press.
Kant, Immanuel. Kant’s Gesammelte Schriften (“Academy Edition”, AA). Königlich Preußische Akademie der Wissenschaften. Cited after volume and page. English translations from the Cambridge University Press edition.
Klosko, George. 2008. Political Obligations. Oxford: Oxford University Press.
Knowles, Dudley. 2010. Political Obligation: A Critical Introduction. London: Routledge.
———. 2012. “Good Samaritans and Good Government.” Proceedings of the Aristotelian Society 112: 161–78.
Kolodny, Niko. “Being Under the Power of Others.” Manuscript. Available at http://sophos.berkeley.edu/kolodny/BeingUnderThePowerOfOthers2.pdf.
Kramer, Matthew. 1998. “Rights without Trimmings.” In A Debate over Rights. Philosophical Enquiries, edited by Matthew Kramer, Nigel Simmonds, and Hillel Steiner, 7–112. Oxford: Clarendon Press.
Kramer, Matthew, and Hillel Steiner. 2007. “Theories of Rights: Is There a Third Way?” Oxford Journal of Legal Studies 27 (2): 281–310.
Krasnoff, Larry. 1998. “Consensus, Stability, and Normativity in Rawls’s Political Liberalism.” Journal of Philosophy 95 (6): 269–92.
Ladenson, Robert. 1980. “In Defense of a Hobbesian Conception of Law.” Philosophy & Public Affairs 9 (2): 134–59.
Landes, William, and Richard Posner. 1979. “Adjudication as a Private Good.” Journal of Legal Studies 8 (2): 235–84.
Lang, Gerald. 2004. “A Dilemma for Objective Act-Utilitarianism.” Politics, Philosophy & Economics 3 (2): 221–39.
Larmore, Charles. 1999. “The Moral Basis of Political Liberalism.” Journal of Philosophy 96 (12): 599–625.
———. 2013. “What Is Political Philosophy?” Journal of Moral Philosophy 10 (3): 276–306.
Lippert-Rasmussen, Kasper. 1996. “Moral Status and the Impermissibility of Minimizing Violations.” Philosophy & Public Affairs 25 (4): 333–51.
———. 2012. “Estlund on Epistocracy: A Critique.” Res Publica 18 (3): 241–58.
Lomasky, Loren. 1987. Persons, Rights, and the Moral Community. New York: Oxford University Press.
Long, Roderick, and Tibor Machan. 2008. Anarchism/Minarchism. Is a Government Part of a Free Country? Aldershot: Ashgate.
Lovett, Frank. 2004. “Can Justice Be Based on Consent?” Journal of Political Philosophy 12 (1): 79–101.
Mack, Eric. 1999. “In Defense of Individualism.” Ethical Theory and Moral Practice 2 (2): 87–115.
———. 2006. “Non-Absolute Rights and Libertarian Taxation.” Social Philosophy and Policy 23 (2): 109–41.
———. 2011. “Nozickian Arguments for the More-than-Minimal State.” In The Cambridge Companion to Nozick’s Anarchy, State, and Utopia, edited by Ralf Bader and John Meadowcroft, 89–115. Cambridge: Cambridge University Press.
Maloberti, Nicolas. 2009. “Libertarianism and the Possibility of the Legitimate State.” Libertarian Papers 1 (5): 1–12.
Manin, Bernard. 1987. “On Legitimacy and Political Deliberation.” Political Theory 15 (3): 338–68.
Marmor, Andrei. 2011. “An Institutional Conception of Authority.” Philosophy & Public Affairs 39 (3): 238–61.
McCabe, David. 2010. Modus Vivendi Liberalism: Theory and Practice. Cambridge: Cambridge University Press.
Miller, David. 2007. National Responsibility and Global Justice. Oxford: Oxford University Press.
———. 2012. “Grounding Human Rights.” Critical Review of International Social and Political Philosophy 15 (4): 407–27.
Morris, Christopher. 1998. An Essay on the Modern State. Cambridge: Cambridge University Press.
———. 2005. “Natural Rights and Political Legitimacy.” Social Philosophy and Policy 22 (1): 314–29.
———. 2008. “State Legitimacy and Social Order.” In Political Legitimization without Morality?, edited by Jörg Kühnelt, 15–32. Dordrecht, London: Springer.
———. 2012. “State Coercion and Force.” Social Philosophy and Policy 29 (1): 28–49.
Murphy, Liam, and Thomas Nagel. 2002. The Myth of Ownership: Taxes and Justice. Oxford: Oxford University Press.
