The Exclusionary Power of Political Directives

Abstract

I defend the exclusionary power of political directives. The prevailing account, which I call the additive account, holds that a legitimate directive only provides a pro tanto obligation for subjects to comply. I show that it falls into a Goldilocks dilemma, giving either insufficient or excessive weight to these obligations. Pace the additive account, I argue that a legitimate directive not only gives subjects a pro tanto reason to comply but also excludes all the reasons bearing on its justifiability regarding subjects’ actions as required by the directive. Unlike Raz, who grounds the exclusionary power of legitimate directives on authorities’ supposedly superior epistemic competence, I justify it by drawing on Kantian political philosophy, which grants states a unique moral standing to make coercive decisions on behalf of their citizenry as a solution to the problem of unilateralism.

Governments order their subjects to drive within speed limits, pay taxes, and serve in wars. Barring philosophical anarchists,¹ most philosophers agree that some governments—e.g., liberal democratic governments—indeed have the right to do those things, and they refer to that right as the right to rule. How to understand the right to rule is a more contentious question. A significant division is between believers in and deniers of political obligation. According to the believers, a government's right to rule correlates with its subjects’ obligation to obey. In contrast, according to the deniers, the right to rule may consist of a variety of moral powers (e.g., to change subjects’ moral liabilities or to use coercion to extract compliance), but the power to impose moral obligations by one's mere say-so is not one of them.²

This paper is mainly concerned with an in-house debate among believers in political obligation regarding the normative power of legitimate directives.³ Two main positions on this issue can be stated as follows:

The additive account: A legitimate directive gives subjects nothing but a reason to comply that amounts to a pro tanto obligation to comply.

The exclusionary account: A legitimate directive not only gives subjects a reason to comply that amounts to a pro tanto obligation to comply but also excludes certain otherwise applicable reasons for them regarding whether to comply.

While many believers in political obligation subscribe to the additive account, some (most prominently, Joseph Raz) hold the exclusionary account.⁴ The additive and the exclusionary accounts agree that legitimate directives impose sui generis obligations on subjects (Addition in short). However, while the exclusionary account further holds that legitimate directives exclude certain otherwise applicable reasons for subjects (Exclusion in short), the additive account denies that. Their disagreement has penetrating practical significances. Consider:

False Conviction: Ian, a defendant, is charged with murder. While he is innocent, the court convicts him by due process and sentences him to life in prison. Samuel, the jailer, receives the order to imprison Ian, but he firmly believes in Ian's innocence. The real murderer, his neighbor, confessed to him about the murder but died in a car crash soon after. Though Samuel has testified in the court for Ian's innocence, the countervailing evidence is overwhelming due to bad epistemic luck. Given the minuscule chance to obtain an acquittal, Ian is likely to die in jail after decades of boredom and suffering.

Suppose that Ian's conviction is erroneous but otherwise impeccable, and the criminal justice system is generally just and effective. Most theorists agree that the court's verdict is legitimate. Prior to Ian's conviction, Samuel has no reason to imprison Ian and should strive to prevent Ian's imprisonment. According to the additive account, the court's legitimate directive gives Samuel nothing but a pro tanto obligation to comply. Thus, after the conviction, whether Samuel ought to abide by the verdict depends on the comparative strength of his political obligation to obey the court's order and his obligation not to jail an innocent person. Since many believe that a person's moral claim against wrongful life imprisonment outweighs any subject's political obligation to comply with a legitimate verdict (Claim Against Wrongful Life Imprisonment > Political Obligation in short), the additive account is likely to imply that Samuel should not imprison Ian. By contrast, according to Raz's exclusionary account, the court's legitimate directive gives Samuel a reason to comply and excludes reasons on which the court is to pronounce, including Ian's innocence.⁵ Thus, Samuel should not act on Ian's innocence after the conviction. Instead, he ought to imprison Ian, period.⁶

The additive and the exclusionary accounts are not only likely to disagree about first-order moral judgments in a wide range of cases, as exemplified by False Conviction. Fundamentally, they offer radically different pictures concerning how legitimate authority interacts with our practical reason. Which picture is correct has broad moral implications for all kinds of personal interactions mediated by political authority. In this paper, I first build up a Goldilocks dilemma for the additive account, namely that it gives either insufficient or excessive weight to political obligation. This dilemma reveals not only a structural problem with the additive account but also a critical feature of legitimate directives, namely that they do not behave as pro tanto reasons that compete with other reasons solely in terms of their weight.

The problem faced by the additive account should motivate a fair hearing for the exclusionary account. Despite Raz's influential work, many theorists do not find his exclusionary account convincing.⁷ This is partly because Raz defends Exclusion mainly based on his service conception of authority,⁸ which grounds authority on its superior epistemic competence and which is subject to a fundamental objection, namely that there is no inherent connection between epistemic superiority and the moral standing to make binding decisions for others.⁹

In contrast to Raz's epistemic justification for authority, Kantians have defended political authority as a solution to the unilateral determination and enforcement of rights. I will demonstrate that the Kantian conception of political authority provides a compelling alternative justification for Exclusion. I will first sketch what I consider the strongest Kantian defense for political authority, which draws not on the indeterminacy of rights—as many Kantians do—but only on reasonable disagreements about rights. I then rely on it to justify Exclusion. Due to space constraints, I cannot fully address the objections to the Kantian position. For readers who support the Kantian view, this paper demonstrates its novel implication, i.e., that Kantians should accept the exclusionary account and reject the additive account. For readers with reservations about the Kantian position, this paper develops one possible justification for Exclusion, leaving it open for others to explore alternatives. My arguments against the additive account are independent of my Kantian justification for Exclusion. If the additive account is not a viable position and Raz's service conception of authority faces serious objections, we have good reason to explore potential alternative arguments for Exclusion. I aim to show that the Kantian approach is a promising candidate.

The paper will unfold as follows. Sections I and II lay out the Goldilocks dilemma for the additive account. Section III draws on Kantian inspirations to justify legitimate authority's exclusive right to rule and then defend my exclusionary account—Thorough Exclusion—on its basis. Section IV revisits the Goldilocks dilemma faced by the additive account and further elucidates a vital qualification of Thorough Exclusion, i.e., that reasons justifying necessity defenses are not excluded. Section V concludes the paper by addressing the paramount concern that Thorough Exclusion has excessively authoritarian implications.

I. The Formidable Power of Legitimate Directives

Consider False Conviction again. As mentioned earlier, many additive theorists think that Samuel is not permitted to jail Ian because of Claim Against Wrongful Life Imprisonment > Political Obligation. Putting aside whether Samuel ought to help Ian escape, all he needs to do to avoid imprisoning an innocent person for life is resign. Forced to choose between imprisoning an innocent person for life and quitting his job, Samuel seems morally required to resign. The fact that the additive theory could support this widely shared and robust moral intuition speaks strongly in its favor. However, in what follows, I will show that we should not readily embrace the claim that Samuel is not permitted to imprison Ian.

My argument starts with an observation that almost any wrong decision made by the political authority is consequential. Consider:

Meals on Wheels: Meals on Wheels, a government-funded program, delivers hot meals to seniors who face malnutrition and social isolation. The government decided to significantly cut its budget. Mary, a civil servant, is in charge of depositing the government funding into the account for Meals on Wheels. She knows that hundreds of thousands of seniors will stop receiving hot meals after she deposits the reduced funds. Moreover, she believes strongly and correctly (by stipulation) that it is wrong—rather than merely uncharitable—for the government to slash the budget of such a worthy program.

