1. A Fact-Based Conception of Legitimacy?

It cannot be overlooked that fact-based views are currently increasingly popular in meta-ethics and epistemology. For ease of reference, I call those views factualist. In relation to practical normativity, in general, factualism holds that that our actions are justified to the extent that they are warranted by normative facts. On the standard interpretation of this view, the relevant normative facts relate to normative practical reasons (Parfit Reference Parfit2011; Scanlon Reference Scanlon2014). A factualist view of epistemic normativity holds that our beliefs are justified only if they are warranted by the facts (Williamson Reference Williamson2000).

A fact-based conception of political legitimacy, as mentioned above, takes political decisions to be legitimate in virtue of their being warranted by normative facts. It is interesting to note that not many political philosophers have advocated factualism about political legitimacy. David Enoch (Reference Enoch2014) probably comes closest to advocating a fact-based conception of legitimate authority, but he does not focus on political legitimacy specifically.

Enoch offers an analysis of legitimate authority that builds on factualism about practical normativity in general. Specifically, he is committed to the view that normative facts, understood as reasons-facts, determine what we should do. Against the background of this commitment, the question he addresses is the following: How can an authority's decision, or ‘say so’, ever make a valid claim about how someone should act? Enoch's answer is that an authority's command is legitimate if it triggers a normative reason for the addressee to act as commanded. A reason is triggered if a change in the nonnormative circumstances makes it the case that a normative reason applies. For example, the command ‘drive on the left’, uttered in an attempt to solve a coordination problem, might trigger a reason to drive on the left. Importantly, for an authority's command to be legitimate, it is not necessary that the normative reason applied all along. One might not have a reason to drive on the left independently of that particular command. Conversely, if an authority's command fails to trigger a normative reason for the addressees to act as commanded, the command is not legitimate. Authorities are constrained in what they can legitimately command.

Translated into the language I am using here, Enoch's analysis supports the key factualist claim that political decisions are legitimate in virtue of being warranted by the normative facts. Enoch explicates the warrant relationship as a triggering relationship: political decisions are legitimate to the extent that they trigger in citizens normative reasons to obey and they have the property of legitimacy in virtue of the reasons-facts that are being triggered by those decisions.

Can the factualist claim sustain a plausible conception of political legitimacy? My answer to this question is negative. The problem with fact-based conceptions of political legitimacy is that they run into an accessibility problem. I call this the accessibility objection to factualism about political legitimacy.

We can begin to see the problem with fact-based conceptions if we bear in mind that factualism about political legitimacy implies that it might not be possible to access the property of political legitimacy in judgment. A particular political decision might be legitimate, on these views, but nobody is in a position to judge that it is or would be. In the context of Enoch's proposal, for example, it is possible that a political decision is legitimate because it has triggered a normative reason for the citizens to obey, but neither the decision-making body nor the citizens are in a position to judge that this is the case.

By way of illustration, consider the Brexit decision-making progress that has been plaguing the United Kingdom for quite some time. Imagine a scenario in which Parliament haphazardly made a series of decisions that concluded in the United Kingdom's revoking Article 50 and committing the country to staying in the European Union. Suppose this was met with widespread protest from citizens who in 2016 voted to leave. But suppose also that this decision would have been the one warranted by the normative facts. The decision would then have been legitimate, on a fact-based conception, even though nobody was in a position at the time to authoritatively judge that it was, in fact, the right decision and even though there was widespread uncertainty and confusion about the legitimacy of this decision in the government and Parliament and among the citizens. Something is clearly off here—accounting for the ground of legitimacy in this way seems to undermine the role that legitimacy plays in assessing political decisions.

Why is the potential lack of access to normative facts a problem for a conception of political legitimacy? Answering this question is less straightforward than it might seem at first. To see this, note that the potential inaccessibility of normative facts is not necessarily seen as a problem for normative properties in other contexts. Consider the property of moral wrongness, for example. Many hold the view that whether an action is morally wrong in a particular context only depends on normative facts. On those views, it makes sense to claim, say, that it is morally wrong for me to blame you for something you did even if I have every reason to believe that what you did is blameworthy. Moral ignorance, on those views, is morally irrelevant (Harman Reference Harman and Schafer-Landau2015).

If lack of access is not necessarily a problem for practical normativity, why is it a problem for political legitimacy? My answer to this question is that legitimacy is a normative property that primarily functions to settle judgments. Not all normative properties function in this way, and lack of access is then not necessarily a problem. This difference in function explains why only conceptions of political legitimacy that secure access in some form are plausible. Conceptions of political legitimacy on which it might not be possible to settle whether a decision is legitimate, by contrast, are not plausible.