Murphy, Mark. 2006. “The Natural Law Rejection of Consent Theory.” In Natural Law in Jurisprudence and Politics, 91–111. Cambridge: Cambridge University Press.
Nagel, Thomas. 1975. “Libertarianism without Foundations.” Yale Law Journal 85 (1): 136–49.
———. 1979. Mortal Questions. Cambridge: Cambridge University Press.
———. 1991. Equality and Partiality. New York: Oxford University Press.
———. 1995. “Personal Rights and Public Space.” Philosophy & Public Affairs 24 (2): 83–107.
Nickel, James. 2005. “Poverty and Rights.” Philosophical Quarterly 55 (220): 385–402.
Nozick, Robert. 1974. Anarchy, State, and Utopia. New York: Basic Books.
Oakeshott, Michael. 1991. Rationalism in Politics and Other Essays. Edited by Timothy Fuller. 2nd Revised edition. Indianapolis: Liberty Fund.
Oberdiek, John. 2008. “Specifying Rights Out of Necessity.” Oxford Journal of Legal Studies 28 (1): 127–46.
O’Neill, Onora. 2005. “The Dark Side of Human Rights.” International Affairs 81 (2): 427–39.
Ord, Toby. 2005. “Consequentialism and Decision Procedures.” B.Phil. thesis, University of Oxford. Available at http://www.amirrorclear.net/files/consequentialism-and-decision-procedures.pdf.
———. 2009. “Beyond Action.” D.Phil. thesis, University of Oxford. Available at http://www.amirrorclear.net/academic/papers/beyond-action.pdf.
Otsuka, Michael. 2003. Libertarianism without Inequality. Oxford: Oxford University Press.
Owens, David. 2011. “The Possibility of Consent.” Ratio 24 (4): 402–21.
Parekh, Bhikhu. 1993. “A Misconceived Discourse on Political Obligation.” Political Studies 41 (2): 236–51.
Parfit, Derek. 1984. Reasons and Persons. Oxford: Clarendon Press.
Perelman, Chaim. 1963. The Idea of Justice and the Problem of Argument. New York: Humanities Press.
Perry, Stephen. 2005. “Law and Obligation.” American Journal of Jurisprudence 50: 263–95.
———. 2012. “Political Authority and Political Obligation.” Oxford Studies in Philosophy of Law, edited by Leslie Green and Brian Leiter, vol. 2, 1-74.
Peter, Fabienne. 2009. Democratic Legitimacy. New York: Routledge.
———. 2016. “Political Legitimacy.” Stanford Encyclopedia of Philosophy (Summer 2016 Edition). http://plato.stanford.edu/archives/sum2016/entries/legitimacy/.
Pettit, Philip. 1987. “Rights, Constraints and Trumps.” Analysis 47 (1): 8–14.
———. 1989. “Consequentialism and Respect for Persons.” Ethics 100 (1): 116–26.
———. 1999. Republicanism: A Theory of Freedom and Government. Oxford: Oxford University Press.
———. 2012. On the People’s Terms: A Republican Theory and Model of Democracy. Cambridge: Cambridge University Press.
Pettit, Philip, and Geoffrey Brennan. 1986. “Restrictive Consequentialism.” Australasian Journal of Philosophy 64 (4): 438–55.
Pitkin, Hanna. 1965. “Obligation and Consent--I.” American Political Science Review 59 (4): 990–99.
———. 1966. “Obligation and Consent--II.” American Political Science Review 60 (1): 39–52.
Postema, Gerald. 1980. “Nozick on Liberty, Compensation, and the Individual’s Right to Punish.” Social Theory and Practice 6 (3): 311–37.
———. 1982. “Coordination and Convention at the Foundations of Law.” Journal of Legal Studies 11 (1): 165–203.
Powell, Benjamin, and Edward Stringham. 2009. “Public Choice and the Economic Analysis of Anarchy: A Survey.” Public Choice 140 (3-4): 503–38.
Quong, Jonathan. 2011. Liberalism without Perfection. Oxford: Oxford University Press.
———. 2013. “On the Idea of Public Reason.” In Companion to Rawls, edited by Jon Mandle and David Reidy, 265–80. Oxford: Wiley.
Railton, Peter. 1984. “Alienation, Consequentialism, and the Demands of Morality.” Philosophy & Public Affairs 13 (2): 134–71.
Rainbolt, George. 2006. The Concept of Rights. Dordrecht: Springer.
Raphael, D. D. 2001. Concepts of Justice. Oxford: Clarendon Press.
Rawls, John. 1955. “Two Concepts of Rules.” Philosophical Review 64 (1): 3–32.
———. 1993. Political Liberalism. New York: Columbia University Press.