Believers in legitimate authority commonly accept that, within reasonable limits, flawed social policies can be legitimate. I take the funding cut in Meals on Wheels to be an example of such flawed but legitimate policies.¹⁰ Thus, Mary's political obligation resembles Samuel's—both being civil servants’ duty to execute consequential, legitimate orders. I will argue that if Samuel must resign when ordered to imprison one innocent person for life, Mary also must resign when ordered to execute this morally flawed budget. My argument relies on three premises. The first is an empirical stipulation that the government can provide for needy seniors without unduly compromising its other moral objectives. This stipulation can easily be satisfied in affluent liberal democracies.

The other two premises are normative but also mostly uncontroversial. The first states the liberal consensus that it is wrong—rather than merely uncharitable—for a government not to assist citizens in dire need when it can do so without unduly compromising its other moral objectives. Unlike imprisoning an innocent person—a paradigmatic case of harming—cutting funding for Meals on Wheels is a form of refraining from aiding. While many believe that harming is harder to justify than withholding aid, they also accept that withholding aid could be wrongful, for example, when parents with proper means leave their children in chronic malnutrition and emotional neglect. Based on the liberal consensus that a state is obligated to assist its needy citizens, leaving seniors in malnutrition and isolation is by nature comparable to parents’ leaving their children in malnutrition and neglect: both are wrong rather than uncharitable.

The second normative premise for my argument is a moderate aggregation principle. According to it, action A could be equally (or even more) wrongful than action B if action A wrongs sufficiently more people to a sufficiently severe degree—even if it wrongs no single person more severely than action B does its victim(s).¹¹ Based on this principle and the additive account, if Mary's compliance with the order would wrong sufficiently more people to a sufficiently severe degree, what Mary is asked to do is no better than what Samuel is asked to do. Assume that depriving needy seniors of the service provided by Meals on Wheels is wrong. Taking away direly needed hot meals and human connection from hundreds of thousands of seniors every day and day after day seems at least as wrongful as imprisoning one innocent person for life. Therefore, if Samuel is morally required to resign when ordered to imprison Ian, Mary is also required to resign when ordered to deposit an unjustly reduced fund for Meals on Wheels.

The moral objection against compliance applies not only to Mary. First, if Mary is required to resign, all of her future replacements are subject to the same requirement because of their similar normative situation.¹² Second, rarely is a flawed government decision insignificant. Unfair tax law may inflict strenuous burdens on middle-class families while allowing the rich to free ride by exploiting its loopholes. A poor public health policy may fail to raise awareness of certain health hazards, thus consigning millions of citizens to chronic pain and premature death. An ill-conceived educational policy can significantly undermine public education and hurt the prospects of many children. If Mary is morally required to resign, civil servants of all stripes are often required to resign when asked to carry out flawed government policies. Thus, if every civil servant and their potential replacements act morally, the democratic will would be disregarded, and the rule of law suspended, almost whenever the state makes a flawed but legitimate decision.¹³

I am not suggesting that any individual civil servant's resignation threatens democracy or the rule of law. Civil servants are typically permitted to resign at will with due notice, and I have no objection against it. Indeed, I hold that when resignations are legally permitted, civil servants are also morally permitted to (and may even have a strong reason to) resign as an expression of conscientious objection when ordered to execute flawed but legitimate policies. Such civic engagement is perfectly consistent with democracy and the rule of law. The problem with the additive account is that it would make resignations morally obligatory even when the state legitimately rejects them to ensure efficient and effective government operation, especially if only minor sanctions are attached to such disobedience.

One might think that civil servants should comply with flawed policies not because their political obligations outweigh other weighty moral considerations but because to comply is to engage in damage control. If Mary makes the reduced deposit for Meals on Wheels, at least many beneficiaries of the program could still receive hot meals even if hundreds of thousands of seniors are left out. Damage control may be a good reason to comply, but I intend to make a stronger claim. Assume that civil servants generally have accurate judgments about intricate moral problems and do not suffer any moral akrasia. Thus, before the new budget causes any damage to Meals on Wheels, all the relevant civil servants resign despite the state's disapproval, which amounts to an unlawful strike. In so doing, they successfully coerce the government to increase the funding for Meals on Wheels and therefore dictate the desirable policy. Such a practice—if normalized—amounts to replacing democracy and the rule of law with the rule of benevolent civil servants who take justice into their own hands. Nevertheless, normalizing this practice is what morality requires, assuming the additive account plus Claim Against Wrongful Life Imprisonment > Political Obligation. Theorists who endorse this package often believe that they give due respect to democracy simply by assigning a bona fide—albeit limited—moral weight to political obligation.¹⁴ However, as I have shown, with a moderate aggregation principle, this package entails that countervailing reasons routinely outweigh citizens’ duty to comply with legitimate directives that fall short of justice. Thus, if we accept this package, we would set democracy up to fail when it is most needed, i.e., to enable the citizens to make consequential collective decisions without subjecting anyone to asymmetric power. Therefore, if we should uphold democracy and the rule of law when political decisions are legitimate, we should reject this package.¹⁵

In sum, to uphold democracy and the rule of law, additive theorists could embrace the notion that subjects’ political obligation to comply with a legitimate verdict outweighs a person's claim against wrongful life imprisonment (Political Obligation > Claim Against Wrongful Life Imprisonment in short). However, to do so is to relinquish the additive account's primary appeal: it helps vindicate defiance against legitimate directives whose defects will generate significant ramifications. Worse still, as I will argue in the next section, it is not even a viable option for the additive account.

II. The Goldilocks Dilemma for the Additive Account

Assume that additive theorists now endorse Political Obligation > Claim Against Wrongful Life Imprisonment. This move would allow them to conclude that Samuel is permitted to imprison Ian rather than required to resign. It also dictates a specific judgment in a related case:

Prison Break 1: Everything remains the same as in False Conviction, except that while Ian is in prison, he thinks to himself: “I do not deserve this fate.” Since he has exhausted juridical channels to prove his innocence, his only hope is to attain a presidential pardon if his case attracts sufficient public attention. He thus escapes the prison and makes an appearance on TV to raise public awareness of his false conviction. He turns himself in soon after he achieves that purpose.

By accepting Political Obligation > Claim Against Wrongful Life Imprisonment, additive theorists should assert that Ian is not permitted to escape in Prison Break 1.¹⁶ This assertion coheres with the law's attitude. The court would not acquit Ian on the charge of escape even if it recognized the real possibility of his being innocent.¹⁷ Backed by the weight of democracy and the rule of law, political obligations appear to be insurmountable within the additive framework.

However, in many cases, countervailing considerations indeed outweigh legitimate directives, and the law acknowledges such possibilities by accepting necessity defenses. For example, despite the legal prohibition against trespassing, it may be permissible to intentionally break into someone else's empty cabin to avoid expected severe harm caused by inclement weather.¹⁸ Courts even acknowledge potential necessity defenses for prisoners who escape from jail to avoid significant harm, such as sexual assault and beatings.¹⁹ This causes complications for Political Obligation > Claim Against Wrongful Life Imprisonment. Consider:

Prison Break 2: Robert commits a murder. He is sentenced to life in prison by due process. While Robert is in prison, a gang of fellow inmates threatens to assault him if he fails to pay a significant sum of “protection” money by a certain deadline. He could not come up with the money. Though he tries to alert the prison authorities, they ignore him. He thus escapes the prison to avoid the looming abuse. He turns himself in soon after he has avoided the threat.