To understand this claim better, it helps to reflect a bit on the nature of normative reasoning. We engage in normative reasoning in both practical and theoretical contexts. And it is important to recognize that normative reasoning has two functions or modes. First, there is what I call the orientation function. In a theoretical context, the orientation function is geared toward what truly is the case—what the world is like. In this context, we might say that truth—a normative property—is the aim of belief. In a practical context, the orientation function is geared toward the normative facts that bear on what we should do.

The orientation function is important, but so is a second function, which I call the settling function. While the orientation function is world-regarding, so to speak, the settling function is agent-regarding. The settling function is geared toward what we should believe and how we should act.

Our normative reasoning tends to move back and forth between these two functions. When trying to answer the question of what we should believe, for example, we are likely to ask ourselves what truly is the case. Vice versa, when asking ourselves what the world is like, we are likely to check what we should believe about what the world is like. In the practical context, we are switching back and forth between considering the normative facts and trying to establish what we should do. The two functions are thus easily confused, but I hope that it is clear to the reader that they are distinct.

Normative properties play different roles in the two functions or modes of normative reasoning, in the following sense. Some normative properties are primarily orientational. Truth, for example, is a normative property that is primarily orientational and so are practical normative properties such as the property of being morally wrong. Other normative properties are not primarily orientational: they play a role in the settling function or mode. In the theoretical context, evidence is not primarily orientational: evidence is what we invoke when trying to settle our beliefs. In the practical context, the properties of being legal or illegal, for example, are not primarily orientational either. They function to settle what we should do. The property of political legitimacy, I contend, is of this kind, too. If the potential legitimacy of political decisions could merely orient political decision making, something would be missing. The property of political legitimacy must be such that it can settle political decision making.

Now, the orientation function of normative reasoning is not undermined by lack of access. It makes sense, for example, to take truth as the aim of belief even if truth remains often inaccessible. Similarly, as just explained above, it can make sense to maintain that an action is morally wrong, even if the agent is not in a position to know that. Normative properties that remain inaccessible do, however, undermine the settling function of normative reasoning. Only properties that we can access can be invoked to establish what we should do or believe.

Against this background, I offer the accessibility objection to fact-based conceptions of political legitimacy. The accessibility objection to fact-based conceptions is that normative facts are the wrong ground of political legitimacy because the property of political legitimacy must be such that it can settle political deliberation (at least temporarily) and, in order to fulfill the settling function, the grounds of political legitimacy must involve the attitudes of at least some citizens. Normative facts, which might remain inaccessible, cannot settle political decision making. If normative facts can play a role at all, they do so through our beliefs about them.

The accessibility objection to factualism about political legitimacy implies that a plausible conception of political legitimacy must rest on grounds that satisfy an access constraint. To say that some access condition has to be satisfied is not to say that the property of legitimacy depends on everybody's judgment about which decision should be made or some similar claim. On some conceptions of political legitimacy, this is how access is interpreted. But this is not the only possible interpretation. The access constraint only implies that at least some citizens, minimally one, must have access to the grounds of legitimacy.

While the accessibility objection rules out factualism, it rules in both belief-based and will-based conceptions of political legitimacy. Will-based and belief-based conceptions of political legitimacy both satisfy the access constraint, but they satisfy it in different ways. In the next two sections, I discuss each of these two families of conceptions of political legitimacy, starting with will-based conceptions.

2. Will-Based Conceptions of Political Legitimacy

As mentioned, on will-based conceptions, political decisions are legitimate in virtue of how they adjudicate between the potentially conflicting wills of the citizens. The will is a practical faculty, involved in deciding what to do.

A paradigmatic example of a will-based conception is Rawls's well-known liberal principle of legitimacy: ‘our exercise of political power is fully proper only when it is exercised in accordance with a constitution the essentials of which all citizens as free and equal may reasonably be expected to endorse in the light of principles and ideals acceptable to their common human reason’ (Reference Rawls1993: 137).

The way in which Rawls specifies the liberal principle of legitimacy contains quite a few technical terms (the explanation of which are beyond the scope of this essay), but the key feature of the principle (in the context of this essay) is the claim that political decisions are legitimate if they are made in accordance with a constitution that is supported by some form of reasoned agreement among the citizens.

The liberal principle of legitimacy is at the core of what has come to be known as a public-reason conception of political legitimacy. Public-reason conceptions treat the property of legitimacy as dependent on the procedural or substantive reasons that are accessible to the citizens. Such public reasons are thought to justify political decisions and this justification is thought to constitute the property of legitimacy.