———. 1999. A Theory of Justice. Revised Edition. Harvard: Belknap Press of Harvard University Press.
Raz, Joseph. 1984a. “On the Nature of Rights.” Mind 93: 194–214.
———. 1984b. “Right-Based Moralities.” In Theories of Rights, edited by Jeremy Waldron, 182–200. Oxford: Oxford University Press.
———. 1984c. “The Obligation to Obey: Revision and Tradition.” Notre Dame Journal of Law, Ethics & Public Policy 1: 139–55.
———. 1988. The Morality of Freedom. New Ed. Clarendon Press.
Renzo, Massimo. 2011. “State Legitimacy and Self-Defence.” Law and Philosophy 30 (5): 575–601.
Riley, Patrick. 1973. “How Coherent Is the Social Contract Tradition?” Journal of the History of Ideas 34 (4): 543–62.
———. 1982. Will and Political Legitimacy: A Critical Exposition of Social Contract Theory in Hobbes, Locke, Rousseau, Kant, and Hegel. Cambridge, Mass.: Harvard University Press.
Ripstein, Arthur. 2004. “Authority and Coercion.” Philosophy & Public Affairs 32 (1): 2–35.
———. 2009. Force and Freedom: Kant’s Legal and Political Philosophy. Harvard, Mass.: Harvard University Press.
Roberson, Christopher. 1998. “The State as Rational Authority: An Anarchist Justification of Government.” Oxford Journal of Legal Studies 18 (4): 617–30.
Rønnow-Rasmussen, Toni. 2015. “Intrinsic and Extrinsic Value.” In The Oxford Handbook of Value Theory, edited by Iwao Hirose and Jonas Olson, 29–43. Oxford: Oxford University Press.
Rossi, Enzo, and Matt Sleat. 2014. “Realism in Normative Political Theory.” Philosophy Compass 9 (10): 689–701.
Sandel, Michael. 1998. Liberalism and the Limits of Justice. Second Edition. Cambridge: Cambridge University Press.
Sanders, John, and Jan Narveson, eds. 1996. For and against the State: New Philosophical Readings. Lanham, Md.: Rowman & Littlefield.
Sangiovanni, Andrea. 2012. “Can the Innate Right to Freedom Alone Ground a System of Public and Private Rights?” European Journal of Philosophy 20 (3): 460–69.
Sartorius, Rolf. 1981. “Political Authority and Political Obligation.” Virginia Law Review 67 (1): 3–17.
Scanlon, Thomas. 2003. “Rights, Goals, and Fairness.” In The Difficulty of Tolerance, 26–41. Cambridge: Cambridge University Press.
———. 2008. Moral Dimensions: Permissibility, Meaning, Blame. Cambridge: Belknap Press of Harvard University Press.
Scheffler, Samuel. 1976. “Natural Rights, Equality, and the Minimal State.” Canadian Journal of Philosophy 6 (1): 59–76.
Schmelzle, Cord. 2012. “Zum Begriff politischer Legitimität.” In Der Aufstieg der Legitimitätspolitik: Rechtfertigung und Kritik politisch-ökonomischer Ordnungen, edited by Anna Geis, Frank Nullmeier, and Christopher Daase, 419–35. Baden-Baden: Nomos.
Schmidtz, David. 1991. The Limits of Government: An Essay on the Public Goods Argument. Boulder, Colo.: Westview Press.
———. 2006. Elements of Justice. Cambridge: Cambridge University Press.
———. 2007. “When Justice Matters.” Ethics 117 (3): 433–59.
Schneewind, J. B. 1996. “Voluntarism and the Foundations of Ethics.” Proceedings and Addresses of the American Philosophical Association 70 (2): 25–41.
Senor, Thomas. 1987. “What If There Are No Political Obligations? A Reply to A. J. Simmons.” Philosophy & Public Affairs 16 (3): 260–68.
Shafer-Landau, Russ. 1995. “Specifying Absolute Rights.” Arizona Law Review 37: 209–25.
Shklar, Judith. 2004. “Liberalism of Fear.” In Political Liberalism: Variations on a Theme, edited by Shaun Young, 149–66. Albany: State University of New York Press.
Shue, Henry. 1988. “Mediating Duties.” Ethics 98 (4): 687–704.
Simmons, John. 1976. “Tacit Consent and Political Obligation.” Philosophy & Public Affairs 5 (3): 274–91.
———. 1979. Moral Principles and Political Obligations. Princeton: Princeton University Press.
———. 1987. “The Anarchist Position: A Reply to Klosko and Senor.” Philosophy & Public Affairs 16 (3): 269–79.