Intuitively, Robert's action may not be wrong, and even courts are open to this possibility. However, additive theorists—with their new commitment—cannot accommodate this possibility. Suppose you break into the prison and have a unique opportunity to save Ian from wrongful life imprisonment or Robert from the physical abuse, all else equal, but you cannot do both. Bracket the morality of your illegal entry into prison. In the additive framework, you certainly should save Ian. It implies that Ian's claim against the wrongful life imprisonment is weightier than Robert's claim against an instance of wrongful physical abuse.²⁰ However, since the courts’ verdicts in both cases are legitimate and the verdict is also correct in Robert's case but not in Ian's, Robert's political obligation to comply is at least as weighty as Ian's. Assuming that Ian's political obligation outweighs his claim against wrongful life imprisonment, Robert's political obligation must outweigh his claim against an instance of wrongful physical abuse. We may summarize the inferences as follows:

1. P1. Ian's political obligation > Ian's claim against wrongful life imprisonment (because it is not permissible for Ian to escape)

2. P2. Ian's claim against wrongful life imprisonment > Robert's claim against an instance of wrongful physical abuse (because a third party should prioritize preventing Ian's wrongful life imprisonment over preventing Robert's wrongful physical abuse)²¹

3. P3. Robert's political obligation ≥ Ian's political obligation (because the only relevant difference is that the verdict is correct in Robert's case but not in Ian's)

4. P4. Robert's political obligation > Robert's claim against an instance of wrongful physical abuse

If P4 is correct, it is morally impermissible for Robert to escape in Prison Break 2. However, such a conclusion is not only counterintuitive but also at odds with the courts’ attitude. Assuming that the court's accommodating attitude regarding Prison Break 2 is justified, we have to reject the package of the additive account plus Political Obligation > Claim Against Wrongful Life Imprisonment.

Additive theorists thus face a dilemma. On the one hand, if they hold Political Obligation > Claim Against Wrongful Life Imprisonment, they must conclude that Ian is not permitted to escape, and neither is Robert. This is unacceptable. Even the court is open to granting Robert permission to escape ex post facto. On the other hand, if they hold Claim Against Wrongful Life Imprisonment > Political Obligation, they must conclude that Robert is permitted to escape, and so is Ian. This amounts to jettisoning the authority of democracy and the rule of law, as I argued in Section I.²² Stated differently, additive theorists lack a Goldilocks principle of the relative weight of political obligations vis-à-vis other weighty moral considerations. Political obligations seem either too strong or too weak.

Prison Break 1 and 2 are not a pair of isolated counterexamples because the Goldilocks dilemma is structural. The additive account must assign enormous weight to political obligations to uphold democracy and the rule of law since legitimate political decisions often fall short of justice and have dire consequences. However, if it assigns such weight to political obligations, it would condemn warranted disobedience based on necessity defenses. This dilemma reveals a puzzling feature of political directives. They could be insurmountable when the stakes are incredibly high but, at the same time, easily overridden when stakes are much lower. The two facets of political directives suggest that they do not operate as normal, pro tanto reasons that compete with other reasons solely in terms of their weight. Instead, political obligations seem vulnerable to the countervailing force of certain reasons but not others.

III. Legitimate Directives and Thorough Exclusion

The exclusionary account offers a principled way to sort out reasons that may defeat political obligations and those that may not. I will lay out my central argument for Exclusion in this section and further explain its qualifications and implications later on. In defending my exclusionary account, I will first establish another thesis, which I call Exclusive Right to Rule, as my stepping stone.

A. The Exclusive Right to Rule

Exclusive Right to Rule: An office-holder has an exclusive right to settle practical issues within the jurisdiction of their office.²³

According to commonsense morality, people have various rights regarding, e.g., their bodies, properties, contracts, and relationships (such as marriage and custody of children). Those rights constitute a sphere of external freedom for the right-holder, i.e., the freedom to set and pursue ends in the outside world without being constrained by others’ contingent will. Insofar as one operates within their rights, they should not be liable to others’ coercion. Furthermore, one's rights impose correlative duties on others. Those duties are, in principle, enforceable. If another person violates one's rights, the rights-violator is liable to necessary and proportional defensive force that aims to stop them and is further liable to compensate the right-holder for the damages ex post facto. Stated differently, rights, which constitute a protected sphere of freedom for the right-holder, are inherently connected with an authorization to use coercion in their defense.

In the absence of authority, everyone is left to assert and enforce their rights unilaterally. Due to normative indeterminacies and epistemic uncertainties, people's unilateral assertions of rights often conflict even if everyone acts in good faith. Potential examples are endless. When two people have an oral agreement but never sign a formal contract, one may believe that a binding contract has been established while the other denies it. When couples seek a divorce, they often disagree about custody arrangements. As people interact, the boundaries of their freedoms are coextensive. Take a divorced couple as an example. The mother has no custody right to be with her child when the father is exercising his right to do so, and vice versa. Because the contours of people's rights are coextensive, one univocal resolution must prevail regarding rights disputes. While normative indeterminacies and epistemic uncertainties are two distinct sources for reasonable disagreements about rights, I will demonstrate that rights disputes per se make legitimate authority morally necessary. Since it is more difficult to justify this claim regarding disputes arising from epistemic barriers alone, I will focus on them.

In defending the moral necessity of authority, let me start by considering an alternative position: authority is dispensable, and disputing parties are bound by and liable to coercion based on the objectively correct resolution, at least when reason dictates a unique solution. I will refer to this proposal as Objectivist Justice. Despite its prima facie attraction, Objectivist Justice does not stand up to scrutiny. Correct resolutions of rights are not automatically binding. Suppose that, in a state of nature, a couple breaks up. Each parent believes they should be the custodial parent based on their reasonable evaluation of their comparative strengths as parents. Assume for now that the actual balance of reasons dictates that the parents spend equal time with their child because, ceteris paribus, this arrangement is best for the child's long-term well-being. While both parents are mistaken, their neighbor—knowing the family for years—makes the right call. Nevertheless, they cannot convince the parents to accept their correct resolution. Suppose that the neighbor, being rich, is able and willing to hire a team of guards to use necessary and proportionate means to coerce the parents to comply. Intuitively, such benevolent intervention is objectionable, demonstrating that pace Objectivist Justice, being correct per se does not entail a moral standing to impose a binding resolution.

The neighbor lacks the standing to impose their correct resolution on the parents, ultimately not because the parents’ views have epistemic merits but because they are equally entitled to be free. Kantians have long advocated that an intuitive and fundamental conception of our freedom is freedom as independence, which is to be one's own master, i.e., “not to be forced to obey the will of another person.”²⁴ When the neighbor forces the parents to obey their will amid their reasonable disagreement about rights, such behavior is incompatible with the parents’ freedom as independence and is thus objectionable.²⁵ More generally, as a fallible reasoner, anyone may make good-faith mistakes about rights regarding any issue. By authorizing the right side to intervene with coercion whenever someone makes a reasonable mistake about rights, Objectivist Justice makes it hardly possible for anyone to enjoy freedom as independence: the enforcement of rights marshaled by the right side is nonetheless private enforcement with “the inevitable side-effect of subjecting us to the wills of others.”²⁶ Furthermore, Objectivist Justice implies that people are, in principle, liable to coercive correction whenever they make reasonable mistakes about rights. Knowing their fallibility without knowing where exactly their mistakes lie, people would live under the constant dread that justified coercion will come to them at any moment by anyone who happens to be right. Such a human condition would be a paradigm of unfreedom.