Besides public-reason conceptions, there are other will-based conceptions of political legitimacy—some of them, like the Rawlsian conception, foregrounding procedural considerations and others foregrounding substantive considerations. Unanimity conceptions take political decisions to be legitimate if they have been made unanimously (or at least near-unanimously) or if they attract unanimous (or near-unanimous) actual or hypothetical consent. The difference between unanimity conceptions and public-reason conceptions is that the latter focus on the reasons that justify political decisions (or the ways in which they are made), whereas the former focus on the citizens’ endorsement of a particular decision. A paradigm example of a unanimity conception of legitimacy can be found in Rousseau's work, but it is popular in contemporary political philosophy as well (Green Reference Green1988; Simmons Reference Simmons2005; Greene Reference Greene, Sobel, Vallentyne and Wall2016). On unanimity conceptions, the property of legitimacy is constituted by the—actual or hypothetical—endorsement of each citizen.

Another important example of will-based conceptions are participation conceptions. According to them, a political decision is legitimate if it has been made in a process that allows for the equal participation of all citizens. To paraphrase Bernard Manin (Reference Manin1987: 352), participation conceptions take political legitimacy to be dependent on the participation of all, that is, on the ability of all to express their will. Good examples are conceptions of political legitimacy that center on democracy, such as deliberative democracy (Manin Reference Manin1987; Cohen Reference Cohen, Bohman and Rehg1997a, Reference Cohen, Bohman and Rehg1997b), egalitarian democracy (Buchanan Reference Buchanan2002; Christiano Reference Christiano2008; Kolodny Reference Kolodny2014a, Reference Kolodny2014b; Viehoff Reference Viehoff2014), and Philipp Pettit's (Reference Pettit2012) equal control view of political legitimacy.

Will-based conceptions of political legitimacy have been defended with a variety of arguments. Most important among them are egalitarian arguments (Christiano Reference Christiano2008; Kolodny Reference Kolodny2014a, Reference Kolodny2014b; Viehoff Reference Viehoff2014), freedom-centered arguments (Gaus Reference Gaus2011; Pettit Reference Pettit2012), and second-personal arguments (Darwall Reference Darwall2010). What these arguments have in common is that they all attach normative significance to the first-personal standpoint. The idea is helpfully captured in Rawls's (Reference Rawls1980: 543) characterization of persons as ‘self-originating sources of valid claims’. Rawls introduced this conception of the person in his essay on Kantian constructivism, and he remained committed to it in later work (Rawls Reference Rawls1993: 32).

This conception of the person has important implications for what makes practical claims valid. Generally speaking, a valid practical claim is a claim about what should be done that carries normative significance. In the context of political decision making, valid claims affect the legitimacy of political decisions, but invalid claims do not. By taking persons to be ‘self-originating sources’ of valid claims, what is denied, in the first instance, is that the validity of a person's claims about what should be done depends on how well they track some normative facts that obtain independently of the person's will-formation processes. Instead, the idea is that a person's capacity to make valid claims about what should be done depends on the exercise of their practical judgment. The claims of persons who exercise their practical judgment well have the potential to make a normative difference and, in the political context, have an impact on which political decisions have the property of legitimacy.

A key premise of will-based conceptions is that the will of potentially all citizens matters for political legitimacy. However, those who accept that persons have the capacity to be a self-originating source of valid claims need not accept that all claims that persons might make are equally valid; some claims may not be valid at all. A person may yield to impulse, and this will undermine her capacity to make a valid claim. To account for this, conceptions that normatively privilege the first-personal standpoint typically understand the capacity to make valid claims as limited to processes of will-formation that qualify as either rational or reasonable, or both. A rational will is one that is suitably consistent. A reasonable will is one that is suitably responsive to other persons as self-originating sources of valid claims. What is important, however, is that neither type of restriction is to be understood in relation to a third-personal standard for the validity of their claims. Instead, rationality and reasonableness are understood as restrictions on the capacity to make valid claims that are internal to a person's will formation.

The problem of political legitimacy that arises if one privileges the normative significance of the first-personal standpoint in this way is the following. If persons are self-originating sources of valid claims about what should be done, we should expect that many of their valid claims about what should be done will be in conflict. This will create a problem if decisions have to be made by or on behalf of people with conflicting wills. Arbitrarily privileging the valid claims of some or restricting those of others will not generate legitimate decisions. So, given this potential for conflicting valid claims about what should be done, how should political decisions be made and which decisions have the normative property of legitimacy? The solution to this problem that will-based conceptions of political legitimacy offer is that legitimate political decisions adjudicate between the wills of the citizens in the right way—by being justified by public reasons, by being based on their unanimous or near-unanimous consent, or by reflecting the outcome of an inclusive decision-making process.

My general objection to will-based conceptions is that their starting premise—that persons are self-originating sources of valid claims—is mistaken. To see why, consider an analogy with the theoretical case first. When it comes to claims about what should be believed, many will hold that the validity of those claims typically depends on third-personal sources—on the substantive truth of the claim, on evidence, on the reliability of the belief formation process, or on similar considerations. Certain types of subjective beliefs, such as the belief that you are in pain, may be an exception. But for most types of beliefs and on most views of epistemic justification, valid claims about what should be believed are not self-originating.