———. 1996. “Philosophical Anarchism.” In For and against the State: New Philosophical Readings, edited by John Sanders and Jan Narveson, 19–39. Lanham, Md.: Rowman & Littlefield.
———. 1999. “Justification and Legitimacy.” Ethics 109 (4): 739–71.
———. 2005a. “Consent Theory for Libertarians.” Social Philosophy and Policy 22 (1): 330–56.
———. 2005b. “The Duty to Obey and Our Natural Moral Duties.” In Is There a Duty to Obey the Law?, by Christopher Wellman and John Simmons, 93–196. Cambridge: Cambridge University Press.
Sinclair, Tom. 2015. “The Power of Public Positions: Official Roles in Kantian Legitimacy.” Presented at the Nuffield Political Theory Workshop, November 16.
Singer, Marcus. 1977. “Justice, Theory, and a Theory of Justice.” Philosophy of Science 44 (4): 594–618.
Smith, Matthew Noah. 2011. “Political Obligation and the Self.” Philosophy and Phenomenological Research 86 (2): 347–75.
Smith, M. B. E. 1973. “Is There a Prima Facie Obligation to Obey the Law?” Yale Law Journal 82 (5): 950–76.
———. 2005. “The Duty to Obey the Law.” In The Blackwell Guide to the Philosophy of Law and Legal Theory, edited by Martin Golding and William Edmundson, 457–66. Malden, MA: Blackwell.
Sobel, David. 2012. “Backing Away from Libertarian Self-Ownership.” Ethics 123 (1): 32–60.
Stark, Cynthia. 1997. “Decision Procedures, Standards of Rightness and Impartiality.” Noûs 31 (4): 478–95.
Stilz, Anna. 2009. Liberal Loyalty: Freedom, Obligation, and the State. Princeton: Princeton University Press.
———. 2011. “Nations, States, and Territory.” Ethics 121 (3): 572–601.
Stocker, Michael. 1976. “The Schizophrenia of Modern Ethical Theories.” Journal of Philosophy 73 (14): 453–66.
Tasioulas, John. 2007. “The Moral Reality of Human Rights.” In Freedom from Poverty as a Human Right: Who Owes What to the Very Poor?, edited by Thomas Pogge, 75–101. Oxford: Oxford University Press.
———. 2010. “Taking Rights out of Human Rights.” Ethics 120 (4): 647–78.
———. 2013a. “Human Dignity and the Foundations of Human Rights.” In Understanding Human Dignity, edited by Christopher McCrudden, 291–312. British Academy.
———. 2013b. “Human Rights, Legitimacy, and International Law.” American Journal of Jurisprudence 58 (1): 1–25.
———. 2013c. “Justice, Equality, and Rights.” In The Oxford Handbook of the History of Ethics, edited by Roger Crisp, 768–93. Oxford: Oxford University Press.
———. 2015. “On the Foundations of Human Rights.” In Philosophical Foundations of Human Rights, edited by Rowan Cruft, Matthew Liao, and Massimo Renzo, 45–70. Oxford: Oxford University Press.
Taylor, Charles. 1979. “What’s Wrong with Negative Liberty?” In The Idea of Freedom: Essays in Honour of Isaiah Berlin, edited by Alan Ryan, 211–29. Oxford: Oxford University Press.
Thompson, Michael. 2004. “What Is It to Wrong Someone? A Puzzle About Justice.” In Reason and Value: Themes from the Moral Philosophy of Joseph Raz, edited by Jay Wallace et al., 333–84. Oxford: Oxford University Press.
Thomson, Judith Jarvis. 1980. “Rights and Compensation.” Noûs 14 (1): 3–15.
———. 1990. The Realm of Rights. Harvard: Harvard University Press.
Timmons, Mark. 1998. “Decision Procedures, Moral Criteria, and the Problem of Relevant Descriptions in Kant’s Ethics.” In Jahrbuch für Recht und Ethik, 5: 389–417. Berlin: Duncker & Humblot.
Valentini, Laura. 2012a. “Assessing the Global Order: Justice, Legitimacy, or Political Justice?” Critical Review of International Social and Political Philosophy 15 (5): 593–612.
———. 2012b. “Kant, Ripstein and the Circle of Freedom: A Critical Note.” European Journal of Philosophy 20 (3): 450–59.
Vallentyne, Peter. 2007. “Libertarianism and the State.” Social Philosophy and Policy 24 (1): 187–205.