In sum, Objectivist Justice is defective because it embraces unilateral enforcement of rights, which is incompatible with people's freedom. While univocal decisions about rights are morally required, no private party has the proper standing to deliver them. This inherent defect of the state of nature is the ultimate justification for political authority. As noted earlier, rights are supposed to constitute people's external freedom and impose enforceable correlative duties on others irrespective of their personal convictions. Only if the boundaries of rights are settled on behalf of everyone involved is no one subjected to others’ contingent will. Thus, rights are a priori premised on the existence of legitimate authority, an omnilateral will in Kantian terms, that validates rights-claims on behalf and in the name of everyone involved.

The totality of people's enforceable rights and obligations constitutes a system of justice. Given that the authority should effectively resolve people's reasonable disagreements about justice on everyone's behalf, it needs to satisfy the following conditions. Formally, this authority must monopolize comprehensive legislative, judiciary, and executive functions for a definitive population regulating both subjects’ interactions (i.e., the private sphere) and its own operation (i.e., the public sphere). Its decisions regarding the private sphere (about, e.g., property, contract, and tort) specify subjects’ private rights and obligations in their mutual dealings. Its decisions regarding the public sphere (about its constitution and maintenance) spell out its jurisdictional rights and citizens’ rights and obligations in sustaining its operation. Both kinds of decisions embody authoritative resolutions of justice. By providing systemic, univocal solutions of justice for its subjects, this authority liberates them from one another's coercive force. Procedurally, this authority must duly incorporate everyone's interest and voice in its decisions, at least at a fundamental level. A person is not free if they are protected from others’ contingent will but subjected to alien domination of the political power, which would be the case if political decisions are unresponsive to their interests and voices. By duly integrating everyone's will and interests, the authority becomes the apparatus of the people's self-governance. Substantively, this authority must offer reasonable resolutions of justice. As I have argued, legitimate authority is primarily justified on the ground that it properly settles people's reasonable disagreements about justice. If its decisions are unreasonable, they cannot be plausibly construed as resolving people's reasonable disputes and therefore lack legitimate authority over the people.

Since a directive satisfies the procedural criterion if and only if it results from a reasonably fair procedure, the procedural criterion—like the substantive criterion—depends on a conception of reasonableness. Ideally, a political directive—regarding either procedural designs or substantive issues—should reflect the balance of reasons for the issue at hand. Officials thus have a stringent duty to reason competently and decide based on their best judgment. Combining this observation with my overarching concern for citizens’ equal freedom, I adopt the following standard of reasonableness:

Reasonableness: A political directive is reasonable if and only if it can result from competent reasoning based on pertinent factual information and moral beliefs constrained by a reasonable conception of justice.²⁷ A conception of justice is reasonable if and only if it upholds an adequate scheme of rights for individuals and sufficiently respects their fundamental status as moral equals.

I do not intend to settle where the defensible thresholds lie regarding competent reasoning, adequate rights schemes, and sufficient respect for citizens’ fundamental equality. This sketchy formula is supposed to capture largely a liberal common ground.²⁸ However, I expect it to have considerable teeth.²⁹

Any agent that satisfies the formal criterion is a de facto authority. If such an authority pervasively fails Reasonableness procedurally or substantively, it lacks general legitimacy. Otherwise, it is generally legitimate. However, a generally legitimate authority can be unreasonable regarding specific issues. Those failures would render the resultant directives illegitimate. In sum, ceteris paribus, an agent that satisfies the aforementioned three criteria sufficiently possesses legitimate authority for the people in general.³⁰ Since—as widely accepted—a liberal democratic state fits the bill, it acquires legitimate general authority for its citizenry. Recall that due to the illegitimacy of unilateral determination and enforcement, no private person has the right to settle reasonable disagreements about justice. Thus, the state is uniquely positioned to do so with its reasonable directives. In this way, the state has the exclusive right to rule. The state exercises its right to rule by creating offices with the mandates to decide practical issues within their jurisdictions. Since the state has an exclusive right to rule, whoever the state duly puts in an office has the exclusive right to settle practical issues within their jurisdiction. Therefore, I have vindicated Exclusive Right to Rule. ³¹

According to Exclusive Right to Rule, legitimate authority has the unique moral standing to demarcate the boundaries of people's rights. Thus, its legitimate directives must possess the moral power to determine people's rights, leading to what Stilz calls “a statist theory of rights.”³² Those directives can render the contours of rights determined where reason leaves open. Within reasonable limits, they can even redefine such contours, defying the precise balance of reason. For example, suppose that the aforementioned divorcing couple—now under a civil condition—turns to the court to resolve their dispute about the custody arrangements. Recall that each parent wants to be the custodial parent based on their reasonable judgments. The court ought to adjudicate the case by considering all the pertinent reasons. Assume henceforth that the actual balance of reasons supports the father's claim: ceteris paribus, his being the custodial parent is best for the child's long-term well-being. However, the court mistakenly but reasonably sides with the mother, mandating that the child lives with her on weekdays and their father on weekends. In light of the court's exclusive right to rule, its legitimate ruling should determine the contours of each parent's custody right, despite being erroneous. This example is by no means exceptional. In general, legitimate laws define citizens’ moral rights by choosing one set of univocal answers to justice-related problems among various reasonable ones on their behalf. This statist theory of rights is moderate. Accordingly, when a directive is unreasonable, it is normatively impotent, which typically constrains how far moral rights shaped by political decisions can diverge from perfect justice.³³

B. My Exclusionary Account

With this moderate statist theory of rights at hand, we can observe first that Exclusive Right to Rule entails Addition. Recall that individuals are always already obligated to respect others’ rights when those rights are properly validated. Since legitimate directives authoritatively define the boundaries of people's rights, subjects have at least a pro tanto moral obligation to comply with those directives. However, I will argue that Addition fails to capture fully the normative power of legitimate directives because Exclusive Right to Rule also implies that those directives possess exclusionary power on top of their moral weight. Specifically, I defend:

Thorough Exclusion: When a legitimate directive requires S to φ, it not only gives S a reason to φ that amounts to a pro tanto obligation to φ but also excludes all the reasons bearing on its justifiability for S—only regarding whether S should φ.