The premise that persons are self-originating sources of valid claims is also mistaken in the practical domain, at least for a broad range of claims. For some practical claims, it must be right to say that persons are self-originating sources of validity. The claim ‘I don't like eating frozen things’ is a good example. But not all practical claims are like this, not even in the political domain. In at least some political contexts, there are correct and incorrect judgments about what should be done and, in the right circumstances, it is possible to develop a sufficiently good grip on what they are. Consider the claim ‘my child should not be vaccinated’ in relation to a highly contagious infectious disease and a harmless vaccine. Or consider the recent decision to enforce a US immigration policy that separates small children from their families. Whatever else we should believe about immigration policies, I think it is sufficiently clear that this particular decision should not have been made. These examples thus illustrate that the validity of at least some practical claims depends on what we should believe about the nonnormative and the normative facts that pertain on what we should do.

Based on this, we can now state the arbitrariness objection to will-based conceptions. The objection, which parallels the well-known arbitrariness objection against constructivist theories of practical normativity (Shafer-Landau Reference Shafer-Landau2003), states that will-based conceptions of political legitimacy support undue arbitrariness in political decision making. Such conceptions wrongly ignore how the validity of some claims about what should be done, even in the political domain, depends on what we should believe about actions warranted by the normative facts in particular contexts. If sufficiently robustly justified beliefs about what should be done are disregarded in political decision making, the resulting decisions will lack legitimacy.

Now, it might be replied that the illegitimacy of such decisions can be accounted for within a will-based conception of legitimacy: what makes such decisions illegitimate is that they fail to treat at least some people as self-originating sources of valid claims, or that they are not the kind of decisions that rational or reasonable persons would support. Does this reply undermine my objection? I do not think so. To show this, I distinguish between two main versions of this reply: the first reply denies that there are third-personal sources for the validity of any claims about what should be done; the second reply allows that the validity of some claims—for example claims about what is owed to persons as self-originating sources of valid claims—has a third-personal source.

The best version of the first reply is that reasonable wills recognize that a decision that encourages or tolerates slavery or genocide, for example, is incompatible with treating all persons as self-originating sources of valid claims (see Rawls Reference Rawls1993). That is so because reasonable wills accept the constraints that the capacity of other persons to be self-originating sources of valid claims imposes on their wills. The problem with this sort of reply, however, is that it seems to be the wrong sort of reply: it might be able to generate the right answer—that such decisions are illegitimate—but for the wrong reasons. What makes such decisions illegitimate is not that reasonable wills will not will it, it is that they go against the correct judgment that those decisions are not what should be done. And we cannot rule out that in other cases, the reply would give the wrong answer: what reasonable wills will need not be what they should will.

The second version of the reply recognizes this problem and argues instead that the normative significance of the first-personal standpoint is grounded in normative facts. Charles Larmore's egalitarian argument in support of a public reason conception is a good example for this sort of reply. Larmore (Reference Larmore2008) maintains that there is a moral principle of equal respect that implies that only decisions that are justified on the basis of reasons that the citizens can share are legitimate. The moral principle of equal respect that grounds his (public reason) conception of political legitimacy is not itself justified on the basis of some form of adjudication between wills, however. Its validity is derived from normative facts about personhood.

My objection to this version of the reply is that even if it were true that there is a moral principle of equal respect (that we know applies), there is no reason to think that this is the only principle supported by normative facts. Instead, we should allow for the possibility that the validity of a whole range of claims about what should be done depends on what we should believe about the normative facts that apply and that legitimate political decision-making should take into account.

My general point is that will-based conceptions of political legitimacy wrongly ignore how the validity of some claims about what should be done, even in the political domain, is not independent of the justification for the beliefs on which they are based. The validity of at least some practical claims depends on third-personal sources of validity. In those circumstances, the problem of political legitimacy is not appropriately described as a problem of how to decide given conflicting wills. And adjudicating between potentially conflicting wills is not the right type of solution to the problem of legitimacy.

3. Belief-based Conceptions of Political Legitimacy

Belief-based conceptions of political legitimacy build on precisely this insight and take the ground of political legitimacy to be sufficiently justified beliefs about what should, in fact, be done. Belief-based and fact-based conceptions of political legitimacy are both premised on there being normative facts that warrant certain decisions. The difference between them is that belief-based conceptions take the ground of legitimacy to be beliefs about the decisions warranted by the normative facts, not the facts as such. Belief-based and fact-based conceptions thus come apart with regard to the role that normative facts play in the conception of legitimacy. Note that the distinction is quite easy to miss. What Estlund (Reference Estlund2008: 99) calls a correctness conception of political legitimacy, for example, appears to cover both fact-based and belief-based conceptions.