———. 2014. “Justice, Interpersonal Morality, and Luck Egalitarianism.” In Distributive Justice and Access to Advantage: G. A. Cohen’s Egalitarianism, edited by Alexander Kaufman, 40–49. Cambridge: Cambridge University Press.
van der Vossen, Bas. 2015. “Imposing Duties and Original Appropriation.” Journal of Political Philosophy 23 (1): 64–85.
Varden, Helga. 2009. “Nozick’s Reply to the Anarchist.” Law and Philosophy 28 (6): 585–616.
———. 2010. “Kant’s Non-Absolutist Conception of Political Legitimacy – How Public Right ‘Concludes’ Private Right in the ‘Doctrine of Right.’” Kant-Studien 101 (3): 331–51.
———. 2011. “Kant’s Non-Voluntarist Conception of Political Obligations: Why Justice Is Impossible in the State of Nature.” Kantian Review 13 (2): 1–45.
Waldron, Jeremy. 1987a. “Nonsense upon Stilts”: Bentham, Burke, and Marx on the Rights of Man. London: Methuen.
———. 1987b. “Theoretical Foundations of Liberalism.” Philosophical Quarterly 37 (147): 127–50.
———. 1988. The Right to Private Property. Oxford: Clarendon Press.
———. 1989. “Rights in Conflict.” Ethics 99 (3): 503–19.
———. 1993. Liberal Rights: Collected Papers, 1981-1991. Cambridge: Cambridge University Press.
———. 1996. “Kant’s Legal Positivism.” Harvard Law Review 109 (7): 1535–66.
———. 1999. “Lex Satis Iusta.” Notre Dame Law Review 75 (5): 1829–58.
———. 2003. “The Primacy of Justice.” Legal Theory 9 (4): 269–94.
———. 2009. “Dignity and Defamation: The Visibility of Hate.” Harvard Law Review 123 (7): 1596–1657.
Wall, Steven. 2002. “Is Public Justification Self-Defeating?” American Philosophical Quarterly 39 (4): 385–94.
———. 2009. “Perfectionism in Politics: A Defense.” In Contemporary Debates in Political Philosophy, edited by Thomas Christiano and John Philip Christman, 99–117. Malden, Mass.: Wiley-Blackwell.
Weinrib, Jacob. 2014. “Authority, Justice, and Public Law: A Unified Theory.” University of Toronto Law Journal 64 (5): 703–35.
Wellman, Christopher. 1996. “Liberalism, Samaritanism, and Political Legitimacy.” Philosophy & Public Affairs 25 (3): 211–37.
———. 2001. “Toward a Liberal Theory of Political Obligation.” Ethics 111 (4): 735–59.
———. 2005. “Samaritanism and the Duty to Obey the Law.” In Is There a Duty to Obey the Law?, by Christopher Wellman and John Simmons, 3–89. Cambridge: Cambridge University Press.
———. 2009. “Rights and State Punishment.” Journal of Philosophy 106 (8): 419–39.
Wenar, Leif. 2005a. “The Nature of Rights.” Philosophy & Public Affairs 33 (3): 223–52.
———. 2005b. “The Value of Rights.” In Law and Social Justice, edited by Joseph Campbell, Michael O’Rourke, and David Shier, 179–211. Cambridge, Mass.: MIT Press.
Wendt, Fabian. 2013. “Peace beyond Compromise.” Critical Review of International Social and Political Philosophy 16 (4): 573–93.
———. 2015a. “Justice and Political Authority in Left-Libertarianism.” Politics, Philosophy & Economics 14 (3): 316–39.
———. 2015b. “Political Authority and the Minimal State.” Social Theory and Practice 42 (1): 97–122.
Wiland, Eric. 2007. “How Indirect Can Indirect Utilitarianism Be?” Philosophy and Phenomenological Research 74 (2): 275–301.
Williams, Bernard. 1973. “A Critique of Utilitarianism.” In Utilitarianism: For and against, edited by J. J. C. Smart and Bernard Williams, 77–150. Cambridge: Cambridge University Press.
———. 2005. “Realism and Moralism in Political Theory.” In In the Beginning Was the Deed: Realism and Moralism in Political Argument, edited by Geoffrey Hawthorn, 1–17. Princeton: Princeton University Press.
Windeknecht, Ryan Gabriel. 2011. “Law Without Legitimacy or Justification? The Flawed Foundations of Philosophical Anarchism.” Res Publica 18 (2): 173–88.
Wolff, Robert Paul. 1970. In Defense of Anarchism. New York: Harper & Row.
Young, Iris Marion. 1990. Justice and the Politics of Difference. Princeton, N.J.: Princeton University Press.
Zhu, Jiafeng. 2012. “Legitimacy as a Mere Moral Power? A Response to Applbaum.” Diametros 33: 120–37.