Here and in what follows, S refers to any subject and φ to any action required by legitimate directives. Two observations are in order. First, I label my thesis Thorough Exclusion because I hold that a legitimate directive excludes all the reasons bearing on its justifiability.³⁴ I will refer to reasons bearing on the justifiability of a legitimate directive as dependent reasons.³⁵ Second, there are two ways to understand Exclusion. According to the justificatory conception, Exclusion first and foremost occurs at the justificatory level: excluded reasons no longer count for or against S's φing. According to the motivational conception, Exclusion only occurs at the motivational level: while S should not act for excluded reasons, those reasons remain relevant to the morality of their φing.³⁶ I hold that we should understand Thorough Exclusion according to the justificatory conception. First, whether Exclusion should occur at the motivational level depends on whether it happens at the justificatory level. If S knows that excluded reasons still bear on the morality of their φing and require them not to φ on balance, it would be irrational for S to φ disregarding those reasons.³⁷ Moreover, while certain reasons are excluded, S is not prohibited from acting for them insofar as they conform with legitimate directives. For example, according to Thorough Exclusion, the legal prohibition of murder reflects and excludes reasons bearing on its justifiability, including that murder is egregiously wrong. However, citizens refraining from murder can undoubtedly be motivated by the wrongness of murder rather than merely by the fact that murder is legally prohibited.³⁸ The justificatory conception of Thorough Exclusion can accommodate this, but the motivational conception cannot. According to the motivational conception, subjects should not act for this excluded reason as they refrain from murder. According to the justificatory conception, the legal prohibition of murder has incorporated and reflects the corresponding stringent moral prohibition, which thus does not provide an extra objection to murder along with the legal prohibition at the normative level. However, typically, the law only cares about subjects’ actions rather than their reasons for their actions, and thus subjects may obey the law “for any motive whatever.”³⁹ Therefore, nothing is problematic in the eyes of the law where citizens refrain from murder for its moral depravity. Those two points show that we must understand Thorough Exclusion according to the justificatory conception if this thesis is defensible at all. The crucial question is whether it is defensible, which will be the focus of the rest of the paper.⁴⁰

For any practical issue within an official's jurisdiction, the official should do their best to weigh pertinent reasons. Nevertheless, their legitimate directives are meant to determine the balance of reasons regardless of whether they in fact strike the right balance. Consider the earlier divorce example. I stipulated that the court arrived at an erroneous but reasonable ruling after considering the pertinent reasons. Still, its decision is supposed to pronounce on where the balance of reasons lies regarding their dispute and determine each parent's custody rights. The court's exclusive right to do so entails that neither parent has the right to act on their own judgment about the balance of reasons but ought to comply with the court's ruling. After the court's ruling, the father should return the child to their mother every Monday. Even if his preferred resolution—i.e., his being the custodial parent—is correct, the court nonetheless has decided incorrectly but legitimately that its adjudication is best supported by pertinent reasons. As Raz puts it, “the original reasons merge into . . . the judgment of a court, which . . . becomes res judicata.”⁴¹ The adjudication replaces the original reasons and serves as the new normative basis for subjects’ actions.

A legitimate directive requiring S to φ is likely based on both dependent reasons counting in favor of S's φing (call them the dependent reasons to φ in short) and the dependent reasons against S's φing (call them the dependent reasons not to φ). In the divorce example, the court designates the mother as the custodial parent and requires the father to bring the child to their mother on Mondays. Reasons supporting the court's decision may include that the mother has more financial resources to foster the child. Reasons against it may include that the father is more devoted to supporting the child's education. A legitimate directive should be able to exclude the dependent reasons to φ and not to φ, fundamentally because it authoritatively determines the balance of those reasons for S regarding whether to φ. Specifically, the dependent reasons not to φ are excluded to rule out the normative basis where subjects appeal to those reasons to justify disobedience. Meanwhile, the dependent reasons to φ are also excluded because otherwise it leads to objectionable double-counting. Since the court's ruling is based on and reflects the normative weight of the fact that the mother has more financial resources to foster the child than does the father, this fact should not be counted again along with the court's ruling regarding whether the father should comply.⁴²

Note that according to my justificatory conception, an asymmetry exists regarding whether Exclusion should occur at the motivational level. Since authority typically does not care about S's reason for action insofar as they comply, S may act for dependent reasons to φ. However, as the dependent reasons not to φ support S to disobey legitimate directives, S should not act for those reasons. In the divorce example, the father may respect the mother's right to be the custodial parent, thinking she can better provide for the child financially if this makes him feel better. However, he cannot act for his superior devotion to the child's education if this may lead him to disobey the ruling.

Below I will further defend Thorough Exclusion by demonstrating its three comparative advantages vis-à-vis the additive account. First, Thorough Exclusion accurately captures what authorities try to do—i.e., to sum up dependent reasons in their decisions—while the additive account flies in the face of their self-image. Raz offers an example where Parliament decides to tax unearned income at a higher rate than earned income, and a person goes to the court claiming that he is unfairly discriminated against because whether one's income is unearned should be irrelevant for the purposes of taxation.⁴³ As Raz observes plausibly, the court may say: “Since the issue of the justice of such a differentiation was clearly decided upon by Parliament, we are precluded from examining the reasons for or against it on their merits. There may be an injustice here, but if so, it is not for us to put it right.”⁴⁴ Generally speaking, when a legislature dictates a law to the judiciary, it intends for the judiciary to issue verdicts based on the law instead of reconsidering the dependent reasons along with the law. The same is true regarding elected executive officials vis-à-vis ordinary civil servants or military superiors vis-à-vis their subordinates. If authorities can be legitimate, the additive account's failure to accommodate their self-perception seems to be a significant weakness.

Second, legal systems commonly sanction disobedience. Such sanctions often seem justified, especially regarding subjects who persistently violate legitimate directives. However, the additive account may have trouble justifying those sanctions. Consider the divorce example again. Assume that the father believes—correctly—that the court has made a mistake. Despite the legitimacy of its ruling, he repeatedly holds on to the child on Mondays. The court therein holds him in contempt and subjects him to fines. Additive theorists may hold either that the father's claim to be the custodial parent outweighs his obligation to comply with the court's ruling or the opposite. However, the court's response is hardly justifiable in either scenario. First, assume that the father's claim to be the custodial parent outweighs his obligation to comply with the court's ruling. He is thus morally permitted to hold on to his child on Mondays. As he has accurately weighed the court's verdict in his deliberation, he shows no contempt for the court. However, assume that the opposite is true. The court's ruling tips the balance of reasons for the father, so he is overall obligated to return the child each Monday. Still, the father may judge incorrectly but reasonably that the court's ruling is not weighty enough to trump the countervailing reasons. In light of the additive account, the court's ruling only provides a pro tanto reason for compliance. Thus, as the father acts on his competent judgment about the new balance of reasons in good faith, he again shows no disrespect for the court despite his error. Therefore, the court's response to the father's disobedience seems morally objectionable in the additive framework. However, it makes perfect sense, according to my exclusionary account. Even though the father has weighed the court's ruling as one consideration among others, he disrespects the court because he fails to respect the verdict as it is meant to be, i.e., an authoritative summary of the reasons that bear on the custody arrangement, which is for the court to determine. Thus, in my account, there is an inherently fitting relation between the father's disobedience and the sanctions the court imposes on him.⁴⁵

Finally, and more fundamentally, while Thorough Exclusion eliminates the unilateral determination of rights, the additive account does the opposite. In the divorce example, after the court legitimately settles the issue for the parents, Thorough Exclusion entails that the father should comply with the court's verdict and respect the mother's moral right to be the custodial parent. By contrast, within the additive framework, the father is supposed to reconsider all the reasons and decide whether the court's decision prevails. Regardless of what conclusion he arrives at in the end, he is unilaterally settling his dispute with the mother regarding their respective custody rights, which he has no moral standing to do—as I have argued. Unilateralism occurs not only when a person defies legitimate authoritative demarcations of rights. Instead, it is fundamentally a structural problem where private persons are supposed to decide by themselves how to resolve rights disputes. Thus, the problem persists even if private persons take legitimate directives into account as one consideration in their deliberation. By prescribing that people resort to private deliberation after authoritative resolutions are in place, the additive account makes the unilateral determination of rights ubiquitous rather than solving it. This final point also helps explain the authority's practices as observed in the first two points. Because only legitimate directives can settle the boundaries of rights without subjecting anyone to unilateral coercion, political authority intends its directives to sum up and replace reasons bearing on them and consequently judges subjects to be disrespectful when they fail to conceive those directives accordingly.