Belief-based conceptions deny that persons are self-originating sources of valid claims and that a conception of political legitimacy should, in the first instance, offer a solution to the problem of how to adjudicate between potentially conflicting wills. Instead, on those conceptions, valid claims about what should be done are those that are appropriately responsive to what we should believe about what should be done. Claims that are not appropriately responsive to that will not be valid.

Understanding the validity of practical claims in this way leads to a characterization of the problem of legitimacy that is not purely practical. The problem that needs solving for political legitimacy is ensuring that political decision-making is appropriately responsive to what we should believe about what should be done. On belief-based conceptions, epistemic considerations are thus relevant for political legitimacy.

I call the entitlement to make binding decisions that is afforded by sufficiently justified beliefs about what should be done normative authority. Normative authority is the entitlement to make a valid claim about what should be done—one that gives oneself and others reasons for action—that is based on sufficiently justified beliefs about which action is favored by the normative facts. The relevant beliefs are those of members of the decision-making body or of a group of experts to which the decision makers defer. Note that while normative authority thus depends on epistemic considerations, it captures an entitlement to make practical claims.

This dependence on epistemic considerations is what distinguishes belief-based conceptions from what I call will-based conceptions. According to will-based conceptions, as discussed above, political decisions are legitimate in virtue of how they adjudicate between the citizens’ conflicting wills. According to belief-based conceptions, by contrast, political decisions are legitimate in virtue of being appropriately responsive to normative authority.

A prime example for a belief-based conception of legitimacy is the one that Joseph Raz (Reference Raz1988Reference Raz2006) put forward and I defend this interpretation against a possible objection elsewhere (Peter Reference Peter, Sadurski, Sevel and Walton2019). The corner stone of Raz's conception of legitimacy is the so-called normal justification thesis. According to this thesis, a political decision is legitimate if it can be shown that it is made by a political authority whose decisions enable people to comply with what they should do anyway, independently of the authority's decisions, better than they would if they relied on their own judgments about what they should do (Raz Reference Raz1988: 53). A political authority, in Raz's account, may be a government or one of its agencies with the capacity to make political decisions. I refer to it as a political decision-making body.

What I call normative authority plays a key role in the normal justification thesis because having normative authority—or being prepared to defer to it—is what enables the political decision-making body to make decisions that enable people to comply better with what they should do anyway. On Raz's picture, a decision-making body will be appropriately responsive to normative authority if its decisions are guided by sufficiently justified beliefs about what should be done and its decisions will be legitimate in virtue of their being guided in this way.

To see the force of the claim that the legitimacy of political decisions depends on whether they reflect sufficiently justified beliefs about what should be done, consider a political decision-making body who, willfully ignoring ample evidence, makes political decisions that are without any doubt contrary to what should be done. By way of example, consider again the recent US decision to enforce an immigration policy that permits separating small children from their families and holding them in cage-like constructions. Belief-based conceptions allow us to question the legitimacy of decisions like this, even in circumstances where a decision has been made in procedurally correct way and satisfies the credentials of democratic authorization. Their diagnosis is that such decisions lack legitimacy because they are not adequately responsive to normative authority.

The clause that the beliefs in question need to be sufficiently justified beliefs is important, but in order to capture a broad range of belief-based conceptions, I will leave open here what counts as a sufficiently justified belief. On some conceptions, political legitimacy requires that political decisions perfectly track correct judgments about what should be done—on those conceptions, political decisions will only be legitimate if the decision-making body knows (in a factive sense) what should be done. On other conceptions, including the Razian, political decisions will be legitimate even if the decision-making body occasionally makes a mistake. We could also conceive of a conception of political legitimacy on which normative authority only obtains if the decision-making body has robust knowledge—for example, knowledge that is overdetermined by the evidence—of what should be done. What belief-based conceptions share is that they take normative authority, as afforded by sufficiently justified beliefs about what should be done, however exactly this clause is interpreted, as the ground of political legitimacy.

In his work on moral and political authority, Stephen Darwall (Reference Darwall2006; Reference Darwall2010) raises an important objection to belief-based conceptions of political legitimacy and to Raz's conception, in particular—call it the standing objection. The objection rests on a claim about the relationship between legitimate decisions and standing to hold accountable. In the context of political legitimacy, the claim is that political decisions will only be legitimate if they are made by a decision-making body with the standing to hold the citizens accountable. A decision-making body without standing to hold those subject to the decision accountable lacks the capacity to make decisions that would bind them. Based on this claim, Darwall then objects that what I call normative authority, even if there were such a thing, is the wrong ground for political legitimacy because normative authority will not give someone standing to hold others accountable. Normative authority may be the basis for giving advice, but not for making binding decision about what to do.

Darwall's argument in support of the standing objection focuses on what must be presupposed for someone to be accountable to, and thus bound by, the claim of another. He argues that there is no accountability to others without an ability to hold oneself accountable for that sort of claim (Darwall Reference Darwall2006: 23). I agree with this part of the argument.