Note that Thorough Exclusion is inherently limited in two ways. First, it only holds when official directives are legitimate. S is morally permitted (or even obligated) to resist illegitimate directives drawing on dependent reasons that those directives grossly fail to respond to. Second, even when directives are legitimate, Thorough Exclusion only applies regarding S's φing. For example, in the divorce case, the court's adjudication normatively determines only that the father ought to return the child on Mondays until the arrangement is lawfully altered. This leaves open for the father to complain or appeal. Similarly, in Meals on Wheels, the authoritative budget is supposed to settle how many public resources are allocated to Meals on Wheels. Mary, the civil servant, or any other citizen, has full discretion to voice their moral objections against the decision through various legal means. Only when subjects defy legitimate directives regarding actions required by those directives do they illegitimately impose their unilateral judgments on others. By contrast, there is no illegitimate assertion of one's own opinion when citizens sort out their disagreements by actively engaging with political decisions in ways consistent with and encouraged by democratic ideals and the rule of law.

IV. The Goldilocks Dilemma Revisited

This section expounds an additional limit of Thorough Exclusion by revisiting the Goldilocks dilemma that besets the additive account. I have argued that an official is uniquely authorized to settle practical issues within their jurisdiction. In this way, they are vested with the moral power to pronounce on the balance of all dependent reasons. Otherwise, their decisions may fail to settle those issues. If—as I have argued—Exclusion is necessary for legitimate directives to settle disputes and eliminate unilateralism, those directives should be able to exclude all the dependent reasons.⁴⁶ Consider False Conviction again. Since Ian's conviction is legitimate, it excludes all the dependent reasons, including both procedural reasons of due process and substantive reasons such as Ian's innocence. As Ian's innocence does not count against Samuel's action as a jailer, he is permitted to imprison Ian. As neither does Ian's innocence count for his own disobedience, he is not permitted to escape in Prison Break 1. Both Samuel and Ian ought to comply with the conviction, not because their political obligation outweighs Ian's claim against wrongful life imprisonment, but because those dependent reasons do not apply to them but exclusively apply to the court regarding their required actions. Thus, though the state wrongly imprisons Ian, its legitimate decision nonetheless binds its citizens.

The fact that a legitimate directive excludes all the dependent reasons regarding S's φing also sets a crucial limit for Exclusion. A directive excludes only dependent reasons, i.e., reasons bearing on its justifiability. For example, a court's verdict only excludes reasons bearing on a defendant's guilt as well as reasons of due process. It does not exclude reasons that have nothing to do with the justifiability of the verdict. Thus, while Ian is not justified to escape citing his innocence in Prison Break 1, Robert in Prison Break 2 may be justified to escape as he faces the pending abuse in prison. The threat Robert faces in prison is not a reason that bears on the court's verdict. While the prison authority should attend to the threat, they unreasonably ignore his alert. Therefore, his reason for action based on the pending abuse is not excluded by any legitimate directive.⁴⁷ Plausibly, his reason to avoid the abuse outweighs his obligation to obey. If Robert has no way to avoid the abuse other than escape, escaping would be justified.

It becomes clear why my exclusionary account avoids the Goldilocks dilemma that besets the additive account. I acknowledge that Ian's claim against wrongful life imprisonment is weightier than Robert's claim against an instance of wrongful physical abuse, and both are weightier than their political obligation to obey the court's order. However, based on Thorough Exclusion, while Ian's claim no longer provides a reason for his escape, Robert's claim does. Therefore, though Ian is not permitted to escape, Robert could be. Thus, my account vindicates judgments about Prison Break 1 and 2 in a way that is consistent with the law's divergent attitudes toward them. In this picture, legitimate directives give subjects a protected pro tanto reason for compliance. Such a reason is pro tanto because it could be outweighed by unexcluded reasons. It is protected because it silences the normative force of some most obvious and extremely weighty countervailing reasons.

No matter which horn of the Goldilocks dilemma additive theorists would take, they have to assert that the court is inconsistent if it condemns Ian's action in Prison Break 1 but accommodates Robert's action in Prison Break 2, especially if the court acknowledges the real possibility of Ian's innocence. By contrast, Thorough Exclusion offers an intuitive explanation of why the court holds different attitudes toward Prison Break 1 and 2. While Ian relies on a reason (i.e., his innocence) that a legitimate directive has settled, Robert appeals to a necessity defense drawing on urgent considerations that have not featured in any legitimate directive. This observation supports an inference to the best explanation in favor of the conjecture that necessity defenses are premised on properly defined exclusionary power of legitimate directives.

This structure of necessity defenses not only explains the law's divergent attitudes toward Prison Break 1 and 2 but also justifies them. It is an inherent imperative for political authority to avoid both feebleness and arrogance. On the one hand, the authority is feeble if its directives give subjects nothing but pro tanto reasons with limited weight. For subjects are supposed to act on their private judgments of the new balance of reasons after legitimate directives are in place. This reintroduces the problem of unilateral determination of rights and casts doubts on the legitimacy of sanctioning disobedience amid reasonable disagreements about justice. Ultimately, the prohibition against Ian's escape is the price for removing unilateral determination and enforcement of justice. On the other hand, the authority is arrogant if it assigns unlimited weight or scope of Exclusion to its otherwise legitimate directives and thus rules out necessity defenses. For example, if the law prohibits prison breaks under imminent threats or trespassing as a necessary means to save lives, it would reveal its indifference toward its subjects and jeopardize its legitimacy. Ultimately, accommodating Robert's escape testifies to the law's due care toward its subjects.

Political authority locates the Goldilocks position between feebleness and arrogance by assigning appropriate weight and properly defined exclusionary power to its legitimate directives. As argued earlier, legitimate directives exclude all the dependent reasons because they are supposed to settle their balance. I submit that excluded reasons, in principle, do not include reasons justifying necessity defenses. In general, the authority has an exclusive right to determine the boundaries of rights and thus to pronounce on all related reasons. Thus, necessity defenses are typically available only under urgent circumstances, where new reasons for action emerge but cannot timely feature in legitimate directives. As we have seen, the fact that laws must leave room for necessity defenses certainly bears on their justifiability. However, since there are too many possible contingencies that the authority cannot foresee, consider, and adjudicate in advance, it cannot (and should not) completely foreclose the question of what urgent conditions justify necessity defenses.⁴⁸ Thus, the law's justice does not depend on its adjudicating reasons that potentially justify necessity defenses. For example, property law must be open to permissible trespassing under urgent circumstances to be justifiable. Nevertheless, the justice of property law does not depend on its enumerating what kind and level of urgency justify trespassing. Therefore, property law does not exclude those considerations. Subjects should be guided by reason in deciding whether a necessity defense is available in a given situation, and the authority may validate or reject their claims ex post facto. By refraining from foreclosing what urgent conditions justify necessity defenses ex ante, the authority further avoids arrogance, demonstrating awareness of its limited foresight.

V. The Danger of Authoritarianism

I want to conclude the paper by addressing one persistent and forceful objection against Exclusion, namely that it is excessively authoritarian. This objection has haunted the exclusionary account since Raz first articulated it.⁴⁹ However, I will address it by specifically considering how it may apply to my account.

A. The Permissibility of Unlawful Resistance

Let me start with where my account embraces political resistance to dispel authoritarian impressions that one may mistakenly attribute to my account. Put aside illegitimate regimes, where my account prescribes no obligation for compliance. I have emphasized that a generally legitimate authority can issue illegitimate directives. Assume that the US has been a legitimate state after abolishing slavery and establishing universal suffrage and basic welfare. Notwithstanding, the Jim Crow laws were clearly illegitimate for their blatant discrimination. Citizens were morally permitted or even obligated to resist Jim Crow laws, with both lawful or unlawful means. Coordinated nonviolent disobedience during the civil rights movement was undoubtedly justified. Indeed, violent unlawful resistance may also be warranted, for example, if protestors were to resort to property sabotage to demonstrate their apt anger and press for change. It is beyond the purview of the paper to determine whether and what unlawful means of resistance are permissible in response to specific illegitimate laws. The crucial point is that illegitimate directives have no moral power over citizens, and thus citizens must resort to private deliberation to determine a permissible means of resistance. If such resistance imposes harm on others, pertinent moral constraints, e.g., necessity and proportionality, would apply.

Citizens often need to break other potentially legitimate laws to force the state to change illegitimate laws. For example, property sabotage as a means to fight against the Jim Crow laws would defy arguably legitimate property laws and thus infringe on other citizens’ morally valid property claims. One may wonder how such unlawful resistance could be permissible given the exclusionary power of legitimate laws. In response, first, publically violating legitimate Law-A as a means to challenge illegitimate Law-B shows no disrespect for Law-A's normative authority. Recall that a legitimate directive only excludes reasons bearing on its justifiability. Assuming that certain property laws—especially to the extent that they were independent of racial segregation—were legitimate during the Jim Crow era, those laws only excluded reasons pertinent to property distributions. Were protestors to violate those laws in pursuing desegregation, they would not be acting on those property-related reasons. Instead, they would be drawing on reasons against segregation, which were not excluded by any legitimate law. Thus, the exclusionary power of property laws constituted no principled objection against property sabotage for the sake of desegregation.

Furthermore, as stated earlier, unlawful resistance is morally permissible only if it satisfies applicable moral constraints, such as necessity and proportionality. Assuming that certain instances of property sabotage were necessary and proportional to desegregation, relevant costs suffered by property owners would be justified as a lesser evil. Nevertheless, because illegitimate laws result from the citizenry's defective self-governance, the citizenry should fairly share the expenses for righting the wrong by using national revenues to compensate private owners for their losses caused by justified unlawful resistance.

B. Objectivism About Legitimacy

Whether my account leaves appropriate space for unlawful resistance crucially depends on how we flesh out Reasonableness sketched in Section III. The more demanding the criterion is, the less authoritarian my account is. The stringency of Reasonableness depends on the thresholds for competent reasoning, adequate rights schemes, and sufficient respect for citizens’ fundamental equality, which I leave open. Nevertheless, as imprecise as it stands, Reasonableness supports a key observation. Political directives can hide their unreasonableness behind seemingly plausible narratives, secrets kept in the name of national security, the lack of public attention, and so forth. For example, voting restrictions in the US carry a prima facie plausible rationale, i.e., combating voter fraud. However, statistics regarding the rarities of voter fraud and the racially skewed voter suppression caused by those restrictions are widely available. I believe that no competent thinker committed to the right to political participation and citizens’ equal status can believe in those restrictions after surveying the statistics. Thus, unlawful resistance—e.g., using expired photo IDs in voter registration—may be perfectly permissible. Similarly, my account may embrace other forms of unlawful disobedience against covertly illegitimate directives, from consequential actions like Daniel Ellsberg's leaking of the Pentagon Papers to quotidian activities like downloading copyrighted materials in defiance of certain unconscionable copyright regulations.

I do not intend to validate these specific examples of unlawful resistance, partly because people may reasonably disagree about whether a political directive passes the threshold for legitimacy. The point is instead that my exclusionary account, combined with defensible criteria of legitimacy, leaves proper room for unlawful resistance, even against directives that do not wear their illegitimacy on their face. However, my acknowledgment of reasonable disagreements about legitimacy apparently engenders a dilemma for my account. If citizens can be bound by objectively legitimate directives despite reasonable disagreements about their legitimacy, why can't objectively correct resolutions of justice bind them despite reasonable disagreements about justice? On the other hand, if citizens are allowed to act on their own reasonable judgments about the legitimacy of political directives, how does the introduction of the state address the problem of unilateral determination?

I embrace the first horn of the dilemma, i.e., citizens are bound by objectively legitimate directives. Call this thesis Objectivist Legitimacy in contrast with Objectivist Justice discussed in Section III, i.e., citizens are simply bound by correct resolutions of justice. I rejected Objectivist Justice because private citizens lack the moral standing to impose their preferred resolution—correct or not—on others who reasonably disagree. While Objectivist Justice licenses unilateral enforcement, which is incompatible with citizens’ equal freedom, Objectivist Legitimacy empowers the state to impose truly legitimate resolutions on all the involved parties. Such law enforcement—authorized by an omnilateral will—is the only form of force compatible with everyone's equal freedom. Stated differently, Objectivist Justice is doomed by its endorsement of unilateral enforcement, and Objectivist Legitimacy is not. This ultimately explains why we should accept the latter but not the former. That being said, when citizens implement Objectivist Legitimacy in good faith, some may mistakenly—albeit reasonably—deny the legitimacy of certain directives and disobey them. In such a case, their disobedience, though impermissible, may be excusable, so the state has reason to at least reduce their punishment.⁵⁰

C. Obedience of Legitimate Directives

I am finally where the rubber hits the road. As Section I shows, I hold that plausible criteria for political legitimacy will leave considerable gaps between legitimacy (where directives are procedurally and substantively reasonable) and justice (where directives represent the balance of reasons). Furthermore, political decisions are seldom inconsequential. The tax code determines our disposable income and wealth. Hate speech regulations specify the boundaries of our free speech. And abortion laws confine women's control of their bodies. Legitimate directives falling short of justice often impose significant, unjustified harms on millions of citizens. However, my exclusionary account entails that citizens ought to comply with legitimate directives under normal circumstances and try to bring changes only through lawful means.

Is this an unacceptable conclusion? I do not think so. First, according to my account, only reasonable political directives have the power to define people's moral rights, i.e., by choosing one reasonable conception of justice among various reasonable ones. Typically, this significantly constrains how far binding laws may diverge from perfect justice. Use women's right to abortion as an example. Assuming for the sake of argument that it is unreasonable and thus illegitimate for the state to ban abortion entirely, a variety of gestational limits for abortion may fall into the range of reasonable options.⁵¹ If a state adopts one of the reasonable options, its decision determines the contours of women's moral right to abortion. Suppose that the adopted gestational limit is shorter than what justice demands. Since it allows a reasonable time window for abortion and is promulgated in advance, it enables women to plan accordingly and thus is unlikely to deprive them of adequate external freedom. Thus, it does not seem too demanding that those who disagree with the law forgo illegal means to fight against it.