But what is the relevant form of accountability? Darwall seems to think that such accountability must be for action and that accountability for action is importantly different from accountability for belief. It is on the basis of this distinction between accountability for action and for belief that he rejects normative authority as a ground of political legitimacy. This part of the argument is premised on a misleading contrast between the practical and the theoretical domains, however.

To see the problem, consider accountability in the two domains. To hold myself accountable for my actions means that I can distinguish between actions I should do and actions that I should not do and to see myself as an appropriate target of blame if I do something I should not do. It seems to me that I can hold myself accountable for my beliefs in the same way, however. To hold myself accountable for my beliefs means that I can distinguish between beliefs I should hold and beliefs I should not hold and to see myself as an appropriate target of blame if I believe something I should not believe. If I believe something I should not believe given the evidence that is at my fingertips, for example, I can accept that I am an appropriate target of blame. I can hold myself accountable for my beliefs in this way even though I do not see myself as a self-originating source of valid claims about what should be believed. Quite to the contrary, in fact. It is because I hold myself accountable to truth (see Railton Reference Railton, Cullity and Gaut2003: 56) that I hold myself accountable for many of my beliefs. So accountability does not entail anything about the source of the validity of the claims.

In addition, accountability for belief is one route to accountability for action. Among the beliefs I can hold myself accountable for are beliefs about how I should act. And those beliefs then form the basis for one type of accountability for my actions: when my actions do not reflect those beliefs, I potentially become an appropriate target for blame. ‘You should have known better’ is what I might say to myself or someone else might say to me.

If we can hold ourselves accountable for actions on epistemic grounds, then Darwall's standing objection to belief-based conceptions of political legitimacy fails. His objection aims to show that normative authority is not the right ground for political legitimacy because practical authority requires that we can hold ourselves accountable for claims about what should be done. But if our accountability for our actions is at least sometimes mediated by what we should believe about what should be done, then this shows that normative authority, which is afforded by sufficiently justified normative beliefs about what should be done, can be a ground for political legitimacy. Normative authority, although afforded by the satisfaction of an adequate epistemic norm, has practical reach.

While I thus do not think that Darwall's objection against belief-based conceptions of political legitimacy such as those of Raz is successful, there is a further objection to them that also targets the central role of normative authority in those conceptions. Because belief-based conceptions of political legitimacy are premised on the availability of normative authority, they run into a problem when there is no normative authority. I call this the epistemic underdetermination objection to belief-based conceptions.

This objection is a serious problem for belief-based conceptions of political legitimacy. There are good reasons to think that in the political domain, it is often not possible to establish who holds normative authority over political decisions—not because there are no normative facts that have implications for what should be done, but because it is difficult to form sufficiently justified beliefs about what should be done and to identify those who hold such beliefs. Political decisions are complex and involve a large number of normative and empirical considerations that bear on what should be done and this means that normative truths about what should be done will often be elusive. In addition, claims about what should be done are hard to confirm in such circumstances and involve invidious comparisons (Estlund Reference Estlund2008) about who does and who does not hold normative authority. Conclusive arguments for why a particular decision is the one that is favored by the normative facts will often be hard to make, as evidenced by our many sophisticated and longstanding disagreements in moral philosophy and beyond. In short, the epistemic circumstances of politics (Peter Reference Peter, Brady and Fricker2016) tend to be such that normative authority is often difficult to establish (see also Waldron Reference Waldron1999 on the circumstances of politics).

If normative authority cannot be established, then reasonable disagreements—disagreements that are the result of conflicting beliefs that cannot be dismissed as simple mistakes or unwillingness to consider the evidence, and the like—are to be expected. Such disagreements will occur if there are limits to accessing, processing, and sharing all the relevant evidence or if the body of shared evidence is so complex that it does not point to a uniquely justified belief. Under those epistemic circumstances, relying on normative authority about what should be done might well not be possible. Instead, there will be a plurality of valid claims about what should be done and no epistemic basis for deciding between them.

I am happy to grant that, even in the political domain, it will be possible to establish normative authority for some political decisions. Political decisions that sanction unnecessary harms to small children, that promote slavery, call for genocide, or incite rape and other forms of violence are good examples for decisions that are illegitimate because they ignore available normative authority. But many, if not most, political decisions will be less clear-cut. The epistemic circumstances of politics are such that for most political decisions, there will not be a decisive normative authority.

In sum, belief-based conceptions of political legitimacy can be adequate only to the extent that normative authority can be established. But the epistemic circumstances of politics tend to undermine the ground of political legitimacy that belief-based conceptions rest on and we cannot assume that normative authority is normally available in the political domain. If there is no decisive normative authority, however, then legitimacy cannot rest on this ground.

4. A Disjunctive Conception of Political Legitimacy

The preliminary conclusion to draw from this discussion of will-based and belief-based conceptions of political legitimacy in the last two sections is that each is facing a serious objection. To recap, the arbitrariness objection states that will-based conceptions fail to accommodate what I call normative authority. People are not always self-originating sources of valid claims. Sometimes, the validity of claims about what should be done, even in the political domain, depends on what we should believe about what should be done. Belief-based conceptions recognize this and take normative authority to be the ground of political legitimacy. As discussed above, however, these conceptions run into a problem when normative authority cannot be established. As the epistemic circumstances of politics make it difficult to establish normative authority, this is a problematic feature of belief-based conceptions of political legitimacy. Belief-based conceptions face an epistemic underdetermination objection.

In the discussion that follows, I put forward a conception of political legitimacy that builds on a hybrid account of the grounds of political legitimacy. The hybrid account I have in mind includes both responsiveness to normative authority and adjudication between valid, but conflicting, claims as grounds of legitimacy. I argue that a conception that builds on such hybrid grounds avoids both the arbitrariness objection and the epistemic underdetermination objection and it should thus be favored over the more familiar belief-based and will-based conceptions. For reasons I explain below, I call this conception of political legitimacy a disjunctive conception. A political decision is legitimate, according to this conception, in virtue of how the decision reflects normative authority, when normative authority can be established, or, when normative authority cannot be established, in virtue of how it adjudicates between the conflicting, but valid claims made by the citizens.

The core idea behind this disjunctive conception is to preserve the key feature of belief-based conceptions of political legitimacy, while supplementing the reliance on normative authority with a requirement to adjudicate between conflicting but valid claims about what should be done in circumstances where there is no decisive normative authority. What needs to be defended, therefore, is (1) the priority that the disjunctive conception assigns to normative authority as a ground of political legitimacy over adjudication and (2) the particular way in which the disjunctive conception views the significance of adjudication between conflicting practical claims.

Before considering a defense of these two claims, it is be helpful first to consider hybrid conceptions of political legitimacy in general. On a hybrid account of the grounds of political legitimacy, there is more than one attribute of political decision making in virtue of which political decisions have the normative property of legitimacy. Importantly, conceptions of legitimacy that build on hybrid grounds will differ with regard to how they combine those attributes, not just with regard to their accounts of what the relevant attributes are. One can thus distinguish between different conceptions of political legitimacy on the basis of how they propose to accommodate multiple grounds of legitimacy. Very generally speaking, a conception of political legitimacy might be conjunctive or disjunctive. It is conjunctive if it takes the multiple grounds to determine jointly the legitimacy of political decisions. It is disjunctive if different grounds determine the legitimacy of political decisions in different circumstances.

Based on this distinction between conjunctive and disjunctive conceptions, I can now begin to defend my claim (1), that normative authority should have priority as a ground of political legitimacy. On a conjunctive conception that builds on the same hybrid account of the grounds of political legitimacy that I am favoring, both normative authority and some form of adjudication matter for legitimacy. The conception of democratic legitimacy that David Estlund develops in Democratic Authority is a good example for such a conjunctive conception. He argues that “[d]emocratic legitimacy requires that the [decision-making] procedure can be held, in terms acceptable to all qualified points of view, to be epistemically the best (or close to it) among those that are better than random” (Estlund Reference Estlund2008: 98). Normative authority explains the emphasis on choosing the decision-making procedure that is epistemically the best, but adjudication between valid claims explains why acceptability to qualified points of view matters as well.

Conjunctive conceptions, by definition, do not assign normative priority to one ground of political legitimacy. This makes them inherently unstable, however, because the two grounds are in tension with one another. The tension is the following (see also Peter Reference Peter, Brady and Fricker2016). Normative authority as a ground of political legitimacy implies that political decision-making should track our best beliefs about what should be done and that decisions can be legitimate even if they clash with the views of individual citizens. Conversely, adjudication between valid claims can only be a plausible ground of political legitimacy if it is presupposed that there is no single valid claim about what should be done. A conception of political legitimacy that rests on hybrid grounds but that does not address the normative division of labor between the two grounds of political legitimacy is thus unstable because it is pulled in two apparently incompatible directions. The disjunctive conception of political legitimacy avoids this problem because it assigns priority to normative authority.

This argument establishes that one ground of legitimacy should have priority. But it does not establish that normative authority should have priority in a disjunctive conception of political legitimacy. What can be said in support of this further claim?

My answer to this question is that the normative primacy of normative authority stems from the normative significance of normative facts. While I have argued that normative facts cannot, by themselves, be a ground of political legitimacy, our best beliefs about them can. If it is accepted that there are normative facts that bear on what we should do, then “epistemic abstinence,” to borrow Raz's term (Reference Raz1990) is not an option: the validity of at least some claims about what should be done depends on what we should believe about what should be done.