Furthermore, forgoing illegal means to fight against legitimate laws does not imply that citizens should do nothing when the law flouts their sense of justice. As an integral dimension of a reasonable regime, a state must allow various lawful means of contestation and dissent to be legitimate. When citizens disagree with legitimate directives, they can protest, contact elected officials, start justice campaigns, go to the courts, organize strikes, and support candidates who share their conceptions of justice. Thorough Exclusion is perfectly compatible with those cherished practices of public deliberation and contestation about the justice of political decisions. Active civic engagement is not only a virtue but also a citizen's moral obligation. As citizens constitute a self-governing body, justice is what we owe to each other. Nevertheless, legitimacy marks the threshold for required compliance. Recall that by providing reasonable resolutions amid citizens’ disagreements, legitimate directives are supposed to free citizens from one another's unilateral interpretation and enforcement of justice. Thus, when directives are legitimate, citizens ought to respect them and only use lawful means to advocate change.

Finally, rarely do those who resort to unlawful resistance understand themselves to act against unjust but legitimate directives. For example, during the civil rights movement, James Baldwin wrote: “white people, who had robbed black people of their liberty and who profited by this theft every hour that they lived, had no moral ground on which to stand. They had the judges, the juries, the shotguns, the law—in a word, power. But it was a criminal power, to be feared but not respected, and to be outwitted in any way whatever.”⁵² It seems a fairly common—though often implicit—understanding that illegitimacy, rather than mere injustice, grounds unlawful resistance. Some may insist that unlawful resistance is permissible even against legitimate directives, such as evading taxes when the tax code is reasonable but unfair. Such tactics are nonetheless too controversial to serve as counterexamples to my account.⁵³

D. False Conviction Again

However, I have not forgotten False Conviction, where innocent Ian is convicted for life by an impeccable judicial procedure. Typically, while legitimate but unjust directives negatively affect millions of people, Reasonableness ensures that they can nonetheless retain adequate external freedom. In contrast, the devastating effects of false convictions concentrate on a few people. According to my account, Ian's innocence no longer constitutes a legitimate reason for him or others to use unlawful means to resist the ruling. Thus, if the verdict cannot be reversed legally, Ian is morally obligated to serve his prison sentence. Despite my abiding discomfort, I submit that this conclusion is forced on us if the rule of law is to solve the unilateralism of rights.

Let me begin with a few observations that may soften this conclusion—but only to a certain extent. First, many false convictions are not like Ian's. Instead, they result from severe procedural or substantive failures of the criminal justice system. My account by no means implies that innocent victims of such illegitimate practices are morally obligated to accept their convictions. The procedural and substantive constraints of reasonableness ensure the rarity of legitimate false convictions. Second, due to the knowing risk of false conviction, a legitimate criminal justice system must have ample space for appeal and retrial.⁵⁴ This further guarantees that legitimate false convictions only reflect persistent bad epistemic luck rather than malice or apathy. Finally, while citizens are required and permitted to comply with false but legitimate convictions, the state should compensate the falsely convicted when they prove their innocence.⁵⁵

Despite the above observations, my conclusion regarding False Conviction remains a bitter pill to swallow. One objection is especially pressing. I hold that citizens are obligated to comply with legitimate directives because they are obligated to refrain from forcing their unilateral judgments of justice on others. However, additive theorists will point out that Ian's claim not to be falsely imprisoned for life significantly outweighs the people's claim not to be subject to his unilateral judgment of justice.

In response, I will argue first that the people's claim against Ian's unilateral resolution and Ian's claim against false imprisonment differ crucially.⁵⁶ While the former is a right, the latter is a mere legitimate interest. This is not because the former is more important than the latter. Instead, it is because the moderate statism of rights defended earlier and feasibility constraints on states’ operations dictate divergent treatments of the two.

As argued earlier, because rights impose coercive constraints on one another's freedom, they must be ratified and enforced by a legitimate authority to avoid morally problematic unilateral coercion. This leads to moderate statism of rights, according to which legitimate states have a moral power to determine rights. Thus, people's rights are affected not only by actual political decisions (when they are reasonable and duly made) but also by institutional constraints on states’ operations. On the one hand, administering criminal justice is part and parcel of a state's jurisdiction due to its coercive nature. Thus, the citizenry has the moral right to see to it that everyone complies with the state's legitimate verdicts. On the other hand, since the most reliable prosecution process has a margin of error, even optimal systems of rights administered by legitimate states cannot grant individuals a moral right to accurate verdicts. Therefore, individuals lack such a right. Stated differently, the fact that even the best criminal justice system cannot afford a right against false but fair convictions rules out the possibility that anyone can enjoy such a right. Therefore, while every citizen has a moral right to a fair trial, no one has a moral right against false imprisonment that results from a fair trial.⁵⁷

Furthermore, granting the citizenry a right to administer criminal justice through the state apparatus while denying Ian a right against false but fair convictions is not to sacrifice Ian for the benefit of the majority. Ian's lack of a moral right against false but fair convictions is ultimately an entailment of his own right to external freedom. His right to external freedom mandates that he enter a legitimate state and enjoy a set of definitive rights administered by the state. Since even the best state cannot grant him a right against false but fair convictions, his right to external freedom dictates that he cannot enjoy a right against false but fair convictions—an observation, mutatis mutandis, equally applicable to everyone else.

That being said, Ian's claim against false imprisonment is a weighty legitimate interest. Thus, I face a further challenge: Why should we prioritize the people's right to administer criminal justice over Ian's weighty legitimate interest against false imprisonment? To respond, it is widely believed that rights typically trump mere legitimate interests. For example, a superrich person should help those in need, e.g., by donating to charities that support promising poor students for a college education. The amount of money means little to the rich person but a lot to the students struggling to afford further education. However, assuming that the property regime is legitimate, the rich person plausibly has the moral right not to donate, which the students or any bystander should not violate even by otherwise innocuous means, e.g., secretly transferring what the rich should donate from the rich's account to the charities’ account.⁵⁸ In terms of legitimate interests, the students’ claim to further education significantly outweighs the rich person's claim to a tiny fraction of their wealth. However, the latter trumps the former because the latter constitutes a right, the former a mere legitimate interest. Ultimately, the same normative dynamic explains Ian's obligation to comply with the false verdict unless the verdict is lawfully nullified. While Ian has no moral right against false imprisonment that results from a fair trial, the people have the moral right not to be subject to his unilateral resolution of criminal justice. Of course, one may further question this widely accepted practice of prioritizing rights. Although rights and their priority have been an entrenched and fundamental fabric of our moral landscape, consequentialism denouncing rights and demanding that we always protect the weightiest legitimate interest remains an influential moral theory. This paper does not aim to settle the debate between consequentialists and rights advocates. Its unique contribution is to show that if one is committed to rights as constituting people's external freedom, one should acknowledge the moral necessity of legitimate states and the exclusionary power of their legitimate directives: they enable rights to perform their supposed moral function.⁵⁹

To conclude, False Conviction, or the tremendous power of legitimate but unjust directives in general, attests to the solemnity of the people's right to collective self-governance through the state. It reminds us that being united as a people under the rule of law is a deadly serious enterprise. The flip side of the people's right to collective self-governance is each citizen's moral obligation to ensure the justice of our collective actions. Our deep discomfort with legitimate but flawed directives should spur us not to reject them but to duly correct them in pursuing a more perfect union.