If it is possible to establish, with sufficient robustness, what should be done, then political decision makers are not entitled to make political decisions that disregard the directive of those with normative authority and decisions that disregard normative authority will not bind the citizens. Defenders of a belief-based conceptions are right to emphasize that point, as discussed above, and I have nothing to add to their arguments in this regard. My argument for my first claim (1), then, that normative authority should have priority, is largely borrowed from defenders of belief-based conceptions and it holds against hybrid conceptions that prioritize the citizens’ wills over normative authority (Viehoff Reference Viehoff2016).

My second claim (2) is that some political decisions will be legitimate in virtue of how they adjudicate between conflicting, but valid claims. My argument for it builds on my objection to belief-based conceptions, which is that normative authority is not likely to be decisive in the typical epistemic circumstances of political decision making. Instead, we should normally expect that even after taking all available evidence is taken into account and without any obvious mistakes in reasoning, there is still a plurality of valid, but conflicting, claims about what should be done. We are thus confronted with a version of the problem of legitimacy that will-based conceptions of legitimacy primarily focus on, namely the problem of how to make political decisions in the face of conflicting valid claims.

Given how pervasive the absence of decisive normative authority is in normal political life, belief-based conceptions leave unanswered an important question: What grounds political legitimacy if valid claims about what should be done conflict? Political decisions will have to be made even if there is no normative authority. In fact, I believe that this is the normal case of political life. For example, suppose there is a reasonable disagreement about how best to legislate abortion. It is not helpful to say that nothing can be said about the legitimacy or illegitimacy of the contested alternatives. A plausible conception of political legitimacy should be able to settle which decisions are and which are not legitimate even under unfavorable epistemic circumstances. And if there is no decisive normative authority, legitimacy can only be established through some form of adjudication between conflicting claims. The ground of political legitimacy, in those circumstances, is then not normative authority, but how we adjudicate between conflicting claims.

Note that we cannot avoid this secondary ground of legitimacy by claiming that all decisions made in the absence of decisive normative authority would be illegitimate. Such a claim would be self-contradictory. On the one hand, it would support decentralized decision making—anarchy. On the other, a move to anarchy would itself amount to a political decision about which we could ask whether it is legitimate, and there is no escaping from that regress.

If we cannot rely on normative authority—because epistemic circumstances do not allow it—a version of the problem of legitimacy that will-based conceptions highlight resurfaces. The problem is that arbitrarily privileging some valid claims over others will not generate legitimate decisions—such arbitrariness will fail to ground an entitlement to make political decisions and it will not have a normative grip on citizens. Will-based conceptions are right to highlight this problem as a central problem of legitimacy, even if they are mistaken about what counts as a valid practical claim.

Admittedly, the disjunctive conception's reliance on will-based conceptions is more limited than its reliance on belief-based conceptions. As discussed above, will-based conceptions treat the problem of how to adjudicate between conflicting wills as normatively fundamental because they rest on a conception of persons as self-originating sources of valid claims. I have rejected this conception of persons and argued in support of a different account of the validity of practical claims, one that depends on what we should believe about what should be done. Adjudication remains an important source of political legitimacy, however, if it is understood as a response to difficulties with establishing normative authority and as applying to the residual plurality of valid claims.

A comparison with John Simmons's (Reference Simmons2001) influential will-based conception of political legitimacy serves to illustrate further the disjunctive conception of legitimacy. Simmons distinguishes between the moral justification for having states and the legitimacy of particular political decisions that states might make. Simmons argues that only political decisions supported by actual consent will be legitimate. To have the property of legitimacy, then, political decisions must be made by the sort of state that is, in fact, morally justified and, in addition, they must be made in a way that adjudicates between potentially conflicting wills in the right way. This conception of political legitimacy is will-based, even though it has a factualist conception of the moral justification of states in the background.

My objection to Simmons's conception is, first and foremost, that it does not give enough weight to normative authority. If there are valid claims about what should be done, refusals to consent may not, pace Simmons, undermine the legitimacy of political decisions. In addition, let me also note that this conception does not give enough weight to the epistemic limitations that might affect the moral justification of states. If there is a plurality of valid claims about what sort of political entities are justified—for example in relation to the extent of personal liberties—then this issue must be settled through some appropriate adjudication between the citizens’ wills, again contrary to what Simmons claims.

5. Concluding Remarks

Rival conceptions of political legitimacy can be helpfully distinguished, and better understood, in relation to the normative ground on which they rest. Focusing on the grounds of political legitimacy has also led me to argue for a hybrid conception of political legitimacy as a more promising alternative to either the will-based or belief-based conceptions of legitimacy that dominate current debates in political philosophy. Last but not least, bringing the hybrid conception into view has allowed me to clarify how political legitimacy depends on both epistemic and practical considerations.